MS with a Series-Coupled Column Ensemble

A pneumatically actuated valve is used to connect the junction point of a series-coupled column ensemble to a ballast chamber containing carrier gas a...
0 downloads 0 Views 81KB Size
Anal. Chem. 2001, 73, 3045-3050

High-Speed GC and GC/MS with a Series-Coupled Column Ensemble Using Stop-Flow Operation Tincuta Veriotti and Richard Sacks*

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

A pneumatically actuated valve is used to connect the junction point of a series-coupled column ensemble to a ballast chamber containing carrier gas at the ensemble inlet pressure in order to periodically stop the carrier gas flow in the first column. When the valve is opened, mixture components, which have migrated across the column junction, are accelerated toward a time-of-flight mass spectrometer that is used as an ensemble detector. Mixture components, which are still in the first column, are frozen in position. This allows for the insertion of time windows into the ensemble chromatogram that can aid in the separation of some overlapping component peaks. The capillary column ensemble (0.18-mm i.d. × 0.18µm film thickness) consists of a 7.0-m length of polar, (trifluoropropyl)methyl polysiloxane column followed by a 7.0-m length of nonpolar dimethyl polysiloxane column. A flame ionization detector located at the column junction point is used to monitor a portion of the effluent from the first column in order to determine the valve timing sequence needed to enhance the separation of component pairs that are separated by the first column but coelute from the column ensemble. When one of the components of a targeted pair has crossed the junction but the other component is still in the first column, the valve is opened, typically for 1-5 s. The stop-flow system is used to enhance the separation of a mixture containing some common essential oil components and a mixture containing some common pesticides. Techniques for the enhancement of selectivity are a major interest in analytical separation science. For gas chromatography, chemical methods of enhancing selectivity involve the design and synthesis of new stationary phases that are uniquely selective for specified applications.1-3 Metric methods involve coupling commercially available columns with different stationary phases. The pressure drops across each of the series-coupled columns are adjusted at the junction point to control the ensemble selectivity for specific applications.4-9 Metric methods are considerably more flexible since the selectivity is readily altered. (1) Rotzsche, H. Stationary Phases in Gas Chromatography; Elsevier: New York, 1991. (2) 2.Guide to Stationary Phases for Gas Chromatography; Analabs: North Haven, CT, 1969. (3) Baiulescu, G.; Ilie, V. A. Stationary Phases in Gas Chromatography; Pergamon Press: New York, 1975. 10.1021/ac001539i CCC: $20.00 Published on Web 05/17/2001

© 2001 American Chemical Society

A novel approach for adjusting selectivity involves a seriescoupled combination of two capillary columns using different stationary phases and a means of applying relatively short pressure pulses to the junction point of the column ensemble.10,11 For component pairs that are separated by the first column but coelute from the column ensemble, a pressure pulse applied after one of the components of a target pair has migrated across the column junction and the other component is still in the first column will result in enhanced separation and resolution of the pair when it elutes from the ensemble. A significant advantage of this approach is that the pulse does not alter the elution order of components that are in the same column during the pressure pulse. This reduces the risk that a pressure pulse used to enhance the resolution of a specified component pair will cause the coelution of a different component pair that is adequately resolved without the pressure pulse. If the pressure at the column junction point decreases during a pressure pulse, the carrier gas velocity in the first column increases and the velocity in the second column decreases. This situation was recently reported for the case of a valve connected between the column junction point and an atmospheric pressure vent.11 The valve is opened for intervals of typically 1-5 s in order to enhance the resolution of targeted peak pairs. The valve is opened when the first component of a peak pair has crossed the column junction while the second component is still in the first column. While very simple and convenient, sample loss through the valve is a problem if the valve is not closed before the second component reaches the junction point. If the pressure at the column junction point during a pulse is greater than the value that would occur at the junction point in the absence of any other connections, the carrier gas velocity decreases in the first column and increases in the second column. With these conditions, sample loss is impossible, and the pulse length can be increased. If the pulse pressure is greater than the GC inlet pressure, the gas flow direction in the first column is reversed, and the contents of the first column are back-flushed toward the GC inlet. This could result in inlet contamination. (4) Deans, D. R.; Scott, I. Anal. Chem. 1973, 45, 1137. (5) Sandra, P.; David, F.; Prood, M.; Diricks, G.; Verstappe, M.; Verzele, M. J High Resolut. Chromatogr. Chromatogr. Commun. 1985, 8, 782. (6) Hinshaw, J. V.; Ettre, L. S. Chromatographia 1986, 21, 561. (7) Akard, M.; Sacks, R. Anal. Chem. 1994, 66, 3036. (8) Akard, M.; Sacks, R. Anal. Chem. 1995, 67, 2733. (9) Sacks, R.; Smith, H.; Nowak, M. Anal. Chem. 1998, 70, 29A. (10) Veriotti, T.; McGuigan, M.; Sacks, R. Anal. Chem. 2001, 73, 279. (11) Veriotti, T.; Sacks, R. Anal. Chem. 2001, 73, 813.

Analytical Chemistry, Vol. 73, No. 13, July 1, 2001 3045

Table 1. Compounds Used in the Test Mixtures peak 1 2 3 4 5 6 7 8 9 10 1 2 Figure 1. Apparatus used for series-coupled column ensemble with stop-flow operation. Key: C1, polar column; C2, nonpolar column; V, pneumatically operated valve; BC, ballast chamber; PC, electronic pressure controller; MS, time-of-flight mass spectrometer; FID, flame ionization detector; I, split GC inlet; CG, carrier gas supply.

If the pulse pressure at the column junction point is equal to the column-ensemble head pressure, the carrier gas flow ceases in the first column (stop-flow operation), and the pulse can be arbitrarily long without the risk of sample loss or inlet contamination. This report considers the use of a pneumatically actuated valve connected between the column junction point and a ballast chamber containing carrier gas at the pressure needed to stop the flow in the first column. The ballast chamber pressure is adjusted empirically to completely stop carrier gas flow in the first column when the valve is open. The stop-flow system is used to enhance selectivity for a mixture containing some common essential oil components and a mixture containing some common pesticides. EXPERIMENTAL SECTION Apparatus. Figure 1 shows the dual-column ensemble and the pressure pulse system used for this study. An HP6890 GC (Hewlett-Packard, Atlanta GA) equipped with an HP7683 autoinjector was used for all experiments. The inlet I of the HP6890 was used in the split mode with electronic pressure control. A time-of-flight (TOF) mass spectrometer MS (Pegasus II, LECO Corp., St Joseph, MI) was used as the primary detector for the column ensemble. The TOFMS uses time array detection to acquire up to 500 full-mass-range spectra/s. The capillary column ensemble (0.18-mm i.d. × 0.18-µm film thickness) consists of C1, a 7.0-m length of polar, (trifluoropropyl)methyl polysiloxane column (DB-200, J&W Scientific, Inc., Folsom, CA) followed by C2, a 7.0-m length of nonpolar 5% phenyl dimethyl polysiloxane column (DB-5, J&W Scientific). Pressure pulses are produced by opening valve V, which is connected between the column junction point and ballast chamber BC containing carrier gas at a controlled pressure. The valve is a pneumatically operated, low-dead-volume device (model MOPV1/50, SGE, Austin, TX). It is operated by a 50-55 psig compressed air source connected through an electronically actuated solenoid valve (model GH3412, Precision Dynamics, Phoenix, AZ). A 4.0cm-long segment of 0.1-mm-i.d. deactivated fused-silica column is used to connect the valve to the column junction point by means of a low-dead-volume all-glass splitter. The ballast chamber is an 3046 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

3 4 5 6 a

compound

peak

Pesticide Mixture R-BHC 12 β-BHC 13 γ-BHC 14 δ-BHC 15 heptachlor 16 aldrin 17 heptachlor eopxide (B) 18 R-chlordane 19 γ-chlordane 20 4,4'-DDe 21

compound dieldrin endrin 4,4'-DDD endosulfan II 4,4'-DDT endrin aldehyde metoxychlor endosulfan sulfate diphenylthiocarbazidea endrin ketone

Essential Oils Mixture 7 1-methyl-2-(1-methylethyl)benzenea R-pinene 8 1-methyl-3-(1-methylethyl)benzenea decane 9 2,3,4-trimethylethyl)benzene R-phellandrene 10 1,8-cineole R-terpinene 11 benzyl alcohol limonene origanene

Impurities.

aluminum cylinder and has a volume of 450 cm3. The ballast chamber pressure is controlled by electronic pressure controller PC (MKS model 640A, MKS Instruments, Andover, MA), which is also connected to a source of purified carrier gas CG. 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. On the basis of standard equations for gas flow through capillary tubes,12 ∼8% of the effluent from the first column is split to this detector. Materials and Procedures. Hydrogen carrier gas is purified by filters for water vapor, oxygen, and hydrocarbons. Injections of 4-methylstyrene (1-µL headspace with 10:1 split ratio) were used with an isothermal column ensemble temperature of 80 °C and a GC inlet pressure of 30 psig to determine the ballast chamber pressure needed to achieve stop-flow conditions in the first column. The two test mixtures used for these studies are listed in Table 1. Compounds identified with asterisks are concomitant species and not added components. The pesticide mixture is a commercially available blend of 20 components at 100 ppm each in a 1:1 hexane/toluene solvent (organochlorine pesticide mix AB-1, Restek, Bellefonte, PA). The mixture was further diluted to 20 ppm for each component using the same solvent mixture. Injection size was 1.0 µL with a 5:1 split ratio. The essential oil components (neat) were obtained from LECO Corp. They were mixed in equal volumes, and the mixture was diluted 1:10 with n-pentane. Injection size was 0.1 µL with a 100:1 split ratio. An inlet temperature of 250 °C was used for both mixtures. For the pesticide mixture, the inlet pressure was 30 psig. Temperature programming from 175 to 300 °C at 50 °C/min beginning at the time of injection was used. For the essential oil mixture, the nominal inlet pressure was 35 psig. Temperature programming from 50 to 130 °C at 50 °C/min beginning at the time of injection was used. (12) Grant, D. Capillary Gas Chromatography; Wiley: New York, 1996.

Figure 2. FID signal for pressure pulses with various ballast chamber pressures. Pressures (psia) are shown to the left of the signals. The nominal inlet pressure was 30 psig.

Data processing for the TOFMS included automated peak finding, spectral deconvolution of overlapping chromatographic peaks, and component characterization by comparison with the NIST mass spectral database. These operations were accomplished with software provided by the manufacturer. A spectral acquisition rate of 25 spectra/s was used for all studies. Data acquisition for the FID, control of the ballast chamber pressure, and operation of the pressure pulse valve were obtained with a 330-MHz PC (Dell, OptiPlex GX1) and a 12-bit A/D board (DT2801, Data Translation, Inc., Marlboro, MA). The interface board was operated with Labtech Notebook software (Laboratory Technologies, Inc., Wilmington, VA). Processing of FID chromatograms was accomplished with Grams/32 software (Galactic Industries, Salem, NH). RESULTS AND DISCUSION System Characterization. With the apparatus shown in Figure 1, if the pressure in the ballast chamber is less than the column junction point pressure with the valve closed, effluent from the first column is split between the second column and the ballast chamber when the valve is open. This manner of sample splitting would result in sample loss and contamination of the ballast chamber. On the basis of calculations using standard equations for gas flow in capillary tubes,12 the valve-closed junction point pressure for the conditions used in this study is 49.7 psia for the essential oil mixture and 46.2 psia for the pesticide mixture. The ballast chamber pressure was always maintained above these values for work with the respective mixtures. When the valve is opened, a small change can be observed in the baseline signal from the FID that is monitoring the column junction point. This change may be a result of the increase in flow with the valve open, or it may be the result of impurities in the ballast chamber (machine oil residue). In either case, this signal change is useful in monitoring valve performance. Figure 2 shows the signal from the FID for various ballast chamber pressures. For this study, a nominal GC inlet pressure of 30 psig

Figure 3. FID signals for injections of 4-methylstyrene with various ballast chamber pressures. Pressures (psia) are shown to the left of the signals. The nominal inlet pressure was 30 psig. The isothermal temperature was 80 °C. See text for details.

was used. The nominal valve-open time was 10.0 s. For a ballast chamber pressure of 44.7 psia, a relatively rectangular pulse is observed with a rise time (10-90% final value) of 1.3 s and a fall time of 4.5 s. For higher ballast chamber pressures, a secondary feature is observed on the falling edge of the pulse. The amplitude of this feature is relatively constant, but its width increases steadily with increasing pressure. The electronic pressure controller in the GC was used to set the inlet pressure to 30.0 psig. Ambient pressure is ∼14.2 psia (neglecting fluctuations in barometric pressure), and thus, a junction point pressure greater than 44.2 psia should reverse the carrier gas flow direction in the first column (back-flush conditions). The first column then fills with gas from the ballast chamber for the duration of the pressure pulse. When the valve is closed, normal flow direction resumes, and ballast chamber gas and its impurities are detected by the FID producing the tailing artifact seen in Figure 2. Back-flush conditions are to be avoided due to the risk of inlet contamination and baseline shifts from the column ensemble. To stop the flow in the first column when the valve is open, the pressure at the column junction point should be equal to the head pressure of the column ensemble (GC inlet pressure). Figure 3 shows the effects of ballast chamber pressure on the elution time of 4-methylstyrene for the first column. The nominal GC inlet pressure was 30 psig, and the isothermal temperature was 80 °C. Numbers to the left of the plots indicate the ballast chamber pressure in psia. Note that pressure drops in the plumbing connecting the ballast chamber to the valve and the valve to the column junction point result in a higher pressure in the ballast chamber than at the column junction point. The pressure pulse waveform also is seen in the FID signals of Figure 3. The vertical lines on the left side of the figure indicate the valve-open and -close times. The nominal pulse width is 5.0 s, but the relatively large fall time extends the time interval for which the carrier gas velocity in the first column is affected by the open valve state. The vertical lines on the right side of the figure show the same time interval beginning at the first-column retention time for the test solute with the valve closed. Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

3047

Figure 5. FID chromatograms of the essential oil component mixture for no-stop-flow pulse (a), and stop-flow pulse widths of 3.0 (b), 6.0 (c), and 12.0 s (d). For chromatograms b-d, the stop-flow valve was opened 27.5 s after sample injection. The inlet pressure was 35 psig, and the ballast chamber pressure was 53 psia. Peak numbers correspond to component numbers in Table 1.

Figure 4. Plots of the retention time change for 4-methylstyrene versus pressure pulse width for the FID (a) and the TOFMS (b). The ballast chamber pressures were 50 (A), 48 (B), 47 (C), and 42 psia (D). Table 2. Correlation Figure 5

Coefficients

and

FID

Slopes

for

MS

pressure, psig

slope

R2

slope

R2

42 47 48 50

0.6209 0.9072 1.0092 1.3111

0.9981 0.9980 0.9992 0.9993

0.6232 0.9064 1.0042 1.3268

0.9973 0.9981 0.9992 0.9990

As the pressure in the ballast chamber increases, the retention time of the 4-methylstyrene steadily increases. The shape of the analyte peak appears to be relativey independent of the pressure in the ballast chamber. For a ballast chamber pressure of 48 psia, the shift in the retention time of the analyte relative to the case with valve remaining closed is about equal to the pulse width. This suggests that the flow is nearly stopped when the valve is open with this ballast chamber pressure. Figure 4 shows plots of the shift in retention time of the 4-methylstyrene peak versus the pulse width using ballast chamber pressures of 50 (A), 48 (B), 47 (C), and 42 pisa (D). The nominal GC inlet pressure was 30 psig, and the isothermal temperature was 80 °C. Data for plots a were obtained from the FID monitoring the effluent from the first column, and data for plots b were obtained from the mass spectrometer monitoring the column ensemble. Table 2 gives slopes and correlation coefficients for the plots. All plots have correlation coefficients greater than 0.9973. For both sets of plots, the slope values are about equal to 1.00 for a ballast chamber pressure of 48.0 psia. Again, this indicates that stop-flow conditions are achieved in the 3048 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

first column. Note that the slopes of these plots are very sensitive to the ballast chamber pressure, thus allowing the easy determination of the pressure needed to achieve stop-flow conditions. Stop-Flow Operation with Temperature Programming. Figure 5 shows FID signals for injections of the essential oil mixture for different valve-open times using a ballast chamber pressure of 53.0 psia and a nominal GC inlet pressure of 35 psig. Note that the ballast chamber pressure was increased by 5.0 psi relative to the value used for Figures 2-4 to achieve stop-flow conditions with the higher inlet pressure used for the essential oil mixture. Peak numbers correspond to the component numbers in Table 1. For chromatogram a, the valve was not opened. For chromatograms b-d, the valve was opened 27.5 s after injection for durations of 3, 6, and 12 s, respectively. At this time, components 1-8 and 10 have eluted from the first column, but components 9 and 11 are still in the first column when the valve is opened. The change in baseline signal when the valve is open is not apparent on the compressed vertical scale used for the relatively high component concentrations. For an isothermal separation (Figure 4), the migration rates of components in the first column after completion of the stopflow pulse return to the values just prior to the pulse. However, for temperature-programmed operation, the column ensemble temperature continues to increase during the stop-flow pulse, and migration rates on the first column are greater after completion of the pulse than just prior to the start of the pulse. This increase in migration rates after completion of the pulse is greater for the longer stop-flow intervals. The time intervals between components 9 and 10 are shown in Figure 5. For case a, where the valve is closed for the entire experiment, the peak apex separation is 2.3 s. For the cases where the valve is opened, the sum of the nominal pulse width and the no-pulse separation differs from the measured peak separation by less than 2% for all nominal pulse widths. This close agreement is obtained under temperature-programmed conditions because component 9 is very near the column junction point when the valve is opened. Thus, the time delay for the detection of

Figure 6. Essential oil chromatograms from the FID (a) and the TOFMS (b) and (c). For chromatogram c, the stop-flow valve was opened for 3.0 s at 27.5 s after sample injection. The inlet pressure was 35 psig, and the ballast chamber pressure was 53 psia. Peak numbers correspond to component numbers in Table 1.

component 9 by the FID is dominated by the width of the stopflow pulse and not by the residual migration time of component 9 in the first column. Note that the separation of components 9 and 11 decreases steadily with increasing pulse width due to the steady increase in column temperature at the termination of the pulse. Stop-Flow Applications. When capillary columns using different stationary phases are connected in series, it is common for some component pairs that are separated by the first column to coelute from the column ensemble. If the carrier gas flow in the first column is stopped when one of the components has crossed the column junction and the other component is still in the first column, the ensemble separation of these components can be made arbitrarily large by control of the stop-flow pulse width. For this approach to be useful, it is necessary for the two components to be adequately resolved by the first column. Figure 6 shows results for the essential oil component mixture described in Table 1. The FID chromatogram a is from the first (polar) column, and the analytical ion chromatograms b and c are from the column ensemble. An analytical ion chromatogram provides information similar to a total ion chromatogram but with appreciably greater signal-to-noise ratio. Some of the mixture components chosen for this study present coelution problems for the flavor and fragrance industries. In particular, limonene (Table 1, peak 6) and R-terpinene (Table 1, peak 5) may coelute on poly(ethylene glycol) columns. Note that these components also coelute (peak 5, 6 in the FID chromatogram) on the (trifluoropropyl)methyl polysiloxane column used as the first column in the ensemble. On 5% phenyl dimethyl polysiloxane columns, R-phellandrene, limonene, and 1,8-cineole may be poorly resolved. All of the target compounds are well resolved with the tandem column ensemble b. However, 2,3,5-trimethylpyrazine (component 9) coelutes with concomitant species 1-methyl-2-(1-methyl)benzene (component 7) and 1-methyl-3-(1-methyl)benzene (component 8). Note that these latter compounds have structural similarities with R-phellandrene and R-terpinene and are often associated with them

Figure 7. Pesticide mixture chromatograms from the FID (a) and the TOFMS (b) without a stop-flow pulse. The inlet pressure was 30 psig, and the ballast chamber pressure was 48 psia. Peak numbers correspond to component numbers in Table 1. Broken vertical lines in chromatogram a indicate the points in the FID chromatogram where stop-flow pulses are used to enhance the resolution of component pairs 2/3, 10/11, and 20/21.

From the FID chromatogram a, it is clear that components 7 and 8 reach the column junction point several seconds before component 9. Thus, if the carrier gas flow in the first column is stopped briefly before component 9 reaches the column junction, these components should be separated in the ensemble chromatogram. The stop-flow valve was opened at 27.5 s, just after component 10 crossed the column junction point. Because of the close proximity of components 7, 8, and 10 in the ensemble chromatogram, the valve was open for 3.0 s so that component 9 would elute after component 10 in the ensemble chromatogram. The result is shown in the analytical ion chromatogram c in Figure 6. Note that all target components are now separated with greater than baseline resolution in ∼55 s. Figure 7 shows the FID chromatogram from column 1 (a) and the TOFMS analytical ion chromatogram from the column ensemble (b) for the pesticide mixture for the no-pulse case. This mixture is more complex than the essential oil mixture, and three peak pairs in the ensemble chromatogram have resolution less than 1.5 (baseline resolution). Components 2 and 3 (β-BHC and γ-BHC, respectively) coelute, as do components 10 and 11 (4,4′DDE and endosulfan I, respectively). Also, component 21 (endrin ketone) shows significant overlap with component 20 (solvent impurity). Component pairs 10, 11 and 20, 21 are adequately separated by the first column to allow for stop-flow resolution enhancement from the column ensemble. Components in peak 2, 3 are isomers and cannot by easily distinguish by MS. In addition, they elute on the tail of the solvent peak, and it is difficult to ascertain their resolution as they elute from the first column. Thus, while a stopflow pulse was used to separate them in time, there may be significant mutual contamination. The broken vertical lines in Figure 7a indicate the positions in the FID chromatogram where stop-flow pulses were used. The first pulse is initiated 28 s after injection. Since the first pulse delays elution from the first column of all components still in the column when the pulse occurs, the second pulse must be delayed from the time (vertical broken line) indicated in Figure 7a. For Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

3049

the third pulse had been delayed until after component 19 had crossed the column junction point. Also note in chromatogram b that the resolution of components 13 and 14 is significantly reduced. For chromatogram c, component 19 elutes from the column ensemble after component 20, and all components in this region of the ensemble chromatogram are very well separated. However, the resolution of components 13 and 14 is unacceptable.

Figure 8. Pesticide mixture analytical ion chromatograms from the TOFMS with three stop-flow pulse. The inlet pressure was 30 psig, and the ballast chamber pressure was 48 psia. Peak numbers correspond to component numbers in Table 1. Broken line boxes indicated the regions of the chromatograms affected by the pulses. The pulse durations for component pairs 2/3, 10/11, and 20/21 were 1.0, 2.0, and 1.0 s, respectively, for (a), 3.0, 6.0, and 3.0 s, respectively, for (b), and 7.0, 10.0, and 7.0 s, respectively, for (c).

an isothermal separation, the migration rates of components in the first column after completion of the first pulse return to the values just prior to the pulse, and the delay for the second pulse is just equal to the width of the first pulse. However, for the temperature-programmed separations used here, the column temperature continues to increase during the first stop-flow pulse, and migration rates in the first column are greater after completion of the first pulse than just prior to the start of the pulse. The appropriate timing for the second and third pulses is easily determined from the FID chromatograms. The results are shown in the analytical ion chromatograms of Figure 8. For chromatogram a, the stop-flow valve was opened for intervals of 1, 2, and 1 s for component pairs 2/3, 10/11, and 20/21, respectively. For chromatogram b, the valve was opened for intervals of 3, 6, and 3 s, respectively, and for chromatogram c, he valve was opened for intervals of 7, 10, and 7 s, respectively. The broken line boxes show the regions of the ensemble chromatogram that were modified by the stop-flow pulses. For chromatogram a, all 20 target compounds are separated with baseline resolution or greater. However, the purity of peaks 2 and 3 is uncertain. The least resolved pair is 13/14. For chromatogram b, the separation of all targeted peak pairs has increased, but component 19 now coelutes with component 20 (solvent impurity). This is because the third pulse was initiated while components 19 and 21 were still in the first column (see FID chromatogram in Figure 7), and the elution order of components 19 and 20 changes in the second column. This coelution of components 19 and 20 would not have occurred if

3050

Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

CONCLUSIONS The column 1 stop-flow method for enhancing selectivity with a series-coupled column ensemble has several significant advantages over other methods, which utilize pressure control at the column junction point for the enhancement of selectivity. First, this method successfully targets individual pairs or small groups of components, and the increased resolution obtained for a targeted pair or group of components does not significantly affect the elution order or resolution of most other mixture components (components that are in the same column during a stop-flow pulse). Second, the stop-flow method eliminates the risk of sample loss and inlet contamination. For more complex mixtures where two or more peak pairs show excessive overlap, multiple stopflow pulses can be used, each one targeting a specific peak pair or small group of peaks. The ballast chamber pressure needed to achieve stop-flow operation in the first column can be determined by plotting the change in retention time versus the pulse width under isothermal conditions for a component that is in the first column at the start of the pulse. The pressure in the ballast chamber then is adjusted to give a slope of 1.0. Under temperature-programmed conditions, the column ensemble continues to heat during a stop-flow pulse, and the component migration rates in the first column after completion of a pulse are greater than just prior to the initiation of the pulse. For this reason, the delay required for the initiation of a second pulse is less than the duration of the first pulse. Since all sources of band broadening except longitudinal diffusion in the carrier gas are absent when carrier gas flow stops in the first column, relatively long stop-flow pulses (greater than 10 s) can be used without significant increases in ensemble peak widths. This may allow the column 1 stop-flow method to be used for other applications such as multiplexing two columns to a single detector. Future studies will attempt to achieve column 1 stopflow conditions by connecting the GC inlet directly to the column junction point through the low-dead-volume valve, thus eliminating the needed for the pressure controller. ACKNOWLEDGMENT The authors gratefully acknowledge LECO Corp., St. Joseph, MI, for use of the Pegasus II TOFMS instrument. Received for review December 29, 2000. Accepted April 10, 2001. AC001539I