Anal. Chem. 2004, 76, 3699-3706
Comprehensive Two-Dimensional Supercritical Fluid and Gas Chromatography with Independent Fast Programmed Heating of the Gas Chromatographic Column Andre Venter and Egmont R. Rohwer*
Chemistry Department, University of Pretoria, South Africa
With comprehensive two-dimensional supercritical fluid and fast, independent temperature-programmed gas chromatography (SFC×GC), a polar column was used in the first dimension to achieve group-type analysis. The eluent of this separation was repetitively sampled and transferred to a fast, resistively heated gas chromatograph to obtain the boiling point distribution over the entire polarity separation. The SFC was operated isothermally with stopped flow to provide a sufficient time span for the GC analysis. The GC analysis had a typical cycle time of 1 min for the system demonstrated here. During this time, the GC column was independently heated at a rate of 450 °C/min to 250 °C and actively cooled again to -50 °C before the next GC injection took place. The analysis of petrochemical samples is presented to illustrate the technique. Comprehensive two-dimensional chromatography is a hyphenated chromatographic technique in which two separation methods that operate with different retention mechanisms are coupled together. Unlike other multidimensional chromatographic arrangements, such as heart cutting, the entire sample eluting from the first separation is analyzed by the second separation and the resolution achieved in the first analysis is conserved throughout subsequent analysis steps.1 The total information that can be obtained from such an arrangement is surprisingly high if the combination consists of completely independent techniques. Any correlation between the retention mechanisms of the two separations leads to the wasteful production of separation space that cannot be used.2 Ultimately, this synentropy or cross-information3 leads to an increase in total analysis time without an increase in information production. Separation mechanisms that are free from synentropy are said to be orthogonal to each other. In such a case, the two separation dimensions produce a rectangular separation space with compounds distributed evenly across the plane. * To whom correspondence should be addressed. E-mail: erohwer@ postino.up.ac.za. (1) Giddings, J. C. In Multidimensional Chromatography; ; Cortes, H. J., Ed.; Marcel Dekker: New York, 1990, pp 1-29. (2) Venkatramani, C. J.; Xu J.; Phillips, J. B. Anal. Chem. 1996, 68, 14861492. (3) Erni, F.; Frei, R. W. J. Chromatogr. 1978, 149, 561-569. 10.1021/ac035538c CCC: $27.50 Published on Web 05/20/2004
© 2004 American Chemical Society
One of the earliest techniques to employ this multidimensional approach used isoelectric focusing (IEF) and gel-gradient electrophoresis.4 The separation is truly orthogonal, with isoelectric point analysis along one axis and molecular weight analysis along another. Today IEF-PAGE analysis is a very powerful technology that is routinely used for protein analysis. Guiochon and coworkers developed a planar column that reminds one very much of a two-dimensional TLC plate, except that the spots were eluted from the plane and observed with a diode array detector.5,6 The first paper to demonstrate comprehensive two-dimensional column chromatography, in which the eluent from one column was fed into the next, used high performance liquid chromatography (HPLC×HPLC) for both separation steps.7 Electrophoresis was also coupled to HPLC.8 This was followed by comprehensive multidimensional gas chromatography (GC×GC) in 1991.9 Two approaches to this technology using moving temperature zones and stationary phase focusing have been demonstrated by using either cryogenic10 or thermal modulation.11,12 A thermal modulation device is available commercially (Zoex Corp., Lincoln, NE). GC×GC instrumentation generally uses a nonpolar column in the first dimension to separate samples according to volatility. Consecutive small sections of this chromatogram are refocused and introduced into a short polar column where each fraction is separated for differences in polarity. By using the same temperature ramp rate for the two columns, orthogonality is achieved by effectively removing the volatility aspect of the retention mechanism in the second column, leaving only the resultant polar interactions. Volatility always plays the biggest part in GC retention mechanisms with secondary interactions, such as polar interactions superimposed on this. Polar and chiral interactions are (4) O’Farrel, P. H. J. Boil. Chem. 1975, 250, 4007-4021. (5) Guiochon, G.; Beaver, L. A.; Gonnord, M. F.; Siouffi, A. M.; Zakaria, M. J. Chromatogr. 1983, 255, 415-437. (6) Guiochon, G.; Beaver, L. A.; Gonnord, M. F.; Siouffi, A. M.; Zakaria, M. Chromatographia 1983, 27, 121-124. (7) Bushey, M. M.; Jorgenson, J. W. Anal. Chem. 1990, 62, 161-167. (8) Bushey, M. M.; Jorgenson, J. W. Anal. Chem. 1990, 62, 978-984. (9) Lui, Z.; Pillips, J. B. J. Chromatogr. Sci. 1991, 29, 227. (10) Kinghorn, R. M.; Marriott, P. J. J. High Resolut. Chromatogr. 2000, 23, 225252. (11) Phillips, J. B.; Gaines, R. B.; Blomberg, J.; et al. J. High Resolut. Chromatogr. 1999, 22, 3-10. (12) Ledford, E. B.; Billesbach, C. J. High Resolut. Chromatogr. 2000, 23, 202204.
Analytical Chemistry, Vol. 76, No. 13, July 1, 2004 3699
influenced by temperature. The strength of these interactions with the stationary phase decreases with an increase in temperature. This leads to a decrease in the contribution of the polar or chiral functional group of the analyte to the retention mechanism at higher analysis temperatures,13 with a subsequent loss in the separation factor due to this polar or chiral functionality. The range of samples that can be analyzed by the GC×GC technique is restricted by the upper temperature limit of the polar column in the second dimension. The maximum attainable final boiling point (FBP) of analyzable samples is limited to 400 °C.14 When the thermal modulator is used, FBPs of samples are further restricted by the fact that the modulator tube (usually the front part of the polar column) should be warmer than the rest of the second column to ensure sharp injection bandwidths. This limits the types of stationary phases that can be used and restrains the scope of GC×GC-type analysis. Even so, this technique has proved to be powerful for detailed analysis of complex samples, as has been demonstrated for the analysis of petrochemical samples.15,16 Attempts at combining HPLC with GC are hampered by the large amount of solvent that needs to be removed, which requires time and can lead to a loss of volatile sample components. Despite refinements in technology, even simple online (heart-cut) HPLC/ GC is seldom accepted as a method of choice.17 The analysis of triglycerides by comprehensive HPLC×GC was recently reported. The triglycerides could be separated into four groups, and individual components could be separated. The HPLC fractions were collected off-line using a fraction collector and injected into a fast GC. Analysis times were typically 11 hours.18 Preliminary results of an online HPLC×GC were also recently presented.19 Supercritical fluid chromatography (SFC) has proved to be highly suitable for the chemical class analysis of complex mixtures.20 Apart from chromatographic advantages of SFC over HPLC, the popular CO2 mobile phase is easy to remove from the analytes when compared to the less volatile solvents typically used in HPLC. Lee et al. demonstrated that the supercritical CO2 mobile phase can also be used as the gaseous mobile phase in the gas chromatographic separation after decompression at the restrictor exit in a comprehensive SFC×GC instrument.21 A cyanopropyl polysiloxane stationary phase was used for group-type separation in the first dimension and a liquid crystal stationary phase in the second dimension for separation by molecular shape. Both columns were in the same oven and, thus, temperature pro(13) Rotzsche; H. Stationary phases in gas chromatography. J. Chromatogr. Library; Elsevier: New York, 1991; Vol. 48, pp 80-83. (14) Beens, J.; Boelens, H.; Tijessen, R.; Blomberg, J. J. High Resolut. Chromatogr. 1998, 21, 47-54. (15) Blomberg, J.; Schoenmakers, P. J.; Beens, J.; Tijssen, R. J. High Resolut. Chromatogr. 1997, 20, 539. (16) Frysinger, G. S.; Gaimes, D. R.; Ledford, E. B. J. High Resolut. Chromatogr. 1999, 22, 195-200. (17) Grob, K. J. Chromatogr., A 2000, 892, 407-420. (18) Janssen, H.; Broers, W.; Steenbergen, H.; Horsten, R.; Floter, E. J. Chromatogr., A 2003, 1000, 385-400. (19) Van Deursen, M. Paper presented at the 7th International Symposium on Hyphenated Techniques in Chromatography and Chromatographic Analysers, Bruges, Belgium, February 2002. (20) Annual Book of ASTM Standards; method D5183; American Society for Testing and Materials: Philadelphia, 1991; Vol. 05.03. (21) Lui, Z.; Ostrovsky, I.; Farnsworth, P. B.; Lee M. L. Chromatographia 1993, 35, 567-573.
3700
Analytical Chemistry, Vol. 76, No. 13, July 1, 2004
grammed at the same rate. This implies that each second dimension gas chromatogram was essentially isothermal for its duration, as is also the case in GC×GC. This had the advantage that the second dimension did not require temperature reequilibration for each next injection delivered by the thermal modulator; however, group separation of the first dimension was compromised by the increasing temperature. Further, due to slow diffusion coefficients, CO2 is not a good mobile phase to use with high-speed gas chromatography.22 While strong on simplicity, this approach to SFC×GC does not bring out the full potential offered by either SFC or GC. Nevertheless, polyaromatic hydrocarbons (PAHs) with up to four rings could be analyzed with good peak shapes using this approach. A low temperature is beneficial for chiral or chemical class separations, and so the SFC separation should preferably be done at a low temperature.23 Temperature programming is a prerequisite for fast GC of complex mixtures spanning a wide volatility range.24 Both these issues are addressed in the proposed approach to SFC×GC described in this paper. Lee, Yang, and Bartle were among the first to suggest the use of direct, resistive heating of capillary columns by an electric current for GC.25 Using resistively heated fast programmable GC, cycle times can be reduced to a minute or less. This can be achieved because of the low energy and time involved in heating and cooling the small thermal mass of the capillary column, as compared to that of the massive conventional GC oven. Direct heating of aluminum-coated26 or silver-painted27 fused-silica capillaries has been demonstrated. Temperature has been measured as a function of the resistance of the column itself, or as a function of a separate sensor wire, inserted into a sleeve together with the column. This was the approach followed by Ehrmann et al.28 as well as in the commercial Flash-GC (Thermedics Corp., Chelmsford, MA), in which the analytical column is placed inside a separate heating sleeve. Commercial units can ramp a 5-m column at rates up to 1200 °C/min;29 however, a separate sleeve increases the thermal mass and thereby increases cycle times due to a slower cooling phase. Analysis times of 30 s for n-alkanes up to C24 are usual, with cool-down times of about 80 s. A temperature-programmable simulated distillation gas chromatograph based on a resistively heated packed column with a cycle time of 3 min is also available commercially (ABB Automation, Lewisburg, WV). Cycle times for simulated distillation are, once again,
