Anal. Chem. 2001, 73, 3340-3347
Dependence of Thermal Mismatch Broadening on Column Diameter in High-Speed Liquid Chromatography at Elevated Temperatures Jonathan D. Thompson,† James S. Brown,‡ and Peter W. Carr*,†
Department of Chemistry, University of Minnesota, Smith and Kolthoff Halls, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, and Department of Chemical Engineering, Georgia Institute of Technology, 778 Atlantic Drive, Atlanta, Georgia 30332
In this paper, we compare a narrow-bore column (2.1mm i.d.) to a conventional-bore column (4.6 mm i.d.) at elevated temperatures under conditions where thermal mismatch broadening is serious and show that narrowbore columns offer significant advantages in terms of efficiency and peak shape at higher linear velocities. We conclude that the so-called thermal mismatch broadening effect is largely due to a radial retention factor gradient and not a radial viscosity gradient. The lower volumetric flow rates inherent with the use of narrower columns lead to lower linear velocity in the heater tubing and longer eluent residence times in the heater. Thus, with the same heater tubing at the same column linear velocity, narrowbore columns give better thermal equilibration between the eluent and the column compared to wider bore columns. This means that high-temperature, ultrafast liquid chromatography no longer requires excessively long preheater tubing to thermally equilibrate the eluent to the column temperature. Consequently, the use of narrowbore columns at high-temperature improves analysis speed and efficiency over wider bore columns. We also discuss the advantages of using liquid heat-transfer media as compared to air as the heat-transfer media. We show that an air bath ought not be used to heat the mobile phase because at high temperature (>80 °C) and high column linear velocity (>1.5 cm/s) the length of tubing needed to heat the mobile phase to column temperature is prohibitively long. Using accurate, empirical heat-transfer correlations, we estimated the length of tubing needed to heat the eluent as a function of the column linear velocity for both air and liquid heat-transfer media. Analysis times in LC are typically long and linear velocities are low (0.25-1.5 cm/s) compared to GC (2-10 cm/s).1 The chief advantage of the use of higher temperatures (>100 °C) in HPLC is that analysis speed can be dramatically improved.2,3 This is a direct result of the 5-10-fold decrease in eluent viscosity upon * To whom correspondence should be addressed: (e-mail)
[email protected]. † University of Minnesota. ‡ Georgia Institute of Technology. (1) Guiochon, G. Anal. Chem. 1980, 52, 2002-2008. (2) Yan, B.; Zhao, J.; Brown, J. S.; Carr, P. Anal. Chem. 2000, 72, 1253-1262. (3) Antia, F. D.; Horvath, C. J. Chromatogr. 1988, 435, 1-15.
3340 Analytical Chemistry, Vol. 73, No. 14, July 15, 2001
increasing the eluent’s temperature from 25 to 200 °C. Consequently, at high-temperature, system pressure drops4-7 and analyte diffusivity is enhanced so that high linear velocities become attainable within the pressure limits of conventional pumps while simultaneously minimizing the broadening associated with the use of high linear velocity. Over a decade ago, Antia and Horvath3 carried out a detailed analysis of the role of temperature in LC and predicted the possibility of achieving dramatic improvements in speed by operating at very high temperatures (150-200 °C). They found that the theoretical analysis time (at constant efficiency and fixed pressure drop) ought to improve by more than a factor of 20 if columns could be operated at high temperatures. High-temperature liquid chromatography is not widely used because there are some severe requirements and implementation problems.2,8,9 The primary limitation is due to the use of silicabased stationary phases in HPLC, which are unstable in aqueous media at temperatures greater than 50-60 °C.10,11 Even though silica phases with improved stability have recently been developed,12-14 the use of temperature as an operating variable in HPLC is certainly underdeveloped and is still far from being fully explored.7,15-17 Commonly used silica-based stationary phases allow for a maximum increase in temperature of only 30-40 °C above ambient conditions and so the benefits in analysis time are at best modest.3 (4) Cramers, C. A.; Rijks, J. A.; Schutjes, C. P. M. Chromatographia 1981, 14, 439-444. (5) Li, J. W.; Hu, Y.; Carr, P. W. Anal. Chem. 1997, 69, 3884-3888. (6) Halasz, I.; Endele, R.; Asshauer, J. J. Chromatogr. 1975, 112, 37-60. (7) Ooms, B. LC-GC 1996, 9, 574-585. (8) Chen, H.; Horvath, C. J. Chromatogr. 1995, 705, 3-20. (9) Chen, H.; Horvath, C. Anal. Methods Instrum. 1993, 1, 213-222. (10) Kikta, E. J.; Grushka, E. Anal. Chem. 1976, 48, 1098-1104. (11) Colin, H.; Diez-Masa, J. C.; Guiochon, G. J. Chromatogr. 1978, 167, 4165. (12) Kirkland, J. J.; Henderson, J. W.; DeStefano, J. J.; Straten, M. A.; Claessens, H. A. J. Chromatogr., A 1997, 762, 97-112. (13) Chloupek, R. C.; Hancock, W. S.; Kirland, B. A.; Boyes, B. E.; Snyder, L. R. J. Chromatogr. 1994, 686, 45-49. (14) Hancock, W. S.; Chloupek, R. C.; Kirkland, J. J.; Snyder, L. R. J. Chromatogr. 1994, 686, 31. (15) Snyder, L. R.; Glajch, J. L.; Kirkland, J. J. Practical HPLC Method Development, 2nd ed.; John Wiley & Sons: New York, 1997. (16) Zhu, P. L.; Dolan, J. W.; Snyder, L. R. J. Chromatogr., A 1996, 756, 41-50. (17) Zhu, P. L.; Snyder, L. R.; Dolan, J. W.; Djordjevic, N. M.; Hill, D. W.; Sander, L. C.; Waeghe, T. J. J. Chromatogr. 1996, 756, 21-39. 10.1021/ac010091y CCC: $20.00
© 2001 American Chemical Society Published on Web 06/15/2001
Figure 1. Block diagram of a simple high-temperature HPLC system.
Second, even though thermally stable metal oxide, polymer, and carbon stationary phases have been developed,18-23 there are problems that arise with the implementation of HPLC at higher temperatures.24-26 To do high-temperature HPLC, it is evident that the eluent and column must be heated (see Figure 1). It is also true that most common types of detectors used in HPLC (absorbance, refractive index) cannot tolerate hot (>60 °C) column effluent. Thus, as shown in Figure 1, when it is too hot, the column effluent must be cooled prior to entering the detector. The chief problem that must be overcome is the extracolumn peak broadening27-29 and increased system back pressure6 due to the tubing needed to heat the eluent to the desired temperature and, when necessary, cool it. Adjusting the eluent temperature without simultaneously introducing overwhelming extracolumn band broadening and excessive pressure drops is quite complicated. This is due to (a) the significant resistances to heat transfer, primarily in the flowing eluent,2 (b) the high heat capacities of aqueous eluents, and (c) high-pressure drops encountered in using long (>10 cm), very narrow ( 60 °C). Under thermally matched conditions, as column temperature increases, the analyte’s diffusion coefficient increases and enhances interphase mass transfer and efficiency. When the eluent is not thermally matched to the column, broadening that comes from the mismatch masks any gain in efficiency that would normally be observed as the temperature is raised. For the wider column, the slopes increase as the temperature increases. This inversion of slopes characterizes temperature mismatch broadening. Even though the magnitude of the thermal mismatch effect is attenuated with a 10-cm preheater, the same trend is observed for Figure 11C. This kind of broadening must
and is in agreement with prior work.2,24,25 The numbers in the parentheses represent the percent percent loss of plates at the same linear velocity. Obviously, the large temperature mismatch originating from the conventional column produces the largest plate height increase.
Figure 12. Calculated temperature mismatch (°C) vs column linear velocity (cm/s) calculated for both 2.1- (O) and 4.6-mm-i.d. (3) columns. The percent increase in plate height relative to 27 °C at 1.75 cm/s is given in parentheses. (A) 5 cm × 0.005 in. i.d. preheater length; (B) 10 cm × 0.005 in. i.d. preheater length.
be avoided and the eluent temperature be properly matched to the column temperature. Thermal Mismatch Should Not Exceed 5 °C. The magnitude of the thermal mismatch can be easily calculated using the same heat-transfer correlation used previously to calculate the length of tubing needed to heat the mobile phase to a certain temperature. Figure 12 shows the calculated temperature mismatch (°C) versus column linear velocity for a column and preheater inserted into a 60 °C stirred silicone oil bath. For the short preheater, the narrow-bore column temperature mismatch does not exceed 5 °C; for the conventional-bore column, the mismatch approaches 25 °C. This shows that a 5 °C temperature mismatch can be tolerated without significant broadening effects
CONCLUSIONS (1) An air bath may be used at low column linear velocities when the eluent is being heated 30-40 °C above ambient. However, high-temperature, high-speed liquid chromatography should not be done in a convecting air bath because the external resistance to heat transfer dominates the overall resistance to heat transfer. (2) When a narrower column is used, the eluent residence time spent in the preheater is significantly longer compared to a wider column at the same column linear velocity. This leads to a greater degree of thermal equilibration and attenuated thermal mismatch broadening. Reduced thermal mismatch broadening allows for more efficient high-temperature, high-speed analysis. (3) The length of tubing needed to heat the eluent from ambient conditions to the column temperature is much less for a narrow column (2.1-mm i.d.) than for a wide column (4.6-mm i.d.). (4) The radial retention factor gradient and not the radial viscosity gradient dominates the temperature mismatch broadening effect. ACKNOWLEDGMENT The authors acknowledge the financial support of the National Institute of Health and Mac-Mod Analytical Inc. for their contribution of the analytical columns used in this study. Received for review January 22, 2001. Accepted May 6, 2001. AC010091Y
Analytical Chemistry, Vol. 73, No. 14, July 15, 2001
3347
