Anal. Chem. 2009, 81, 7632–7638
Confocal Raman Microscopy Investigation of the Wetting of Reversed-Phase Liquid Chromatographic Stationary Phase Particles Jennifer L. Gasser-Ramirez and Joel M. Harris* Department of Chemistry, University of Utah, 315 South 1400 East Salt Lake City, Utah 84112-0850 Wetting phenomena in reversed-phase liquid chromatographic (RPLC) stationary phase particles strongly influence the retention of solutes with high water-content mobile phases. To better understand these phenomena, this work reports the spectroscopic observation of the wetting of the interior of individual C18-silica stationary phase particles with acetonitrile-water mobile phase solutions by confocal Raman microscopy. It was found that the pores of dry C18 silica do not wet when the concentration of acetonitrile is below 0.12 mol fraction (28% by volume). It was also found that there is a wetting hysteresis, where particles that had been previously exposed to acetonitrile solutions above the wetting transition remain filled with solution even at much lower concentrations of acetonitrile in the surrounding solution. Contact angles of acetonitrile-water solutions were measured at a planar C18-modified silica surface and used to predict the capillary wetting of the particles based on the Young-Laplace equation. The solution composition at the wetting transition detected by Raman microscopy is higher in acetonitrile concentration than predicted by the Young-Laplace equation, which may be due to the presence of a vapor or air gap at the interface between the hydrophobic pores and aqueous solution. Further evidence of this behavior is found in water porosimetry results, which show wetting pressures ∼5 times greater than predicted by the Young-Laplace equation and are consistent with only 50% of a water interface being in contact with the C18 surface. This fraction increases to 80% at an acetonitrile concentration of 0.12 mol fraction, leading to spontaneous and irreversible wetting of the hydrophobic pores. Reversed-phase chromatographic separations using highly aqueous mobile phase compositions are difficult to model, where abnormal and irreproducible retention behavior is often observed.1 These aqueous mobile phase solutions often do not reliably wet the pores of n-alkane chain functionalized stationary phase unless the RPLC column is pressurized, forcing the mobile phase into the pores of the stationary phase.1-3 The difficulty of wetting the hydrophobic RPLC stationary phase can lead to peak tailing, * Corresponding author. E-mail:
[email protected]. (1) Li, Z.; Rutan, S. C.; Dong, S. Anal. Chem. 1996, 68, 124–129. (2) Walter, T.; Iraneta, P.; Capparella, M. Waters Corporation, 1997.
7632
Analytical Chemistry, Vol. 81, No. 18, September 15, 2009
retention loss, and baseline instability, resulting in poor reproducibility of separations.1-3 These problems can be addressed and avoided by understanding the wetting interactions between the hydrophobic stationary phase and the aqueous mobile phase. Wetting of RPLC stationary phases has been studied using optical transmittance,1 water porosimetry,4,5 chromatographic retention,2,3 and fluorescence probe methods.6-9 Molecular dynamics and Monte Carlo simulations have also been used to better understand the structure of water in nanoscale, hydrophobic pores.10-12 Vibrational spectroscopy provides direct, in situ analytical methods, which do not require the addition of probe molecules. These methods are well suited to the investigation of the wetting of the RPLC stationary phase particles with mobile phase solutions. Infrared13-15 and Raman16-28 spectroscopy have been used previously to characterize the conformation of n-alkane chain(3) Walter, T. H.; Iraneta, P.; Capparella, M. J. Chromatogr., A 2005, 1075, 177–183. (4) Fadeev, A. Y.; Eroshenko, V. A. J. Colloid Interface Sci. 1997, 187, 275– 282. (5) Helmy, R.; Kazakevich, Y.; Ni, C.; Fadeev, A. Y. J. Am. Chem. Soc. 2005, 127, 12446–12447. (6) Okamoto, K.; Shook, C. J.; Bivona, L.; Lee, S. B.; English, D. S. Nano Lett. 2004, 4, 233–239. (7) Jayaraman, K.; Okamoto, K.; Son, S. J.; Luckett, C.; Gopalani, A. H.; Lee, S. B.; English, D. S. J. Am. Chem. Soc. 2005, 127, 17385–17392. (8) Zhong, Z.; Geng, M. L. Anal. Chem. 2007, 79, 6709–6717. (9) Zhong, Z. Doctoral Dissertation, University of Iowa, Iowa City, IA, 2007. (10) Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999, 103, 4570– 4577. (11) Giaya, A.; Thompson, R. W. J. Chem. Phys. 2002, 117, 3464–3475. (12) Luzar, A. J. Phys. Chem. B 2004, 108, 19859–19866. (13) Srinivasan, G.; Kyrlidis, A.; McNeff, C.; Muller, K. J. Chromatogr., A 2005, 1081, 132–139. (14) Sander, L. C.; Callis, J. B.; Field, L. R. Anal. Chem. 1983, 55, 1068–1075. (15) Srinivasan, G.; Neumann-Singh, S.; Muller, K. J. Chromatogr., A 2005, 1074, 31–41. (16) Ho, M.; Cai, M.; Pemberton, J. E. Anal. Chem. 1997, 69, 2613–2616. (17) Ho, M.; Pemberton, J. E. Anal. Chem. 1998, 70, 4915–4920. (18) Doyle, C. A.; Vickers, T. J.; Mann, C. K.; Dorsey, J. G. J. Chromatogr., A 2000, 877, 25–39. (19) Doyle, C. A.; Vickers, T. J.; Mann, C. K.; Dorsey, J. G. J. Chromatogr., A 2000, 877, 41–59. (20) Pemberton, J. E.; Ho, M.; Orendorff, C. J.; Ducey, M. W. J. Chromatogr., A 2001, 913, 243–252. (21) Ducey, M. W., Jr.; Orendorff, C. J.; Pemberton, J. E.; Sander, L. C. Anal. Chem. 2002, 74, 5576–5584. (22) Ducey, M. W., Jr.; Orendorff, C. J.; Pemberton, J. E.; Sander, L. C. Anal. Chem. 2002, 74, 5585–5592. (23) Orendorff, C. J.; Ducey, M. W., Jr.; Pemberton, J. E. J. Phys. Chem. A 2002, 106, 6991–6998. (24) Orendorff, C. J.; Ducey, M. W., Jr.; Pemberton, J. E.; Sander, L. C. Anal. Chem. 2003, 75, 3360–3368. (25) Orendorff, C. J.; Ducey, M. W., Jr.; Pemberton, J. E.; Sander, L. C. Anal. Chem. 2003, 75, 3369–3375. 10.1021/ac901037d CCC: $40.75 2009 American Chemical Society Published on Web 08/25/2009
functionalized stationary phases. Raman spectroscopy is better suited than infrared spectroscopy for in situ characterization of these materials because the silica support and surrounding water do not give rise to strong Raman scattering, thus making possible the study of the interface chemistry of actual stationary phase particles. Although Raman scattering is a weak effect, interfacial analysis of chromatographic supports is possible because these high surface area materials yield strong signals from the large fraction of molecules at the solid-liquid interface. Confocal Raman spectroscopy has recently been adapted to examine the composition and structure of organic modifiers at the C18-silica/solution interface.29 A confocal Raman microscope with a high numerical aperture objective offers the ability to analyze small sample volumes ( 90°, high pressure is required to drive the mobile phase liquid into the hydrophobic pores.2,3 In order to examine the utility of the Young-Laplace equation in predicting the wetting behavior of the RPLC stationary phase, we measured contact angles for a series of solutions of acetonitrile and water ranging from 0 to 1.0 mol fraction of acetonitrile on monomeric C18-modified and TMCS end-capped fused silica surfaces, and the results are plotted in Figure 3. As expected, the contact angle between the solution and the fused silica surface decreases as the concentration of acetonitrile in the solution increases, because the surface tension of the solution decreases and acetonitrile solvates the n-alkane chains, reducing the Gibb’s free energy at the interface. These data were used to predict the required pressure of wetting of 4 (±0.5) nm radius (determined from mercury intrusion porosimetry) hydrophobic pores of the RPLC stationary phase using the Young-Laplace equation. Values used for the solution surface tension were tabulated in the CRC Handbook of Chemistry and Physics35 and were corrected for nonideal mixing.36 The predicted pressures for pore wetting are also shown in Figure 3. When the contact angle is less than 90°, the Young-Laplace equation predicts that the porous stationary phase will spontaneously fill with mobile phase solution. This condition is met for a planar surface when the concentration of acetonitrile is greater than approximately 0.036 mol fraction or 10% by volume. The Raman spectroscopy results obtained for the pore-wetting transition, however, show that wetting of the pores occurs when the concentration of acetonitrile is approximately 0.12 mol fraction, nearly 4 times greater than the concentration predicted by the Young-Laplace equation. These results suggest that the Young-Laplace equation does not accurately predict the (35) In CRC Handbook of Chemistry and Physics, Internet version; Lide, D. R., Ed.; Taylor and Francis: Boca Raton, FL, 2007. (36) Katz, E. D.; Ogan, K.; Scott, R. P. W. J. Chromatogr. 1986, 352, 67–90.
7636
Analytical Chemistry, Vol. 81, No. 18, September 15, 2009
Figure 4. Water porosimetry data for C18 - modified porous silica showing two wetting transitions. At a low pressure of 63 kPa (9 psi), the volume between the particles is filled. At much higher pressure of ∼20 MPa, the internal pores of the RPLC material are filled. To quantify this transition, the data between 12 and 30 MPa were fit to a sigmoid function to the wetting pressure, ∆P ) 20.1 ( 0.1 MPa (2920 ( 10 psi); see inset.
wetting behavior of the nanometer pores of the RPLC stationary phase based on contact angles measured on a planar surface. Other researchers5,7 have also found evidence that the Young-Laplace equation is not applicable by extrapolation of behavior on planar surfaces to the nanoscale pores of the RPLC media. Helmy and co-workers5 determined using water porosimetry that the Young-Laplace equation underestimates the pressure required to force water into nanoscale hydrophobic pores, where the wetting pressure is several times greater than that predicted by the Young-Laplace equation. The validity of the Young-Laplace equation for nanoscale pores has also been tested using fluorescent probe molecules.6,7,9 Okamoto et al.6 developed a method to study the wetting of C18-modified silica nanotubes using fluorescence microscopy to observe the mobility of the fluorophor DiIC18. Jayaraman et al.7 applied this method and showed that 30 nm, C18-functionalized silica nanotubes do not fill with aqueous solution at the concentrations of methanol predicted by the Young-Laplace equation. Instead, they found that the concentration of methanol required to wet the pores was 0.5 mol fraction, 10 times greater than that predicted by the Young-Laplace equation. To test the wetting pressure of the particular RPLC stationary phase used in this study, water intrusion porosimetry data were collected on a sample of the C18-silica particles and results are plotted in Figure 4. The data show that there are two pressures at which the wetted volume increases, corresponding to two different populations of pores. The increase in volume that occurs at around 63 kPa (9 psi) corresponds to the wetting and filling of the volume between the particles. At much higher pressures, there is another increase in wetted volume, corresponding to filling of the internal pores of the C18-silica material. To quantify this transition, the data points between 12 and 30 MPa were fit to a sigmoid function to determine the inflection point on the
curve, which corresponds to a wetting pressure of ∆P ) 20.1 ± 0.1 MPa (2920 ± 10 psi). Note that this pressure exceeds the typical operating pressures of HPLC (10 MPa or 1500 psi), indicating that the pores of this material would not be wetted by pure water under typical HPLC operating conditions. This result is comparable to other measured wetting pressures for hydrophobic functionalized silicas4,5 but is greater than expected based on the water contact angle for a planar functionalized silica surface. For 4 nm radius pores, the water contact angle for a C18functionalized and end-capped planar silica surface (Figure 3), θ ) 97 ± 1°, and the surface tension of water, γ ) 72 mN/m, the wetting pressure predicted by eq 1, ∆P ) 4.0 MPa, is ∼5 times smaller than the measured value. Although differences in surface morphology and inhomogeneity (from residual surface silanols) between the functionalized planar substrate and the stationary phase particles could contribute to differences in wetting, a greater effect is likely the effect of small hydrophobic pores. Helmy et al.5 have suggested that the disparity between a predicted and observed wetting pressure for small hydrophobic pores can be explained by the existence of a hydrophobic gap or a low density vapor between the liquid and the hydrophobic surface. High-resolution X-ray microscopy has been employed to determine that the magnitude of the hydrophobic gap on a planar silica substrate functionalized with n-alkane chains is between 1 and 6 Å.37 These experimental findings are consistent with molecular dynamics simulations of interactions between liquid water and the hydrophobic interface for small (
