Anal. Chem. 1996, 68, 124-129
Wetting of Octadecylsilylated Silica in Methanol-Water Eluents Zengbiao Li,†,‡ Sarah C. Rutan,*,† and Shannian Dong*,‡
Department of Chemistry, Virginia Commonwealth University, 1001 West Main Street, Richmond Virginia 23284-2006, and Department of Analytical Chemistry, Beijing Medical University, Beijing 100083, People’s Republic of China
The wetting of an octadecylsilylated silica in methanolwater mixtures was studied by optical transmittance, visual observations, and measurements of the retention of model compounds. The octadecylsilylated silica particles remain wetted as the methanol content is decreased from 100% (v/v) to 20% (v/v). With the methanol content in the range of 20% (v/v) to 10% (v/v), the octadecylsilylated silica particles are still wetted, but the degree of solvation of the C18 chains decreases with decreasing methanol concentration. The octadecylsilylated silica particles are not wetted when the methanol content in the mobile phase is lower than 10% (v/v). After equilibration with water, the octadecylsilylated silica particles remain nonwetted until the methanol content reaches 65% (v/v). The nonwetted phase showed significantly reduced chromatographic retention. A much longer equilibration time is required when the octadecylsilylated silica particles are not well solvated or nonwetted. The retentive behavior of the column will depend on the history of mobile-phase exposure. In reversed-phase liquid chromatography, a stationary phase is composed of not only the bonded alkyl chains but also the sorbed mobile-phase solvent molecules and residual silanols. Stationary-phase composition and properties will therefore change with the eluent composition. Wetting of the stationary phase is strongly dependent on the eluent composition.1 It may affect column efficiency,2 dead volume,3 equilibration time,4,5 interactions between residual silanols and basic solutes,4 adsorption of pairing agents in ion pair chromatography,6 and solvent migration rate in reversed-phase thin-layer chromatography.7 Abnormal Chromatographic Behavior. In water-rich eluents, abnormal chromatographic behavior is often observed.5,8,9-14 A very long time is required to bring a column to equilibrium if water alone is used as the eluent.5,8,12 With water as the eluent, †
Virginia Commonwealth University. Beijing Medical University. (1) Scott, R. P. W.; Kucera, P. J. Chromatogr. 1977, 142, 213-232. (2) Foley, J. P.; May, W. E. Anal. Chem. 1987, 59, 110-115. (3) Gutnikov, G.; Hung, L.-G. Chromatographia 1984, 19, 260-265. (4) Engelhardt, H.; Dreyer, B.; Schmidt, H. Chromatographia 1982, 16, 1117. (5) Scott, R. P. W.; Simpson, C. F. J. Chromatogr. 1980, 197, 11-20. (6) Dreux, M.; Lafosse, M.; Agbo-Hazoume, P. Chromatographia 1984, 18, 1517. (7) Guiochon, G.; Ko ¨ro¨si, G.; Siouffi, A. J. Chromatogr. Sci. 1980, 18, 324329. (8) Engelhardt, H.; Mathes, D. J. Chromatogr. 1977, 142, 311-320. (9) Melander, W. R.; Horva´th, Cs. Chromatographia 1984, 18, 353-361.
a plot of column equilibration time versus C18 bonding density shows a maximum at bonding density of about 2.9 µmol/m2.12 The addition of 3% 1-propanol in water significantly reduces column equilibration time.12 The sorption capacity of a C18 column with water as the eluent decreases with time, but can be restored by washing with methanol.13 Conventional retention models cannot predict retention in water-rich eluents.9,10 In mobile phases with less than 10% methanol or acetonitrile, the increase of retention of nonpolar compounds with decreasing organic modifier content in the mobile phase is much less than expected.14 Retention may even decrease with decreasing organic modifier content in water-rich eluents.5,10,11 If changes in the stationary phase with changing eluent are considered in the retention model, the prediction is much better.10,15 In addition, column efficiency decreases in water-rich eluents.2 All of these phenomena are presumed to reflect the state of wetting of the stationary phase. When the organic modifier content in the eluent is lower than a certain value, preequilibration of the column with organic modifier leads to a longer retention time than preequilibration with water.4 This indicates that the stationary phase may exist in different states even when in contact with the same eluent. Gilpin et al. observed that, with pure water as the eluent, a bonded phase may exist in a different state after heating.16-18 Even the rate of cooling after heating may affect the properties of bonded phases.19 Wetting. The wetting of a solid by a liquid, involving the formation of a solid-liquid interface, is usually described by the contact angle θ. The contact angle can be related to surface tensions by Young’s equation,
γSV ) γSL + γLV cos θ
(1)
where γSV, γSL, and γLV are the surface tensions of the solidvapor interface, the solid-liquid interface, and the liquid-vapor interface, respectively. On the molecular level, the wetting of octadecylsilylated silica is influenced by chain-chain interactions, chain-eluent interactions, and residual silanol-eluent interac-
‡
124 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
(10) Schoenmakers, P. J.; Billiet, H. A. H.; de Galan, L. J. Chromatogr. 1983, 282, 107-121. (11) Gilpin, R. K.; Gangoda, M. E. J. Chromatogr. Sci. 1983, 21, 352-361. (12) Cole, L. A.; Dorsey, J. G. Anal. Chem. 1990, 62, 16-21. (13) Jandera, P.; Kuba´t, J. J. Chromatogr. 1990, 500, 281-299. (14) Hsieh, M.-M.; Dorsey, J. G. J. Chromatogr. 1993, 631, 63-78. (15) Martire, D. E.; Boehm, R. E. J. Phys. Chem. 1983, 87, 1045-1062. (16) Gilpin, R. K.; Squires, J. A. J. Chromatogr. Sci. 1981, 19, 195-199. (17) Gilpin, R. K.; Gangoda, M. E.; Krishen, A. E. J. Chromatogr. Sci. 1982, 20, 345-348. (18) Yang, S. S.; Gilpin, R. K. J. Chromatogr. 1987, 394, 295-303. (19) Hammers, W. E.; Verschoor, P. B. A. J. Chromatogr. 1983, 282, 41-58. 0003-2700/96/0368-0124$12.00/0
© 1995 American Chemical Society
tions.18,20 However, the macroscopic wetting phenomenon, i.e., the contact angle, has not been well correlated with interactions on the molecular level or the nanoscopic structure at the mobilephase/stationary-phase boundary.21 On the molecular scale, it is more appropriate to refer to the boundary as an interphase region,22,23 which has a finite width and is composed of the bonded alkyl chains and sorbed mobile-phase solvent molecules. The composition and width of the interphase region will depend on the mobile-phase composition. The interphase region may not be homogeneous. Therefore, a more accurate description of the boundary may be a relatively gradual concentration (and therefore energy) gradient rather than an abrupt discontinuity at the stationary-phase/mobile-phase interface. Wetting Studies for Alkyl Bonded Phases. Scott and Kucera titrated a bonded-phase suspension in an organic solvent to study what they termed wettability.1 About 50% water is needed to bring a persistent film of the alkyl bonded-phase particles to the surface of the suspension if methanol is used as the initial solvent.1 Engelhardt and Mathes8 and Welsch et al.20 tested wettability by titrating a bonded-phase suspension in water. For the bonded phases used extensively in reversed-phase liquid chromatography, about 60% (v/v) methanol is required to wet the bonded phases in this suspension.8,20 Less organic modifier is required to wet a bonded phase with a lower surface coverage of alkyl groups.1,20 The length of the bonded alkyl chains also affects wettability.20 Riedo et al.24 studied the wetting properties of bonded phases by measuring the capillary rise of various organic solvents in glass capillaries with alkyl bonded inner surfaces. From measurements of capillary rise in glass capillaries treated with octadecyldimethylsilanol, Guiochon et al.7 obtained contact angles for acetonitrile, methanol, ethanol, and aqueous mixtures of methanol and ethanol with less than 50% (v/v) of water. The contact angles for these liquids are in the range from 15° to 70°. Contact angles for aqueous mixtures of methanol and ethanol on bonded alkyl phases were also calculated from solvent migration rates on reversedphase thin-layer chromatographic plates.7 For example, the contact angles for mixtures with 30% (v/v) water in methanol and 50% (v/v) water in ethanol are around 80°. Recently, Montgomery et al. published a series of articles concerning the wetting of octadecyl-derivatized silica plates by measurements of contact angles and fluorescence depolarization.25-28 The plates are not wetted by pure water and show a contact angle of 93°. Aqueous eluents with 20% methanol or 5% 1-propanol have contact angles of 65° and 69°, respectively.25 The effects of the sorbed short-chain alcohols,25 long-chain n-alcohols,28 and sodium dodecyl sulfate26,27 on the orientation and ordering of the bonded octadecyl chains were probed using a hydrophobic fluorophor, p-bis(o-methyl)styrylbenzene. (20) Welsch, T.; Frank, H.; Vigh, Gy. J. Chromatogr. 1990, 506, 97-108. (21) Heslot, F.; Fraysse, N.; Cazabat, A. M.; Levinson, P.; Carles, P. In Wetting Phenomena; De Coninck, J., Dunlop, F., Eds.; Springer-Verlag: Berlin, 1990; pp 41-48. (22) Jaycock, M. J.; Parfitt, G. D. Chemistry of Interfaces; Ellis Horwood Ltd.: Chichester, U.K., 1981; Chapters 1 and 5. (23) Dill, K. A. J. Phys. Chem. 1987, 91, 1980-1988. (24) Riedo, F.; Czencz, M.; Liardon, O.; Kova´ts, E. Sz. Helv. Chim. Acta 1978, 61, 1912-1941. (25) Montgomery, M. E., Jr.; Green, M. A.; Wirth, M. J. Anal. Chem. 1992, 64, 1170-1175. (26) Montgomery, M. E., Jr.; Wirth, M. J. Anal. Chem. 1992, 64, 2566-2569. (27) Montgomery, M. E., Jr.; Wirth, M. J. Langmuir 1994, 10, 861-869. (28) Montgomery, M. E., Jr.; Wirth, M. J. Anal. Chem. 1994, 66, 680-684.
While most of the abnormal chromatographic phenomena cited above appear in water-rich eluents or in water, wetting studies concerning contact angles were conducted in eluents with relatively high concentrations of organic modifier, except for the experiments performed by Montgomery et al. mentioned above, which utilized high water content solvent mixtures.25 Extremely slow equilibration and low reproducibility may be the reason for the lack of wetting studies in water-rich eluents7. In most cases, alkyl-derivatized silica plates and capillary inner surfaces, instead of octadecylsilylated silica particles, have been used in wetting studies. Though the wetting of bonded alkyl phases has been studied for a long time, a clear and complete picture of wetting under chromatographic conditions has yet to appear. None of the methods that have been applied to the study of the wetting of planar surfaces can be used to study porous particles. In our preliminary experiments, it was found that the bonded octadecyl phase in eluents with different compositions showed different degrees of transparency. Therefore, optical transmittance of light through bonded octadecyl particles packed in a flow cell is used here to study the wetting of the stationary phase. As the composition of the interphase region changes, the refractive index will also change, leading to a change in the optical transmittance of the system. In this paper, the wetting of a bonded octadecyl phase, HDG C18H37, is studied by the retention of model compounds, visual observations, and optical transmittance. The transitions between different wetting states and the effects of wetting on column equilibration time are also investigated. EXPERIMENTAL SECTION Materials and Chemicals. LiChrosorb SI 100 (30 µm) was purchased from E. Merck (Darmstadt, Germany). HDG C18H37 [125-150 µm, specific surface area of 250-400 m2/g, and 24.09% C (w/w)] was a gift from Tianjin No. 2 Chemical Manufacturer (Tianjin, China). Methanol of spectrophotometric grade was from EM Science (Gibbstown, NJ). Retention Measurements. An SYB-J peristaltic pump (Beijing Qingyun Equipment Corp., Beijing, China) and a UV-120-02 UVvisible spectrophotometer (Shimadzu, Japan) with an 8-µL flow cell as a detector were used in retention measurements. 0.25 g of HDG C18H37 was packed into a 50 × 6.5 mm glass column. The column was rinsed with methanol before every experiment. After equilibration with an eluent, 1 mL of a 10 µg/mL solution of a model compound in the eluent was pumped through the column at a flow rate of 0.44 mL/min. This procedure was originally designed for studying solidphase extraction and is different from regular chromatographic practice, but this approach is convenient and also allows the effect of wetting on chromatographic retention to be investigated. Measurements of Optical Transmittance. All optical transmittance measurements were made on a Shimadzu UV-265 spectrophotometer equipped with an integrating sphere attachment. A flow cell with a path length of 1 mm was packed with about 0.05 g of the stationary phase of interest. After the stationary phase was equilibrated with a mobile phase, the transmittance of the stationary-phase/mobile-phase system at 550 nm was measured. The transmitted light is highly diffuse, so an integrating sphere attachment was used for these measurements. Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
125
Figure 1. Plot of the transmittance of HDG C18H37 versus the composition of a methanol-water eluent. The arrows indicate the direction of equilibration. The solid and dashed curves were obtained from downward and upward equilibration experiments, respectively. The solid curve is separated into five regions, labeled with Arabic numbers, while Roman numerals are used to label the four regions of the dashed curve.
RESULTS AND DISCUSSION Abnormal Retention Behavior on HDG C18H37 in Methanol-Water Eluents. The retention of barbital increases with a decrease of the methanol content in the mobile phase, φ, if φ is larger than 7%. However, a further decrease in the methanol content results in a decrease in retention. Similar results were observed for caffeine and sulfamethoxazole with retention maxima occurring at methanol contents of 7% and 9%, respectively. Although structurally simpler solutes could have been selected, for example, benzene, the retention in high water content mobile phases is prohibitively long, even for the short column used here. Visual Observations of the Wetting of HDG C18H37. The HDG C18H37 material was packed in a short glass column. After the column was rinsed with methanol, 10 mL of each of the eluents (specified below), each with a decreasing volume ratio of methanol to water, was passed through the column successively. After each passage, the appearance of the HDG C18H37 material was noted. When eluents with volume ratios of 80:20, 70:30, 30:70, 20:80, and 10:90 (methanol-water) were passed through the column, the particles had a translucent appearance. This was taken as evidence that the particles were still wetted by these eluents. When the eluent with a volume ratio of 7:93 was passed through the column, the particles had an intermediate appearance between translucent and opaque white. When the eluent with a volume ratio of 4:96 was passed through the column, the particles had an opaque white appearance. This was taken as evidence that the particles were not wetted by the eluent. When the eluent with a volume ratio of 1:99 or pure water was passed through the column, the particles also had an opaque white appearance. Wetting and Wetting Hysteresis Observed by Optical Transmittance Measurements. A plot of the transmittance of HDG C18H37 versus the composition of the methanol-water eluent is shown in Figure 1. The stationary phase was first equilibrated with methanol. Then it was equilibrated with mobile phases containing decreasing amounts of methanol, specified as downward equilibration. The percent transmittance was measured once the value for transmittance had stabilized after changing to a new mobile phase. The data from downward equilibration are shown 126 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
by the solid curve in Figure 1. After equilibration with water, the stationary phase was equilibrated with mobile phases containing increasing amounts of methanol, and this process is called upward equilibration. The dashed curve in Figure 1 represents the data from the upward equilibration experiment. The transmittance of a bare silica, LiChrosorb SI 100, in methanol-water eluents with different methanol contents was measured to determine the change in transmittance induced solely by the change in the refractive index of the mobile phase. For bare silica, the transmittance varies by only 4% over the entire composition range, with a maximum in transmittance at a methanol content of 55% (v/v). As the methanol content decreases, the curve from downward equilibration shows five distinct regions, as shown in Figure 1. In region 1, as φ changes from 100% to 55%, the transmittance value for HDG C18H37 increases gradually. This is followed by a slight decrease over region 2, as φ changes from 55% to 20%. The changes in transmittance for HDG C18H37 in these two regions resemble those observed for bare silica. It can be concluded that there is no dramatic change in the wetting of the bonded alkyl layer over these mobile-phase compositions for downward equilibration. In region 3, as φ decreases from 20% to 10%, the transmittance of HDG C18H37 increases, which is not consistent with changes observed in the refractive index of the mobile phase. The magnitude of this increase is small. The transmittance value decreases by a large amount in region 4, as φ decreases from 10% to 6%, followed by a pronounced increase in region 5, as φ decreases further from 6% to 0%. These dramatic changes in the transmittance value cannot be caused by changes in the refractive index of the mobile phase. They must be caused by significant changes in the interphase region. By combining the results from the retention measurements, visual observations, and optical transmittance measurements, it is concluded that HDG C18H37 is not wetted in regions 4 and 5. The bonded alkyl chains collapse on the surface of the silica support25,29 to form a hydrocarbon film. This significantly reduces the volume of the stationary phase and the retention of solutes. It promotes an adsorption rather than a partitioning retention mechanism. The mobile-phase composition separating regions 3 and 4, 10%, is defined as the nonwetting limit. In regions 1, 2, and 3, with the methanol content in the mobile phase larger than the nonwetting limit, the stationary phase is considered to be wetted, but in region 3 the degree of solvation of C18 chains may decrease as the methanol content in the mobile phase is reduced. Literature studies show that when φ in a methanol-water eluent is between 30% and 80%30 or between 50% and 90%,31 the composition of the absorbed solvent layer on octadecylsilylated silica does not change significantly with the eluent composition. The amount of absorbed methanol is equivalent to a monolayer.30,31 In region 2, however, the amount of sorbed methanol increases with φ.30 With increasing φ, the curve from the upward equilibration experiment can be divided into four regions, labeled I-IV, as shown in Figure 1. In region I, as φ increases from 0% to 50%, the transmittance value for the HDG C18H37 phase decreases gradually. There is no dramatic change in the refractive index of (29) Lochmu ¨ ller, C. H.; Hunnicutt, M. L. J. Phys. Chem. 1986, 90, 4318-4322. (30) Slaats, E. H.; Markovski, W.; Fekete, J.; Poppe, H. J. Chromatogr. 1981, 207, 299-323. (31) Hammers, W. E.; Meurs, G. J.; de Ligny, C. L. J. Chromatogr. 1982, 246, 169-189.
the interphase region. We interpret this behavior to mean that the HDG C18H37 particles remain nonwetted. There is an abrupt drop in the transmittance value over region II, as φ increases from 50% to 60%, followed by a sharp rise in transmittance in region III, when φ increases from 60% to 65%. In region IV, as φ increases from 65% to 100%, the curve overlaps with the one observed during the downward equilibration experiment, providing evidence that the HDG C18H37 phase is wetted under these conditions. The mobile phase composition separating regions III and IV, 65%, is defined as the rewetting limit. There are two different procedures for the wettability tests based on titration methods described in the literature (see introduction). In one of these procedures, a bonded-phase suspension in an organic solvent is titrated with water.1 This is analogous to the downward equilibration experiment, but the results from this wettability test are more consistent with the results from our upward equilibration experiment. For the other wettability test, a bonded-phase suspension in water is titrated with an organic solvent.8,20 This approach is analogous to our upward equilibration experiments. The wettability obtained in such a test, around 60% methanol for common C18 phases,8,20 is in good agreement with the rewetting limit described here. In Figure 1, when the methanol content in the mobile phase is lower than 65%, a hysteresis in the transmittance of the HDG C18H37 phase is observed. Different transmittance values for the stationary phase are observed with the same mobile phase, depending on the prior conditioning of the stationary phase. With the same mobile phase, the difference in transmittance can only be caused by a difference in the condition of the stationary phase, i.e., a difference in the wetting of the stationary phase. Engelhardt et al.4 observed that, in an eluent with less than 40% methanol, the dead volume and the retention of 1-butanol on octadecylsilylated silica depended on prior conditioning of the column. A larger retention and dead volume were obtained when the column was conditioned with methanol than with water. But no further studies were conducted, and these phenomena were not completely explained. On the basis of the results of our experiments, it can be deduced that 40% is the rewetting limit of their octadecylsilylated silica and their octadecylsilylated silica was not wetted by pure water. Although the relationship between optical transmittance and the wetting of the stationary phase cannot be precisely determined, the optical transmittance is clearly very sensitive to the change in the wetting of the stationary phase, due to changes in the refractive index of the interphase region. Monitoring of the Equilibration Process by Optical Transmittance. Until now, the equilibration process has only been monitored indirectly by measuring the retention of model compounds. When a stationary phase reaches equilibrium, all its physical and chemical properties, including solute retention and optical transmittance, should become constant. In this work, we show that the equilibration of the stationary phase can also be evaluated by measuring its optical transmittance. When the mobile phase was changed from methanol to a methanol-water mixture, the equilibration process was monitored by recording the transmittance value of the HDG C18H37 phase as a function of time. The equilibration processes for the methanol-water mixtures in regions 2, 3, and 4, with 30%, 15%, and 7% methanol, respectively, are plotted in Figure 2. The flow rate is 0.25 mL/ min for all the measurements described in this section.
Figure 2. Plot of the transmittance of HDG C18H37 versus time. The mobile phase was changed from methanol to a methanol-water eluent with 30%, 15%, and 7% methanol. Arabic numbers are used to label different regions of the equilibration curves, following the pattern for the downward equilibration curve shown in Figure 1.
Figure 3. Dependence of the equilibration time required for HDG C18H37, from methanol to a methanol-water eluent, on the composition of the eluent. The flat part of the curve is plotted in the inset with a different ordinate scale.
The equilibration curves in Figure 2 follow the same trend observed in Figure 1 for downward equilibration. The different regions defined for the downward equilibration curve in Figure 1 can also be found in the corresponding equilibration curves in Figure 2. This implies that the same sequence of events resulting in changes in the wetting of the stationary phase is followed by the two processes. To show the details of the curves, the plots in Figure 2 present the equilibration processes for the first 60 min. The complete equilibration process may take a much longer time. The time required to bring the HDG C18H37 phase to a constant transmittance value, or equilibrium, from methanol to a methanol-water mixture, is plotted versus the mobile-phase composition in Figure 3. From the inset in Figure 3, it can be seen that it took about 1 min to equilibrate the HDG C18H37 phase when the mobile phase was changed from methanol to an eluent in region 1. For eluents in region 2, the equilibration time increases with decreasing methanol content, but the time required for equilibration is still less than 5 min in this region. In regions 3, 4, and 5, with φ less than 20%, the equilibration time increases rapidly with decreasing methanol content. From the dramatic increase in the equilibration time from region 2 to region 3, it can be concluded that the solvation or wetting of the HDG C18H37 phase is significantly different in these two regions. Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
127
Figure 4. Plot of the transmittance of HDG C18H37 versus time. The mobile phase was changed from methanol to water, at a flow rate of 0.25 mL/min. After 4.35 mL of water was passed through the flow cell, the flow was stopped.
Figure 5. Dependence of the final constant transmittance on the volume of water passed through the flow cell before the flow was stopped.
Effect of Flow and Temperature on the Equilibration Process as the Eluent Is Switched from Methanol to Water. As shown in Figure 3, it takes a very long time to equilibrate the HDG C18H37 phase with water as the mobile phase. During the equilibration process, methanol is replaced by water, and in addition, we propose that the C18 chains on the silica surface are reorganizing to form a more energetically favorable configuration. We questioned whether these two processes occur over the same time frame. If not, how much longer does it take the C18 chains to reorganize after the methanol is replaced with water? To answer these questions, the following experiments were carried out. For the measurement of the equilibration time from methanol to water in Figure 3, the mobile phase was kept at a flow rate of 0.25 mL/min during the whole equilibration process. If the flow of water is stopped before the stationary phase reaches equilibrium, the optical transmittance of the HDG C18H37 particles keeps changing, as shown in Figure 4. Monitoring of the transmittance is continued until a constant value is finally reached. The dependence of the final constant transmittance on the volume of water passed through the cell before the flow is stopped is shown in Figure 5. The final constant transmittance does not change as long as the volume of water passed through the cell is larger than 4.35 mL. The final constant transmittance at this point, 66.2%, is within the experimental error of the constant transmittance when 128 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
Figure 6. Effect of temperature on the equilibration process when the eluent was switched from methanol to water. The mobile phase flow rate was 0.25 mL/min.
the flow is not stopped, which is 66.0 ( 0.2%. It can be concluded that methanol is completely replaced by water after 4.35 mL of water is passed through the cell. Mobile-phase replacement is a relatively fast process. We propose that the slow change in the transmittance after mobile-phase replacement is caused by C18 chain reorganization. The equilibration process, from methanol to water, was monitored at both 23 and 48 °C. The equilibration curves for the first 100 min are shown in Figure 6. The mobile phase was kept at a flow rate of 0.25 mL/min for both equilibration processes in Figure 6. The difference between the final constant transmittance values at the two temperatures is small. We interpret this to mean that the difference in the final configuration of the chains of the HDG C18H37 phase at the two temperatures is not significant. As shown in Figure 6, it takes much less time to reach equilibrium at 48 °C than at 23 °C. One possible explanation is that chain reorganization is a slow process and requires a significant activation energy. An increase in temperature can significantly accelerate this chain reorganization process. Hydrocarbon films formed by association among bonded alkyl chains in contact with water can be transformed to a different state after heating.16-19 Gilpin et al.16-18 believed that bonded alkyl chains rearrange to an extended state upon heating and remain in such a configuration after cooling. A different surface model was proposed by Hammers and Verschoor19 with convincing experimental evidence. They believed that a rough hydrocarbon film is formed because the degree of ordering among bonded alkyl chains is low. After heating, the bonded alkyl chains rearrange to form an extended pattern. During cooling, a smoother hydrocarbon film is formed because of the more ordered associations among the bonded alkyl chains. Retention times before heating were larger than the ones obtained after heating.16,19 Retention times on a column with a higher cooling rate after heating were larger than the ones on a column with a lower cooling rate.19 Two elution peaks were obtained for one solute if the column was rapidly cooled while a solute was on the column.19 All of this evidence strongly favors the latter model. In our experiment, no evidence for differences in chain organization at room temperature, 23 °C, and elevated temperature, 48 °C, was observed. A temperature higher than 48 °C may be required to rearrange C18 chains to an extended state.
Transition between Different Wetting States. From the above observations of the wetting of the HDG C18H37 phase, it can be concluded that this phase can be changed from a more wetted state to a less wetted state simply by equilibrating the HDG C18H37 phase with the eluent corresponding to the less wetted state. However, it may not always be changed from a less wetted state to a more wetted state simply by equilibrating the phase with the eluent corresponding to the more wetted state because of the hysteresis phenomenon. After equilibration with water, this phase cannot be transformed to a more wetted state with a mobile phase containing less than 50% methanol, as demonstrated by the upward equilibration curve shown in Figure 1. The transition between different wetting states was studied further by additional retention measurements, as described below. The states corresponding to eluent compositions of 1% and 3% methanol were taken as the less wetted state and the more wetted state, respectively. After rinsing with methanol, the column was equilibrated with the eluent containing 1% methanol. The column was then equilibrated with a state transition eluent, containing a specified percentage of methanol. The column was finally equilibrated with the eluent containing 3% methanol. The retention of barbital in this eluent was then determined. The retention of barbital in the eluent containing 3% methanol after exposure to a state transition eluent that cannot convert the HDG C18H37 phase from the less wetted state to the more wetted state is expected to be smaller than for a state transition eluent able to achieve the conversion. The retention of barbital in the eluent containing 3% methanol versus the methanol content in the state transition eluent is plotted in Figure 7. From Figure 7 it can be concluded that only eluents with more than 65% methanol, which corresponds to the rewetting limit in Figure 1, can completely convert the HDG C18H37 phase from the less wetted state to the more wetted state. Several experimental variables may affect the specifics of the wetting process as discussed in this paper. These include the pressure of the system and the nature of the stationary phase. Liquid chromatographic columns are typically exposed to substantially higher pressures than those that we can apply in our spectrophotometric flow cell. Column pressure may affect the wetting of the stationary phase in two ways, by affecting the contact angle and the penetration of the mobile phase into the pores of the stationary phase. The effect of pressure on the contact angle can be obtained from the dependence of the surface tension on pressure. Ideally, the change of the surface tension with pressure is proportional to the change in molar volume when a molecule goes from the bulk to the interface region.32 However, this simple relationship is complicated by adsorption on the (32) Adamson, A. W. Physical Chemistry of Surfaces, 4th ed.; Wiley-Interscience: New York, 1982; p 56.
Figure 7. Plot of the retention of barbital in a methanol-water mobile phase containing 3% (v/v) methanol versus the methanol content of the state transition eluent (see text for explanation).
interface and the transfer of molecules from one phase to another.32 The penetration of the mobile phase into the pores of the stationary phase is also affected by pressure through capillary action, caused by the pressure difference across the curved surface of a meniscus, ∆P, which can be estimated from the surface tension of the eluent, γLV, the contact angle, θ, and the radius of the pore, r,
∆P ) 2γLV cos θ/r
(2)
Additionally, different octadecylsilylated silicas have different degrees of surface coverage with C18 groups; the availability of silanols on the silica surface undoubtedly plays an important role in the wetting characteristics of a particular stationary phase. However, the general trends reported in the literature, and observed in these studies, now supported by optical transmittance measurements, support the same general model of the wetting process. ACKNOWLEDGMENT The authors acknowledge the National Science Foundation (Grant CHE-9318484) for support of this research and Peter W. Carr, University of Minnesota, for his valuable comments during the preparation of the manuscript. Received for review July 6, 1995. Accepted October 16, 1995.X AC950666P X
Abstract published in Advance ACS Abstracts, November 15, 1995.
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
129
