Glass Interface - American Chemical Society

Department of Chemical Engineering, The UniVersity of Texas at Austin, ... Surfactant and Colloid Group, Department of Chemistry, UniVersity of Hull, ...
2 downloads 0 Views 384KB Size
Download PDF
Langmuir 2006, 22, 2161-2170

2161

Wetting Phenomena at the CO2/Water/Glass Interface Jasper L. Dickson,† Gaurav Gupta,† Tommy S. Horozov,‡ Bernard P. Binks,‡ and Keith P. Johnston*,† Department of Chemical Engineering, The UniVersity of Texas at Austin, Austin, Texas 78712, and Surfactant and Colloid Group, Department of Chemistry, UniVersity of Hull, Hull HU6 7RX, U.K. ReceiVed October 7, 2005. In Final Form: December 16, 2005 A novel high-pressure apparatus and technique were developed to measure CO2/water/solid contact angles (θ) in situ for pressures up to 204 bar. For two glass substrates with different hydrophilicities, θ increased significantly with CO2 pressure. As the pressure was increased, an increase in the cohesive energy density of CO2 caused the substrate/ CO2 and water/CO2 interfacial tensions (γ) to decrease, whereas the water/substrate γ value increased. θ for the more hydrophobic substrate was predicted accurately from the experimental water/CO2 γ value and an interfacial model that included only long-range forces. However, for the more hydrophilic substrate, short-range specific interactions due to capping of the silanol groups by physisorbed CO2 resulted in an unusually large increase in the water/substrate γ value, which led to a much larger increase in θ than predicted by the model. A novel type of θ hysteresis was discovered in which larger θ values were observed during depressurization than during pressurization, even down to ambient pressure. Effective receding angles were observed upon pressurization, and effective advancing angles were observed upon depressurization on the basis of movement of the three-phase contact line. The greater degree of hysteresis for the more hydrophilic silica can be attributed in part to the capping of silanol groups with CO2. The large effects of CO2 on the various interfacial energies play a key role in the enhanced ability of CO2, relative to many organic solvents, to dry silica surfaces as reported previously on the basis of FTIR spectroscopy (Tripp, C. P.; Combes, J. R. Langmuir 1998, 14, 7348-7352).

Introduction CO2 is a particularly appealing green solvent because it is inexpensive, relatively nontoxic, and nonflammable and has a critical temperature of only 31 °C. Because it has a low polarizability per unit volume, solubilities are often limited, and only a small number of processes can be carried out in a homogeneous CO2-based phase. However, CO2 offers many opportunities for heterogeneous systems with CO2 at one or more interfaces.1 Advancements in the field of colloid and interface science in CO2 have led to a fundamental understanding of the stabilization of colloidal dispersions,2 including microemulsions,3-5 miniemulsions,6,7 macroemulsions,8-10 polymer latexes,11 and * Corresponding author. E-mail:[email protected]. Phone: 512-4714617. Fax: 512-475-7824. † The University of Texas at Austin. ‡ University of Hull. (1) Johnston, K. P.; da Rocha, S. R. P.; Lee, C. T.; Li, G.; Yates, M. Z. Colloid and interface science for CO2-based pharmaceutical processes. Drugs Pharm. Sci. 2004, 138, 213-245. (2) Shah, P. S.; Hanrath, T.; Johnston, K. P.; Korgel, B. A. Nanocrystal and Nanowire Dispersability in Supercritical Fluids. J. Phys. Chem. B 2004, 108, 9574-9587. (3) Harrison, K.; Goveas, J.; Johnston, K. P. Water-in-Carbon Dioxide Microemulsions with a Fluorocarbon-Hydrocarbon Hybrid Surfactant. Langmuir 1994, 10, 3536-3541. (4) Fulton, J. L.; Pfund, D. M.; McClain, J. B.; Romack, T. J.; Maury, E. E.; Combes, J. R.; Samulski, E. T.; DeSimone, J. M.; Capel, M. Aggregation of Amphiphilic Molecules in Supercritical Carbon Dioxide: A Small-Angle X-ray Scattering Study. Langmuir 1995, 11, 4241-4249. (5) Johnston, K. P.; Harrison, K. L.; Clarke, M. J.; Howdle, S. M.; Heitz, M. P.; Bright, F. V.; Carlier, C.; Randolph, T. W. Water-in-carbon dioxide microemulsions: An environment for hydrophiles including proteins. Science 1996, 271(5249), 624-626. (6) Psathas, P. A.; Janowiak, M. L.; Garcia-Rubio, L. H.; Johnston, K. P. Formation of Carbon Dioxide in Water Miniemulsions Using the Phase Inversion Temperature Methodology. Langmuir 2002, 18, 3039-3046. (7) Dickson, J. L.; Psathas, P. A.; Salinas, B.; Ortiz-Estrada, C.; Luna-Barcenas, G.; Hwang, H. S.; Lim, K. T.; Johnston, K. P. Formation and Growth of Waterin-CO2 Miniemulsions. Langmuir 2003, 19(12), 4895-4904. (8) Singley, E. J.; Liu, W.; Beckman, E. J. Phase Behavior and Emulsion Formation of Novel Fluoroether Amphiphiles in Carbon Dioxide. Fluid Phase Equilib. 1997, 128, 199-219.

inorganic silica12-14 and metal15-17 dispersions in CO2. Recently, these CO2-based systems have been utilized in such novel applications as particle assembly by drop-casting,18 free meniscus coating,19 photoresist drying,20 surfactant-based cleaning and drying of low-k dielectric insulators,21 and chemical mechanical planarization (CMP).22,23 In addition, the low viscosity and (9) Lee, C. T.; Psathas, P. A.; Johnston, K. P.; deGrazia, J.; Randolph, T. W. Water-in-Carbon Dioxide Emulsions: Formation and Stability. Langmuir 1999, 15, 6781-6791. (10) Psathas, P. A.; da Rocha, S. R. P.; Lee, C. T.; Johnston, K. P.; Lim, K. T.; Webber, S. E. Water-in-Carbon Dioxide Emulsions with PDMS based Block Copolymer Ionomers. Ind. Eng. Chem. Res. 2000, 39(8), 2655-2664. (11) DeSimone, J. M.; Maury, E. E.; Menceloglu, Y. Z.; McClain, J. B.; Romack, T. J.; Combes, J. R. Dispersion polymerizations in supercritical carbon dioxide. Science 1994, 265(5170), 256-259. (12) Yates, M. Z.; Shah, P. S.; Lim, K. T.; Johnston, K. P. Steric Stabilization of Colloids by Poly(dimethylsiloxane) in Carbon Dioxide: Effects of Cosolvents. J. Colloid Interface Sci. 2000, 227, 176-184. (13) Sirard, S. M.; Castellanos, H. J.; Hwang, H. S.; Lim, K. T.; Johnston, K. P. Steric Stabilization of Silica Colloids in Supercritical Carbon Dioxide. Ind. Eng. Chem. Res. 2004, 43(2), 525-534. (14) Dickson, J. L.; Shah, P. S.; Binks, B. P.; Johnston, K. P. Steric Stabilization of Nanoparticles in Liquid Carbon Dioxide at the Vapor Pressure. Langmuir 2004, 20, 9380-9387. (15) Holmes, J. D.; Bhargava, P. A.; Korgel, B. A.; Johnston, K. P. Synthesis of Cadmium Sulfide Q Particles in Water-in-Carbon Dioxide Microemulsion. Langmuir 1999, 15, 6613-6615. (16) Shah, P. S.; Holmes, J. D.; Doty, C.; Johnston, K. P.; Korgel, B. A. Steric Stabilization of Nanocrystals in Supercritical CO2 Using Fluorinated Ligands. J. Am. Chem. Soc. 2000, 122, 4245-4246. (17) McLeod, M. C.; McHenry, R. S.; Beckman, E. J.; Roberts, C. B. Synthesis and Stabilization of Silver Metallic Nanoparticles and Premetallic Intermediates in Perfluoropolyether/CO2 Reverse Micelles. J. Phys. Chem. B 2003, 107, 26932700. (18) Shah, P. S.; Novick, B. J.; Hwang, H. S.; Lim, K. T.; Carbonell, R. G.; Johnston, K. P.; Korgel, B. A. Kinetics of Nonequilibrium Nanocrystal Monolayer Formation: Deposition from Liquid Carbon Dioxide. Nano Lett. 2003, 3(12), 1671-1675. (19) Novick, B. J.; DeSimone, J. M.; Carbonell, R. G. Deposition of Thin Polymeric Films from Liquid Carbon Dioxide Using a High-Pressure FreeMeniscus Coating Process. Ind. Eng. Chem. Res. 2004, 43, 515-524. (20) Goldfarb, D. L.; de Pablo, J. J.; Nealey, P. F.; Simons, J. P.; Moreau, W. M.; Angelopoulos, M. Aqueous-based photoresist drying using supercritical carbon dioxide to prevent pattern collapse. J. Vac. Sci. Technol. B: Microelectron. Nanometer Struct. 2000, 18(6), 3313-3317.

10.1021/la0527238 CCC: $33.50 © 2006 American Chemical Society Published on Web 02/02/2006

2162 Langmuir, Vol. 22, No. 5, 2006

interfacial tension of CO2 coupled with its high wettability makes it especially applicable for processes such as the infusion of22 precursors into nanometer-sized pores for nanoscale casting24 and for the infusion of nanoparticles into mesoporous silica.25 Supercritical CO2 with zero surface tension can be used in drying to avoid capillary collapse of various materials due to Laplace pressures. It can also be used to facilitate diffusion into block copolymer templates without perturbing the template morphology.26 The expansion of CO2 into various new fields involving solid interfaces requires a fundamental understanding concerning wetting at the solid/CO2 interface. The adsorption of CO2 on hard solid surfaces including activated carbon,27,28 silica,29 and zeolites30 has been investigated with various experimental techniques. In pressure regions where CO2 is highly compressible, large CO2 excess layers are observed on solid substrates. Even at CO2 activities as low as 0.5, the van der Waals interaction with the surface, which is more polarizable than CO2, can lead to a large excess of CO2 molecules over the range of a few molecular diameters. For CO2 sorption in a polymer (soft surface), an anomalous maximum in the swelling peak has been observed near the critical pressure in polystyrene (PS) latex particles by dynamic light scattering,31 in poly(methyl methacrylate) (PMMA) thin films by ellipsometry,32 and in styrenebutadiene random copolymer films by neutron reflectivity.33 The anomalous maximum was attributed to inhomogeneities in the concentration at the free surface of the polymer due to the high compressibility.32 The presence of this excess CO2 layer near the interface is responsible for the unusually rapid dynamics observed in the lowering of the glass transition temperature (Tg) for thin films,34 for polymer interdiffusion,35 and for the welding of PS latex particles to form films.36 The measurement of θ values (21) Zhang, X.; Pham, J. Q.; Martinez, H. J.; Wolf, J.; Green, P. F.; Johnston, K. P. Water-in-Carbon Dioxide Microemulsions for Removing Postetch Residues from Patterned Porous Low-k Dielectrics. J. Vac. Sci. Technol. B: Microelectron. Nanometer Struct. 2003, 21, 2590-2598. (22) Bessel, C. A.; Denison, G. M.; DeSimone, J. M.; DeYong, J.; Gross, S.; Schauer, C. K.; Visintin, P. M. Etchant Solutions for the Removal of Cu(0) in a Supercritical CO2-based “Dry” Chemical Mechanical Planarization Process for Device Fabrication. J. Am. Chem. Soc. 2003, 125, 4980-4981. (23) Visintin, P. M.; Carbonell, R. G.; Schauer, C. K.; DeSimone, J. M. Chemical Functionalization of Silica and Alumina Particles for Dispersion in Carbon Dioxide. Langmuir 2005, 21, 4816-4823. (24) Wakayama, H.; Inagaki, S.; Fukushima, Y. Nanoporous Titania Synthesized by a Nanoscale Casting Process in Supercritical Carbon Dioxide. J. Am. Ceram. Soc. 2002, 85, 161-164. (25) Gupta, G.; Shah, P. S.; Zhang, X.; Saunders: A. E.; Korgel, B. A.; Johnston, K. P. Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with Supercritical Carbon Dioxide. Chem. Mater. 2005, 17, 6728-6738. (26) Pai, R. A.; Humayun, R.; Schulberg, M. T.; Sengupta, A.; Sun, J. N.; Watkins, J. J. Mesoporous Silicates Prepared Using Preorganized Templates in Supercritical Fluids. Science 2004, 303, 507-511. (27) Chen, J. H.; Wong, D. S. H.; Tan, C. S.; Subramanian, R.; Lira, C. T.; Orth, M. Adsorption and Desorption of Carbon Dioxide onto and from Activated Carbon at High Pressures. Ind. Eng. Chem. Res. 1997, 36, 2808-2815. (28) Humayun, R.; Tomasko, D. L. High-resolution adsorption isotherms of supercritical carbon dioxide on activated carbon. AIChE J. 2000, 46(10), 20652075. (29) di Giovanni, O.; Dorfler, W.; Mazzotti, M.; Morbidelli, M. Adsorption of Supercritical Carbon Dioxide on Silica. Langmuir 2001, 17, 4316-4321. (30) Gao, W.; Butler, D.; Tomasko, D. L. High-Pressure Adsorption of CO2 on NaY Zeolite and Model Prediction of Adsorption Isotherms. Langmuir 2004, 20, 8083-8089. (31) Otake, K.; Webber, S. E.; Munk, P.; Johnston, K. P. Swelling of Polystyrene Latex Particles in Water by High-Pressure Carbon Dioxide. Langmuir 1997, 13, 3047-3051. (32) Sirard, S. M.; Ziegler, K. J.; Sanchez, I. C.; Green, P. F.; Johnston, K. P. Anomalous Properties of Poly(methyl methacrylate) Thin Films in Supercritical Carbon Dioxide. Macromolecules 2002, 35, 1928-1935. (33) Koga, T.; Seo, Y. S.; Zhang, Y.; Shin, K.; Kusano, K.; Nishikawa, K.; Rafailovich, M. H.; Sokolov, J.; Chu, B.; Peiffer, D.; Occhiogrosso, R.; Satija, S. K. Density-Fluctuation-Induced Swelling of Polymer Thin FIlms in Carbon Dioxide. Phys. ReV. Lett. 2002, 89, 125506/1-125506/4. (34) Pham, J. Q.; Sirard, S. M.; Johnston, K. P.; Green, P. F. Pressure, Temperature and Thickness Dependence of CO2-Induced Devitrification in Polymer Films. Phys. ReV. Lett. 2003, 175503/1-175503/4.

Dickson et al.

Figure 1. Schematic illustrating the CO2/water/solid contact angle. γWC, γSC, and γSW represent the water/CO2, solid/CO2, and solid/ water interfacial tensions, respectively.

offers an opportunity to shed further light into the equilibrium behavior of these excess layers. Of particular interest in a variety of fields, for example, microelectronics processing37 and enhanced oil recovery, is how exposure to CO2 alters the wettability of a substrate. In a study of the dewetting of PS thin films from 5 to 100 nm thick, the θ value of PS droplets on SiOx/Si substrates was estimated in supercritical (sc) CO2 from the macroscopic dimensions of the droplet.38 θ was considerably higher in CO2 than in air, indicating that CO2 decreased the wettability, through effects of CO2 on interfacial potentials.38 Yokoyama and Sugiyama39 measured the air/water/solid θ values on various surfaces of fluorinated block copolymers that had been annealed in sc carbon dioxide. On average, θ for the surfaces annealed in sc CO2 was 8° higher than the values for surfaces annealed in a vacuum. This increase in θ was attributed to the ability of sc CO2 to penetrate and swell the block copolymer films.39 Another example involving a solid/CO2 interface is the stabilization of emulsions consisting of water and CO2 with solid nanoparticles.40 The stability of these emulsions was observed to be highly dependent on the particle hydrophilicity and its subsequent θ value at the water/CO2 interface.40 A schematic of the CO2/water/solid θ value is shown in Figure 1. The relationship between θ and the three interfacial energies shown in Figure 1 is given by Young’s equation

cos(θ) )

γSC - γSW γWC

(1)

where γSC, γSW, and γWC are the solid/CO2, solid/water, and water/CO2 interfacial energies, respectively. Of the three interfacial energies, only γWC can be measured experimentally. Using various methods, including capillary rise41 and pendant drop,42,43 γWC has been measured as a function of both temperature (35) Koga, T.; Seo, Y. S.; Hu, X.; Shin, K.; Zhang, Y.; Rafailovich, M. H.; Sokolov, J. C.; Chu, B.; Satija, S. K. Dynamics of Polymer Thin Films in Supercritical Carbon Dioxide. Europhys. Lett. 2002, 60, 559-565. (36) Abramowitz, H.; Shah, P. S.; Green, P. F.; Johnston, K. P. Welding Colloidal Crystals with Carbon Dioxide. Macromolecules 2004, 37, 7316-7324. (37) Gorman, B. P.; Orozco-Teran, R. A.; Zhang, Z.; Matz, P. D.; Mueller, D. W.; Reidy, R. F. Rapid repair of plasma ash damage in low-k dielectrics using supercritical CO2. J. Vac. Sci. Technol. B: Microelectron. Nanometer Struct. 2004, 22, 1210-1212. (38) Meli, L.; Pham, J. Q.; Johnston, K. P.; Green, P. F. Polystyrene thin films in CO2. Phys. ReV. E 2004, 69, 051601/1-051601/8. (39) Yokoyama, H.; Sugiyama, K. Surface Hydrophobicity of Fluorinated Block Copolymers Enhanced by Supercritical Carbon Dioxide Annealing. Langmuir 2004, 20, 10001-10006. (40) Dickson, J. L.; Binks, B. P.; Johnston, K. P. Stabilization of Carbon Dioxide-in-Water Emulsions Using Silica Particles. Langmuir 2004, 20, 79767983. (41) Chun, B. Y.; Wilkinson, T. Interfacial Tension in High-Pressure Carbon Dioxide Mixtures. Ind. Eng. Chem. Res. 1995, 34, 4371-4377. (42) da Rocha, S. R. P.; Harrision, K. L.; Johnston, K. P. Effect of Surfactants on the Interfacial Tension and Emulsion Formation Between Water and Carbon Dioxide. Langmuir 1999, 15, 419-428.

Wetting Phenomena at the CO2/Water/Glass Interface

Langmuir, Vol. 22, No. 5, 2006 2163

Table 1. Effect of Dichlorodimethylsilane (DCDMS) on the Air/Water/Solid Contact Angle substrate

DCDMS conc (M)

advancing θa (deg)

receding θa (deg)

θb (deg)

37% SiOH 12% SiOH

1 ( 10-3 1 ( 10-2

69 ( 2 97 ( 1

64 ( 2 94 ( 1

71 ( 2 98 ( 2

a Contact angle measured using a drop shape analysis system, DSA 10 (Kruss). b Contact angle measured using the high-pressure apparatus shown in Figure 2.

and pressure. Because of the compressible nature of CO2, modest changes in the temperature and/or pressure have been found to have a pronounced effect on γWC. Chi et al.44 measured θ for CO2/water/coal systems as a function of CO2 pressure to shed light on a new process for cleaning coal. The results showed an increase in θ from 84° to 120° as the CO2 pressure was increased from atmospheric pressure to 62 bar at room temperature. In liquid CO2, θ increased further to 145°. The heterogeneous nature of the coal substrate precluded a fundamental analysis of θ in terms of the relevant surface energies. The objectives of this study were to measure CO2/water/solid θ values on two glass substrates with different hydrophilicities versus pressure in a novel apparatus and to understand how longrange interactions and specific short-range interactions with the silanol groups influence the relevant interfacial energies and θ. In a compressible fluid, the cohesive energy density of the solvent can be changed simply by varying the pressure without changing the chemical structure of the solvent. The substrates were glass slides that had been hydrophobically modified with different amounts of dichlorodimethylsilane (DCDMS). In all cases, exposing the glass slides to CO2 resulted in a significant increase in θ, indicating an increase in the hydrophobicity of the substrates. This phenomenon is discussed in terms of CO2 adsorption on the substrate and the effect of CO2 pressure on the three relevant interfacial energies, γSC, γSW, and γWC. A novel type of θ hysteresis was discovered by measuring the CO2/water/solid θ value upon both pressurizing and depressurizing the system. An interfacial model based only on long-range forces is compared with the experimental θ values to estimate the role of the shortrange specific interactions as a function of the silanol concentration on the surface. Experimental Section Materials. The solid substrates used in this study were glass microscope slides (Chance Propper Ltd.). Using a previously described technique,45 the hydrophilicity of the glass slides was altered through silanization with various amounts of dichlorodimethylsilane (DCDMS, Fluka). The glass slides were first placed in a solution of DCDMS in dry cyclohexane (99.7%, Prolabo) under a N2 atmosphere for 1 h. The supernatant was decanted and replaced with pure cyclohexane multiple times to remove any unreacted silane. Finally, the hydrophobically modified glass slides were washed with absolute ethanol (Analar grade) and placed in an oven at 110 °C for 30 min. The hydrophilicity of the modified substrates was characterized by measuring the air/water/solid θ value using a drop-shape analysis system (DSA 10, Kruss). Table 1 summarizes the effect of the DCDMS concentration on the advancing and receding θ values. (43) Hebach, A.; Oberhof, A.; Dahmen, N.; Kogel, A.; Ederer, H.; Dinjus, E. Interfacial Tension at Elevated Pressures-Measurements and Correlations in the Water + Carbon Dioxide System. J. Chem. Eng. Data 2002, 47, 1540-1546. (44) Chi, S. W.; Morsi, B. I.; Klinzing, G. E.; Chiang, S. H. Study of Interfacial Properties in the Liquid CO2-Water-Coal System. Energy Fuels 1988, 2, 141145. (45) Horozov, T. S.; Aveyard, R.; Clint, J. H.; Binks, B. P. Order-Disorder Transition in Monolayers of Modified Monodisperse Silica Particles at the OctaneWater Interface. Langmuir 2003, 19, 2822-2829.

Figure 2. (A) Schematic of high-pressure apparatus used to measure the CO2/water/solid contact angle. The apparatus, which was adapted from Psathas et al.,36 consists of a light source, a variable-volume high-pressure view-cell, an optical rail, a CCD camera, and a computer. (B) Enlarged front view of the measurement cell with a port at the bottom for a removable stage. The air/water/solid θ value increased as the DCDMS concentration increased. The advancing θ value (measured through water) for the more hydrophobic glass slide, was 97° ( 1°, whereas the advancing θ value for 37% SiOH (10-3 M) was 69° ( 2°. As discussed below, the silanol (SiOH) coverages for the two substrates were estimated to be 12% and 37%, respectively. For simplicity, we will refer to the more hydrophobic and less hydrophobic glass slides as the 12% SiOH surface and the 37% SiOH surface, respectively, even though these are only approximate values. Deionized water (Nanopure II, 16 µS/cm) was used in all experiments, and research-grade CO2 (Matheson) was filtered through an oxygen trap prior to use. High-Pressure Contact-Angle Measurement. A schematic of the apparatus used to measure the CO2/water/solid θ values is shown in Figure 2A. It consisted of a 2.5-in.-o.d., 11/16-in.-i.d. stainless steel variable-volume view cell, an optical rail for proper alignment, a light source (Chiu Technical Corporation, model HG-DM), a Sony CCD camera (model XC-73CE), and a computer. The high-pressure cell was fitted with a moveable piston to allow for the cell volume to be varied. Figure 2B shows an enlarged front view of the variablevolume view cell, which was equipped with two sapphire windows (5/8-in. diameter × 1/8-in. thickness) mounted on the side of the cell at 180° (Swiss Jewel) and a stainless steel mounting stage with a circular face with a diameter of 3/8 in. The glass slide (approximately 0.31 in. × 0.31 in.) was mounted to the stage using two small screws. The high-pressure seals for the side windows and the mounting stage were achieved using Teflon O-rings compressed against a flat surface on the cell. To verify that the substrate had been mounted parallel to the optical rail, the difference in θ between the left and right edges of the drop was examined. Typically, the difference between these values was less than 1°, indicating that the substrate was level.

2164 Langmuir, Vol. 22, No. 5, 2006 To measure θ, a water drop, typically 8-10 µL, was first placed on the modified glass slide with a 10-µL syringe (Hamilton Co.). Using the CCD camera, the shape of the water drop was then recorded, and the images were analyzed using a software package from KSV Instruments Ltd. (Helsinki, Finland) to estimate the air/water/solid θ value. The software package used the Young-Laplace method to estimate θ by fitting the complete contour of the sessile drop as a function of interfacial and gravity forces. The pure-component densities for water and CO2 were used in estimating θ. Altering the densities to account for the mutual miscibility of water and CO2 changed θ by less than 0.5°. The measured air/water/solid θ values for both 12% SiOH and 37% SiOH surfaces agreed well with the advancing θ values reported in Table 1. With the water drop still on the glass slide, the high-pressure cell was sealed, and CO2 vapor was slowly introduced into the view cell using a manual high-pressure syringe pump (Ruska). Before the measurement cell was sealed, a small amount of excess water was first placed in the rear of the high-pressure cell to ensure saturation of the CO2 phase. All of the measurements of θ were conducted at room temperature, 23 ( 1 °C. The system pressure, measured using a Sensotec pressure transducer, was increased from atmospheric pressure to the vapor pressure (61.2 bar at 23 °C) in equal increments of 3.4 bar. At the vapor pressure, the backside of the piston was slowly pressurized using the same manual syringe pump in order to condense the CO2 vapor in the front side of the piston containing the drop. It is important to note that the transition from the vapor to the liquid phase did not disturb the position of the denser water drop on the glass slide. At each desired pressure, the drop contour was recorded and analyzed to estimate θ. To investigate any hysteresis effects, θ was also measured upon depressurizing the system. The system was depressurized by opening a vent valve and slowly releasing CO2 from the cell. These θ measurements were repeated at least three times on multiple slides of the same hydrophilicity, and the reproducibility was (2°. For a given droplet measured multiple times, the reproducibility in θ was (0.5°. All of the θ measurements are presented in tabular form in the Supporting Information. Binary Water/CO2 Interfacial Tension Measurement. The water/CO2 interfacial tension was measured as a function of CO2 pressure using high-pressure pendant-drop tensiometry. The apparatus used to measure γWC is discussed elsewhere.46 Briefly, water drops were formed on the tip of a modified silica capillary46 (Western Analytical products) (180-µm o.d., 50-µm i.d.) inside a variablevolume view cell. At each pressure, the recorded images of the pendant drop were analyzed using a software package from KSV Instruments Ltd. (Helsinki, Finland) in order to solve the Laplace equation for γWC. The measured γWC values are provided in the Supporting Information.

Experimental Results Effect of CO2 Pressure on the Contact Angle. θ for water on both the 12% SiOH and 37% SiOH substrates was observed to change significantly upon the addition of CO2 into the highpressure cell. Figure 3 is a photograph of a water drop on a 12% SiOH substrate at 23 °C and various pressures. θ increased significantly as the system pressure was increased from atmospheric pressure to 136 bar. As shown in Figure 4, the θ values for both substrates increased monotonically as the pressure was increased from atmospheric pressure to the vapor pressure (61.2 bar at 23 °C). In this pressure range, the CO2/water/ 37% SiOH surface θ value increased from 71° ( 2° to 99° ( 2°, and the θ value for the 12% SiOH surface increased from 98° ( 2° to 141° ( 3°. Further pressurizing the system above the vapor pressure resulted in only a moderate in(46) Psathas, P. A.; Sander, E. A.; Lee, M. Y.; Lim, K. T.; Johnston, K. P. Mapping the stability and curvature of emulsions of H2O and supercritical CO2 with interfacial tension measurements. J. Dispersion Sci. Technol. 2002, 23(13), 65-80.

Dickson et al.

Figure 3. Photographs of the CO2/water/solid contact angle for the 12% SiOH surface at 23 °C and (A) 1 bar, (B) 41 bar, (C) 61 bar (liquid CO2), and (D) 136 bar. The vapor pressure is 61 bar.

crease in θ. Once in the liquid region, the CO2/water/solid θ value was fairly independent of CO2 pressure up to 204 bar (Figure 4A). At pressures of tens of bars, CO2 becomes highly nonideal, and fugacity coefficients are far from unity. Thus, although θ was experimentally measured as a function of pressure, a more thermodynamically meaningful parameter for representing the data is the CO2 activity. The activity for a pure solvent is related to the pressure by

a)

φP φ P

sat sat

(2)

where φ is the fugacity coefficient at pressure P and φsat is the fugacity coefficient at the vapor pressure (Psat), as calculated from an accurate equation of state for CO2.47 Figure 4B shows θ vs CO2 activity for both the 37% SiOH and 12% SiOH surfaces at 23 °C. Whereas θ increased significantly as the pressure was increased to the vapor pressure (activity ) 1), θ was nearly independent of the activity in the liquid region. The driving force for the wetting of a surface can be characterized by the spreading coefficient.48,49 In its simplest terms, the spreading coefficient (S) represents the difference in the surface energy of a dry substrate (in our case, without water) and the surface energy of a wet substrate49

S ) (Esubstrate)dry - (Esubstrate)wet

(3)

Complete wetting, or S > 0, results when the energy of the wet surface is smaller than that of the dry surface. For the replacement of a solid/CO2 interface by solid/water and water/CO2 interfaces, (47) Ely, J. F. An equation of state model for pure CO2 and CO2 rich mixtures. Proc., Annu. ConV.-Gas Process. Assoc. 1986, 65th, 185192. (48) Rosen, M. J. Surfactants and Interfacial Phenomena, 2 ed.; John Wiley & Sons: New York, 1989. (49) de Gennes, P. G.; Brochard-Wyart, F.; Quere, D. Capillarity and Wetting Phenomena: Drops, Bubbles, Pearls, WaVes; Springer: New York, 2004.

Wetting Phenomena at the CO2/Water/Glass Interface

Langmuir, Vol. 22, No. 5, 2006 2165

Figure 6. Relationship between the water/CO2 interfacial tension and cos(θ) at 23 °C.

Figure 4. (A) Effect of CO2 pressure on θ at 23 °C. (B) Effect of CO2 activity on θ. The activity was calculated using eq 2.

Figure 5. Spreading coefficient as a function of CO2 activity at 23 °C. The CO2 activity is related to the pressure through eq 2.

the spreading coefficient can be written as

S ) γSC - (γSW + γWC)

(4)

By rearranging eqs 1 and 4, S can also be expressed as48

S ) γWC[cos(θ) - 1]

(5)

Defined as such, the spreading coefficient can only be zero (complete wetting, θ ) 0) or negative (partial wetting, θ > 0). Figure 5 illustrates that S, determined from experimental values of γWC and θ (eq 5), is negative for both substrates, indicating partial wetting at all CO2 activities at 23 °C. In these experiments, the droplet equilibrated with the other phases, and θ did not

change over 1 h, indicating that S is essentially an equilibrium value. The presence of dissolved water in the CO2 phase might influence the value of γSC. In each case, S became less negative as the CO2 activity was increased. A less negative S corresponds to an increase in wetting and a decrease in θ. In systems with ideal-gas vapors, such as air, the γ value between the droplet and the vapor phase is independent of pressure. However, in this study, the CO2/water/solid θ value for the 12% SiOH and 37% SiOH surfaces actually increased even though S became less negative at higher pressures. The decrease in the magnitude of S was due to the lowering of γWC with increased pressure. As the pressure was increased from 10 to 200 bar at 23 °C, γWC decreased from approximately 65 to 20 mN/m. This decrease was enough to offset the increase in θ [more negative values of cos (θ)] and result in a decrease in the magnitude of S. The spreading coefficient will be discussed in terms of γSC, γSW, and γWC in further detail below. A large number of wettability studies beginning with Zisman50 have focused on measuring the air/liquid/solid θ value on lowenergy plastic surfaces. On a given substrate, cos(θ) increased as the surface tension of the wetting liquid (γLV) decreased.50-53 By extrapolating the curve to cos(θ) ) 1, the critical surface tension for the solid, γc, could be determined.50 A somewhat related concept can be investigated in CO2. Figure 6 shows a plot of cos(θ) vs γWC for both substrates in CO2. In previous θ studies, the development of these plots required measurements of θ for multiple liquids on a given substrate. However, in this study, Figure 6 was generated by measuring θ at various CO2 pressures for a given water droplet. As the pressure was increased to 204 bar, γWC decreased from 72 to 20 mN/m (Supporting Information). This decrease in γWC is consistent with previous studies of the water/CO2 interface.41-43 Figure 6 shows that cos(θ) for both substrates decreased (θ increased) as γWC decreased, opposite to what has been shown previously for air/liquid/solid systems where the liquid was changed to decrease γLV.50-53 A key difference between this study and earlier studies is that the CO2based system contained two liquid components, or CO2 as a (50) Zisman, W. A. Relation of the equilibrium contact angle to liquid and solid constitution. In AdVances in Chemistry Series 43; Gould, R. F., Ed.; American Chemical Society: Washington, DC, 1964; pp 1-51. (51) Li, D.; Neumann, A. W. Contact Angles on Hydrophobic Solid Surfaces and Their Interpretation. J. Colloid Interface Sci. 1992, 148, 190-200. (52) Kwok, D. Y.; Li, D.; Neumann, A. W. Fowkes’ Surface Tension Component Approach Revisited. Colloids Surf. A: Physiochem. Eng. Aspects 1994, 89, 181191. (53) Kwok, D. Y.; Wu, R.; Li, A.; Neumann, A. W. Contact angle measurements and interpretation: Wetting behavior and solid surface tensions for poly(alkyl methacrylate) polymers. J. Adhes. Sci. Technol. 2000, 14, 719-743.

2166 Langmuir, Vol. 22, No. 5, 2006

Figure 7. (A) CO2/water/solid contact angle for 37% SiOH measured during ()) pressurization and (2) depressurization of the system. (B) CO2/water/solid contact angle for 37% SiOH during ()) first exposure to CO2, (2) repressurization immediately after exposure to CO2, and (0) repressurization 1 week after exposure to CO2. All contact-angle measurements were done at 23 °C. For clarity, the error bars are not shown in part B. Typical error was (2°.

nonideal gas, whereas the vapor phase in all previous studies was an ideal gas or noninteracting component. Contact-Angle Hysteresis in CO2. Contact-angle hysteresis is a phenomenon of significant interest in wettability studies. Hysteresis arises from various factors such as surface roughness, surface deformations, and surface heterogeneity and is often characterized by the difference between the advancing and receding θ values.48,54 At ambient pressure, the advancing θ value can be measured while slowly adding liquid to a sessile drop using a syringe tip that is in contact with the drop. However, for the experimental apparatus illustrated in Figure 2, once the sessile drop was placed on the substrate and the high-pressure cell was pressurized, it was not possible to add liquid to the drop to measure the advancing θ. Thus, we attempted to find new techniques for measuring effective advancing and receding θ values in situ at high pressure. We introduce an alternative type of θ hysteresis in which θ is measured upon pressurizing the cell from atmospheric pressure to the vapor pressure (activity ) 1) and then depressurizing back to atmospheric pressure. Figure 7A shows the CO2/water/ 37% SiOH surface θ value as a function of CO2 activity upon both pressurization and depressurization. In most cases, the cell was only pressurized to the vapor pressure before being depressurized, because most of the change in θ occurred in the vapor region relative to the liquid region. As the system was (54) Johnson, R. E.; Dettre, R. H. Wetting of Low-Energy Surfaces. In Wettability; Berg, J. C., Ed.; Marcel Dekker: New York, 1993; pp 1-73.

Dickson et al.

Figure 8. (A) CO2/water/solid contact angle for 12% SiOH measured during (() pressurization and (0) depressurization of the system. (B) CO2/water/solid contact angle for 12% SiOH during (() first exposure to CO2 and (0) repressurization immediately after exposure to CO2. All contact-angle measurements were done at 23 °C. For clarity, the error bars are not shown. Typical error was (2°.

pressurized from atmospheric pressure to the vapor pressure, θ increased from 71° ( 2° to 99° ( 2°. Upon depressurization to atmospheric pressure, θ decreased from 99° ( 2° to 79° ( 1°. Although θ decreased as the system was depressurized, it was noticeably higher at all pressures during depressurization than during pressurization. To our knowledge, this type of hysteresis with pressure has not been reported previously. Immediately after depressurization was completed, a second pressurization and depressurization cycle was performed. The θ values were within 2° of those measured during the first cycle for each step, indicating little change. Of particular interest was the observation that θ after depressurization to atmospheric pressure was 8° higher than the initial θ value. To investigate this hysteresis further, a new drop was placed on the same 37% SiOH substrate immediately after it had been depressurized to atmospheric pressure, and θ was remeasured as a function of CO2 activity. Upon repressurization, θ for the second drop was noticeably higher than that for the first drop below a CO2 activity of about 0.8 (Figure 7B). However, at higher CO2 activities, the measured θ values for the two cases were identical. Interestingly, after the same 37% SiOH substrate had been placed in a desiccator for a week, θ returned to its original value upon repressurization. Unlike the case for the 37% SiOH surface, the CO2/water/ 12% SiOH surface system did not exhibit hysteresis. As illustrated in Figure 8A, θ was essentially identical upon both pressurization and depressurization. To fully verify this lack of hysteresis, the system was pressurized to 204 bar and then slowly depressurized

Wetting Phenomena at the CO2/Water/Glass Interface

Langmuir, Vol. 22, No. 5, 2006 2167

back to atmospheric pressure. A new drop was then placed on the same substrate, and the process was repeated. The contact angle for this second drop was identical to that of the first drop (Figure 8B). This procedure was repeated three times, and no changes in θ were observed. For both of the SiOH surfaces, a novel type of effective advancing and receding θ was defined on the basis of movement of the three-phase contact line. Receding angles were observed upon pressurization as θ decreased, and advancing angles were observed upon depressurization as θ increased. For the 37% SiOH surface, the difference between the advancing and receding θ values was significant, whereas the difference vanished for the 12% SiOH surface.

Discussion Both substrates shown in Figure 4 became more hydrophobic as the CO2 activity was increased, as reflected in the increase in θ. Despite the similarities in the activity effect, the mechanisms for the increase in θ will be shown to be quite different, as suggested by the differences in hysteresis. Before presenting separate discussions for each substrate, we begin with a description of models for the relevant interfacial energies. Estimation of Surface Properties in CO2. In eq 1, the only two terms that can be determined experimentally are θ and γWC. There are no direct methods available for measuring γSC and γSW. However, various approaches have been proposed to estimate these two interfacial energy terms. One commonly used approach is to use combining rules along with the surface component properties for the three γ values.55 As first proposed by Fowkes,56 γ (usually measured against air) for any phase can be expressed as a sum of components that arise because of the various types of intermolecular forces

γ ) γd + γh + γdi + γq + ...

(6)

where γd, γh, γdi, and γq are the interfacial tension components due to dispersion forces, hydrogen bonding, dipole interactions, and quadrupole interactions, respectively. For simplicity, eq 6 is often expressed in terms of long-range forces, excluding specific short-range interactions, as

γ ) γd + γp

(7)

where γp is the polar component of the total interfacial tension and represents the sum γh + γdi + γq. Using a geometric-mean combining rule, the interfacial tension between any two phases, R and β, can be expressed in terms of the component properties as follows55

γRβ ) γRd + γRp + γβd + γβp - 2xγRdγβd - 2xγRpγβp (8) Clint55

Binks and have previously provided a detailed procedure through which eqs 1, 7, and 8 can be used to predict the γ components for the substrate, γsd and γsp. Here, γsd and γsp are defined to be dependent on the physical nature of only the surface and droplet phase, and not the continuous liquid phase, as verified elsewhere.55 For given γsd and γsp values, Binks and Clint55 predicted the oil/water/solid θ values for a range of oils on a given substrate, and the theoretical predictions were shown to be in good agreement with experiment. (55) Binks, B. P.; Clint, J. H. Solid Wettability from Surface Energy Components: Relevance to Pickering Emulsions. Langmuir 2002, 18, 12701273. (56) Fowkes, F. M. Additivity of intermolecular forces at interfaces. I. Determination of the contribution to surface and interfacial tensions of dispersion forces in various liquids. J. Phys. Chem. 1963, 67, 2538-2541.

Table 2. Calculated Surface Propertiesa property θAW (deg) θOW (deg) γSd (mN/m) γSp (mN/m) γSO (mN/m) γSW (mN/m) γSA (mN/m)

37% SiOH

12% SiOH

71 91 24.5 12.1 12.3 13.2 36.6

98 120 16.2 3.0 3.2 29.2 19.2

a For the calculations, γWA was taken as 71.9 mN/m with γWd ) 21.5 mN m-1 and γWp ) 50.4 mN m-1. The oil used was heptane, with γOd ) 19.8 mN m-1 and γOp ) 0.0 mN m-1. The interfacial tension component values were taken from the literature.55

An approach similar to that of Binks and Clint55 was taken in this study to predict the CO2/water/solid θ value as a function of CO2 activity. To determine γsd and γsp for the two types of substrates, the water/solid θ value was first measured under both air and heptane. Using these experimentally measured θ values (Table 2) and the previously determined polar and dispersive components for water and heptane (γWd ) 21.5 mN m-1, γWp ) 50.4 mN m-1, γOd ) 19.8 mN m-1, γOp ) 0.0 mN m-1),55 γSd and γSp were calculated from eqs 1, 7, and 8. The surfaceenergy components for both substrates along with the calculated values for γSO, γSW, and γSA are summarized in Table 2. On the basis of the heptane/water/solid θ values, the silanol (SiOH) coverages for the two substrates were estimated to be 12% and 37%, respectively.55 Earlier studies of θ for the air/liquid/solid interface for alkanes on fluorinated surfaces showed that the adsorption of air molecules at the vapor/solid interface was negligible and that solid/air interfacial tension (γSV) could be assumed to be constant for each surface.51,52 In contrast, the physics is much richer for hydrophilic silica in the presence of CO2 as a result of its large cohesive energy density.57 To estimate the oil/water/solid θ value, it is necessary to know γod, γop, γsd, and γsp. In the previous study by Binks and Clint,55 all of the oils used were liquids at atmospheric pressure. By measuring γOA and γOW, γod and γop could easily be determined from eqs 7 and 8. In the present study, the “oil” is CO2 and is a gas at atmospheric pressure. Because air is miscible with CO2, the CO2/air interfacial tension (γCA) cannot be measured experimentally. However, a value can be assigned for this hypothetical interface from the solubility parameter of pure CO2.58,59 The total solubility parameter (δ) of pure CO2 is given rigorously by

δ2 )

hig - RT - h + Pν ν

(9)

where hig is the enthalpy of CO2 at ideal-gas conditions and h and ν are the enthalpy and molar volume, respectively, of CO2 at a specified T and P. The interfacial tension is related to the solubility parameter by59

γ ) kδ2ν1/3

(10)

where k ) 1.8 × 10-9 mol1/3.60 Previous studies have shown (57) Tripp, C. P.; Combes, J. R. Chemical Modification of Metal Oxide Surfaces in Supercritical CO2: The Intreaction of Supercritical CO2 with the Adsorbed Water Layer and the Surface Hydroxyl Groups of a Silica Surface. Langmuir 1998, 14, 7348-7352. (58) Hildebrand, J. H.; Scott, R. L. The Solubility of Nonelectrolytes, 3rd ed.; Reinhold Publishing Corporation: New York, 1950. (59) Barton, A. F. M. Handbook of Solubility Parameters and Other Cohesion Parameters; CRC Press: Boca Raton, FL, 1983. (60) Siow, K. S.; Patterson, D. The prediction of surface tensions of polymer liquids. Macromolecules 1971, 4, 26-30.

2168 Langmuir, Vol. 22, No. 5, 2006

that quadrupole-quadrupole interactions can account for as much as 20% of CO2’s solubility parameter,61,62 while the remaining 80% is from the dispersive interactions. Thus, from eqs 7 and 10, it is straightforward to show for pure CO2 that γd/γ ) 0.82/(0.82 + 0.22) ) 0.94 and similarly γp/γ ) 0.06. This approach allowed for both γCd and γCp to be estimated at any specified T and P, even in the vapor region. A plot of γCA as a function of CO2 activity is provided in the Supporting Information. It is important to note that changing the relative contributions of the dispersive (γCd) and polar (γCp) components of γCA had only a minimal effect on the predicted θ values. As illustrated in the Supporting Information, the hypothetical CO2 vapor/ air and CO2 liquid/air γ values at 23 °C and 63 bar were estimated to be approximately 1.5 and 10 mN/m, respectively. Using these values along with eq 8, the CO2 vapor/liquid γ value at an activity of 1 and a temperature of 23° C was calculated to be 3.75 mN/m. This rough estimation is in fair agreement with experimental values on the order of 1 mN/m at 25 °C.63 Once the energy components for both the solid and CO2 were estimated, γSW and γSC were calculated using the appropriate form of eq 8

Dickson et al.

Figure 9. Estimated γSC as a function of CO2 activity at 23 °C.

γSW ) γSd + γSp + γWd + γWp - 2xγSdγWd - 2xγSpγWp (11) γSC ) γSd + γSp + γCd + γCp - 2xγSdγCd - 2xγSpγCp (12) In this simple model, the calculated value for γSW was independent of the CO2 density, because all of the terms in eq 11 were constant. The solid/water interactions can be expected to change because of the increasing solubility of CO2 in water with increasing pressure. As discussed in greater detail below, γSW can change because of specific interactions between surface silanol groups and water or CO2. However, in the simple expression in eq 11, the values of γSW are constant. The calculated values for the 37% SiOH and 12% SiOH surfaces were 13.2 and 29.2 mN/m, respectively (Table 2). The larger γSW value for the more hydrophobic surface, 12% SiOH, is consistent with the larger air/water/solid θ value. Figure 9 shows the effect of CO2 activity on the experimentally measured γWC value and the value of γSC calculated according to eq 12 for both substrates. Figure 9 indicates a decrease in γSC with increased CO2 activity on the basis of eq 12. As the cohesive energy density of CO2 increases at elevated pressures, becoming closer to that of the substrate, the interactions between the solid and CO2 become more favorable. Consequently, γSC decreases. However, γSC did not decrease as significantly as did the experimentally measured γWC. With the calculated γSW and γSC values from eqs 11 and 12 and the experimentally measured γWC value, the CO2/water/ solid θ can be predicted as a function of CO2 activity. Previous molecular dynamic simulation studies64 have indicated the complexities of the structure of the water/CO2 interface. Consequently, eq 8 was not capable of accurately predicting (61) O’Neill, M. L.; Cao, Q.; Fang, M.; Johnston, K. P.; Wilkinson, S. P.; Smith, C. D.; Kerschner, J. L.; Jureller, S. H. Solubility of Homopolymers and Copolymers in Carbon Dioxide. Ind. Eng. Chem. Res. 1998, 37, 3067-3079. (62) Prausnitz, J. M.; Lichtenthaler, R. N.; de Azevedo, E. G. Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed.; Prentice Hall: Upper Saddle River, NJ, 1999. (63) Quinn, E. L. The surface tension of liquid carbon dioxide. J. Am. Chem. Soc. 1927, 49, 2704-11. (64) da Rocha, S. R. P.; Johnston, K. P.; Westacott, R. E.; Rossky, P. J. Molecular Structure of the Water-Supercritical CO2 Interface. J. Phys. Chem. B 2001, 102, 12092-12104.

Figure 10. Comparison between experimentally measured contact angles for (() 37% and (2) 12% SiOH and the (0) theoretically predicted values (long-range forces only) at 23 °C.

γWC, and thus, the experimental values were used in the calculations of θ. Figure 10 shows the predicted θ values along with the experimentally measured values for both substrates. Although the predicted and experimental θ values for the 12% SiOH surface agreed quite well, the model was not able to predict the increase in θ accurately for the 37% SiOH surface, particularly at high activities. To further examine the wetting mechanism, γSW was calculated as a function of CO2 activity for both substrates using the experimentally measured θ and γWC values along with the values of γSC predicted using eq 12. γSW was calculated by rearranging eq 1 as follows

γSW ) γSC - γWC cos(θ)

(13)

As shown in Figure 11, γSW was relatively constant at ∼25-30 mN/m for the 12% SiOH surface. However, it increased from approximately 12 to 25 mN/m with activity for the 37% SiOH surface, indicating that the solid/water interactions became increasingly less favorable. In view of these differences, the behavior for each substrate is discussed in detail in the following two sections. For both substrates, large changes in slope and curvature are observed in S in Figure 5 for CO2 activities between approximately 0.8 and 1.1. In this activity region, the excess adsorption of CO2 is known to be large on hard surfaces, and large anomalies are observed in swelling and refractive indices of polymer thin films, as discussed in the Introduction. These types of concentration fluctuations in a fluid with high compressibility and free volume are likely to produce unusual behavior in S because of large

Wetting Phenomena at the CO2/Water/Glass Interface

Figure 11. Estimated γSW as a function of CO2 activity at 23 °C. γSW estimated using eq 13.

excess concentrations of CO2. These fluctuations have previously been seen to cause a large minimum in γWC at a temperature just above the critical temperature.42 A change in curvature is evident in γWC in Figure 9 at 23 °C with the approach to an activity of unity. It is likely that an excess CO2 layer on the surface might cause an unusual decrease in γSC. However, the large excess concentrations do not appear to cause unusual changes in curvature in θ despite the unusual behavior for S. Perhaps there is a cancellation of these effects for the various three surface energies in the expression for cos(θ). 12% SiOH Surface. The lower SiOH coverage for the 12% SiOH surface might be expected to simplify the theoretical description by limiting the specific interactions between the water drop and the surface. Because of the limited number of silanol groups available on the 12% SiOH substrate, only a minimal amount of CO2 might be expected to cap these hydrophilic sites. With few silanol sites, θ is governed primarily by long-range forces and not specific interactions. Consequently, it is reasonable to assume that the addition of CO2 has a relatively small effect on γSW as was shown above in Figure 11. Instead, the decrease in the magnitude in S for the 12% SiOH surface at elevated CO2 pressures (Figure 5) can be shown to be influenced more strongly by a large decrease in γWC and a smaller decrease in γSC, as shown in Figure 9. As summarized in Table 1, the air/water/12% SiOH surface θ was greater than 90°. For θ > 90°, the decrease in γWC will increase θ and cause cos(θ) to become more negative. In addition, the decrease in γSC and a weakly varying value of γSW will make the numerator more negative, which will also make cos(θ) become more negative. Thus, both the numerator and denominator contributed to the experimentally observed more negative values of cos(θ) as the CO2 activity was increased. As illustrated in Figure 4b, θ increased more rapidly as the CO2 activity was increased above 0.6. This phenomenon is due in part to the formation of an excess “liquidlike” layer of CO2 on the solid, as discussed earlier, as well as a large decrease in γWC (Supporting Information). The low interfacial energy of liquid CO2, which contributes to a decrease in γSC with pressure, promotes dewetting (increase in θ) of the surface by water. Because high-energy fluids, for example, water, do not tend to spread on low-energy surfaces, the presence of a low-energy CO2 layer on silica will cause the water/solid θ value to increase in order to decrease contact between these two phases. In addition, the decrease in γWC will cause θ to increase further above 90° to increase the interfacial area between water and CO2. These factors appear to be captured with reasonable quantitative accuracy by the long-range model in Figure 10. The predictive model does

Langmuir, Vol. 22, No. 5, 2006 2169

not include specific interactions with silanol groups or utilize any θ data, indicating that the behaviors of γSC and γSW are governed primarily by long-range forces. However, specific interactions between water and CO2 are included by utilizing available binary experimental γWC values. The limited number of silanol groups on the 12% SiOH surface and the dominant role of long-range forces can be shown to explain the lack of hysteresis in θ (Figure 8a) for this substrate. For a small amount of CO2 physisorbed on the silanol sites, there is a small probability for CO2 molecules to become kinetically trapped on silanol sites during depressurization. If larger numbers of silanol sites were capped by CO2 molecules during depressurization versus initial pressurization, higher values of θ would have been observed during depressurization versus pressurization. Instead, hysteresis was not observed, consistent with the small numbers of silanol groups for the 12% SiOH surface. The pH of water changes only slightly with pressure at 25 °C from 2.83 to 2.80 with an increase in pressure from 70 to 200 atm.65 However, the pH is 3.8 for a pressure of 1 bar. For the 12% SiOH surface, θ decreased from 96° to 95° as the pH of an aqueous droplet in air at ambient pressure was reduced from 7 to 4. However, it dropped to 85° at pH 3 and did not change for a further change to pH 2. The shift in θ from pH 4 to 3 is consistent a point of zero charge of SiO2 of 3, measured by several other techniques.66 In this study, θ increased as the CO2 pressure was increased from a starting value of zero, despite the decrease in pH from 7 to 3.8 and eventually 2.8. Because it shifted in the opposite direction with pH, as in the above case for droplets in air at ambient pressure, the role of pH in changes in θ was very minor. 37% SiOH Surface. As summarized in Table 1, the air/water/ 37% SiOH surface θ value was less than 90°, indicating that the surface was slightly hydrophilic. From eq 1, a θ value less than 90° indicates that γSC > γSW. In this situation, which took place for all but the highest activities of CO2, a decrease in γWC alone would result in a decrease in θ and an increase in cos(θ), assuming that the numerator of eq 1, γSC - γSW, remained constant. However, Figure 4 shows the opposite trend in θ for the 37% SiOH surface. The θ value actually increased with activity, as γSC decreased (Figure 9) and γSW increased (Figure 11). The decreases in γSC were somewhat similar for the two substrates, as shown in Figure 9. However, a large increase in γSW was observed for the 37% SiOH surface when the activity increased from 0 to 0.15 as the CO2 molecules capped the silanol groups and displaced water. It was the increase in γSW coupled with the decrease in γSC and γWC that resulted in a decrease in the magnitude of S at elevated pressures for the 37% SiOH surface. The physisorption of CO2 on the silanol groups with an increase in pressure makes the surface less hydrophilic, and the subsequent increase in γSW resulted in a decrease in cos(θ) despite the decrease in γWC. Eventually, γSC became smaller than γSW, and θ became larger than 90°. In this region, a decrease in γWC increased θ in order to increase the water/CO2 interfacial area. On the basis of FTIR spectroscopy, Tripp and Combes57 reported that CO2 vapor at 50 °C was capable of physisorbing onto the exposed silanol groups on the surface of deuterated silica for pressures as low as 3.6 bar. The results also indicated increased adsorption at higher pressures.57 This adsorption of CO2 can be expected to reduce the hydrophilicity of the substrate (65) Toews, K. L.; Shroll, R. M.; Wai, C. M.; Smart, N. G. pH-Defining Equilibrium between Water and Supercritical CO2. Influence on SFE of Organics and Metal Chelates. Anal. Chem. 1995, 67(22), 4040-3. (66) Bourikas, K.; Vakros, J.; Kordulis, C.; Lycourghiotis, A. J. Phys. Chem. B 2003, 107, 9441-9451.

2170 Langmuir, Vol. 22, No. 5, 2006

by removing sites for specific interactions with water, consistent with our observation of an increase in θ. The short-range, specific interactions with silanol groups were also responsible for the difference between the predicted and experimentally measured θ values (Figure 10). A slight increase in γSW with an increase in CO2 activity is expected because of long-range forces, as was seen for the 12% SiOH surface because of the low surface energy of an excess CO2 layer. However, the increase in γSW for the 37% SiOH surface is even larger than expected from long-range forces due to the capping of silanol groups by CO2. Consequently, the model underpredicts θ. Such an increase in γSW was not present for the 12% SiOH surface, because it had much fewer silanol groups. This physisorption of CO2 molecules onto the silanol sites on the surface can also explain the hysteresis in θ observed for the 37% SiOH surface (Figure 7a). In a recent study by McCool and Tripp,67 IR spectroscopy was used to investigate the desorption of CO2 from the silanol groups on fumed nonporous hydrophilic silica. IR spectra for silica were taken (1) immediately after contact with CO2 at 50 °C and 200 bar and (2) 1 month after contact. After 1 month, the CO2 peak had decreased by nearly a factor of 2 as CO2 molecules desorbed from the surface. However, the peak intensity was still fairly strong, indicating that a significant amount of CO2 was still adsorbed at the solid surface. The fact that CO2 does not rapidly desorb from silica surfaces explains the larger θ values during depressurization relative to pressurization. Although some of the CO2 molecules desorbed from the surface as the CO2 activity was lowered, a significant fraction remained physisorbed on the silanol groups. These CO2 molecules remaining on the surface produced the observed “lag” or hysteresis in θ during depressurization. Even after the system had been fully depressurized to atmospheric pressure, enough CO2 remained on the surface to result in an increase in the air/ water/solid θ value by 8°. After 1 week, all of the excess CO2 had completely desorbed from the silica surface, and θ returned to its original value (Figure 7b). Compared to the more hydrophilic silica in the previous FTIR study,67 the 37% SiOH substrate had fewer surface silanol groups. It is therefore not surprising that it took less than a week, as opposed to a month, for the CO2 to fully desorb from the surface. The cycle for the hysteresis experiments always started from a droplet on a glass surface at atmospheric pressure. The effective receding angle was measured upon pressurization as the contact line moved inward. The effective advancing angles were observed upon depressurization as the three-phase contact line moved outward. In this case, the advancing droplet traversed a surface that was previously exposed to water during the pressurization cycle. The system appeared to reach a metastable state, as the difference between the advancing and receding angles for a given pressure was the same upon the first and second cycles, where each cycle consisted of a pressurization and depressurization step. For the 37% SiOH surface, the difference between advancing and receding angles was significant, but the difference vanished for the 12% SiOH surface. Thus, it appeared that the capping (67) McCool, B.; Tripp, C. P. Inaccessible Hydroxyl Groups on Silica Are Accessible in Supercritical CO2. J. Phys. Chem. B 2005, 109, 8914-8919.

Dickson et al.

of the SiOH groups by CO2 played a key role in the hysteresis by maintaining a significant hydrophobicity of the surface upon depressurization.

Conclusions The novel high-pressure technique described herein has been utilized successfully to measure CO2/water/solid contact angles including hysteresis during pressurization and depressurization. The θ values for water on both the 12% SiOH and 37% SiOH substrates increased significantly as the CO2 pressure was increased from atmospheric pressure to the vapor pressure (61.2 bar) at 23 °C and beyond. A novel type of θ hysteresis was discovered in which effective advancing θ values measured during depressurization exceeded effective receding θ values measured during pressurization on the basis of movement of the threephase contact line. The greater degree of hysteresis for the less hydrophobic silica can be attributed in part to the capping of silanol groups with CO2. For both substrates, the change in the surface energy was larger for γWC than for γSC or γSW, and the decreases in γSC were similar. Although both substrates exhibited an increase in θ with CO2 pressure, the role of shortrange specific interactions was quite different. A simple model, which addressed only the long-range forces, predicted θ for the surface with a small silanol concentration of 12% and, thus, relatively few specific intermolecular interactions. In contrast, it underpredicted θ for the 37% SiOH surface, as physisorption of CO2 displaced water from the surface silanol groups, as seen previously by FTIR spectroscopy.57 The capping of silanol groups by CO2 increased γSW to a greater extent than for the 12% SiOH surface. This interfacial study provides fundamental insight into earlier observations of Tripp and Combes57 that CO2 is more effective than many organic solvents in removing water from silica, even when water is more soluble in the organic solvent. The specific short-range interactions between CO2 and the substrate result in a decrease in γSC and an increase in γSW, both of which favor desorption of water from the surface. The ability of CO2 to aid dewetting of surfaces and to prevent wetting even after depressurization to ambient pressure is of great practical interest in fields such as enhanced oil recovery, cleaning and drying of surfaces including low-k dielectric materials in microelectronics, and processing of polymer thin films. Further studies of the fundamental aspects of wetting in CO2-based systems will aid the development of these applications. Acknowledgment. This material is based on work supported in part by the Department of Energy Office of Basic Energy Sciences (DE-FG02-04ER15549), the Welch Foundation, International Sematech, and the Separations Research Program at the University of Texas. Supporting Information Available: Information concerning the water/CO2 and CO2/air interfacial tensions. This material is available free of charge via the Internet at http://pubs.acs.org. LA0527238