Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of water and solvents at liquid/solid interfaces by sum frequency generation vibrational spectroscopy Yang, Zheng 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2011_fall_yang_zheng.pdf [ 1.56MB ]
Metadata
JSON: 24-1.0059841.json
JSON-LD: 24-1.0059841-ld.json
RDF/XML (Pretty): 24-1.0059841-rdf.xml
RDF/JSON: 24-1.0059841-rdf.json
Turtle: 24-1.0059841-turtle.txt
N-Triples: 24-1.0059841-rdf-ntriples.txt
Original Record: 24-1.0059841-source.json
Full Text
24-1.0059841-fulltext.txt
Citation
24-1.0059841.ris

Full Text

Studies of Water and Solvents at Liquid/Solid Interfaces by Sum Frequency Generation Vibrational Spectroscopy by  Zheng Yang B.Sc., Peking University, Beijing, 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  May 2011 © Zheng Yang, 2011  Abstract  This dissertation studies the surface chemistry of water and organic solvents at liquid/mineral interfaces using IR-visible sum frequency generation (SFG) vibrational spectroscopy. The solvents studied include pentane, heptane, tetradecane and toluene, and the minerals include silica and mica. These liquid/mineral interfaces are relevant for environmental and industrial processes, such as ice nucleation and oilsands extraction. Structures of water at water/silica interfaces were studied in the presence of alkali chloride in solution. Perturbations of the interfacial water structures were observed with NaCl concentrations as low as 1x10-4 M. Different alkali cations produce different magnitudes of perturbation, with K+ > Li+ > Na+. This order was explained by the different effective ionic radii and electrostatic interactions between the cations and silica surfaces. The adsorption of water at solvent/silica interfaces was studied at room temperature. A water layer without detectable free OHs was discovered at toluene/silica interface. This water layer showed resistance against further adsorption of water molecules and was very stable at room temperature. However, similar structure of water was not observed at heptane/silica interfaces. Water structures on mica were studied with atmospherically relevant sulphuric acid concentrations. Experimental data showed that ordered water structures on mica completely disappeared when the concentration of sulfuric acid reached 5 mol/L. The results partially explain why sulfuric acid coatings influence the ice nucleation properties of mineral dust particles in the atmosphere. ii  The competitive adsorption of toluene and n-alkane at solvent/silica interfaces was studied. The surface coverage of toluene for toluene-pentane, toluene-heptane, and toluene-tetradecane mixtures were measured over the complete mole fraction range from 0 to 1. Overall, toluene competes favorably on silica, but the molar adsorption free energy of alkanes increases as the chain length increases. Finally, experiments were conducted to study the effect of interfacial water on bitumen liberation from mineral surfaces in water. The bitumen liberation rate increases when the water content between bitumen and the mineral increases. The liberation also highly depends on the surface properties of minerals. At the same water content, the rate of bitumen displacement on different mineral surfaces is: freshly cleaved mica > rinsed mica > silica.  iii  Preface  A version of Chapter 3 has been published as: Yang, Z.; Li, Q. F.; Chou, K. C. Structures of Water Molecules at the Interfaces of Aqueous Salt Solutions and Silica: Cation Effects. Journal of Physical Chemistry C 2009, 113, 8201-8205. The project was initiated by Dr. Chou. Dr. Li gave valuable suggestions for optimizing experimental setup and data analysis. My major contributions are: literature survey, design of the research project, data acquisition, interpretation of the observations, and preparing the manuscripts. A version of Chapter 4 has been published as: Yang, Z.; Li, Q. F.; Gray, M. R.; Chou, K. C. Structures of Water Molecules at Solvent/Silica Interfaces. Langmuir 2010, 26, 16397-16400. The project was initiated by Dr. Chou and Dr. Gray. Dr. Li helped in optimizing the experimental setup. My major contributions are: literature survey, design of the research project, data acquisition, explanation of the observations, preparing the manuscripts. A version of Chapter 5 will be submitted for publication: Yang, Z.; Bertram, A. K.; Chou, K. C. Why do Sulfuric Acid Coatings Influence the Ice Nucleation Properties of Mineral Dust Particles in the Atmosphere? The project was initiated by Dr. Bertram and Dr. Chou. Also Dr. Bertram and Dr. Chou contributed to the design of the project, and editing of the manuscript. My major contributions are: literature survey, design of the research project, data acquisition, explanation of the observations, preparing the manuscripts. A version of Chapter 6 has been published as: Yang, Z.; Li, Q. F.; Hua, R.; Gray, M. R.; Chou, K. C. Competitive Adsorption of Toluene and n-Alkanes at Binary iv  Solution/Silica Interfaces. Journal of Physical Chemistry C 2009, 113, 20355-20359. The project was initiated by Dr. Chou and Dr. Gray. Dr. Chou developed the computer program used to fit the SFG spectra. Dr. Li calculated the orientation distribution of the CH3 groups. Dr. Hua obtained the SFG vibrational spectra of toluene in ppp polarization configuration. My major contributions are: literature survey, design of the research project, data acquisition. A version of Chapter 7 will be submitted for publication. Yang, Z.; Bailey, G.; Gray, M. R.; Chou, K. C. Effect of Interfacial Water Content on Bitumen Liberation from Silica and Mica Surfaces. The project was initiated by Dr. Chou and Dr. Gray. Gwen Bailey developed the computer program used to analyze the contact angle of bitumen droplet in water. My major contributions are: literature survey, design of the research project, data acquisition, explanation of the observations, preparing the manuscripts.  v  Table of Contents  Abstract..............................................................................................................................ii Preface................................................................................................................................iv Table of Contents..............................................................................................................vi List of Figures....................................................................................................................ix Abbreviations..................................................................................................................xiv Acknowledgments............................................................................................................xv Dedication………...........................................................................................................xvii Chapter 1 Introduction.....................................................................................................1 1.1 Overview..................................................................................................................2 1.2 Water Structures on Water/Mineral Interfaces........................................................4 1.3 Solvent Adsorption on Minerals..............................................................................5 1.4 Water’s Role in Oil Sands Processing.....................................................................7 Chapter 2 IR-visible Sum Frequency Generation Spectroscopy...................................9 2.1 Introduction............................................................................................................10 2.2 Theory of Sum Frequency Generation...................................................................11 2.3 Fresnel Factors.......................................................................................................15 2.4 Molecular Orientation at the Interface...................................................................17 2.5 Experimental Instrumentation................................................................................19 vi  Chapter 3 Structures of Water Molecules at the Interfaces of Aqueous Salt Solutions and Silica: Cation Effects................................................................................................21 3.1 Introduction............................................................................................................22 3.2 Experimental Section.............................................................................................25 3.3 Results and Analysis..............................................................................................28 3.4 Conclusions............................................................................................................36 Chapter 4 Structures of Water Molecules at Solvent/Silica Interfaces......................37 4.1 Introduction............................................................................................................38 4.2 Experimental Section.............................................................................................40 4.3 Results and Analysis..............................................................................................41 4.4 Conclusions............................................................................................................49 Chapter 5 Why do Sulfuric Acid Coatings Influence the Ice Nucleation Properties of Mineral Dust Particles in the Atmosphere....................................................................50 5.1 Introduction............................................................................................................51 5.2 Experimental Section.............................................................................................54 5.3 Results and Analysis..............................................................................................56 5.4 Conclusions............................................................................................................64 Chapter 6 Competitive Adsorption of Toluene and n-Alkanes at Binary Solution/Silica Interfaces.................................................................................................65 6.1 Introduction............................................................................................................66  vii  6.2 Experimental Section.............................................................................................68 6.3 Results and Analysis..............................................................................................70 6.4 Conclusions............................................................................................................83 Chapter 7 Effect of Interfacial Water Content on Bitumen Liberation from Silica and Mica Surfaces............................................................................................................84 7.1 Introduction............................................................................................................85 7.2 Experimental Section.............................................................................................86 7.2.1. Sum frequency generation spectroscopy measurements..........................................86 7.2.2. Dynamic and equilibrium contact angle measurements..........................................87  7.3 Results and Analysis..............................................................................................89 7.3.1. Sum frequency generation spectroscopy experiments.............................................89 7.3.2. Dynamic and equilibrium contact angle study.........................................................90  7.4 Conclusions............................................................................................................96 Chapter 8 Conclusion......................................................................................................97 References.......................................................................................................................102  viii  List of Figures  Figure 2.1 SFG setup in the reflection geometry. ............................................................11 Figure 2.2 The definition of orientation angle for methyl group. Orientation angle (θ) between surface normal and methyl group, axis c is the principal axis of the methyl group. The ab plane is perpendicular to axis c. The ac plane contains one C-H bond. ...............18 Figure 2.3 Layout of OPG/OPA. DM, RM and BS represent the dichromatic mirror, reflective mirror, and beam splitter, respectively. ............................................................20 Figure 3.1 Schematic layout of the SFG vibrational spectroscopy. The frequency of the visible beam was fixed at 532 nm and the frequency of the IR beam was scanned from 2800 – 3900 cm-1. The 532 nm and IR beams were overlapped both spatially and temporally on the bottom surface of the solution. ............................................................26 Figure 3.2 SFG spectra from the water/silica interfaces with different concentrations of three alkali chloride solutions: (a) NaCl, (b) LiCl, and (c) KCl. The solid lines are fitting curves derived using two Lorentzian peaks at 3200 cm-1 and 3400 cm-1. ........................30 Figure 3.3 Amplitudes (Aq) as functions of the concentration of alkali chloride solutions for water/silica interfaces obtained by fitting the spectra in Figure 3.2 with Equation (3.2). ..................................................................................................................................32 Figure 4.1 SFG spectra of heptane/silica interfaces with dry (open circles) and watersaturated (solid circles) heptane. The peaks between 2800 and 3000 cm-1 are associated with C-H vibrations of heptane, the peaks at 3200 and 3400 cm-1 are vibrational peaks associated with hydrogen-bonded water, and the 3660/3680 cm-1 peak is assigned to the free O-H stretching mode. .................................................................................................42 ix  Figure 4.2 SFG spectra of toluene/silica interfaces with dry (open circles) and watersaturated (solid circles) toluene. The peaks between 2800 and 3100 cm-1 are associated with C-H vibrations of toluene, the broad peak centered at 3150 cm-1 is associated with hydrogen-bonded water, and the 3600 cm-1 peak is assigned to the free O-H stretching mode. .................................................................................................................................44 Figure 4.3 Adsorption of water on a regular water layer with free OHs. The open circles represent the SFG spectrum of a toluene/silica interface with dry toluene on an initially moistened silica surface. In this regular water layer, the 3200 cm-1, 3400 cm-1, and free OH peaks are all present. Water can condense on such a regular water surface, as indicated by the solid-circle spectrum, which was taken 30 min. after the toluene was saturated with water. .........................................................................................................45 Figure 4.4 Schematic drawing of the “water-resistant” water structure. In the process of water adsorption, the water molecules minimize their energy by forming the maximal number of hydrogen bonds with either the neighbouring water molecules or the silanol groups (SiOH) on silica. At a certain stage, very few free OHs are available. The absence of free OH groups significantly reduces further water adsorption. ..................................47 Figure 5.1 Panel (a) schematic showing the range of RH values over which uncoated and coated mineral dust particles are good ice nuclei. The figure is specifically created for illite particles based on the work of Chernoff et al. A similar behavior was also observed for kaolinite. The temperature applicable for the ice nucleation studies was approximately 237 K. Panel (b) and (c) show the pH and the molarity of the sulfuric acid coating, assuming the coating is in equilibrium with the relative humidity. ..................................52 Figure 5.2 SFG spectra of D2O/mica interfaces with D2SO4 concentrations of (a) 0 M, (b)  x  5x10-6 M, (c) 5x10-5 M, (d) 5x10-4 M, (e) 5x10-3 M, (f) 5x10-2 M, (g) 5x10-1 M and (h) 5 M. The inset shows the schematic layout of the spectroscopic setup. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. ................57 Figure 5.3 The fitted amplitudes (a) and frequencies (b) of water peaks in the SFG spectra with various D2SO4 concentrations. Triangles represent “ice-like” peaks and dots represent “liquid-like” peaks. ............................................................................................59 Figure 5.4 Schematics of water molecules and hydrated sulfate (bisulfate) ions at solution/mica interfaces with different concentrations of H2SO4 solutions. (a) Pure H2O/mica interface with pH~6. The surface is highly negatively charged. Water molecules are more ordered near the charged surface. (b)With a low concentration of H2SO4, the surface charge decreases as the surface is protonated, and the anions interact with the mineral surface. (c) With a high concentration of H2SO4 (for example 5M), water molecules are well captured by the sulfate/anions in the solution, and few water molecules are freely available for the mica surface. In (a) and (b) the dashed lines separate the ordered water molecules from the water molecules without order. Above the dashed lines, water molecules have good order because of negative surface potential induced by the mica surface. In (b) and (c) H+ adsorb on the mica surfaces. The dashed circles represent hydrated sulfate ions (in low H2SO4 concentration) or hydrated bisulfate anion (in high H2SO4 concentration). Water molecules in the dashed circle are parts of the hydrated anions and move together with the core anions. ..........................................63 Figure 6.1 Schematic layout of the spectroscopic setup. The frequency of the visible beam was fixed at 532 nm, and the frequency of the IR beam was tunable. The 532 nm and IR beams were overlapped both spatially and temporally on the top surface of the  xi  solution. The thickness of the solvent layer is 3 mm. .......................................................68 Figure 6.2 (A) SFG vibrational spectra of toluene in ssp and ppp polarization configurations. The ssp and ppp spectra are offset from each other by 0.5 arbitrary units for clarity. (B) Calculated  ( 2) χ ssp 2) χ (ppp  for CH3 symmetric (solid line) and asymmetric (dashed  line) modes as a function of the tilting angle. The orientational distribution of the CH3 groups was assumed to be a delta function. The solid circle indicates the measured  ( 2) χ ssp 2) χ (ppp  value of 4.4 which corresponds to a tilting angle of 25°. .................................................71 Figure 6.3 SFG vibrational spectra of toluene-pentane mixtures on silica with toluene B = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. ........................73 volume fraction φtoluene  Figure 6.4 SFG vibrational spectra of toluene-heptane mixtures on silica with toluene B volume fraction φtoluene = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. ........................74  Figure 6.5 SFG vibrational spectra of toluene-tetradecane mixtures on silica with toluene B = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. ......................75 volume fraction φtoluene  Figure 6.6 Adsorption isotherms of toluene on silica for binary mixtures of pentane– toluene (■), heptane–toluene (●), and tetradecane–toluene (▲). The solid curves are fitting curves using Equation (6.8). ...................................................................................78 Figure 7.1 Schematic layout of the spectroscopic setup. The frequency of the visible beam was fixed at 532 nm, and the frequency of the IR beam was tunable. The 532 nm and IR beams were overlapped both spatially and temporally at the bottom surface of silica plate. ........................................................................................................................87  xii  Figure 7.2 SFG vibrational spectra of water adsorbed on silica surface as a function of relative humidity (RH). .....................................................................................................90 Figure 7.3 Sequential images extracted from a video of bitumen liberation from silica at time equal to (a) 0, (b) 20, (c) 80 sec. ...............................................................................91 Figure 7.4 Measured dynamic contact angles for bitumen on (a) silica, (b) rinsed mica, and (c) cleaved mica. ...............................................................................................................92  xiii  Abbreviations  AFM  Atomic force microscopy  BS  Beam splitter  CHWP  Clark hot water process  DM  Dichromatic mirror  KTP  KTiOPO4  OPG/OPA  Optical parametric generator/amplifier  PMT  Photomultiplier tube  PZC  Point of zero charge  RH  Relative humidity  RM  Reflective mirror  SFG  Sum frequency generation  SHG  Second harmonic generation  STM  Scanning tunneling microscopy  YAG  Yttrium aluminum garnet  xiv  Acknowledgments  Foremost I would like to express my deep and cordial gratitude to my advisor, Professor Keng Chang Chou for his continuous support and encouragement throughout my PhD study in UBC. Professor Chou has always been an inspiring, kind and helpful mentor. His profound knowledge in laser and spectroscopy has been of great value to me and inspires me to be a better researcher. I enjoy every discussion with him, for advice from him always helps me overcome difficulties not only in research but also in daily life. His great insights in science, logical way in thinking and communication, optimism in life, kindness and patience as an advisor and courteous attitude to people enlighten me on characteristics that one should have to be successful. I hold sincere gratitude to Dr. Qi Feng Li, who has been in our group as a postdoctoral fellow since I just joined the research group. He offered me a lot of help in both science and life. His knowledge and skills on optics help me immensely during the experiments. I will never forget the accomplishments and failures we experienced together. I am also truly grateful to Rui Hua, who joined our group one year earlier than me as a graduate student. I really enjoyed the interesting discussions with him. And I will never forget his encouragement and support. I would like to thank former and current group members, Sherry Wu, Gwen Bailey, Bonnie Leung, Henry Tang, Zhen Wei Wang and Zhe Wang, for sharing their knowledge, for giving support and help.  xv  In the end, I would to thank Jingyi and my parents who experienced both my excitements and frustrations, for their warm support during the past years. I pray deeply for a healthy and enjoyable life for them.  xvi  Dedication  Dedicate this thesis to Jingyi and my parents.  xvii  Chapter 1  Introduction  1  1.1 Overview Surface phenomena have attracted the interest of numerous researchers in a wide range of disciplines. Much effort has been devoted to studying the properties of interfaces, from simple systems such as the adsorption of gases at inert surfaces, to more complicated systems such as liquid/liquid or solid/liquid interfaces. Solid/liquid interfaces, in particular, are of great importance in surface science. They are not only ubiquitous in both the environment and industry, but also highly relevant to numerous important processes, including cleaning, wetting, adhesion, lubrication, etching, corrosion, electrochemical reactions, and oil recovery.1 It is expected that the properties of liquid molecules on a solid surface are very different from those in the bulk.2,3 This is because the structure of the interfacial liquid can be more readily disturbed by the attractive/repulsive interactions between a solid surface and liquid molecules. Detailed investigations of solid/liquid interfaces have been hindered by a lack of effective experimental probes, which are needed to provide molecular-level information on the chemical composition and geometric structure of the functional species at buried liquid interfaces. High surface specificity is necessary to separate the signal of interfacial molecules from a huge amount of molecules in the bulk. Optical techniques such as attenuated total reflection spectroscopy, infrared spectroscopy, and ellipsometry have been used to probe buried liquid interfaces, but they are not intrinsically very surfacespecific and cannot distinguish signals from the bulk. Scanning tunneling microscopy (STM) and atomic force microscopy (AFM) may be applied to thin films of liquids on solid substrates but have difficulty in producing clear microscopic images of liquid surfaces because of molecular movement in most cases.4-6 Recently, nonlinear optical  2  spectroscopic techniques such as sum-frequency generation (SFG) have become the preferred techniques to study liquid interfaces.7-10 Being highly surface-specific and applicable to all interfaces accessible by light, SFG has been repeatedly demonstrated to be the most versatile and powerful analytical tool for investigating liquid interfaces. SFG, as a second-order nonlinear optical process, is forbidden under the electricdipole approximation in the bulk media with inversion symmetry but allowed at surfaces or interfaces where the inversion symmetry is broken.11 Therefore, this technique is highly interface-specific with submonolayer sensitivity. Currently, SFG is the only technique that can provide vibrational spectra of buried liquid interfaces. In this dissertation SFG is used to study the structures and properties of water and hydrocarbons at various liquid/mineral interfaces.  3  1.2 Water Structures on Water/Mineral Interfaces The water/mineral interface is of great importance in surface science since it is ubiquitous in both the environment and industry and arises in many processes such as contaminant migration, ice nucleation, soil formation, microorganism growth, lubrication, and oil recovery.1,12-19 Generally, when water molecules encounter a solid surface, they will try to reorient to form as many hydrogen bonds as possible to minimize their energy, which leads to an ordered structure adjacent to the surface. For example, water structures at a hydrophilic solid surface, such as silica, are understood to be a mixture of a more ordered and a less ordered hydrogen-bond network, which are associated with vibrational peaks at 3200 and 3400 cm-1, respectively.20 The 3200 cm-1 peak is due to water molecules with a full complement of hydrogen bonds in a tetrahedral arrangement and has often been referred to as “ice-like” since it exhibits a vibrational peak position similar to the solid phase of water, ice. The peak at 3400 cm-1 is broad and assigned to water molecules with either asymmetric hydrogen bonds or bifurcated hydrogen bonds and is called “water-like” because the resonance is found at approximately the same frequency as the infrared absorption of water molecules in the bulk liquid.10,20 The detailed interfacial water structures depend on the property of the surface and the aqueous solution, such as the hydrophilicity/hydrophobicity of the surface, surface charge, the species and concentration of ions in solution and so on.10,21 Chapters 3, 4 and 5 of this dissertation present studies of liquid/mineral interfaces relevant to petrolic, edaphic and atmospheric processes, which include the water structures at aqueous salt solutions/silica interfaces, solvent/silica interfaces, and high concentration sulphuric acid solution/mica interfaces.  4  1.3 Solvent Adsorption on Minerals In contrast to the vast amount of studies focusing on the structures of water at the interfaces,10,20 there are fewer works studying solvents at solid interfaces using SFG.8,22,23 An important conclusion from previous works is that the polarity of the interfacial region plays a key role in controlling the structure and orientation of adsorbed organic molecules.24-27 If there is more than one species of organic solvent existing in the liquid phase, they will compete for adsorption at the solid interface. The competitive adsorption of organic solvents (usually hydrocarbons) at liquid/solid interfaces is important for many industrial and scientific processes, such as oilsands processing, petroleum recovery, contamination removal, heterogeneously catalyzed reactions, thin film materials production and many extraction techniques.28-38 Generally the competitive adsorption depends on the molecular properties of solvents, their respective mole fractions, and the strength of attractions to the solid surface, so at a binary solution/solid interface, it is expected that the surface chemical composition is different from the bulk composition because of the different interaction strength with the surface.39 In many cases, the competitive adsorption of solvents at liquid/solid interfaces is a critical factor in determining the effectiveness of a technological process. In the 1950’s and 1960’s, the adsorption isotherms for binary mixtures at liquid/solid interfaces were studied by various immersion methods, and a number of theories were developed to elucidate the thermodynamic properties of pure liquids and binary mixtures over solid surfaces.40,41 Despite this effort, the problem was not completely resolved because the macroscopic measurements were indirect, and the theories required molecular-level information about adsorbates as input parameters.41,42 Even with modern technologies, it  5  remains challenging to selectively probe the interfacial region between a liquid and a solid and to directly measure the surface coverage of a particular component at a liquid/solid interface. SFG provides a new opportunity to directly measure the adsorption isotherms at solvent/mineral interfaces.  6  1.4 Water’s Role in Oil Sands Processing Oil sands (also known as bituminous sands or tar sands) are found in large amounts in many countries throughout the world. It is estimated that there are at least 170 billion barrels (27×109 m3) of bitumen that can be recovered from the Athabasca oil sand deposits in Canada.43 The oil sands contain naturally occurring mixtures of sand, clay, water, and a dense and extremely viscous form of petroleum referred to as bitumen. At present, the technology used to extract bitumen from the oil sands deposits is a variation on the Clark hot water process (CHWP) where a lower processing temperature (35-55 °C) is used.44 The oil sands conditioning stage and bitumen recovery step include: (1) bitumen displacement along a sand grain, (2) bitumen detachment, (3) bitumen droplet attachment to an air bubble and (4) bitumen flotation in separation vessels. A prerequisite to the bitumen extraction process is liberation of bitumen from the mineral surface, which requires a huge amount of water. Therefore, understanding the mechanisms behind the displacement of oil by water on minerals is important for bitumen recovery. Many investigations have been done using dynamic contact angle measurements to study bitumen displacement and detachment from solid surfaces in the presence of water containing salt, surfactants and clays at different temperatures and pH values.45-48 Previous studies have also suggested that the interfacial water content between mineral surfaces and bitumen is one of the most important factors affecting bitumen recovery.49 However, the detailed mechanism remains poorly understood. To study the mechanism, both macroscopic (liberation rate, dynamic and static contact angle) and microscopic (water amount at mineral surfaces under different relative humidity, structures of water molecules adsorbed on mineral surfaces) information is  7  needed. The study in chapter 7 provides a molecular-level understanding for the effect of interfacial water on the liberation process.  8  Chapter 2  IR-visible Sum Frequency Generation  9  2.1 Introduction Since the first experimental demonstration of surface vibrational spectroscopy via infrared-visible sum-frequency generation (SFG) published in 1987,50 SFG spectroscopy has become a versatile spectroscopic technique to study the properties of various surfaces and interfaces. SFG is extremely surface-specific and sensitive, because as a secondorder nonlinear optical process, SFG is forbidden in a medium with inversion symmetry under the electric-dipole approximation, but is allowed at a surface or interface where the inversion symmetry is broken. In the IR-visible SFG process, two pulsed laser beams (one visible with frequency ωVis and one infrared with frequency ω IR ) are overlapped spatially and temporally at an inferface, and an SFG with a frequency of ωSF = ωvis + ω IR is generated. The intensity of SFG is enhanced when the frequency of the IR beam is resonant with the vibrational modes of molecules at the interface. Detailed theoretical descriptions of SFG process are available in the literature.51-53 In this chapter, a brief introduction of SFG technique is presented.  10  2.2 Theory of Sum Frequency Generation A typical SFG setup in a reflection geometry is shown in Figure 2.1. The visible and IR beams, with frequencies ωvis and ω IR respectively, are overlapped in time and space at an interface. The SFG beam generated at the interface has a frequency of ω SF = ω vis + ω IR , dictated by the conservation of energy between the incoming and outgoing photons.  SFG  Visible  IR  Medium 1 Medium 2  Figure 2.1. SFG setup in the reflection geometry.  Conservation of momentum (phase-matching condition) is also required for the photons involved in the SFG process: k SF = k vis + k IR  (2.1)  where k vis and k IR are the wave vectors of the incident beams. By combining the conservations of energy and momentum, we can determine the direction of the SFG beam by the relation: n SF k SF sin θ SF = nvis k vis sin θ vis + n IR k IR sin θ IR  (2.2)  11  where θ i is the angle of the indicated light to the surface normal and k i is the absolute value of the wave vector of each light. As a nonlinear optical process, the second-order nonlinear polarization P ( 2 ) (ωSF ) , which generates the SFG output, and the two incident electric fields E (ωvis ) and E (ω IR ) are related as follows: P ( 2 ) (ωSF ) = ε 0 χ ( 2 ) : E (ωvis ) E (ω IR )  (2.3)  where χ ( 2 ) is known as the second-order susceptibility and ε 0 is the electric permittivity of free space. The SFG intensity is proportional to the square of P ( 2 ) (ωSF ) , therefore the measurable SFG intensity I SF is proportional to the intensities of the two incident laser (2) beams and the square of the effective surface second-order susceptibility χ eff .  The sum frequency intensity in the reflected direction is given by:54 I SF =  2 ωSF  2  8ε 0 c cos θ SF 3  2  χ eff( 2 ) I vis I IR  (2.4)  with  χ eff( 2 ) = [L(ωSF ) ⋅ eˆ SF ] : χ ( 2 ) : [L(ωvis ) ⋅ eˆvis ][L(ωIR ) ⋅ eˆ IR ]  (2.5)  Here, ωi is the frequency of the visible, IR and SFG beams, eˆi is the unit polarization vector of the optical field at ωi , and L(ωi ) is the tensorial Fresnel factor at frequency ωi . The macroscopic second-order susceptibility χ ( 2 ) is a third rank tensor which consists of 27 elements χ ijk( 2 ) . The values of the components are the property of the medium and are invariant under symmetry operations. The number of non-zero χ ( 2 )  12  elements is often reduced because of symmetry constraints. For example, in a bulk solution with a centrosymmetric environment, χ ( 2 ) is invariant under inversion symmetry, but the electric field and polarization must change signs as they are vectors. Based on equation (2.3), the inversion operation gives: χ ( 2 ) = − χ ( 2 ) . So χ ( 2 ) must equal to 0 in a centrosymmetric medium. This is the reason why SFG is forbidden in a medium with inversion symmetry. However, the inversion symmetry is naturally broken at the surface. There are only four nonequivalent and non-vanishing χ ( 2 ) values for an isotropic surface. With the lab coordinates chosen such that z is along the interface normal and x is ( 2) ( 2) ( 2) ( 2) ( 2) ( 2) in the incident plane, the non-vanishing χ ( 2 ) are: χ xxz , χ xzx , χ zxx , = χ yyz = χ yzy = χ zyy ( 2 ) 52 and χ zzz . These four components can be deduced by measuring SFG with four different  input and output polarization combinations: ssp, (the sum frequency, visible, and infrared beams are s-polarized, s-polarized, and p-polarized, respectively), sps, pss, and ppp.  χ ijk( 2 ) is the macroscopic orientational average of the microscopic molecular (2) hyperpolarizability β lmn of all the interfacial molecules: (2) χ ijk( 2 ) = N s ∑ β lmn (iˆ ⋅ lˆ )( ˆj ⋅ mˆ )( kˆ ⋅ nˆ )  (2.6)  l ,m ,n  N s denotes the number density of the interfacial molecules. The subscripts i, j, and k refer to the lab coordinates, and the subscripts l, m, and n, refer to the axes for the molecular coordinate system. (iˆ ⋅ lˆ)( ˆj ⋅ mˆ )( kˆ ⋅ nˆ ) is the coordination transformation from  molecular fixed coordinates to laboratory fixed coordinates, and  indicates an average  over the molecular orientation distribution. (2) (2) and βlmn consist of a sum of resonant terms and a nonresonant term: Both χ ijk  13  (2) χ ijk( 2 ) = χ NR ,ijk + ∑  υ  Aυ ,ijk  ω IR − ωυ − iΓυ  (2) (2) = β NR β lmn ,lmn + ∑  υ  βυ ,lmn ω IR − ωυ − iΓυ  (2.7)  (2.8)  In the above equations Aυ ,ijk , ωυ and Γυ are the amplitude of the transition moment, frequency of the transition and line width of the vibrational mode υ respectively. The transition moment amplitude βυ ,lmn is related to the Raman and IR properties of the vibrational mode, βυ ,lmn is nonzero only when both the Raman polarizability tensor and the IR transition moment are nonzero. Therefore, a SFG active vibrational mode must be both Raman and IR active. When the frequency of a vibrational ωυ is resonant with the IR frequency ω IR , the denominator in Equation (2.7) becomes small and χυ( 2,ijk) will be enhanced.  14  2.3 Fresnel Factors  For isotropic surfaces, SFG experiments can be conducted in different input and output polarization combinations such as ssp, sps, pss, and ppp. Equation (2.5) can be rewritten for an isotropic interface:  χ eff( 2 ),ssp = Lyy (ωSF ) Lyy (ωvis ) Lzz (ωIR ) sin θ IR χ yyz  χ eff( 2 ),sps = Lyy (ωSF )Lzz (ωvis )Lyy (ωIR ) sin θ vis χ yzy χ eff( 2 ), pss = Lzz (ωSF ) L yy (ωvis ) L yy (ω IR ) sin θ SF χ zyy  (2.9)  χ eff( 2 ), ppp = − Lxx (ωSF ) Lxx (ωvis ) Lzz (ω IR ) cosθ SF cosθ vis sin θ IR χ xxz − Lxx (ωSF ) Lzz (ωvis ) Lxx (ω IR ) cosθ SF sin θ vis cosθ IR χ xzx + Lzz (ωSF ) Lxx (ωvis ) Lxx (ω IR ) sin θ SF cosθ vis sin θ IR χ zxx + Lzz (ωSF ) Lzz (ωvis ) Lzz (ω IR ) sin θ SF sin θ vis sin θ IR χ zzz  The Fresnel factors depend on the refractive indices of the media and the t experimental geometry. For isotropic surfaces, only the diagonal elements of L (ω i ) need  to be considered: Lxx (ωi ) =  2n1 (ωi ) cosγ i n1 (ωi ) cosγ i + n2 (ωi ) cosθ i  L yy (ωi ) =  2n1 (ωi ) cosθ i n1 (ωi ) cosθ i + n2 (ωi ) cosγ i  2n2 (ωi ) cosθ i Lzz (ωi ) = n1 (ωi ) cosγ i + n2 (ωi ) cosθ i  (2.10) ⎛ n1 (ωi ) ⎞ ⎜⎜ ⎟⎟ n ' ( ω ) i ⎠ ⎝  2  15  In the above equations, ni is the refractive index of medium i as shown in Figure 2.1, θ i is the incident angle, and γ i is the refracted angle. n' (ω i ) is the refractive index of the interfacial layer. Since the interfacial layer is only one or a few monolayers thick, n' (ω i ) can be difficult to measure.55 It is therefore an usual practice that n' (ω i ) is chosen to be the bulk refractive index of the material at the interface.  16  2.4 Molecular Orientation at the Interface  As discussed above, SFG measures the macroscopic second-order susceptibility χ ( 2 ) which is related to the microscopic molecular hyperpolarizability β ( 2 ) through  average molecular orientation as shown in Equation (2.6). β ( 2 ) is an intrinsic property of a molecule and the components of β ( 2 ) can be determined by the symmetry of the molecule. If β ( 2 ) is known, then the average orientation of the functional group can be deduced from the measurement of χ ( 2 ) using Equation (2.6). For example, the methyl group has C3v symmetry and the symmetric stretch mode is cylindrically symmetric with symmetry axis along  cˆ , so there are only two  nonvanishing independent components in β ( 2 ) : β aac = β bbc and β ccc . Therefore, from Equation (2.6) we find for an azimuthally isotropic surface:56  χ xxz ,s = χ yyz ,s = N s β ccc [〈cosθ 〉(1 + r ) − 〈cos3 θ 〉(1 − r )] 1 2  χ xzx ,s = χ yzy ,s = χ zxx ,s = χ zyy ,s = N s β ccc [〈cosθ 〉 − 〈cos3 θ 〉 ](1 − r ) 1 2  (2.12)  χ zzz ,s = N s β ccc [r 〈cosθ 〉 + 〈cos 3 θ 〉(1 − r )] Here β aac = rβ ccc and r is the hyperpolarizability ratio which is equal to 2.3 for the methyl group.52,57 θ is the tilt angle which is the angle of the molecular symmetry axis respect to  zˆ  cˆ  with  in the laboratory coordinate system as shown in Figure 2.2:  17  Figure 2.2. The definition of orientation angle for methyl group. Orientation angle (θ) is between surface normal and methyl group, axis c is the principal axis of the methyl group. The ab plane is perpendicular to axis c. The ac plane contains one C-H bond. The absolute value of N s is difficult to determine, so we can cancel it by measuring the ratios of independent nonvanishing χ ( 2 ) components. Then we can find the orientation θ by assuming a δ -function distribution for θ.  18  2.5 Experimental Instrumentation  As shown in Figure 2.3, the SFG experiments were carried out with a visible 532 nm beam and an IR beam tunable from 1800 to 4000 cm-1, which were generated using a Nd:YAG (yttrium aluminum garnet) laser system (1064 nm, 10 Hz, and 25 ps). The 532 nm beam was produced in a KTiOPO4 (KTP) crystal by second harmonics generation (SHG) from the fundamental 1064 nm beam. A portion of the 532 nm beam was used to pump a home-made KTP optical parametric generator/amplifier (OPG/OPA) and a KTiOAsO4 (KTA) difference frequency generation (DFG) system which mixes the idler output of the OPG/OPA system and the fundamental 1064 nm beam. Both the visible and IR beams had a pulse duration of 25 ps and a repetition rate of 10 Hz. The input angles were set to 45° and 56° with respect to the surface normal for the visible and the IR beams, respectively. The input energy was ~250μJ/pulse for the visible beam with a spot size of 2mm on the sample. For the IR beam the input energy was ~200μJ/pulse with a spot size of 0.5mm on the sample. The IR and visible input beams were overlapped both spatially and temporally at the sample. The SFG output passed through a series of bandpass filters to eliminate the background light, and then was detected by a HAMAMATSU R3896 photomultiplier tube (PMT). The signal intensity was recorded by a gated integrator (SR 280, Stanford Research Systems Inc.) and digitized by a computer. The frequency of IR beam was calibrated by the absorption lines of polystyrene film (Thermo Electron Corp.). All SFG spectra presented in this dissertation were normalized against the SFG spectrum of a z-cut quartz.  19  RM  532 nm RM  RM  SHG Nd: YAG laser  RM  25 ps, 10 Hz, 40mJ/pulse  DM  KTP  OPG/OPA KTP KTP  BS BS 532 nm  RM DM  1300~1800 nm 1064 nm  RM  RM  KTA DFG  Tunable IR 2000~4000 cm-1  Figure 2.3. Layout of OPG/OPA. DM, RM and BS represent the dichromatic mirror, reflective mirror, and beam splitter, respectively.  20  Chapter 3  Structures of Water Molecules at the Interfaces of Aqueous Salt Solutions and Silica: Cation Effects*  * A version of this chapter has been published. Yang, Z.; Li, Q. F.; Chou, K. C. Structures of Water Molecules at the Interfaces of Aqueous Salt Solutions and Silica: Cation Effects. Journal of Physical Chemistry C 2009, 113, 8201-8205.  21  3.1 Introduction  Buried aqueous interfaces play an important role in many natural and industrial processes. Among the various aqueous interfaces, water/silica interfaces, which affect contaminant migration, ice nucleation, soil formation, and microorganism growth, have been of great interest.12-19 Since it is experimentally challenging to probe a buried interface, current understanding of buried aqueous interfaces remains limited. For this reason, a large number of theoretical simulations of water interfaces have been carried out to provide microscopic information, such as the density profiles and orientations of water molecules at interfaces.4,58-66 Although the calculations provide qualitative information, the simulation results are often limited by the capacity of computers, which are insufficient for determining the detailed structure of interfacial water under practical conditions involving large numbers of molecules. Previously, the structure of water molecules at buried aqueous interfaces were studied using X-ray spectroscopy,67-69 electron diffraction,70 second harmonic generation (SHG),71 and sum-frequency generation (SFG) vibrational spectroscopy11,13,72. Generally, water structures at a hydrophilic solid surface, such as silica, are understood to be a mixture of a more ordered and a less ordered hydrogen-bond network, which are associated with vibrational peaks at 3200 and 3400 cm-1, respectively.20 In any case, the detailed structure of water molecules on silica is not yet completely understood. Experimentally, the structure of water on silica has been shown to depend on the surface charges. Silica surfaces possess negative charges because of the deprotonation of surface silanol groups (SiOH). The surface charges, which create a surface electric field, induce polar ordering of interfacial water molecules. Ong et al. studied water/silica  22  interfaces with various pH values using second harmonic generation (SHG).71 They concluded that water molecules near the silica interface were polarized by the interfacial electric field and responsible for the observed SHG. A few years later, Du et al. studied OH vibrations of water molecules at water/fused quartz interfaces using IR-visible SFG vibrational spectroscopy.72 Their results showed that the orientation of the OH bonds and the ordering of interfacial water molecules are strongly affected by their electrostatic interaction with the deprotonated surface silanol groups. In a recent study on water/quartz interfaces, Shen and his coworkers indicated that two different surface sites exist with different deprotonation pK values on crystalline quartz.73 The peak at 3200 cm-1 seemed to be associated with surface sites that have higher pK values, and the peak at 3400 cm-1 was closely associated with surface sites having lower pK values. The detailed mechanism for the formation of these two different sites is not yet understood. Little is known about the structure of interfacial water molecules under perturbations by cations. Alkali cations are particularly important, as they are the most abundant cations in natural water. Previous studies at air/water interfaces with NaCl solutions showed that a reduced ion density was present near the surface, and the ions produced little effect on the surface water structure, for molar fractions up to 0.036 (~ 2 M).74,75 The environment of an air/water interface is very different from that of a water/silica interface, because cations can interact with the silica surface via electrostatic interactions.76 In this paper, we present studies of water structures on silica surfaces using IR-visible sum frequency generation vibrational spectroscopy. Significant perturbation of the water structures was observed with a relatively low concentration of NaCl in water. Further, different alkali cation species (Na+, Li+, and K+) showed different degrees of  23  impact on the interfacial water structures. The electrostatic interactions between the hydrated cations and the silica surface as well as the effective ionic radii of the cations need to be considered to explain the observed phenomena.  24  3.2 Experimental Section  The visible and tunable IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a home-made KTP optical parametric generator/amplifier (OPG/OPA) pumped by the 532 nm beam. The 532 nm and IR beams were overlapped, both spatially and temporally, on the sample, as shown in Figure 3.1. The laser fluence was approximately 2 mJ/cm2 per pulse for the visible beam and 5 mJ/cm2 per pulse for the IR beam. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. The SFG intensity was detected by a photomultiplier tube after spatial filtering by an aperture, and spectral filtering by a bandpass filter. The SFG intensity was normalized against that from z-cut quartz. Each spectrum shown in the current study was an average of 10 scans in a 10 cm-1 step, and each scan was obtained by averaging the SFG intensity of 40 laser shots at each step. Fused silica plates, with a thickness of 3mm, were cleaned with a commercial cleaning agent (extran AP12) for 3 min. It was immersed in a 50/50 (v/v) HNO3/H2SO4 solution for ~12 hours, followed by rinsing in pure water (resistivity > 18.2 MΩ⋅cm, Millipore). Alkali chloride salts (> 99.8%, certified ACS reagents; purchased from Sigma Aldrich) were used to prepare solutions with different concentrations. No organic contamination in the salt solutions was observed in the SFG spectra between 2700 cm-1 and 3000 cm-1. All data presented in the current study were collected within a period of  25  two weeks. During the experimental period, the silica substrates were mostly kept either in air or acidic solutions to avoid surface quality changes due to prolonged exposures to pure water, as previously reported by Li et al.77 The SFG spectrum of the pure water/silica interface was also monitored at the beginning and the end of the experiment for each electrolyte to ensure that the quality of the silica surface stayed consistent during the experimental period.  Figure 3.1. Schematic layout of the SFG vibrational spectroscopy. The frequency of the visible beam was fixed at 532 nm and the frequency of the IR beam was scanned from 2800 – 3900 cm-1. The 532 nm and IR beams were overlapped both spatially and temporally on the bottom surface of the solution.  After the pure water/silica SFG spectrum was measured, the SFG spectra of a series of NaCl solutions with concentrations of 1x10-4 M, 5x10-4 M, 1x10-3 M, 1x10-2 M, and 1x10-1 M were measured in the sequence from the lowest concentration to the highest concentration. For each solution with a particular concentration, the cell and silica plates were rinsed thoroughly with the solution before the spectroscopic measurements. The 26  rinsing process ensured that the bulk solution in the cell had the desired electrolyte concentration. For each concentration, five scans were collected in a time period of ~1 hour during which no change of the SFG spectrum was observed. Then the cell and silica plates were cleaned with acids as described above for measuring next electrolyte solutions (LiCl, for example). The measurements of different electrolytes were carried out using the same silica substrate within a period of one week, and the experiments were repeated under the same condition in the second week with freshly prepared solutions and the same silica substrate. There was no observable difference in the SFG spectra in comparison to those collected in the previous week. It also confirmed that the surface quality of the silica substrate had stayed consistent during the whole experiments. Finally, the 10 spectral scans of the same electrolyte with the same concentration (five scans in the first experiment and five scans in the repeated experiment) were averaged to improve the signal-to-noise ratio. As the pH values of solutions can affect the surface water structure, great attention has been made to monitor the pH values of the solutions to assure that the observed changes of SFG spectra were not a pH effect. The pH value of water was 7 when freshly obtained from the Milli-Q system. However, it is known that water exposed to air is mildly acidic because water readily absorbs carbon dioxide from the air. It ultimately leads to a pH of approximately 5.7.78,79 All salts studied in this paper, such as LiCl, NaCl, and KCl, are known as neutral salts, which cause little change of the pH values in aqueous solutions. The pH values of all solutions used in the current study were around 5.74 with a standard deviation of 0.05. A table presenting the pH value for each individual solution was given in the supporting document.  27  3.3 Results and Analysis  Figure 3.2A shows the SFG spectra of water/silica interfaces with pure water and aqueous NaCl solutions of 1x10-4 M, 5x10-4 M, 1x10-3 M, 1x10-2 M, and 1x10-1 M. The spectra of water exhibit two peaks at 3200 cm-1 and 3400 cm-1.11,72 As the concentration of NaCl increased, the intensity of both peaks decreased. It has been interpreted that the peak at 3200 cm-1 represents OH in a more ordered hydrogen-bond network, and the peak at 3400 cm-1 represents a less ordered hydrogen-bond network.10,20,80 It is known that silanol groups on silica play a critical role in determining the molecule adsorption on the surface in an aqueous solution.76 The dissociation of protons from the surface silanol groups creates negative charges on a silica surface. The reaction can be described as follows: –  -SiOH + mH2O = -SiO -- mH2O + H+  (3.1)  where -SiOH is the surface silanol group and m describes the number of water molecules associated with the -SiO–.81,82 The silanol groups are estimated to have a surface density of ~5 × 1014 cm-2 on the silica surface, which is equivalent to one silanol group per 20 Å2.76,83 When the pH value of the aqueous solution increases, the silanol groups become deprotonated, and the surface charge increases. Experimentally, both the 3200 cm-1 and 3400 cm-1 peaks were observed to increase with increasing pH value, suggesting that a larger surface electrical field induces a better ordered hydrogen-bond network on silica.11,73,84 Previous studies by Ong et al., using second harmonic generation (SHG), showed that there were two types of silanol groups at a water/silica interface, with pK values of 4.9 and 8.5, populating 19% and 81% of the surface area, respectively.71 Further studies by Ostroverkhov et al., using a phase-sensitive SFG technique, indicated  28  that the 3200 cm-1 peak is associated with surface sites that have a higher pK value, and the 3400 cm-1 peak is associated with surface sites with a lower pK value.73 To obtain quantitative information, the spectra in Figure 3.2 were fitted by two Lorentzian peaks at 3200 cm-1 and 3400 cm-1: I (ω SFG ) ∝ χ  (2) NR ,ijk  +  ∑ω  q =1, 2  Aq ,ijk IR  2  −ω q +iΓ q  (3.2)  ( 2) is the nonresonant contribution; ω IR is the frequency of the input infrared where χ NR  laser beam; q represents the qth vibrational mode; Aq is the amplitude; Γ q is the width; and ω q is the resonant frequency. The fitting curves are shown in Figure 3.2 as solid lines, and the fitted amplitudes Aq are plotted in Figure 3.3. Parts B and C of Figure 3.2 show the SFG spectra collected with LiCl and KCl solutions. The magnitudes of the decreases are different when the cation species are changed. Similar experiments with different anions, such as NaBr and NaI, were also carried out, but, within measurement error, the spectra of NaBr and NaI were the same as those of the NaCl solutions. (Spectra of NaBr and NaI solutions are not shown.) Therefore, the cation, which interacts with the silica surface via electrostatic interactions, is the key factor in perturbing the hydrogen-bond network at the water/silica interface.  29  Figure 3.2. SFG spectra from the water/silica interfaces with different concentrations of three alkali chloride solutions: (a) NaCl, (b) LiCl, and (c) KCl. The solid lines are fitting curves derived using two Lorentzian peaks at 3200 cm-1 and 3400 cm-1.  30  The interaction of a hydrated cation with a silica surface has been shown to promote negative charge development.76 The reaction can be described as: –  -SiOH + nH2O·M+ = -SiO -- nH2O·M++ H+  (3.3)  where M+ denotes a cation and n describes the number of water molecules solvating M+. –  The notation, -SiO -- nH2O·M+, indicates that cations are located at a small but finite distance from the silica surface. Overall, the surface charges created by the cations are mostly neutralized by the cations themselves. Therefore, the surface charges developed by the cations are not expected to enhance the ordering of surface water molecules, or the observed SFG intensity. The pH of the point of zero charge for amorphous silica is about 2 - 3.85 With a –  solution pH value of 5.7, the silica surface is negatively charged. The surface -SiO groups, described in Equation (3.1), interact with the cations via electrostatic interaction:76 –  –  -SiO -- mH2O + nH2O·M+ = -SiO -- nH2O·M++ mH2O  Kassc  (3.4)  where Kassc is the equilibrium constant. As a result of the electrostatic interaction, the cations reduce the surface electrical field. Overall, the ordering of the original hydrogenbond network is perturbed because of the decrease in the surface electrical field and the replacement of the more ordered water molecules by the hydrated cations. Qualitatively, this mechanism explains that the SFG intensity of interfacial water peaks decreases as the salt concentration increases.  31  Figure 3.3. Amplitudes (Aq) as functions of the concentration of alkali chloride solutions for water/silica interfaces obtained by fitting the spectra in Figure 3.2 with Equation (3.2).  32  As shown in Figures 3.2 and 3.3, for the same concentration, the SFG intensities of interfacial water in alkali chloride solutions are in the order: Na+ > Li+ > K+. To explain the different behaviors for Li+, Na+, and K+, both the equilibrium constant Kassc in Equation (3.4) and the effective ionic radii of the hydrated cations need to be considered. First, the equilibrium constant Kassc describes the interaction between the surface -SiO– groups and the hydrated cation nH2O·M+. Previous studies by Dove et al. showed that the equilibrium constant Kassc for alkali chloride solutions is in the order: Kassc, KCl > Kassc, NaCl > Kassc,LiCl.86 A larger equilibrium constant Kassc indicates a higher density of “-SiO–-nH2O·M+” on the silica surface; thus, a larger perturbation of the water structures, and consequently, a smaller SFG intensity. Therefore, if only the equilibrium constant Kassc is considered, the SFG peak intensity would be in the order: Li+ > Na+ > K+. Second, the sizes of the hydrated cations should be considered. Ions in aqueous solution are hydrated. Generally, the smaller and higher-charged ions attract more water molecules. The hydrated ionic radii of Li+, Na+, and K+ are approximately 0.6 nm, 0.4 nm, and 0.3 nm, respectively.87 In an aqueous solution, the radius of a hydrated Li+ is roughly twice that of a hydrated K+. Therefore, a hydrated Li+ displaces more ordered H2O at the surface. Consequently, if only the effective ionic radii are considered, one would expect that the SFG intensity of water be in the order: K+ > Na+ > Li+. Overall, the balance between these two effects gives the observed SFG intensity in the order: Na+ > Li+ > K+. The SFG intensity is lower for K+ and Li+ solutions because K+ has a stronger interaction with the surface -SiO– groups and Li+ has a larger effective ionic radius. Nevertheless, these two effects, even though they perturb the interfacial water ordering in different ways, are not totally independent. The size of a hydrated cation is determined by the electrostatic  33  interaction between the cation and water molecules, and the strength of the interaction between the hydrated cation and the silica surface depends on the size of the hydrated cation.88-91 The structures of the water of hydration remain as an active research area, and many theoretical studies of the detailed structures can be found in the references. 92-97 As shown in Figure 3.3, the peak at 3200 cm-1 is more vulnerable to the cation perturbation. The amplitude of the 3200 cm-1 peak experiences a ~ 20% decrease with a concentration of 1x10-4 M, while the 3400 cm-1 peak does not have an significant decrease until 1x10-2 M. Additionally, the perturbation for the 3200 cm-1 peak reaches its saturation at a concentration of 0.01 M, while the perturbation for the peak at 3400 cm-1 is not saturated until 0.1 M. Eventually, both peaks have a 50% decrease in their amplitudes  Aq , which is equivalent to a 75% decrease in the measured SFG intensity. Since the peak at 3200 cm-1 is more vulnerable to the cation perturbation, the water structure associated with this peak is more likely to exist in the region where the surface electrical field is higher. As described above, cations interact with the silica surface via electrostatic interaction, and therefore, the surface number density of cations is higher in the region where the surface electrostatic field is higher. Previous studies have suggested that the silanol groups with lower pK values are isolated silanol groups because they are relatively easier to dissociate.71,76,98 Vincinal silanol groups, which locate closely and can be coupled to each other through hydrogen bonds, have higher pK values. The vincinal silanol groups are likely to create a higher local surface charge density because of their higher local number density. Therefore, vincinal silanol groups are more likely to create a higher local electrical field on the silica surface, consequently a better ordered structure and attract more cations to the surface. This model is consistent  34  with previous observations by Ostroverkhov et al., indicating that the peak at 3200 cm-1 is associated with surface sites that have a higher pK value.73  35  3.4 Conclusions  IR-visible sum frequency vibrational spectroscopy was applied to study the structure of water molecules at water/silica interfaces with the presence of alkali chloride in the solutions. Significant perturbations of the interfacial water structures were observed with a 1x10-4 M NaCl solution. The cations play a key role in perturbing the hydrogen-bond network at the water/silica interfaces as they interact with the silica surface via electrostatic interaction. Different alkali cation species produce different degrees of perturbation in the order: K+ > Li+ > Na+. This order can be explained by considering the electrostatic interaction between the cations and silica surfaces and the effective ionic radii of the cations. The peak at 3200 cm-1 experiences lager perturbation suggesting that the more ordered structure at 3200 cm-1 is associated with the vincinal silanol groups, which produce a higher surface electrical field and have a higher pK value compared to isolated silanol groups.  36  Chapter 4  Structures of Water Molecules at Solvent/Silica Interfaces*  * A version of this chapter has been published. Yang, Z.; Li, Q. F.; Gray, M. R.; Chou, K. C. Structures of Water Molecules at Solvent/Silica Interfaces. Langmuir 2010, 26, 16397-16400.  37  4.1 Introduction  Surface chemistry at solvent/mineral interfaces plays a critical role in many industrial and environmental processes, such as oil sand processing28,29,34,38, petroleum recovery32, contamination removal30,33,35,37, and many other extraction techniques31,36. Water molecules, either existing naturally or being added in a process, form a thin layer on the mineral surface and may affect the effectiveness of the process. For example, previous studies have suggested that the loss of water from the mineral interfaces can significantly decrease the yield of oil sand extraction.99 Despite its importance, the behaviour of water molecules at solvent/mineral interfaces is poorly understood, mainly because it is technically challenging to probe the buried interfaces. Recent developments in sum frequency generation (SFG) vibrational spectroscopy have proven that SFG is an effective technique to study buried liquid interfaces.20,72 Previous studies at water/silica interfaces using SFG have shown that the structure of water is strongly affected by the surface charges on silica.84 In addition to the liquid-water vibrational peak at 3400 cm-1, an “ice-like” peak located at 3200 cm-1 was observed at water/silica interfaces. The “icelike” peak grows significantly with higher surface charge density, suggesting that the surface charges on silica induce a better ordered water network with increased hydrogen bonds. On the other hand, the SFG spectra of water at hydrophobic interfaces exhibit a narrow peak near 3650 cm-1, which is associated with the free OH dangling bonds, and a broad figure between 3000 and 3600 cm-1, which is associated with hydrogen-bonded OH. Studies by Scatena et al. at hexane/water interfaces have shown that the hydrogen bonding between water molecules at the interface is weaker than that at air/water interfaces.100 Therefore, the structure of water at solvent/silica interfaces will depend on  38  the interactions of water with the hydrophilic mineral surfaces and the hydrophobic solvents. In this paper, the adsorption of water at toluene/silica and heptane/silica interfaces was studied using SFG. Toluene and heptane were chosen because of their popularity in laboratories and industries. The properties of toluene and heptane are also fundamentally interesting for a comparison, as toluene is aromatic and heptane is a straight-chain alkane. Silica was used in the current study because of its abundance in nature. The surface chemistry of silica is important for oil sand processing28,29,34,38 and liquid chromatography101-104. We observed that the structure of water adsorbed at the solvent/silica interface was strongly affected by the organic solvent. We discovered a highly hydrogen-bonded water structure at the toluene/silica interface with no detectable free OHs. Surprisingly, this structure of water showed resistance against further adsorption of water and was extremely stable at room temperature. On the other hand, a similar structure was not observed at heptane/silica interfaces.  39  4.2 Experimental Section  The visible and tunable IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a KTP optical parametric generator/amplifier (OPG/OPA) pumped by the 532 nm beam. The 532 nm and IR beams passed through the silica window and were overlapped, both spatially and temporally, at the solvent/silica interface. The laser fluence was approximately 2 mJ/cm2 per pulse for the visible beam and 5 mJ/cm2 per pulse for the IR beam. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. The SFG intensity was detected by a photomultiplier tube after spatial filtering by an aperture, and spectral filtering by a bandpass filter. The SFG intensity was normalized against that from a z-cut quartz. Optically-smooth fused silica windows (ISP optics, USA) with a thickness of 3mm and sample cell made of glass (capacity: 1 ml) were cleaned with a commercial cleaning agent (extran AP12) for 3 min. They were then immersed in a mixture of sulfuric acid (98%) and nochromix reagent (GODAX Laboratories, Inc.) for 24 hours, followed by rinsing in pure water (resistivity > 18.2 M⋅cm, Millipore), and finally dried at 100℃ for 3 h to remove residual surface water. After these treatments, the silica plates were kept in heptane or toluene (HPLC grade, Fisher) to prevent further water adsorption on the surface. All spectra were taken at room temperature.  40  4.3 Results and Analysis  Figure 4.1 shows the SFG spectra of heptane/silica interfaces with dry (open circles) and water-saturated (solid circles) heptane. The peaks between 2800 and 3000 cm-1 are associated with C-H vibrational modes from heptane. The detailed study of heptane/silica interfaces has been reported previously.105 The current analysis will focus on the water peaks between 3000 and 3700 cm-1. The dry heptane/silica interface contains a small amount of water, and a free OH peak is visible at 3680 cm-1. The water-saturated spectrum was collected at ~30 min after adding a small water droplet in the cell to saturate the water content in heptane. The solubility of water in heptane is only ~ 0.056% (mole fraction) at 25°C.106-108 However, the adsorption of water on silica is favorable because the silanol groups (either SiOH or SiO–) on the silica surface interact with water molecules strongly. With the water-saturated heptane, three major water peaks can be identified at 3200, 3400, and 3660 cm-1 as shown in Figure 4.1 (solid circles). The peaks at 3200 and 3400 cm-1 have been commonly observed at water/silica interfaces.20 These two peaks are located at the same positions as the IR absorption peaks of bulk ice and liquid water and hence are sometimes labeled as “ice-like” and “liquid-like” peaks, respectively.20 It has been proposed that the 3200 cm-1 peak represents more ordered water molecules with symmetric tetrahedral coordination, while the 3400 cm-1 peak represents the more disordered asymmetrically bonded molecules.80,100 The free OH peak at 3660 cm-1 has been widely observed at water surfaces in contact with hydrophobic environments, such as air/water11 or oil/water100 interfaces. Therefore, the 3660 cm-1 peak in Figure 4.1 represents free OHs in contact with the hepane phase. The spectra of water at heptane/silica interfaces, as shown in Figure 4.1, are basically consistent with our  41  current understanding of interfacial water molecules and show a combination of SFG spectra from water at a typical water/hydrophilic-surface and a typical water/hydrophobic-surface interfaces.  SFG Intensity (arb. u.)  16 14 12 10 8 6 4 2 0 2800  3000  3200  3400  3600  3800  -1  Wavenumber (cm )  Figure 4.1. SFG spectra of heptane/silica interfaces with dry (open circles) and watersaturated (solid circles) heptane. The peaks between 2800 and 3000 cm-1 are associated with C-H vibrations of heptane, the peaks at 3200 and 3400 cm-1 are vibrational peaks associated with hydrogen-bonded water, and the 3660/3680 cm-1 peak is assigned to the free O-H stretching mode.  Figure 4.2 shows the SFG spectra of toluene/silica interfaces with dry (open circles) and water-saturated (solid circles) toluene. The vibrational peaks between 2800 and 3100 cm-1 are associated with toluene. The detailed peak assignments for toluene have been reported previously and can be found in the reference.105 Again, a small amount of water was present at the interface with the dry toluene and showed a free OH  42  peak near 3600 cm-1. The position of the free OH peak is red-shifted and broadened compared to that at heptane/silica interfaces (Figure 4.1). The redshift of the free OH mode indicates that the solvent does interact with the interfacial water molecules. When water was added into the dry toluene, water molecules adsorbed on silica and formed a new structure which had not been previously observed. The solubility of water in toluene is ~ 0.28% (mole fraction) at 25°C.109,110 As expected, the water in toluene adsorbed on the silica surface, and the vibrational peak associated with water appeared, as shown in Figure 4.2 (solid circles). Surprisingly, the adsorbed water produced a very different spectrum compared to that at the heptane/silica interface shown in Figure 4.1 (solid circles). At the toluene/silica interface, water formed a broad peak centered at 3150 cm-1. Based on our current understanding of water surfaces, two important features are missing in the water-saturated spectrum shown in Figure 4.2. First, the free OH peak, which exists at an oil/water interface, is no longer detectable. Second, the weakly bonded water peak at 3400 cm-1, which exists at the water/silica interface, is also missing. The spectrum suggested that water formed a highly-hydrogen bonded structure at the toluene/silica interface with nearly no free OHs. An important character of a water surface without free OHs is that the interaction between the water surface and an incoming water molecule is significantly reduced.111 The water structure shown in Figure 4.2 (solid circles) was found to be very stable in the water-saturated toluene. For more than 12 hours, the spectrum remained unchanged, and the water surface did not accumulate more water to become a regular liquid water layer similar to that at the heptane/silica interface. For comparison, the same experiment was carried out on an initially moistened silica surface (by exposing to N2 of 60% relative  43  humidity), where the 3200 cm-1, 3400 cm-1, and free OH peaks were all present initially, as shown in Figure 4.3 with open circles. The solid circles in Figure 4.3 show the SFG spectrum taken 30 min after water was added in toluene. It was clear that, on this regular water surface, further adsorption of water continued when the water content in toluene was saturated.  SFG Intensity (arb. u.)  10 8 6 4 2 0 2800  3000  3200  3400  3600  3800  Wavenumber (cm-1)  Figure 4.2. SFG spectra of toluene/silica interfaces with dry (open circles) and watersaturated (solid circles) toluene. The peaks between 2800 and 3100 cm-1 are associated with C-H vibrations of toluene, the broad peak centered at 3150 cm-1 is associated with hydrogen-bonded water, and the 3600 cm-1 peak is assigned to the free O-H stretching mode.  44  SFG Intensity (arb. u.)  6  4  2  0 2800  3000  3200  3400  3600  3800  -1  Wavenumber (cm )  Figure 4.3. Adsorption of water on a regular water layer with free OHs. The open circles represent the SFG spectrum of a toluene/silica interface with dry toluene on an initially moistened silica surface. In this regular water layer, the 3200 cm-1, 3400 cm-1, and free OH peaks are all present. Water can condense on such a regular water surface, as indicated by the solid-circle spectrum, which was taken 30 min. after the toluene was saturated with water.  The water-saturated spectrum in Figure 4.2 (solid circles) suggests that the formation of a water surface with no free OHs is possible without an atomically smooth surface. On an atomically smooth mica surface, it has been proposed that the condensed water molecules may form a two-dimensional (2D) ice-like structure which is stabilized via the interaction between water molecules and the solid surface.112,113 This structure was latter confirmed by SFG vibrational spectroscopic measurements for D2O on mica, with a strong “ice-like” peak located 2375 cm-1 accompanied by the disappearance of 45  both the “liquid-like” peak at 2510 cm-1 and the free O-D peak at 2740 cm-1.114 A similar 2D flat ice structure was also observed at -160 °C on an atomically smooth crystalline Pt(111) surface, on which free OHs were not present because all water molecules bound either directly to the metal surface or to each other through the in-layer hydrogen bonds. This 2D ice layer was reported by Kimmel et al. as a hydrophobic water monolayer because it depressed further growth of ice on the surface.111 The structure shown in Figure 4.2 (solid circles) shares many similar properties with these 2D water structures, such as the disappearance of both the “liquid-like” and free OH peaks and the depression of further water adsorption. However, one major difference was that the silica surface used in the current study was not atomically smooth. The optically flat silica surface was mechanically polished and has a roughness of a several nanometers under an atomic force microscope. The rough surface cannot provide a lattice matching condition to form a 2D ice structure similar to that on an atomically smooth mica or Pt surface. A new mechanism is needed to explain the observed phenomena. A possible explanation is that the surface defects (scratches, cavities, or pits) on silica may help the adsorbed water molecules to minimize the number of free OHs. When water molecules adsorb on the silica surface one by one to fill up the defects, the water molecules will minimize their energy by forming the maximal number of hydrogen bonds. At a certain stage, the OHs bind either to other water molecules or to the silanol groups (protonated or deprotonated) on silica, as demonstrated in Figure 4.4, and these nanometer holes are then “sealed” for further water adsorption. This model explains the long-term stability of the highly hydrogen-bonded structure shown in Figure 4.2 (solid circles). On the other hand, if the defects are initially overfilled with too many available  46  free OHs, as shown in Figure 4.3 (open circles), the free OH peaks will always exist on the regular water layer, and then further adsorption of water is favorable. Although the “water-resistant” water structure centered at 3150 cm-1 has a vibrational frequency near that of ice, it is unlikely to have an icelike structure with symmetric tetrahedral coordination because the tetrahedral coordination would have to be highly distorted to fit in the defect sites, which explains the broad width of the peak. Nevertheless, the large red shift of the OH stretching frequency, in comparison to free OHs, indicates that the strength of the hydrogen bonds is likely comparable to that of ice.  Figure 4.4. Schematic drawing of the “water-resistant” water structure. In the process of water adsorption, the water molecules minimize their energy by forming the maximal number of hydrogen bonds with either the neighbouring water molecules or the silanol groups (SiOH) on silica. At a certain stage, very few free OHs are available. The absence of free OH groups significantly reduces further water adsorption.  47  Toluene plays an important role in stabilizing the “water-resistant” water surface. When experiments were carried out at an air/silica interface with the relative humidity (RH) increasing from ∼0%to 100%, the free-OH peak was observed at all RH values, indicating the water surface without free OHs either could not form or was unstable at an air/silica interface. Therefore, toluene plays a critical role in minimizing the number of free OHs. Our data also indicated that the hydrophobicity of toluene was not the only driving force to minimize the number of free OHs on the water surface, as we did not observe a similar structure at heptane/silica interfaces. Further experimental and theoretical studies will be needed to better determine whether the shape of toluene or the stronger interaction between toluene and water also plays a role in forming the water layer without free OHs.  48  4.4 Conclusions  The adsorption of water at solvent/silica interfaces was studied using SFG vibrational spectroscopy. Experimental data showed that the interactions between solvents and water molecules can significantly change the interfacial water properties. We discovered a water layer between toluene and silica with no detectable free OHs. The water layer without free OHs showed resistance against further adsorption of water molecules. However, this special structure of water was not observed at heptane/silica interfaces, at which free OHs were always observed.  49  Chapter 5  Why do Sulfuric Acid Coatings Influence the Ice Nucleation Properties of Mineral Dust Particles in the Atmosphere*  * A version of this chapter has been published. Yang, Z.; Bertram, A. K.; Chou, K. C. Why do Sulfuric Acid Coatings Influence the Ice Nucleation Properties of Mineral Dust Particles in the Atmosphere? Journal of Physical Chemistry Letters 2011, 2, 1232-1236.  50  5.1 Introduction  Mineral dust particles are abundant in the atmosphere. Common minerals found in aerosolized dust include quartz, illite, muscovite, chlorite, kaolinite, and calcite.115-125 During their lifetime in the atmosphere, mineral dust particles can become coated with inorganic material such as sulfuric acid. 126,127 Recently several studies have investigated the effect of sulfuric acid coatings on the ice nucleation properties of mineral dust particles.118,128-133 Laboratory studies with supermicron particles have shown that uncoated minerals particles such as illite and kaolinite are good ice nuclei both below and above water saturation.134,135 But once the supermicron particles are coated with sulfuric acid they are only good ice nuclei at close to and above 100% relative humidity (RH) for temperatures around 235 K.129,132 Experiments with submicron mineral particles also show that particles coated with sulfuric acid are poor ice nuclei below water saturation, but can act as ice nuclei above water saturation (although the ice nucleation ability may be reduced above water saturation compared to the uncoated case due to the permanent loss of some active sites).133 This difference between uncoated and coated mineral dust particles in terms of ice nucleation ability is illustrated in Figure 5.1a, for the specific case of supermicron illite particles from the work of Chernoff et al.132 Figures 5.1b and 5.1c show the pH and H2SO4 concentration of the coating as a function of RH, calculated from the Aerosol Inorganic Model.136 A comparison of panel a, b and c illustrate that coated mineral dust particles are poor ice nuclei when the coating material has a low pH (less than 1) and a high acid concentration (H2SO4 concentration > 0.1 M).  51  Figure 5.1. Panel (a) schematic showing the range of RH values over which uncoated and coated mineral dust particles are good ice nuclei. The figure is specifically created for illite particles based on the work of Chernoff et al.132 A similar behavior was also observed for kaolinite. The temperature applicable for the ice nucleation studies was approximately 237 K. Panel (b) and (c) show the pH and the molarity of the sulfuric acid coating, assuming the coating is in equilibrium with the relative humidity.  Despite the evidence that sulfuric acid coatings influence the ice nucleation properties of mineral dust particles, a molecular-level understanding is still lacking. A molecular-level understanding is preferred so that laboratory results can be confidently 52  extrapolated to the atmosphere. In the following we investigate why sulfuric acid coatings influence the ice nucleation properties of mineral dust particles at molarities > 0.1 M. We probed the structure of water at the mineral aqueous acid interface as a function of the sulfuric acid concentration using IR-visible sum frequency generation (SFG) vibrational spectroscopy, which is a highly surface-specific optical technique for studying liquid/solid interfaces.20 We specifically focused on differences in water structures for concentrations less than and greater than 0.1 M. For these studies mica was chosen for the mineral interface since mica has a surface structure similar to illite which is reported to make up more than 50% percentage of dust in some regions.137 Additionally, atomically flat mica surfaces are relatively easy to obtain by cleaving a mica sheet.  53  5.2 Experimental Section  The visible and IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a KTP optical parametric generator/amplifier (OPG/OPA) pumped by the 532 nm beam. The 532 nm and IR beams were overlapped, both spatially and temporally, on the sample, as shown in the inset of Figure 5.2. The energy of the laser beams were ~ 200 uJ/pulse for both the visible and IR beams. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. The SFG intensity was detected by a photomultiplier tube after spatial filtering by an aperture and spectral filtering by a bandpass filter. Each spectrum shown in the current study is an average of 6 scans in a 10 cm-1 step, and each scan is an average of 50 laser shots per step. Deuterium oxide D2O (99.9% deuterated water; Sigma-Aldrich) was used, instead of H2O, to avoid the IR absorption in mica due to the O-H stretching at 3620 cm-1.138 Deuterated sulfuric acid D2SO4 (Sulfuric acid-d2, 96-98 wt.% deuterated, solution in D2O, 99.5 atom% deuterated, Sigma-Aldrich) was used to prepare solutions for the SFG measurements. Mica sheets were cleaved right before the SFG measurements. A series of D2SO4 solutions with concentrations of 0, 5x10-6, 5x10-5, 5x10-4, 5x10-3, 5x10-2, 5x10-1 mol/L were measured in the sequence from the lowest concentration to the highest concentration. For each solution, the spectroscopic cell and mica sheet were rinsed thoroughly with the solution before the spectroscopic measurements. The rinsing process  54  ensured that the concentration of D2SO4 in the cell is the same as the solution used. For each concentration, 6 scans were collected in a period of ~1 hour during which no significant change of the SFG spectrum was observed. All spectra presented in the paper were taken using the same mica sheet. All experiments were carried out at room temperature.  55  5.3 Results and Analysis  Figure 5.2 shows the SFG vibrational spectra of deuterium oxide (D2O)/mica interfaces with D2SO4 concentrations of 0, 5x10-6, 5x10-5, 5x10-4, 5x10-3, 5x10-2, 5x10-1, and 5 M. Deuterated water and sulfuric acid was used, instead of H2O, to avoid the IR absorption peak of mica at 3620 cm-1.138 The spectra of D2O exhibit two peaks located near 2375 and 2550 cm-1, which represent the peaks at 3200 and 3400 cm-1, respectively, for interfacial H2O.11 The redshift of the D2O peaks with respect to those of H2O is a result of the isotopic substitution. These two peaks are sometimes labeled as the ‘‘icelike’’ and the ‘‘liquid-like’’ peaks, indicating that their peak positions are similar to those of ice and liquid water, respectively.20,72 It is generally accepted that the former represents water in a more ordered hydrogen-bonding network, and the latter represents a less ordered hydrogen-bonding structure.10,20 The detailed water structures associated with these two peaks are not completely understood, but it was proposed that the 2375 cm-1 peak represents water molecules with symmetric tetrahedral coordination, while the 2550 cm-1 peak represents the asymmetrically bonded molecules.80,100  56  Figure 5.2. SFG spectra of D2O/mica interfaces with D2SO4 concentrations of (a) 0 M, (b) 5x10-6 M, (c) 5x10-5 M, (d) 5x10-4 M, (e) 5x10-3 M, (f) 5x10-2 M, (g) 5x10-1 M and (h) 5 M. The inset shows the schematic layout of the spectroscopic setup. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively.  To obtain a more quantitative analysis, the spectra in Figure 5.2 were fitted with Lorentzian lineshape I (ω SFG ) ∝ χ  ( 2) NR  +∑ q  Aq ω IR −ω q +iΓ q  2 ( 2) , where χ NR is the nonresonant  57  contribution, Aq is the amplitude, Γ q is the width, and ω q is the resonant wavenumber for the qth vibrational mode. The fitting curves were shown in Figure 5.2 (solid lines), and the fitting amplitudes and resonant wavenumbers were shown in Figure 5.3. The best fits were obtained with a negative Aq for the “ice-like” peak and a positive Aq for the “liquid-like” peak. These assignments are consistent with the phase-sensitive SFG measurements at low pH reported by Ostroverkhov et al.73 As shown in Figure 5.2 and 5.3a, when the concentration of D2SO4 increased, both the ‘‘liquid-like’’ and the ‘‘icelike’’ peaks intensities decreased. With a D2SO4 concentration of 5 M (pH ~ -0.7), both peaks completely vanished, indicating that the ordered hydrogen-bonding network of water no longer exists. Figure 5.3b shows that the frequencies of the water peaks are pH dependent. In general, both the surface potential of mica and dipole-dipole coupling between water molecules can shift the frequency139, but studies of the frequency shifts are beyond the focus of the current study. The drastic reduction of the ordered water structures (Figure 5.2) with D2SO4 concentrations of 0.5 M is consistent with previous observations by Eastwood et al. and Chernoff et al. indicating H2SO4-coated mineral surfaces with H2SO4 concentrations > 0.1 M are poor ice nuclei (see Figure 5.1).129,132,135 The results suggest that the ordered surface water structures may have played an important role in the ice nucleation process. Although the drastic reduction and eventual disappearance of the ordered water structures offers a microscopic explanation for why H2SO4-coated mineral surfaces are poor ice nuclei, the question remains to be answered is why the ordered surface water structures is drastically reduced with a high acid concentration. Based on our current understanding, there are three possible effects responsible for the disappearance of ordered water  58  structures: (a) the decrease of mica surface potential, (b) the adsorption of sulfates on the surface, and (c) solvations of sulfates in water. These effects are discussed below.  Figure 5.3. The fitted amplitudes (a) and frequencies (b) of water peaks in the SFG spectra with various D2SO4 concentrations. Triangles represent “ice-like” peaks and dots represent “liquid-like” peaks.  It is known that a decrease in the surface potential leads to a less ordered water structure on material surface.84,140-142 The surface potential on a mineral surface is pH dependent. The structure of muscovite mica constituted of octahedral hydroxyl-aluminum 59  sheets lying between two silicon tetrahedral layers. One in four silicon atoms is substituted by an aluminum atom in the silicon tetrahedral layer, and the substitution results in a negative charge which is neutralized by K+ ions located between the silicon tetrahedral layers. When mica is cleaved, a cleavage plane happens in the potassium layer. The K+ ions are equally distributed between the two surfaces, and the surface is overall neutral.143 When placed in water, hydrated potassium ions dissociate from the mica surface,144,145 and the surface, which consists of Si, Al, and O connected by Si-O and Al-O bonds, become negatively charged, then partially neutralized by H+ ions in water. The surface charge of mica is therefore dependent on the pH of the solution. Overall, the surface charge decreases when the pH value decreases. The phenomena have also been observed on other mineral surfaces.71 Measurements for the point of zero charge (PZC) of mica were not conclusive, but it is generally accepted that the PZC of muscovite mica is less than 3.146 One of the difficulties in measuring the PZC for mica is the increased solubility of lattice aluminium ions at low pH,147 which creates negative surface charges on the surface. However, it was clearly observed that the increased concentration of hydrogen ion at low pH decreased the surface potential of mica, which can be attributed to the protonation of the Si-O groups at the surface to form ionizable surface silanol groups.146 Therefore, the decrease of the surface potential on mica at lower pH is one of the mechanism responsible for the decrease of the water peaks in Figure 5.2. When the mica surface approached neutral or positively charged, the adsorption of anions on mica surface may also affect the structure of water on mica. It is known that the adsorption of sulfate on mineral surface is mainly associated with Al and Fe oxy-  60  hydroxides and with allophanic constituents.148,149 Sulfate adsorption generally increases when the pH decreases.148,150,151 The vibrational peaks of sulfate are too weak to be measured directly by the current optical setup. To further study whether the disappearance of the ordered water structure was related to any particular properties of sulfate, the same experiment with 5 M of hydrochloric acid (HCl) was carried out for comparison. The experiment showed the SFG peaks of water also disappeared with 5 M of HCl solution. Therefore, the vanishing of the ordered water structures was likely not linked to any particular properties of sulfates. However, the anions should interact with the mineral surface when the surface approached neutral or become positively charged. With a high concentration of H2SO4, the solvations of sulfates anions (SO42-) become an important factor affecting the water structure on mica. Previous ab initio studies by Cannon et al, showed that 13 water molecules are present in the first solvation shell of SO42-.152 With a H2SO4 concentration of 5 M, the water-sulfate mole ratio is roughly 8:1. At this concentration, the mica surface must compete for water with the sulfate ions in the bulk solution. The phenomenon can affect ice nucleation processes in two ways. First, the number of water molecules available to the mica surface is reduced. Second, as the mica and sulfates are competing for water molecules, most water molecules are well captured either by the surface charges on mica or by the anions in the solution. The situation creates an energy barrier for ice nucleation because formations of ice nuclei would have to overcome the electrostatic interactions to rearrange the ordering of water molecules. When the concentration of sulfuric acid increases, the above three mechanisms all work against ordering of water molecules on mica surface. When pH decreases, the  61  surface potential decreases, and consequently the ordering of interfacial water decreases which was indicated by the decrease of the SFG peaks. As the surface potential decreases, the adsorption of anions on the mica surface becomes significant, and the adsorbed anions displace ordered water molecules on the surface. Finally, when the concentration of sulfuric acid reaches a critical concentration, in which the solvation of sulfate ions consumes large amount of water, the ordered water structure disappears. At this stage, nearly all water molecules are captured by the anions, and few water molecules are freely available to the mica surface. The process is illustrated in Figure 5.4 showing water molecules and hydrated sulfate (bisulfate) anions at solution/mica interfaces with increasing concentrations of H2SO4.  62  Figure 5.4. Schematics of water molecules and hydrated sulfate (bisulfate) ions at solution/mica interfaces with different concentrations of H2SO4 solutions. (a) Pure H2O/mica interface with pH~6. The surface is highly negatively charged. Water molecules are more ordered near the charged surface. (b)With a low concentration of H2SO4, the surface charge decreases as the surface is protonated, and the anions interact with the mineral surface. (c) With a high concentration of H2SO4 (for example 5M), water molecules are well captured by the sulfate/anions in the solution, and few water molecules are freely available for the mica surface. In (a) and (b) the dashed lines separate the ordered water molecules from the water molecules without order. Above the dashed lines, water molecules have good order because of negative surface potential induced by the mica surface. In (b) and (c) H+ adsorb on the mica surfaces. The dashed circles represent hydrated sulfate ions (in low H2SO4 concentration) or hydrated bisulfate anion (in high H2SO4 concentration). Water molecules in the dashed circle are parts of the hydrated anions and move together with the core anions.  63  5.4 Conclusions  The structures of water on mica surfaces in the presence of H2SO4 with atmospheric relevant concentrations were studied using SFG vibrational spectroscopy. We found that ordered water structures significantly decreased with 0.5 M of H2SO4 and disappeared with 5 M of H2SO4. The study provided a molecular-level understanding for previous laboratory studies showing minerals particles coated with sulfuric acid are relatively poor ice nuclei. The observed phenomenon was explained by a combined effect of the decreased mica surface potential at low pH, the adsorption of sulfates on mica, and the lack of free water molecules in high concentration of acidic solution.  64  Chapter 6  Competitive Adsorption of Toluene and nAlkanes at Binary Solution/Silica Interfaces*  * A version of this chapter has been published. Yang, Z.; Li, Q. F.; Hua, R.; Gray, M. R.; Chou, K. C. Competitive Adsorption of Toluene and n-Alkanes at Binary Solution/Silica Interfaces. Journal of Physical Chemistry C 2009, 113, 20355-20359.  65  6.1 Introduction  The competitive adsorption of hydrocarbons at liquid/mineral interfaces plays a critical role in many industrial and environmental processes, such as oilsands processing28,29,34,38, petroleum recovery32, contamination removal30,33,35,37, and many extraction techniques31,36. Toluene and alkanes are particularly important because they are the most commonly used solvents for both industrial and scientific applications. At a binary solution/mineral interface, it is expected that the surface chemical composition is different from the bulk composition because of the different interaction strength with the surface.39 In many cases, the competitive adsorption of solvents at liquid/solid interfaces is a critical factor determining the effectiveness of a technological process. In the 1950’s and 1960’s, the adsorption isotherms for binary mixtures at liquid/solid interfaces were studied by various immersion methods, and a number of theories were developed.40,41 Despite this effort, the problem was not completely resolved42, because the macroscopic measurements were indirect, and the theories require molecular-level information about adsorbates as input parameters.41 Even with modern technologies, it remains challenging to directly measure the surface coverage of a particular component at a liquid/solid interface. Recent developments in IR-visible sum frequency generation (SFG) vibrational spectroscopy have shown that SFG is an effective technique to obtain molecular-level information at buried liquid interfaces, and many studies have been done at water interfaces.8 Compared with water interfaces, little is known about solvent/solid interfaces.153 SFG vibrational spectroscopy has been used for a broad range of studies by a rapidly growing number of research groups.20,51,154-156 As a second-order nonlinear  66  optical process, SFG is forbidden in centrosymmetric media, such as liquids, but the symmetry is broken at an interface. For this reason, SFG is highly surface-specific and capable of measuring surface vibrational spectra under ambient conditions. For liquid surfaces, it has been shown that SFG from a water surface is dominated by the top monolayer.157 With the monolayer sensitivity and short probing depth, SFG provides a new opportunity to directly measure the adsorption isotherms at solvent/mineral interfaces. In this chapter, SFG vibrational spectroscopy was used to study the competitive adsorption of toluene and alkanes on silica, which is one of the most abundant minerals and has also been widely used in liquid chromatography101-104. The adsorption process at a liquid/solid interface is significantly different from that at a gas/solid interface because there is no empty site at a liquid/solid interface. A change in the bulk composition results in the replacement of one component by the other component. The Langmuir isotherm has often been used to describe the adsorption process for dilute solutions at liquid/solid interfaces153,158-163, but it is a good approximation only for strongly adsorbed molecules. In the current study, the adsorption free energies for the solvents are of the same order of magnitude. Therefore, the Langmuir isotherm is not suitable for the current study. In this paper, we obtained the surface coverage of toluene using SFG and used the Everett isotherm to fit the measured adsorption isotherm over the complete toluene fraction range.  67  6.2 Experimental Section  The visible and IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with an output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a KTP optical parametric generator/amplifier pumped by the 532 nm beam. The 532 nm and IR beams were overlapped, both spatially and temporally, on the sample, as shown in Figure 6.1. The energy of the laser beams were ~ 200 μJ/pulse for both the visible and IR beams. The SFG intensity was detected by a photomultiplier tube and normalized against that from a z-cut quartz. Each spectrum shown in the current study was an average of four scans in a 5 cm-1 step, and each scan was obtained by averaging the SFG intensity of 40 laser shots at each step.  Figure 6.1. Schematic layout of the spectroscopic setup. The frequency of the visible beam was fixed at 532 nm, and the frequency of the IR beam was tunable. The 532 nm and IR beams were overlapped both spatially and temporally on the top surface of the solution. The thickness of the solvent layer is 3 mm. 68  Fused silica plates, with a thickness of 3 mm, were cleaned with a commercial cleaning agent (extran AP12) for 3 min. Then, they were immersed in a 50/50 (v/v) HNO3/H2SO4 solution for ~12 hours, followed by rinsing in pure water (resistivity > 18.2 M⋅cm) and finally dried at 100°C for 2 hours to remove residual surface water. After these treatments the silica plates were kept in toluene to prevent further water adsorption on the surface. Toluene, pentane, heptane and tetradecane (Fisher; HPLC grade) were used as received to prepare mixtures with different volume fractions. The SFG spectrum of the pure toluene/silica interface was monitored at the beginning and the end of the experiment for each toluene-alkane mixture to ensure that the sample had stayed consistent during the experimental period. After the pure toluene/silica SFG spectrum was measured, the SFG spectra of a B = 0.8, 0.6, 0.4, 0.2, series of toluene-heptane mixtures with toluene volume fraction φtoluene  and 0 were measured in the sequence from the highest toluene fraction to the lowest. For each binary mixture with a particular toluene fraction, the cell and silica plates were rinsed thoroughly with the mixture before the spectroscopic measurement. The rinsing process ensured that the bulk mixture in the cell had the intended toluene fraction. For each toluene fraction, four scans were collected in a period of 30 minutes during which no change of the SFG spectrum was observed. Then the cell and silica plates were cleaned with acids as described above for experiments with a different alkane. All spectra were taken at room temperature.  69  6.3 Results and Analysis  The SFG vibrational spectra of toluene/silica interfaces in ssp (SFG, visible, and IR polarizations are s-, s-, and p-polarized, respectively) and ppp are shown in Figure 6.2A. The peaks are assigned as follows: 2860 cm-1 and 2875 cm-1 to the combination/overtone modes, 2920 cm-1 to the symmetric stretch of the CH3, 3022 cm-1 to the ν20a CH stretching mode of the phenyl group, and 3075 cm-1 to the ν2 CH stretching mode of the phenyl group.164 Previously, Hommel et al. have observed a peak at 2945 cm1  at air/toluene surface using SFG and assigned the peak to the CH3 asymmetric mode.  However, the ssp and ppp intensity ratio of the 2945 cm-1 peak in Figure 6.2A is not consistent with the CH3 asymmetric mode. Assuming the orientation distribution of the CH3 groups is a delta function, the ratio of the second order nonlinear susceptibility in ssp and ppp configurations  ( 2) χ ssp 2) χ (ppp  can be calculated as a function of the CH3 tilting angle52,  and the results are shown in Figure 6.2B. (The detailed calculation is available in the supporting materials.) For the CH3 asymmetric mode, the calculated ratio  ( 2) χ ssp 2) χ (ppp  is always  less than 1, even with a finite distribution width, but the spectra in Figure 6.2A indicates a ratio of  ( 2) χ ssp  χ  ( 2) ppp  ~ 3. Therefore, the 2945 cm-1 peak could be a Fermi resonance associated  with the symmetric mode, instead of an asymmetric mode. For the CH3 symmetric mode, the measured ratio  ( 2) χ ssp 2) χ (ppp  is ~ 4.4, which corresponds to a tilting angle of 25° with respect  to the surface normal, as indicated by the solid circle and the dotted lines in Figure 6.2B. This tilting angle is consistent with the molecular dynamic simulation for adsorption of 70  toluene on silica, showing the plane of the phenyl ring mostly adopts an upright geometry with a tilting angle of about 30° with respect to the surface normal because of the interaction of its π electrons with the silica surface.165  Figure 6.2. (A) SFG vibrational spectra of toluene in ssp and ppp polarization configurations. The ssp and ppp spectra are offset from each other by 0.5 arbitrary units for clarity. (B) Calculated  ( 2) χ ssp 2) χ (ppp  for CH3 symmetric (solid line) and asymmetric (dashed  line) modes as a function of the tilting angle. The orientational distribution of the CH3 groups was assumed to be a delta function. The solid circle indicates the measured  ( 2) χ ssp 2) χ (ppp  value of 4.4 which corresponds to a tilting angle of 25°.  71  Figure 6.3a-f show the ssp SFG spectra of toluene-pentane mixtures with the bulk B = 1, 0.8, 0.6, 0.4, 0.2, and 0, respectively. Previously, toluene volume fraction φtoluene  Selfer et al. studied the adsorption of alkanes on silica and showed that hexadecane lies flat on the silica surface.22 When the axis of the CH3 group is along the surface, the asymmetric peak will dominate, and the symmetric peak will be missing. If the axis of the CH3 group is along the surface normal, the situation will be reversed. The spectrum of pure pentane on silica in Figure 6.2f shows two peaks at 2857 cm-1 and 2951 cm-1, which are consistent with the CH2 symmetric and CH3 asymmetric modes, respectively.166 Figure 6.3a-f show that the peak intensities of the toluene ν20a and ν2 modes decrease as B φtoluene decreases. The spectra shown in Figure 6.3 were collected in the sequence from  B B = 1) to the lowest ( φtoluene = 0). Within the the highest toluene volume fraction ( φtoluene  measurement error, similar spectra were obtained for experiments carried out in a reversed order. Therefore, the competitive adsorption process is reversible, and the surface composition depends only on the bulk mole fraction of toluene, not on the history of the system. Experiments were repeated for toluene-heptane and toluene-tetradecane binary mixtures on silica, as shown in Figures 6.4 and 6.5, respectively. While the spectrum of pure heptane on silica (Figure 6.4f) is similar to that of pure pentane on silica in Figure 6.3f, significant CH3 symmetric peak at 2875 cm-1 and CH2 asymmetric peak at 2920 cm1  were observed at tetradecane/silica interfaces.166 Therefore, there is an increasing  conformational disorder for longer alkane chains.  72  Figure 6.3. SFG vibrational spectra of toluene-pentane mixtures on silica with toluene B = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. volume fraction φtoluene  73  Figure 6.4. SFG vibrational spectra of toluene-heptane mixtures on silica with toluene B = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. volume fraction φtoluene  74  Figure 6.5. SFG vibrational spectra of toluene-tetradecane mixtures on silica with toluene B = (a) 1, (b) 0.8, (c) 0.6, (d) 0.4, (e) 0.2, and (f) 0. volume fraction φtoluene  75  Since the spectra of both toluene and alkanes have CH peaks, the CH peaks are not good indicators for the adsorbed chemical species on silica. On the other hand, the ν20a and ν2 peaks from the phenyl group are the unique signature of toluene and allow us to quantitatively study toluene absorbed on the silica surface. Because the intensity of the ν20a peak is higher than that of the ν2 peak, the following analysis will focused on the ν20a peak to obtain the absorption isotherm of toluene on silica. To obtain quantitative information, further theoretical analysis of the SFG spectra is required. The detailed theoretical background of SFG can be found in the references.8 Briefly, the SFG intensity is given by I (ω SFG ) ∝ [L(ω SFG ) ⋅ e(ω SFG )] ⋅ χ ( 2 ) : [L(ωvis ) ⋅ e(ωvis )][L(ω IR ) ⋅ e(ω IR )] ⋅ I (ωvis ) ⋅ I (ω IR ) 2  (6.1) where L(ωi ) is the tensorial Fresnel coefficient, and e(ωi ) is the unit polarization vector, χ ( 2 ) is the surface nonlinear susceptibility tensor, and I (ωvis ) and I (ω IR ) are the  intensities of incident visible and IR beams, respectively. The surface nonlinear (2) can be expressed as susceptibility χ ijk  (2) χ ijk( 2 ) = χ NR ,ijk + ∑ q  Aq ,ijk ω IR −ω q +iΓ q  (6.2)  (2) where χ NR ,ijk describes the nonresonant contribution, and Aq ,ijk , ωq , and Γ q are the  amplitude, frequency, and damping constant of the qth vibrational mode, respectively. The amplitude Aq ,ijk in the lab coordinate is related to the molecular hyperpolarizability  α q,lmn in the molecular coordinates:  76  Aq ,ijk = n S  ∑α  ^ ^  q ,lmn  ^  ^  ^  ^  (i ⋅ l )( j⋅ m )( k ⋅ n )  (6.3)  l ,m ,n  where n S is the surface number density, and the angular brackets refers to an average over the molecular orientation. If the orientation of the molecule on a surface is not strongly dependent on the coverage, the amplitude Aq ,ijk is proportional to the surface density n S . In the current study, the orientation of toluene was verified using the ratio of the ν20a peak in ssp and ppp configurations.153 The amplitude ratio  Appp Assp  was found to be  ~ 0.42 and independent of the toluene coverage. Therefore, it is feasible to correlate the amplitude of the ν20a peak to the surface number density of toluene on silica. The SFG spectra in Figures 6.3-6.5 were fitted using Equation (6.1) and (6.2) to obtain the amplitude Aq ,ijk . Calculating the Fresnel coefficients described in Equation (6.1) requires the refractive index of the mixture nmix . In general, nmix changes with the mixture composition. The refractive index of a mixture follows a “mixture rule”167  nmix = φ1n1 + φ2 n2  (6.4)  where φi and ni are the volume fraction and the refractive index of component i , respectively. Although small deviations from Equation (6.4) have been reported,168 the small deviations are insignificant for the current study. The refractive indices for toluene, pentane, heptane and tetradecane are 1.4963, 1.357, 1.38, and 1.428, respectively.169 The amplitude Aq ,ijk was then calibrated using Equation (6.1) and (6.4) so that Aq ,ijk is proportional to the surface number density n s . The toluene surface coverage θ c ≡  nS S nmax  ,  77  S where nmax is the maximum density adsorbed on the surface when only toluene is  presented in the solution, was then derived using the amplitude of the ν20a peak. Figure 6.6 shows the toluene surface coverage as a function of the bulk mole fraction.  Figure 6.6. Adsorption isotherms of toluene on silica for binary mixtures of pentane– toluene (■), heptane–toluene (●), and tetradecane–toluene (▲). The solid curves are fitting curves using Equation (6.8).  As shown in Figure 6.6, toluene competes favorably against pentane but the advantage decreases as the chain length of alkane increases. This is consistent with the conclusion that alkanes on average lie flat on the silica surface. In this geometry, the molar adsorption energy of alkane increases as the chain length increases. To gain better insight into the competitive adsorption process, a theoretical model is needed. The wellknown Langmuir equation is not a good description for the adsorption isotherm at liquid/solid interface because there is no empty site at a liquid/solid interface. A number 78  of theories have been developed for the adsorption isotherm at binary liquid/solid interfaces over the complete mole fraction range.41,42,170-173 It has been shown that, for adsorption on a homogeneous surface from an ideal miscible binary liquid with component 1 and 2, the surface mole fraction of component 1 can be written as42  x1S =  K1 x1B 1 + ( K1 − 1) x1B  (6.5)  with ⎧ Δ a μ °1 − Δ a μ ° 2 ⎫ ⎧ ( μ °1S − μ °1B ) − ( μ ° 2S − μ ° 2B ) ⎫ K1 = exp ⎨− ⎬ = exp ⎨− ⎬ RT RT ⎩ ⎭ ⎩ ⎭  (6.6)  where x1B and x 2B are the bulk mole fractions for component 1 and 2, respectively, x1S and x 2S are the surface mole fractions for component 1 and 2, respectively, R is the gas constant, T is the temperature, μ °iB and μ °iS are the chemical potential (or partial molar Gibbs free energy) of component i in its standard state for the bulk and surface, respectively, and Δ a μ °i ≡ μ °iS − μ °iB is the chemical potential change associated with the adsorption of component i . The above expression was derived by Everett171, and it is often called the Everett isotherm. If component 1 is a molecule strongly adsorbed on the surface, or  − ( Δ a μ °1 − Δ a μ ° 2 ) >> RT , one gets K 1 >> 1 . In this case, Equation (6.5) can be approximated by  x1S =  K1 x1B 1 + K1 x1B  (6.7)  This expression is analogous to the Langmuir equation. However, the Langmuir equation is a good approximation only for a strongly adsorbed component. As the adsorption free 79  energies of toluene and alkanes are comparable, this approximation is not valid for the current study. The Everett isotherm will be used in the following analysis. The amplitude of SFG peaks, as described in Equation (6.3), measures the surface number density n S , instead of the surface mole fraction x1S shown in Equation (6.5). Therefore, it is desirable to express Equation (6.5) in terms of surface coverage θ c . Equation (6.5) can be rewritten as42  θc ≡  n1S βK1 x1B = S nmax, 1 + ( β K1 − 1) x1B 1  (6.8)  with  β≡  S n max, 2 S n max, 1  (6.9)  Equation (6.8) and (6.9) indicate that the adsorption isotherm is governed by the values of  β and K1 . The value of β describes the relative footprint on the surface for component 1 and 2, and the value of K1 is determined by the adsorption free energy difference between the two components. The Everett isotherm as described in Equation (6.8) allows a more quantitative analysis for the meassured isotherms in Figure 6.6. The values of β and K1 are coupled in Equation (6.8). Without additional information, the measured adsorption isotherms in Figure 6.6 are not sensitive to the individual value of β and K1 . Using β K1 as a single parameter to fit the measured adsorption isotherms in Figure 6.6, we obtained the best fits with β K1 = 1.69, 0.659, and 0.296 for pentane, heptane and tetradecane mixtures, respectively. To estimate K1 , further assumptions must be made for β . The value of β ,  80  defined as β ≡  S n max, 2 S n max, 1  , is mainly determined by the sizes of the molecules because a  S larger molecule has a larger footprint on the surface and a smaller value of n max . The  space that a molecule occupies can be estimated using the molar volume. The molar volumes Vm of toluene, pentane, heptane and tetradecane are 106.29, 115.26, 146.51, and ⎛V ⎞ 260.3 mL/mol. To a first approximation, the value of β should scale as ⎜⎜ m , 2 ⎟⎟ ⎝ Vm ,1 ⎠  2/3  , which  corresponds to a random adsorption geometry on silica. With this assumption, the values of β are 0.95, 0.81 and 0.55 for toluene-pentane, toluene-heptane, and toluenetetradecane mixtures, respectively. However, molecules adsorbed on a surface have a preferred orientation, and the β values need to be corrected with an orientation factor. As described above, the tilting angle of toluene with respect to the surface normal was estimated at θ ~ 25°, and the alkanes mostly lie flat on the silica surface with θ ~ 90°. With a preferred orientation, the average footprint of molecules on the surface scales as sin θ sin θ  oriented  where the angle brackets denote the orientational average  random  ∫ f (θ ) sin θ ⋅ dΩ with ∫ f (θ ) ⋅ dΩ  f (θ ) denoting the orientational distribution function. Assuming  the orientational distribution function is a delta function with θ = 25° for toluene and θ = 90° for alkane, the values of β become 0.40, 0.34 and 0.23 for toluene-pentane, tolueneheptane, and toluene-tetradecane mixtures, respectively. Then the best fit for K1 were obtained with Δ a μ ° toluene − Δ a μ ° pen tan e = −3.4 ± 0.3 kJ/mol, Δ a μ ° toluene − Δ a μ ° hep tan e = −1.8  81  ± 0.3 kJ/mol, and Δ a μ ° toluene − Δ a μ ° tetradecane = −0.84 ± 0.3 kJ/mol. The best-fit curves are shown in Figure 6.6. In all cases, toluene competes favorably against the alkanes for the adsorption on silica. However, the adsorption free energy of alkane increases as the chain length increases.  82  6.4 Conclusions  IR-visible sum frequency vibrational spectroscopy was applied to study the competitive adsorption of toluene and n-alkanes at binary solution/silica interfaces. The surface coverage of toluene on silica for toluene-pentane, toluene-heptane, and toluenetetradecane mixtures were obtained using the measured SFG peaks intensity of toluene. The competitive adsorption processes are reversible, and the surface coverage of toluene only depends on the toluene molar fraction in the binary mixture, not on the history of the mixture in contact with the silica. The measured adsorption isotherms fitted well with the Everett isotherm over the complete mole fraction range. The estimated molar adsorption free energy of toluene is 3.4 ± 0.3 kJ/mol, 1.8 ± 0.3 kJ/mol, and 0.84 ± 0.3 kJ/mol higher than pentane, heptanes, and tetradecane, respectively. Overall, toluene competes favorably on silica against the alkanes, and the molar adsorption free energy difference between toluene and alkane decreases as the chain length of the alkane increases.  83  Chapter 7  Effect of Interfacial Water Content on Bitumen Liberation from Silica and Mica Surfaces*  * A version of this chapter will be submitted for publication. Yang, Z.; Bailey, G.; Gray, M. R.; Chou, K. C. Effect of Interfacial Water Content on Bitumen Liberation from Silica and Mica Surfaces. 2011  84  7.1 Introduction  Bitumen liberation from mineral surfaces is the prerequisite for a bitumen extraction process.44 Understanding the mechanisms of oil displacement by water on minerals is important for oil sand recovery. Many investigations on bitumen displacement and detachment from the solid surface in the presence of water containing salt, surfactants and clays at different temperature and pH have been completed using dynamic contact angle measurements.45-48,174 Dynamic contact angle measurements provide macroscopic information such as rate of liberation, the dynamic, and the equilibrium contact angles in the bitumen liberation processes. Previous studies have also suggested that the interfacial water content between mineral surfaces and bitumen is one of the most important factors affecting bitumen recovery.49 However, the detailed mechanism remains poorly understood. To study the effect of interfacial water on bitumen liberation both macroscopic (liberation rate, dynamic and static contact angle) and microscopic (water amount at mineral surfaces under different relative humidity, structures of water molecules adsorbed on mineral surfaces) information is needed. In this chapter, we investigate the effect of the interfacial water content on bitumen liberation from silica and mica surfaces in an aqueous environment using both sum frequency generation (SFG) vibrational spectroscopy and dynamic contact angle measurements. The objective of the current study is to provide a molecular-level understanding for the effect of interfacial water on the liberation process.  85  7.2 Experimental Section  7.2.1. Sum frequency generation spectroscopy measurements The visible and tunable IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a KTP optical parametric generator/amplifier (OPG/OPA) pumped by the 532 nm beam. The 532 nm and IR beams were overlapped, both spatially and temporally, on the silica/vapor interface, as shown in Figure 7.1. The laser fluence was approximately 2 mJ/cm2 per pulse for the visible beam and 5 mJ/cm2 per pulse for the IR beam. The polarizations of the beams were s-, s-, and ppolarized for SFG, visible, and IR, respectively. The SFG intensity was detected by a photomultiplier tube after spatial filtering by an aperture, and spectral filtering by a bandpass filter. The SFG intensity was normalized against that from a z-cut quartz. Each spectrum shown in the current study is an average of 5 scans in a 10 cm-1 step, and a scan represents an average of 50 laser shots per step. Fused silica plates, with a thickness of 3mm, were cleaned thoroughly with a commercial cleaning agent (extran AP12). It was then immersed in a mixture of sulfuric acid (98%) and nochromix reagent (GODAX Laboratories, Inc.) for 24 hours, followed by rinsing in pure water (resistivity > 18.2 M⋅cm, Millipore) and finally drying at 100°C for 2h. Then it was put on the top of a chamber which was made of Teflon and cleaned thoroughly with a 50/50 (v/v) HNO3/H2SO4 mixture. The humidity in the chamber was controlled by adjusting the flow rate of N2 gas bubbling through a container filled with water and was measured with  86  a hygrometer (model: Omega RH82). The humidity was adjusted from 1% relative humidity (RH) to 99% RH with an accuracy of ±1% RH.  Figure 7.1. Schematic layout of the spectroscopic setup. The frequency of the visible beam was fixed at 532 nm, and the frequency of the IR beam was tunable. The 532 nm and IR beams were overlapped both spatially and temporally at the bottom surface of silica plate.  7.2.2. Dynamic and equilibrium contact angle measurements The dynamic and equilibrium contact angle measurements were conducted on three different substrates: silica plate, rinsed muscovite mica sheet and freshly cleaved muscovite mica sheet. The silica plate was cleaned in the same way as described above. The mica sheet was cleaved on both sides, followed by rinsing in pure water (resistivity > 18.2 M⋅cm, Millipore) and blow-drying with nitrogen gas. Experiments were performed on freshly cleaved muscovite mica sheet, while the mica sheet was just cleaved on both sides. In each contact angle measurement the substrate was placed into a glass container with a top cover for 30 min. The relative humidity in the container was also adjusted in the same way as above. After that 11 drops (~0.3ml) of diluted bitumen (Vtoluene/Vbitume = 3/7) were placed on the surface of substrate and kept in the glass container for 2 min to let the three phases (substrate, diluted bitumen and vapor) reach equilibrium. The bitumen became a thin layer and formed a disk. Then the substrate with a bitumen layer on the top was quickly placed 87  into a chamber of water. The temperature of water in the chamber was maintained at 25± 0.5 °C by a hot plate for all experiments. And the depth of water was always kept at 1.5 cm. The bitumen displaced by water contracted uniformly along the inward radial direction and finally the thin layer of bitumen became a droplet and the equilibrium condition was reached. For all RH values, the bitumen droplet remained attached on the silica plate and rinsed mica sheet, but left the surface of freshly cleaved mica sheet immediately. The whole recession process was recorded by a Nikon D90 camera with a 60mm macro lens.  88  7.3 Results and Analysis  7.3.1. Sum frequency generation spectroscopy experiments Figure 7.2 shows the SFG vibrational spectra of the water molecules on silica under various degrees of relative humidity. The spectrum is similar to the SFG spectrum of the water/air interface and it exhibits three peaks at 3200cm-1, 3400cm-1 and 3700cm-1.10 The two peaks at 3200cm-1 and 3400cm-1 are associated with an ordered and disordered hydrogen-bonding network. They have been called “ice-like” and “water-like” peaks respectively because ice has an IR adsorption at 3200cm-1 and water has an IR adsorption at 3400cm-1.10,20,80 The sharp peak at 3700cm-1 corresponds to the O-H stretch of the nonhydrogen-bonded O-H groups of water which is also known as the “free O-H” stretch.10 The SFG intensities of ice-like and water-like peaks are very small in the spectrum with 1% RH and is indicative of few water molecules adsorbed on the silica surface at 1% RH. However, there is a relatively large and sharp peak at 3750cm-1. This peak originates from the O-H stretch of SiOH groups on silica surface.175 At 28% RH O-H stretch modes associated with the hydrogen-bonded O-H groups can be observed in the spectrum. The peak of the SiOH OH stretch becomes weaker because of the formation of hydrogen bonds between SiOH and adsorbed water molecules. Another sharp peak at 3700cm-1 corresponding to the free O-H stretch of water begins to emerge. As the relative humidity further increases, the intensity of the water peaks becomes larger, indicating that more water molecules are adsorbing onto the silica surfaces. This result is similar to previous SFG experiments done on the mica/D2O vapour interface except that a 2D water structure formed on the mica surface at 90% RH.114  89  Figure 7.2. SFG vibrational spectra of water adsorbed on silica surface as a function of relative humidity (RH).  7.3.2. Dynamic and equilibrium contact angle study Figure 7.3 shows a sequence of frames extracted from a video. The pictures show the shape of a bitumen droplet on the silica plate when exposed to water. The recorded videos have 24 frames per second. Each frame was extracted and analyzed with a home-made Matlab program to obtain the dynamic contact angle. The dynamic contact angles were plotted as a function of time, as shown in Figure 7.4. For all videos, the time axis has been reset as zero when the contact angle equals to 60°.  90  Figure 7.3. Sequential images extracted from a video of bitumen liberation from silica at time equal to (a) 0 , (b) 20, (c) 80 sec.  Figure 7.4 (a) shows the dynamic contact angle for bitumen on the silica surface as a function of time with various relative humidity. In general, the dynamic contact angles increase with time until they reach equilibrium values, which are the static contact angles. The value of the static contact angle is an indication of the tendency for the bitumen droplet to detach from the mineral surface. As shown in Figure 7.4 (a), when RH equals to 0%, 25% and 46%, the changing rate of dynamic contact angle is basically the same. However, when the RH increased to 70%, the changing rate was significantly larger. When the RH increased to 90%, the changing rate was very similar to that at RH = 70%. It can also be observed that under different RH the static contact angles of bitumen on the silica surface are different. The static contact angle is 115° when RH is equal to 0%, 25% and 46%, while it is 122° when RH is equal to 70% and 90%. The results suggest that the increase of water content on the silica surface caused an increase in the changing rate of dynamic contact angle and an increase of static contact angle of bitumen, although the effect of water content is small for RH up to 46%.  91  (a)  (b)  (c)  Figure 7.4 Measured dynamic contact angles for bitumen on (a) silica, (b) rinsed mica, and (c) cleaved mica.  92  Figure 7.4 (b) shows the dynamic contact angle for bitumen on the rinsed mica surface as a function of time with various relative humidity. Similarly, both the changing rate of dynamic contact angle of bitumen and the static contact angle of bitumen on rinsed mica surface increased with RH. However, the time scale is much shorter compared to the silica surface. Furthermore, the static contact angles on the rinsed mica surface are larger compared to silica. A larger static contact angle indicates a higher tendency for bitumen to detach from the surface. On the rinsed mica surface, the increase of changing rate of dynamic contact angle due to the increase of interfacial water content was obvious at RH = 50%, unlike the silica experiment in which this increase was observed when RH is as high as 70%. Figure 7.4 (c) shows the time dependence of the dynamic contact angle of bitumen on the freshly cleaved mica surface with various RH. The dynamic contact angles reached equilibrium in several seconds, which is an order of magnitude faster than those on rinsed mica. As the RH increases, it is observed that the changing rates of dynamic contact angle have little difference, and so do the static contact angles. Based on the experimental data, it can be concluded that more water content on the silica and rinsed mica surface help the liberation of bitumen and increased the rate of the process. However, the properties of the mineral surface play a more critical role in the liberation process. The changing rate of dynamic contact angle showed an order-ofmagnitude difference for mica and silica. On three different mineral surfaces the changing rate of dynamic contact angle of bitumen is: cleaved mica > rinsed mica > silica. To explain the huge differences among the changing rates of dynamic contact angle on three different minerals, the surface charge of the mineral surfaces and the formation of 93  hydrogen bond between the mineral surfaces and bitumen should be considered. On one hand, bitumen does not carry any net charge, but if there are some net charges on the mineral surfaces, then the attractive electrostatic force between bitumen and the mineral surfaces will occur. The attractive force will slow down the liberation of bitumen from the mineral surfaces when they are exposed to an aqueous environment. On the other hand, as a complex mixture, bitumen bears various types of natural surfactants on its surface, such as R-NH2 and R-COOH.176,177 These functional groups can form hydrogen bonds with the mineral surfaces,178 and the hydrogen bonds will also decrease the liberation rate of bitumen in water. So the larger the surface charge density is on mineral surfaces and the more hydrogen bonds can form between bitumen and mineral surfaces, the slower the liberation of bitumen will be. The functional groups on the silica surface are basically silanol groups and the surface density of silanol groups is estimated to be ∼5 × 1014 cm-2.76,83 These silanol groups have two forms: -SiOH and –SiO-. The SiO- groups create a charge density of ~0.05 e- per nm2 at pH = 6. Both the -SiOH and –SiO- can form hydrogen bond with the functional groups on bitumen surface. The surface properties of mica are different from silica. The negative surface charges resulted from O- are neutralized by K+ after the mica is cleaved. However, after rinsing with water, 1% of K+ on the mica surface dissolves and the mica surface will carry negative charges with a density of ~0.015 e- per nm2 which is smaller than that on silica surface.179 These O- groups can also form hydrogen bonds with bitumen, but the density of hydrogen bonds between mica and bitumen is smaller than that between silica and bitumen. So this explains why bitumen on the mica surface has a larger liberation rate compared to silica. As for freshly cleaved mica, there is neither surface charge on the surface nor hydrogen bond formation between mica and bitumen, thus the force between cleaved mica  94  and bitumen should be the smallest and it explains why bitumen on mica has the largest liberation rate in our experiments performed on three different mineral surfaces. The surface charge density and hydrogen bond formation can also explain the phenomenon that water content on silica and rinsed mica surfaces increases the bitumen dynamic contact angle changing rate. In the silica experiment, for example, silica has a negative surface charge and the silanol groups can form hydrogen bonds with bitumen. At RH = 0% bitumen contacts the silica surface directly, so the attractive electrostatic force should be largest and the number of hydrogen bonds between bitumen and silanol groups should also be largest. As the RH increases, the water content on the silica surface increases. The negative charge on the silica surface will be partially shielded by water molecules so the attractive force will decrease. Because of the existence of water molecules on the silica surface, bitumen interact less with the silanol groups. Therefore, the number of hydrogen bonds will also decrease. Consequently, the water content on the silica surface will help the liberation of bitumen. The explanation for the rinsed mica experiments is similar to that for silica. For the cleaved mica surface, there is no surface charge on the surface, and no hydrogen bond formation between cleaved mica and bitumen. Therefore, the water content has little effect on the bitumen liberation rate, and this is consistent with what we observed in the experiment.  95  7.4 Conclusions  SFG vibrational spectroscopy and dynamic contact angle measurements experiments were applied to study the effect of interfacial water content on bitumen liberation from mineral surfaces in water. The time dependence of the dynamic contact angle of bitumen on the silica, freshly cleaved mica, and rinsed mica were compared with different water content on mineral surfaces. The bitumen liberation rate increases with higher water content on the mineral surfaces for silica and rinsed mica. At the same water content, the rate of bitumen liberation on different mineral surfaces is: freshly cleaved mica > rinsed mica > silica. To explain the differences in bitumen liberation on the three minerals and to understand the role of water content on mineral surfaces on bitumen liberation, the surface charge density of the mineral surfaces and the formation of hydrogen bond between the mineral surfaces and bitumen were considered.  96  Chapter 8  Conclusion  97  This thesis presents studies of petrolic, edaphic and atmospheric relevant liquid/mineral interfaces using SFG, which include water structures at aqueous salt solutions/silica interfaces, solvent/silica interfaces, and sulphuric acid solution/mica interfaces, effects of interfacial water content on bitumen liberation from silica and mica surfaces, and competitive adsorption of toluene and n-alkanes on silica interfaces. The cations play a key role in perturbing the hydrogen-bond network at the water/silica interfaces as they interact with the silica surface via electrostatic interaction (Chapter 3). Significant perturbations of the interfacial water structures were observed with a 1x10-4 M NaCl solution. Different alkali cation species produce different degrees of perturbation in the order: K+ > Li+ > Na+. This order can be explained by considering the electrostatic interaction between the cations and silica surfaces and the effective ionic radii of the cations. Our current SFG setup cannot provide the information on the orientation of water molecules. Phase-sensitive SFG vibrational spectroscopy is needed for further experiments to study how the orientation of water molecules responds to the concentration changes of alkali cations. Water molecules adsorbed at the toluene/silica interface form a highly H-bonded layer with no detectable free OHs (Chapter 4). The water layer without free OHs showed resistance against further adsorption of water molecules. However, this special water structure was not observed at heptane/silica interfaces, at which free OHs were always observed. The experimental data showed that the interactions between solvents and water molecules can significantly change the interfacial water properties. However, the reason why toluene aided the formation of “water-resistant” structure is still unclear. More  98  experiments with different solvents and minerals may reveal the mechanism for the formation of this “water-resistant” water layer. The structures of water on the mica surfaces in the presence of H2SO4 with atmospheric relevant concentrations were studied (Chapter 5). We found that the ordered water structure disappeared when the concentration of H2SO4 reached 5 mol/L. The observed phenomenon was a combined effect of the decreased mica surface potential at low pH, the adsorption of sulfates on mica, and the lack of free water molecules in highly concentrated acidic solution. The results have offered a microscopic understanding for why H2SO4-coated mineral surfaces are poor ice nuclei. The good ice nucleation ability above liquid water saturation is correlated with the presence of structured water suggesting that structured water at the interface may be needed for efficient heterogeneous ice nucleation. More experiments could be done to confirm this explanation. Water structures on efficient ice nuclei minerals (such as kaolinite and muscovite) and poor ice nuclei minerals (such as quartz and calcite) can be studied by SFG. If the proposed mechanism is correct, water will form ordered structure on efficient ice nuclei minerals, while disordered water structure will be found on poor ice nuclei minerals. However, some minerals are not flat or transparent. In this case, SFG signal will be scattered, and the signal lost could be a potential issue. Besides water/solid interfaces, competitive adsorption of toluene and n-alkanes at binary solution/silica interfaces was studied (Chapter 6). The surface coverage of toluene on silica for toluene-pentane, toluene-heptane, and toluene-tetradecane mixtures was obtained using the measured SFG peaks intensity of toluene. The competitive adsorption processes are reversible, and the surface coverage of toluene only depends on the toluene  99  molar fraction in the binary mixture, not on the history of the mixture in contact with the silica. The measured adsorption isotherms fitted well with the Everett isotherm over the complete mole fraction range. Overall, toluene competes favorably on silica against the alkanes, and the molar adsorption free energy difference between toluene and alkane decreases as the chain length of the alkane increases. Theoretically the surface coverage of any molecule on a solid surface accessible by light can be measured with a similar method as long as the molecule has a unique vibration peak. In reality most organic molecules have similar vibration peaks. However, if the target molecule can be deuterated, this method will still work. SFG and dynamic contact angle measurements experiments were applied to study the effect of relative humidity (RH) on bitumen displacement by an aqueous phase on mineral surfaces (Chapter 7). The time variation of the dynamic contact angle of bitumen on the silica plate, freshly cleaved mica pieces, rinsed mica pieces were compared with different water content on mineral surfaces. The bitumen displacement rate increases with higher water content on mineral surfaces. At the same water content, the rate of bitumen displacement on different mineral surfaces is: freshly cleaved mica > rinsed mica > silica. To explain the differences in bitumen displacement on the three minerals and to understand the role of water content on bitumen displacement on mineral surfaces, the surface charge density of mineral surfaces and the formation of hydrogen bonds between the mineral surfaces and bitumen should be considered. In our current setup the visible beam is 532 nm which can stimulate our bitumen samples to give off strong fluorescence. Therefore, it is impossible to detect the SFG signal from water molecules between minerals and bitumen. In the future study, the visible beam with a longer wave length can  100  be used to minimize the fluorescence from bitumen, then information of water molecules between minerals and bitumen can be gathered.  101  References (1) Adamson, A. W.; Gast, A. P. Physical chemistry of surfaces, 6th ed.; Wiley: New York, 1997. (2) Evans, R.; Marconi, U. M. B. J. Chem. Phys. 1987, 86, 7138. (3) Israelachvili, J.; Wennerstrom, H. Nature 1996, 379, 219. (4) Cerda, J.; Michaelides, A.; Bocquet, M. L.; Feibelman, P. J.; Mitsui, T.; Rose, M.; Fomin, E.; Salmeron, M. Phys. Rev. Lett. 2004, 93. (5) Mitsui, T.; Rose, M. K.; Fomin, E.; Ogletree, D. F.; Salmeron, M. Science 2002, 297, 1850. (6) Tyrrell, J. W. G.; Attard, P. Phys. Rev. Lett. 2001, 87. (7) Shen, Y. R. Nature 1989, 337, 519. (8) Miranda, P. B.; Shen, Y. R. J. Phys. Chem. B 1999, 103, 3292. (9) Eisenthal, K. B. Chem. Rev. 1996, 96, 1343. (10) Richmond, G. L. Chem. Rev. 2002, 102, 2693. (11) Du, Q.; Freysz, E.; Shen, Y. R. Science 1994, 264, 826. (12) Fisk, J. D.; Batten, R.; Jones, G.; O'Reilly, J. P.; Shaw, A. M. J. Phys. Chem. B 2005, 109, 14475. (13) Gurau, M. C.; Kim, G.; Lim, S. M.; Albertorio, F.; Fleisher, H. C.; Cremer, P. S. Chemphyschem 2003, 4, 1231. (14) Hassanali, A. A.; Singer, S. J. J. Phys. Chem. B 2007, 111, 11181. (15) Helms, C. R.; Deal, B. E.; Electrochemical Society. Electronics Division.; Electrochemical Society. Dielectric Science and Technology Division. The Physics and chemistry of SiO b2 s and the Si-SiO b2 s interface 2; Plenum Press: New York, 1993. (16) Kim, J.; Kim, G.; Cremer, P. S. Langmuir 2001, 17, 7255. (17) Kim, J.; Kim, G.; Cremer, P. S. J. Am. Chem. Soc. 2002, 124, 8751. (18) Konek, C. T.; Musorrafiti, M. J.; Al-Abadleh, H. A.; Bertin, P. A.; Nguyen, S. T.; Geiger, F. M. J. Am. Chem. Soc. 2004, 126, 11754. (19) O'Reilly, J. P.; Butts, C. P.; I'Anson, I. A.; Shaw, A. M. J. Am. Chem. Soc. 2005, 127, 1632. (20) Shen, Y. R.; Ostroverkhov, V. Chem. Rev. 2006, 106, 1140. (21) Blokzijl, W.; Engberts, J. B. F. N. Angew. Chem. Int. Ed. 1993, 32, 1545. (22) Sefler, G. A.; Du, Q.; Miranda, P. B.; Shen, Y. R. Chem. Phys. Lett. 1995, 235, 347. (23) Miranda, P. B.; Pflumio, V.; Saijo, H.; Shen, Y. R. J. Am. Chem. Soc. 1998, 120, 12092. (24) Motokura, K.; Tada, M.; Iwasawa, Y. J. Am. Chem. Soc. 2007, 129, 9540. (25) Brindza, M. R.; Walker, R. A. J. Am. Chem. Soc. 2009, 131, 6207. (26) Brindza, M. R.; Ding, F.; Fourkas, J. T.; Walker, R. A. J. Chem. Phys. 2010, 132. (27) Steel, W. H.; Walker, R. A. Nature 2003, 424, 296. (28) Acevedo, S.; Ranaudo, M. A.; Garcia, C.; Castillo, J.; Fernandez, A. Energy Fuels 2003, 17, 257. (29) Arabshahi, S. H.; Ackerson, M. D.; Rye, W. C.; Babcock, R. E. Chem. Eng. Commun. 1990, 89, 195. 102  (30) Grandel, S.; Dahmke, A. Biodegradation 2004, 15, 371. (31) Hubert, A.; Wenzel, K. D.; Manz, M.; Weissflog, L.; Engewald, W.; Schuurmann, G. Anal. Chem. 2000, 72, 1294. (32) Kenchington, J. M.; Phillips, C. R. Energy Sources 1981, 5, 317. (33) Khaitan, S.; Kalainesan, S.; Erickson, L. E.; Kulakow, P.; Martin, S.; Karthikeyan, R.; Hutchinson, S. L. L.; Davis, L. C.; Illangasekare, T. H.; Ng'oma, C. Environ. Prog. 2006, 25, 20. (34) Long, Y. C.; Dabros, T.; Hamza, H. Fuel 2004, 83, 823. (35) Luo, Y.; Zou, L.; Hu, E. Environ. Technol. 2006, 27, 359. (36) Saim, N.; Dean, J. R.; Abdullah, M. P.; Zakaria, Z. J. Chromatogr. A 1997, 791, 361. (37) Scullion, J. Naturwissenschaften 2006, 93, 51. (38) Stasiuk, E. N.; Schramm, L. L. Fuel Process. Technol. 2001, 73, 95. (39) Fowkes, F. M. Ind. Eng. Chem 1964, 56, 40. (40) Blackburn, A.; Kipling, J. J. J. Chem. Soc. 1954, 3819. (41) Everett, D. H. Pure Appl. Chem. 1986, 58, 967. (42) Oscik, J. Adsorption; Chichester: New York, 1982. (43) Burrowes, A.; Marsh, R.; Ramdin, N.; Evans, C. Alberta Energy and Utilities Board 2007. (44) Clark, K. A.; Pasternak, S. D. Ind. Eng. Chem. 1932, 24, 1410. (45) Basu, S.; Kanda, W. C.; Nandakumar, K.; Masliyah, J. H. Ind. Eng. Chem. Res. 1998, 37, 959. (46) Basu, S.; Nandakumar, K.; Lawrence, S.; Masliyah, J. Fuel 2004, 83, 17. (47) Basu, S.; Nandakumar, K.; Masliyah, J. H. J. Colloid Interface Sci. 1996, 182, 82. (48) Basu, S.; Nandakumar, K.; Masliyah, J. H. Colloids Surf., A 1998, 136, 71. (49) Dang-Vu, T.; Jha, R.; Wu, S. Y.; Tannant, D. D.; Masliyah, J.; Xu, Z. H. Energy Fuels 2009, 23, 2628. (50) Zhu, X. D.; Suhr, H.; Shen, Y. R. Phys. Rev. B 1987, 35, 3047. (51) Lambert, A. G.; Davies, P. B.; Neivandt, D. J. Appl. Spectrosc. Rev. 2005, 40, 103. (52) Zhuang, X.; Miranda, P. B.; Kim, D.; Shen, Y. R. Phys. Rev. B 1999, 59, 12632. (53) Du, Q.; Superfine, R.; Freysz, E.; Shen, Y. R. Phys. Rev. Lett. 1993, 70, 2313. (54) Shen, Y. R. The principles of nonlinear optics; J. Wiley: New York, 1984. (55) Casson, B. D.; Bain, C. D. Langmuir 1997, 13, 5465. (56) Hirose, C.; Akamatsu, N.; Domen, K. J. Chem. Phys. 1992, 96, 997. (57) Zhang, D.; Gutow, J.; Eisenthal, K. B. J. Phys. Chem. 1994, 98, 13729. (58) Frenkel, J. Kinetic Theory of Liquids; Dover: New York, 1955. (59) Lee, C. Y.; McCammon, J. A.; Rossky, P. J. J. Chem. Phys. 1984, 80, 4448. (60) Rossky, P. J.; Lee, S. H. Chemica Scripta 1989, 29A, 93. (61) Lee, S. H.; Rossky, P. J. J. Chem. Phys. 1994, 100, 3334. (62) Jedlovszky, P.; Vincze, A.; Horvai, G. J. Mol. Liq. 2004, 109, 99. (63) Jedlovszky, P.; Vincze, A.; Horvai, G. PCCP 2004, 6, 1874. 103  (64) Puibasset, J.; Pellenq, R. J. M. J. Chem. Phys. 2003, 119, 9226. (65) Puibasset, J.; Pellenq, R. J. M. PCCP 2004, 6, 1933. (66) Benjamin, I. J. Chem. Phys. 2004, 121, 10223. (67) Toney, M. F.; Howard, J. N.; Richer, J.; Borges, G. L.; Gordon, J. G.; Melroy, O. R.; Wiesler, D. G.; Yee, D.; Sorensen, L. B. Nature 1994, 368, 444. (68) Reedijk, M. F.; Arsic, J.; Hollander, F. F. A.; de Vries, S. A.; Vlieg, E. Phys. Rev. Lett. 2003, 90. (69) Chu, Y. S.; Lister, T. E.; Cullen, W. G.; You, H.; Nagy, Z. Phys. Rev. Lett. 2001, 86, 3364. (70) Ruan, C. Y.; Lobastov, V. A.; Vigliotti, F.; Chen, S. Y.; Zewail, A. H. Science 2004, 304, 80. (71) Ong, S. W.; Zhao, X. L.; Eisenthal, K. B. Chem. Phys. Lett. 1992, 191, 327. (72) Du, Q.; Freysz, E.; Shen, Y. R. Chem. Phys. Lett. 1994, 72, 238. (73) Ostroverkhov, V.; Waychunas, G. A.; Shen, Y. R. Phys. Rev. Lett. 2005, 94. (74) Liu, D. F.; Ma, G.; Levering, L. M.; Allen, H. C. J. Phys. Chem. B 2004, 108, 2252. (75) Raymond, E. A.; Richmond, G. L. J. Phys. Chem. B 2004, 108, 5051. (76) Iler, R. K. The Chemistry of Silica; Wiley: New York, 1979. (77) Li, I.; Bandara, J.; Shultz, M. J. Langmuir 2004, 20, 10474. (78) Kendall, J. J. Am. Chem. Soc. 1916, 38, 2460. (79) Kendall, J. J. Am. Chem. Soc. 1916, 38, 1480. (80) Schrodle, S.; Richmond, G. L. J. Phys. D: Appl. Phys. 2008, 41, 14. (81) Bousse, L.; Meindl, J. D. Acs Symposium Series 1986, 323, 79. (82) Sahai, N.; Sverjensky, D. A. Geochim. Cosmochim. Acta 1997, 61, 2827. (83) The Colloid Chemistry of Silica; Bergna, H. E., Ed.; American Chemistry Society: Washington, DC, 1994. (84) Ostroverkhov, V.; Waychunas, G. A.; Shen, Y. R. Chem. Phys. Lett. 2004, 386, 144. (85) Kosmulski, M. Chemical Properties of Material Surfaces; CRC Press, 2001. (86) Dove, P. M.; Craven, C. M. Geochim. Cosmochim. Acta 2005, 69, 4963. (87) Kielland, J. J. Am. Chem. Soc. 1937, 59, 1675. (88) Malati, M. A.; Estefan, S. F. J. Colloid Interface Sci. 1966, 22, 306. (89) Malati, M. A. Surf. Coat. Technol. 1987, 30, 317. (90) Kosmulski, M. Ber. Bunsen-Ges. Phys. Chem 1994, 98, 1062. (91) Kosmulski, M. Colloids Surf., A 1994, 83, 237. (92) Perera, L.; Berkowitz, M. L. J. Chem. Phys. 1991, 95, 1954. (93) Lee, H. M.; Kim, J.; Lee, S.; Mhin, B. J.; Kim, K. S. J. Chem. Phys. 1999, 111, 3995. (94) Bock, C. W.; Markham, G. D.; Katz, A. K.; Glusker, J. P. Theor. Chem. Acc. 2006, 115, 100. (95) Thomas, A. S.; Elcock, A. H. J. Am. Chem. Soc. 2007, 129, 14887. (96) Morais-Cabral, J. H.; Zhou, Y. F.; MacKinnon, R. Nature 2001, 414, 37.  104  (97) Zhou, Y. F.; Morais-Cabral, J. H.; Kaufman, A.; MacKinnon, R. Nature 2001, 414, 43. (98) Chuang, I. S.; Maciel, G. E. J. Phys. Chem. B 1997, 101, 3052. (99) Liu, J. J.; Xu, Z. H.; Masliyah, J. Energy Fuels 2005, 19, 2056. (100) Scatena, L. F.; Brown, M. G.; Richmond, G. L. Science 2001, 292, 908. (101) Taylor, D. R.; Maher, K. J. Chromatogr. Sci. 1992, 30, 67. (102) Crego, A. L.; Gonzalez, A.; Marina, M. L. Crit. Rev. Anal. Chem. 1996, 26, 261. (103) Vissers, J. P. C.; Claessens, H. A.; Cramers, C. A. J. Chromatogr. A 1997, 779, 1. (104) Vervoort, R. J. M.; Debets, A. J. J.; Claessens, H. A.; Cramers, C. A.; de Jong, G. J. J. Chromatogr. A 2000, 897, 1. (105) Yang, Z.; Li, Q. F.; Hua, R.; Gray, M. R.; Chou, K. C. J. Phys. Chem. C 2009, 113, 20355. (106) Bittrich, H. J.; Gedan, H.; Feix, G. Z. Phys. Chem. 1979, 260, 1009. (107) Polak, J.; Lu, B. C. Y. Can. J. Chem. 1973, 51, 4018. (108) Schatzberg, P. J. Phys. Chem. 1963, 67, 776. (109) Glasoe, P. K.; Schultz, S. D. J. Chem. Eng. Data 1972, 17, 66. (110) Gregory, M. D.; Christia.Sd; Affsprun.He. J. Phys. Chem. 1967, 71, 2283. (111) Kimmel, G. A.; Petrik, N. G.; Dohnalek, Z.; Kay, B. D. Phys. Rev. Lett. 2005, 95, 166102. (112) Hu, J.; Xiao, X. D.; Ogletree, D. F.; Salmeron, M. Science 1995, 268, 267. (113) Odelius, M.; Bernasconi, M.; Parrinello, M. Phys. Rev. Lett. 1997, 78, 2855. (114) Miranda, P. B.; Xu, L.; Shen, Y. R.; Salmeron, M. Phys. Rev. Lett. 1998, 81, 5876. (115) Pruppacher, H. R.; Klett, J. D. Microphysics of clouds and precipitation, 2nd rev. and enl. ed.; Kluwer Academic Publishers: Dordrecht ; Boston, 1997. (116) DeMott, P. J.; Sassen, K.; Poellot, M. R.; Baumgardner, D.; Rogers, D. C.; Brooks, S. D.; Prenni, A. J.; Kreidenweis, S. M. Geophys. Res. Lett. 2003, 30. (117) Hung, H. M.; Malinowski, A.; Martin, S. T. J. Phys. Chem. A 2003, 107, 1296. (118) Archuleta, C. M.; DeMott, P. J.; Kreidenweis, S. M. Atmos. Chem. Phys. 2005, 5, 2617. (119) Mangold, A.; Wagner, R.; Saathoff, H.; Schurath, U.; Giesemann, C.; Ebert, V.; Kramer, M.; Mohler, O. Meteorol. Z. 2005, 14, 485. (120) Sassen, K. Geophys. Res. Lett. 2002, 29. (121) Sassen, K. Nature 2005, 434, 456. (122) DeMott, P. J.; Cziczo, D. J.; Prenni, A. J.; Murphy, D. M.; Kreidenweis, S. M.; Thomson, D. S.; Borys, R.; Rogers, D. C. PNAS 2003, 100, 14655. (123) Sassen, K.; DeMott, P. J.; Prospero, J. M.; Poellot, M. R. Geophys. Res. Lett. 2003, 30. (124) Toon, O. B. Nature 2003, 424, 623. (125) Twohy, C. H.; Poellot, M. R. Atmos. Chem. Phys. 2005, 5, 2289. (126) Wiacek, A.; Peter, T. Geophys. Res. Lett. 2009, 36.  105  (127) Sullivan, R. C.; Guazzotti, S. A.; Sodeman, D. A.; Prather, K. A. Atmos. Chem. Phys. 2007, 7, 1213. (128) Cziczo, D. J.; Froyd, K. D.; Gallavardin, S. J.; Moehler, O.; Benz, S.; Saathoff, H.; Murphy, D. M. Environ. Res. Lett. 2009, 4. (129) Eastwood, M. L.; Cremel, S.; Wheeler, M.; Murray, B. J.; Girard, E.; Bertram, A. K. Geophys. Res. Lett. 2009, 36. (130) Gallavardin, S. J.; Froyd, K. D.; Lohmann, U.; Moehler, O.; Murphy, D. M.; Cziczo, D. J. Aerosol Sci. Technol. 2008, 42, 773. (131) Knopf, D. A.; Koop, T. J. Geophys. Res. Atmos. 2006, 111. (132) Chernoff, D. I.; Bertram, A. K. J. Geophys. Res. Atmos. 2010, 115. (133) Sullivan, R. C.; Petters, M. D.; DeMott, P. J.; Kreidenweis, S. M.; Wex, H.; Niedermeier, D.; Hartmann, S.; Clauss, T.; Stratmann, F.; Reitz, P.; Schneider, J.; Sierau, B. Atmos. Chem. Phys. 2010, 10, 11471. (134) Zimmermann, F.; Ebert, M.; Worringen, A.; Schutz, L.; Weinbruch, S. Atmos. Environ. 2007, 41, 8219. (135) Eastwood, M. L.; Cremel, S.; Gehrke, C.; Girard, E.; Bertram, A. K. J. Geophys. Res. Atmos. 2008, 113. (136) Clegg, S. L.; Brimblecombe, P.; Wexler, A. S. J. Phys. Chem. A 1998, 102, 2137. (137) Glaccum, R. A.; Prospero, J. M. Mar. Geol. 1980, 37, 295. (138) Serratosa, J. M.; Bradley, W. F. Nature 1958, 181, 111. (139) Stamenkovic, V.; Chou, K. C.; Somorjai, G. A.; Ross, P. N.; Markovic, N. M. J. Phys. Chem. B 2005, 109, 678. (140) Gragson, D. E.; Richmond, G. L. J. Am. Chem. Soc. 1998, 120, 366. (141) Watanabe, M.; Brodsky, A. M.; Reinhardt, W. P. J. Phys. Chem. 1991, 95, 4593. (142) Gavish, M.; Wang, J. L.; Eisenstein, M.; Lahav, M.; Leiserowitz, L. Science 1992, 256, 815. (143) Xu, L.; Lio, A.; Hu, J.; Ogletree, D. F.; Salmeron, M. J. Phys. Chem. B 1998, 102, 540. (144) Claesson, P. M.; Herder, P.; Stenius, P.; Eriksson, J. C.; Pashley, R. M. J. Colloid Interface Sci. 1986, 109, 31. (145) Herder, P. C.; Claesson, P. M.; Herder, C. E. J. Colloid Interface Sci. 1987, 119, 155. (146) Scales, P. J.; Grieser, F.; Healy, T. W. Langmuir 1990, 6, 582. (147) Lyons, J. S.; Furlong, D. N.; Healy, T. W. Aust. J. Chem. 1981, 34, 1177h. (148) Chao, T. T.; Harward, M. E.; Fang, S. C. Soil Sci. Soc. Am. Proc. 1964, 28, 632. (149) Parfitt, R. L. Adv. Agron. 1978, 30, 1. (150) Zhang, G. Y.; Zhang, X. N.; Yu, T. R. J. Soil Sci. 1987, 38, 29. (151) Courchesne, F.; Hendershot, W. H. Can. J. Soil Sci. 1989, 69, 337. (152) Cannon, W. R.; Pettitt, B. M.; McCammon, J. A. J. Phys. Chem. 1994, 98, 6225. (153) Zhang, L. N.; Liu, W. T.; Shen, Y. R.; Cahill, D. G. J. Phys. Chem. C 2007, 111, 2069.  106  (154) Gracias, D. H.; Chen, Z.; Shen, Y. R.; Somorjai, G. A. Acc. Chem. Res. 1999, 32, 930. (155) Shultz, M. J.; Schnitzer, C.; Simonelli, D.; Baldelli, S. Int. Rev. Phys. Chem. 2000, 19, 123. (156) Chen, Z.; Shen, Y. R.; Somorjai, G. A. Annu. Rev. Phys. Chem. 2002, 53, 437. (157) Morita, A.; Hynes, J. T. Chem. Phys. 2000, 258, 371. (158) Blum, L.; Huckaby, D. A. J. Chem. Phys. 1991, 94, 6887. (159) Zhu, B. Y.; Gu, T. R. Adv. Colloid Interface Sci. 1991, 37, 1. (160) Chang, C. H.; Franses, E. I. Colloids Surf., A 1995, 100, 1. (161) Zhang, Z. H.; Tsuyumoto, I.; Takahashi, S.; Kitamori, T.; Sawada, T. J. Phys. Chem. A 1997, 101, 4163. (162) Cosman, N. P.; Roscoe, S. G. Langmuir 2004, 20, 1711. (163) Mohan, D.; Pittman, C. U.; Steele, P. H. J. Colloid Interface Sci. 2006, 297, 489. (164) Hommel, E. L.; Allen, H. C. Analyst 2003, 128, 750. (165) Choi, P. private communication, 2009. (166) Macphail, R. A.; Strauss, H. L.; Snyder, R. G.; Elliger, C. A. J. Phys. Chem. 1984, 88, 334. (167) Heller, W. J. Phys. Chem. 1965, 69, 1123. (168) Brocos, P.; Pineiro, A.; Bravo, R.; Amigo, A. PCCP 2003, 5, 550. (169) Dean, J. A.; Lange, N. A. Lange's handbook of chemistry, 15th ed.; McGraw-Hill: New York :, 1999. (170) Fu, Y.; Hansen, R. S.; Bartell, F. E. J. Phys. Colloid Chem. 1948, 52, 374. (171) Everett, D. H.; Munn, R. J. Trans. Faraday Soc. 1964, 60, 1951. (172) Ash, S. G.; Everett, D. H.; Findeneg.Gh. Trans. Faraday Soc. 1970, 66, 708. (173) Ash, S. G.; Bown, R.; Everett, D. H. J. Chem. Thermodyn. 1973, 5, 239. (174) Basu, S.; Nandakumar, K.; Masliyah, J. H. J. Colloid Interface Sci. 1998, 205, 201. (175) Desyatov, I. V.; Paukshtis, E. A.; Mashkina, A. V. React. Kinet. Catal. Lett. 1990, 41, 161. (176) Liu, J. J.; Zhou, Z.; Xu, Z. H. Ind. Eng. Chem. Res. 2002, 41, 52. (177) Liu, J. J.; Xu, Z. H.; Masliyah, J. Langmuir 2003, 19, 3911. (178) Smith, C.; Chatergoon, L.; Whiting, R. Analyst 1996, 121, 373. (179) Pashley, R. M. J. Colloid Interface Sci. 1981, 80, 153.  107  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0059841/manifest

Comment

Related Items