Molecular Bonding and Interactions at Aqueous Surfaces as...

Molecular Bonding and Interactions at Aqueous Surfaces as Probed byVibrational Sum Frequency Spectroscopy

G. L. Richmond

Department of Chemistry, University of Oregon, Eugene, Oregon 97403

Received September 25, 2001

ContentsI. Introduction 2693II. Vibrational Sum Frequency Spectroscopy 2694

A. Principles and Concepts 2694B. Experimental Considerations 2697

III. Hydrogen Bonding and Structure of Water atAqueous Surfaces

2698

A. Vapor/Water Interface 2698B. Organic/Water Interfaces 2701C. Water at Solid Surfaces 2703

1. Surface Melting of Ice 27032. Water Adsorbed at Solid Substrate

Surfaces2704

D. Effect of Adsorbates and Ions on SurfaceWater Structure and Bonding

2706

1. Alkyl Surfactants Adsorbed at theVapor/Water Interface

2706

2. Alkyl Surfactants Adsorbed at Organic/Water Interfaces

2708

3. Solutes, Acids, and Salts in the AqueousPhase

2709

IV. Adsorbate Structure and Bonding at AqueousSurfaces

2709

A. Surfactants at Vapor/Water Interfaces 2709B. Water-Soluble Solutes Adsorbed at the

Vapor/Water Interface2714

C. Surfactants Adsorbed at Organic/WaterInterfaces

2716

1. Charged Alkyl Surfactants 27162. Biomolecules 2718

D. Surfactants and Adsorbates at Solid/AqueousInterfaces

2720

E. Electrochemical Interfaces 2720F. Polymer Surfaces 2721

V. Summary and Conclusions 2721VI. Acknowledgment 2722VII. References 2722

I. IntroductionAqueous surfaces and interfaces are important in

many physical, chemical, and biological processes inour world. The adsorption, dissolution, and reactionof atmospheric gases at the surfaces of atmosphericaerosols and oceanic waters play a key role in thecomposition of our atmosphere and the sustainabilityof plant and animal species in land waters. The

transport and exchange of ions and solutes across theinterface between an aqueous phase and hydrophobicbiomolecular assemblies underlies some of the mostimportant processes in living plants and animals.Membrane formation, protein folding, and micelleformation all involve, and are often controlled by,bonding interactions with water molecules at theirsurfaces. The unique physical, chemical, and biologi-cal properties of aqueous surfaces arise from thestrong hydrogen bonding that occurs between watermolecules and the asymmetry in this otherwisetetrahedral bonding coordination that results fromthe termination of the bulk water phase. Althoughthere has been increased experimental and theoreti-cal effort in recent years focused on developing amolecular picture of the structure and bonding ofwater layers to other solid, liquid, and gaseous media,consensus on the details of interfacial bonding hasbeen slow or nonexistent in many areas. The adsorp-tion of ions, surfactants, and solute molecules atthese interfaces adds a level of complexity to the

Geraldine Richmond holds the Richard M. and Patricia H. Noyes Professorof Chemistry position at the University of Oregon. She received her Ph.D.degree under the mentorship of George Pimentel at the University ofCalifornia, Berkeley, in 1980. From 1980 to 1985 she was on the facultyat Bryn Mawr College and moved to the University of Oregon in 1985 asan associate professor. Richmond is recognized for her fundamentalstudies of the structure, dynamics, and bonding characteristics of surfacesand interfaces. Her research group uses a combination of linear andnonlinear optical methods, thermodynamic measurements, and theory tocharacterize interactions at aqueous surfaces, metal and semiconductorsurfaces in contact with liquids and adsorbates, and liquid/liquid interfaces.Richmond has received several recent honors for these studies includingthe 2002 ACS Spectrochemical Analysis Award, the 2001 Oregon Scientistof the Year, and the 1996 Olin-Garvan Medal of the ACS and has beena Fellow of the American Physical Society since 1993.

2693Chem. Rev. 2002, 102, 2693−2724

10.1021/cr0006876 CCC: $39.75 © 2002 American Chemical SocietyPublished on Web 07/27/2002

interfacial picture that is only beginning to be ad-dressed in molecular level experimental and theoreti-cal efforts.

A major factor in the slow progress made inunderstanding hydrogen bonding and adsorption ataqueous surfaces is the paucity of experimentalstudies that selectively probe in situ these aqueoussurfaces and interfaces. In the realm of surfacescience, the field is in the dark ages relative to whatwe know about adsorption of molecules on solidsurfaces under more controlled vacuum conditions.Much of the progress for aqueous surfaces has comefrom theoretical efforts over the past several dec-ades,1-15 but even these studies are limited to rela-tively simple descriptions of the interactions betweenwater molecules, solutes, and adsorbates at aqueoussurfaces. It is clear that the challenge to fullyunderstand surface water hydrogen bonding andadsorption at aqueous interfaces will continue intothe future due to the complexity of interactionspossible at these interfaces. However, the challengeis worth pursuing with vigor because of the centralrole that aqueous surfaces play in all aspects of ouratmospheric and oceanic environment, in the mostimportant biochemical processes in our body includ-ing respiration, ion transport, and protein folding,and in technological areas that range from oil extrac-tion to semiconductor processing.

What is clearly needed in order to make progressin this area are more experimental studies of theseinterfaces that can selectively probe the interfacialregion and coupling of these experimental studieswith theoretical efforts. Recent advances in a numberof newly developed experimental techniques forspecifically studying liquid surfaces bode well for thefuture.16-32 The focus of this review is on the contri-bution of vibrational sum frequency spectroscopy(VSFS) to our understanding of the bonding andstructure at aqueous surfaces.33 As a relatively newvibrational spectroscopic method, it is showing par-ticular promise for measuring the molecular spec-troscopy of water and adsorbed molecules at liquidinterfaces because of its inherent ability to discrimi-nate between molecules that reside in that thin mo-lecular layer that defines a surface and those mol-ecules in the centrosymmetric bulk media. Whereasa decade ago the technique was used only by ahandful of investigators, its use is rapidly growingas more investigators are discovering the uniqueinformation that can be gained from its applicabilityto a wide range of systems. This review provides anoverview of the studies that have thus far employedVSFS to study aqueous surfaces and interfaces, witha particular focus on the structure and hydrogenbonding of water at the surface of neat water andaqueous solutions studied at the vapor/water inter-face and the interface between two immiscible liquids(organic/water interface). It highlights many of therecent studies in the field as a means of demonstrat-ing the type of information obtainable. In addition,it seeks to provide the reader with a sense of wherefuture development is needed in the technique andanalysis methods in order for VSFS to reach thegeneral applicability as a vibrational spectroscopic

method that vibrational infrared (IR) and Ramanspectroscopies have attained.

The article begins with a description of the tech-nique, conceptual underpinnings, experimental con-siderations, and issues that must be considered inthe analysis of VSF data. This is followed by a sectionthat focuses on VSF studies of the neat vapor/waterinterface. The neat vapor/water interface is one of thesimplest systems to examine from a chemical per-spective, and yet there are many unknowns aboutthe molecular structure and bonding at this interfaceand the related vapor/ice interface. This is followedby an overview of what has been learned throughVSF studies about water structure hydrogen bondingat an interface between water and an immiscibleliquid, referred to as the organic/water interface. Aswill be shown, there are distinct differences andsimilarities in the nature of hydrogen bonding atthese two interfaces that will play an important rolein our understanding of many processes in livingorganisms and in our environment. Water structur-ing next to a hydrophobic surface is central to issuesof wettability, protein folding, chemical separations,and oil extractionsto name only a few. Beyond theseneat water surface studies, this review will examinehow the hydrogen bonding at water surfaces isaltered by the presence of surfactants at the surfaceand by variation of the composition of the aqueousphase by addition of salts, acids, and bases.

The second portion of the review examines themolecular structure of adsorbates at aqueous sur-faces. This is divided into two parts: the first focuseson small solute molecules adsorbed at aqueous sur-faces, with atmospherically important molecules be-ing the most prevalent systems examined. In thesestudies the molecular orientation and bonding ofsulfur- and nitrogen-containing solutes have beenexamined. The second part discusses VSF studies ofalkyl surfactants adsorbed at both the vapor/waterand organic/water interfaces. A wide variety of com-mercially important surfactants have been examinedin these studies with the primary focus on measuringconformational ordering and orientation of alkylportions of the surfactants. Recent studies of biologi-cal surfactants, namely, phospholipids, will also bedescribed. Phospholipids are important componentsof biological membranes, and understanding theirbehavior at these aqueous interfaces is important forunderstanding a host of important biological pro-cesses.

II. Vibrational Sum Frequency Spectroscopy

A. Principles and ConceptsVibrational spectroscopy has long been recognized

as an important tool for measuring molecular struc-ture in gases, liquids, and solids. This is as true formolecules at surfaces as it is for molecules in bulkmedia. Vibrational sum frequency spectroscopy is atechnique that is uniquely suited for measuring thevibrational spectrum of molecules at surfaces.33 In aVSFS experiment, light pulses from a visible laserbeam and a tunable IR laser beam are coincident intime and space at the interface. Figure 1 provides

2694 Chemical Reviews, 2002, Vol. 102, No. 8 Richmond

an illustration of the experiment for two differentsystems, the vapor/water interface (above) and theliquid/liquid interface (below). The high-intensityelectric fields of the incident laser beams induce acoherent nonlinear polarization in the molecules atthe interface, and this oscillating nonlinear polariza-tion at the sum of the two frequencies is the sourceof SF light collected. When the infrared source istuned through the spectral region of interest, coin-cidence between the photon energy and the energyof the molecular vibrational mode results in a reso-nant enhancement in the SF response. It is similarto the optically simpler second harmonic generation(SHG) process that involves the summation of twofixed frequency light beams.31,33-36 The surface speci-ficity arises from the second-order nature of theresponse. Under the dipole approximation, this second-order response is forbidden in media possessinginversion symmetry.34,37 At the interface between twocentrosymmetric media there is no inversion centerand the SF process is allowed. This inherent surfacesensitivity makes it an advantageous method ofstudying the vibrational spectroscopy of molecules atsurfaces over other linear vibrational spectroscopiessuch as infrared or Raman spectroscopy. The disad-vantage to date over these other methods has beenthe difficulty in measuring VSF spectra over a broadIR frequency range (due to limitations in IR lasersystems used) and the weak SF signals due to thehigher order nature of the response.

The sum frequency intensity is proportional to thesquare of the surface nonlinear polarization. Thispolarization is induced by the two incident beams andgives rise to sum frequency generation; PSFG isdependent on the surface nonlinear susceptibilityøs

(2)(ωsfg ) ωvis + ωir) given by

with øNR(2) and øR

(2) the nonresonant and resonant

parts of øs(2) respectively, and γν is the relative phase

of the νth vibrational mode. The nonresonant partof øs

(2) depends primarily on the polarizability of themolecules at the interface and must be included inthe full spectral analysis. The resonant term arisesfrom a coincidence in frequency between the tunableinfrared light and a vibrational mode in the moleculeof interest. The coherent nature of the VSF responseleads to a more complex expression for the sampleresponse and a more complicated deconvolution ofoverlapping peaks than in linear spectroscopy. Be-cause the nonlinear susceptibility generally has animaginary and a real part, each resonant term in thesummation is associated with a relative phase, γν,which describes the interference between overlappingvibrational modes. The resonant term, øRν

(2) is de-pendent on the number of molecules N and theorientationally averaged molecular hyperpolarizabil-ity ⟨âν⟩ in the following way

Given the expressions in eqs 1 and 2, the squareroot of the sum frequency intensity is shown todepend on the number of molecules giving rise to theresponse. Their average orientation can be derivedthrough the âν term. These two contributions can beused to determine the orientation of interfacialmolecules and changes in orientation under variousexperimental conditions.

The enhancement in the sum frequency responsethat occurs when the frequency of IR radiation isresonant with a sum frequency active vibration arisesfrom the molecular hyperpolarizability, âν. Thisenhancement in the molecular polarizability is givenby

if one assumes a Lorentzian distribution of vibra-tional energies and that the dipole approximationholds. In this expression AK is the IR transitionmoment, MIJ is the Raman transition probability, ωνis the resonant mode frequency, and Γν is the naturalline width of the transition. Since sum frequencyactive modes must be both IR and Raman active, anymolecule with an inversion center cannot be sumfrequency active.

In the presence of a charge at the surface thatinduces an electrostatic field E0 at the interface, anadditional third-order factor can contribute to theoverall SF response described in eq 1. This third-order polarization term, P(3)

SFG takes the form of

and contains the electrostatic field dependence of thenonlinear polarization induced at the interface. Thisadditional term has been shown to be important ina number of VSF studies where ionic species ad-sorb at a water surface.38-42 The third-order contri-bution to the nonlinear polarization results from

Figure 1. Schematic of the VSF method as applied tovapor/water and organic/water interfaces. P and S cor-respond to the polarization of the light either parallel orperpendicular to the incident plane, respectively.

ISFG ∝ |PSFG|2 ∝ |øNR(2) + ∑

ν|øRν

(2)eiγν|2IVisIIR (1)

øRν(2) ) N

εo⟨âν⟩ (2)

øRν(2) ∝

AKMIJ

ων - ωir - iΓν(3)

P(3)SFG ) ø(3):EVisEIRE0 (4)

Molecular Bonding and Interactions at Aqueous Surfaces Chemical Reviews, 2002, Vol. 102, No. 8 2695

several factors, namely, the electronic nonlinearpolarizability, R(3), the alignment of the interfacialwater molecules by the electrostatic field Eo, and themagnitude of the electrostatic field. The presence ofa large electrostatic field aligns the interfacial watermolecules beyond the first few water layers and thusremoves the centrosymmetry over this region, therebyallowing more water molecules to contribute to thenonlinear polarization.43

As with any vibrational spectroscopy, spectralfitting of the data requires an appropriate choice ofline shape for the vibrational peaks. The simplestapproach to fitting the data is to use eq 3, whichassumes a Lorentzian distribution of energies andallows for interferences between nearby vibrationalmodes. Unfortunately, the broad “wings” character-istic of a Lorentzian distribution often overestimatethe amount of overlap, and therefore the amount ofinterference, between widely separated peaks. Par-ticularly useful, although computationally more in-tensive, is the line shape profile described by Goateset al.44 Similar to a Voigt profile, this line shapeexpression is a convolution of eq 3 and a Gaussiandistribution to account for inhomogeneous broaden-ing.44,45 This profile is the most appropriate for theanalysis of spectra exhibiting significant overlap ofmodes of different phases.

Although the complex nature of the nonlinearresponse makes data analysis more complicatedrelative to linear spectroscopies, the interferencebetween different vibrations can be exploited toprovide orientational information if a complete analy-sis of the VSF spectrum is employed that takes intoaccount the phase relationships of the contributingvibrational modes to the sum frequency response.46,47

In particular, it is possible to constrain the averageorientation of the molecules at the surface by relatingthe macroscopic second-order susceptibility øIJK,ν

(2) ofthe system to the molecular hyperpolarizabilities,âlmn,ν of the individual molecules at the interface.48,49

Interference effects can be understood by use of amore a rigorous expression for the molecular hyper-polarizability, âν

where l, m, and n represent the molecular inertialaxes (a, b, and c); rlm and µn represent the Ramanand dipole vibrational transition elements, respec-tively, for a particular vibrational mode. An energydiagram illustrating the interference between differ-ent vibrational modes is shown in Figure 2. Theinterference of different nearby vibrations in VSFSis analogous to the double-slit experiment. In bothcases, particles passing through indistinguishableintermediate states give rise to distinct interferencepatterns.50

In the VSF experiment, the macroscopic observableis øs

(2). This represents the sum of the molecularhyperpolarizabilities, âν, over all vibrational modesand all of the molecules at the interface and includesinformation about the orientation of each molecule.Orientational information is obtained from the ex-

perimental spectra through consideration of therelationship between the observed Cartesian compo-nents of the macroscopic second-order susceptibilityøIJK,ν

(2) and the corresponding spectroscopically activecomponents of the molecular hyperpolarizability âlmn.This is accomplished through an Euler angle rotationof the molecular axis system into the laboratory axissystem as defined through the use of the rotationalmatrix µIJK:lmn. The general expression for the trans-formation from a molecular-fixed axis system to alaboratory-fixed system is

The indices I, J, and K represent the lab framecoordinates X, Y, or Z observed in a specific experi-ment. The indices l, m, and n run through themolecular coordinates a, b, and c. The orientation ofthe molecular axis system in the lab frame is definedby the transformation tensor, µIJK:lmn, through theEuler angles θ, æ, and ø. If the signs of âlmn,ν andøIJK,ν

(2) are known, then the average orientation of themolecules can be constrained by analyzing how thesign of the transformation tensor changes withrespect to the angles θ, æ, and ø. An excellentresource for these transformation equations is givenby Hirose et al.48,49 For vapor/water and liquid/liquidinterfaces discussed here, the signs of the øIJK,ν

(2)

terms are determined through a comprehensive fitof the observed sum frequency spectra to eqs 1 and3, as described below in the discussion of water, andthe signs of the âlmn,ν components can be determinedthrough ab initio calculations.51,52

The polarizations of the incident and outgoingfields can be used in different combinations to providefurther information about the orientation of interfa-cial molecules. Given that the interface between twoisotropic bulk media is isotropic in the plane of theinterface (i.e., it has C∞v symmetry), the surfacesusceptibility øs

(2) which is a 27-element tensor cangenerally be reduced to the following four indepen-dent nonzero elements

âlmn,ν )⟨g|rlm|ν⟩⟨ν|µn|g⟩ωIR - ων + iΓν

(5)

Figure 2. Energy diagram representative of a typical sumfrequency transition exhibiting interference between dif-ferent vibrations. In simplistic terms, the sum frequencyresponse can be viewed as a combination of a resonantinfrared transition with a nonresonant Raman transition.The overall process leaves the molecule in its originalground state. Interference patterns arise when differentinfrared transitions overlap, giving rise to multiple indis-tinguishable “paths” that give rise to the same sumfrequency response. (Reprinted with permission from ref50. Copyright 2000 American Chemical Society.)

øIJK,ν ) ∑lmn

µIJK:lmn ‚ âlmn,ν (6)


where z is the direction normal to the interface. Inprinciple, the four independent values can be deter-mined after the acquisition of sum frequency spectraunder four different polarization combinations (SSP,SPS, PSS, PPP) with the polarizations listed in orderof decreasing frequency (sum frequency, visible, IR).P-polarized light is defined as having its electric fieldvector parallel to the plane of incidence, whereasS-polarized light is polarized perpendicular to theincident plane (see Figure 1). The SSP polarizationcombination probes modes with IR transition mo-ment components perpendicular to the interfacialplane, SPS and PSS polarization combinations probemodes that have IR transition moment componentsin the plane of the interface, and the PPP polarizationcombination probes all components of the allowedvibrations.

Additional factors that must be taken into accountwhen describing the SF response are the linear andnonlinear Fresnel factors that describe the transmis-sion and reflection of the three light beams at thesurface, with their values dependent on the opticalgeometry of the experiment. Fresnel factors are thereflection and transmission coefficients for electro-magnetic radiation at a boundary and depend on thefrequency, polarization, and incident angle of theelectromagnetic waves and the indices of refractionfor the media at the boundary.47,53 The dependenceof the intensities on the Fresnel coefficients underthe four polarization schemes are

where the subscript i stands for x or y and f and f̃are the linear and nonlinear Fresnel factors, respec-tively.

B. Experimental ConsiderationsThere are numerous types of laser systems and

detection methods used to conduct VSF experimentsthat can be found in the literature. Only generalfeatures of the experimental setup are provided here.Because of the higher order nature of the VSFresponse and the relatively low polarizability of mostliquids, the VSF signal from aqueous surfaces isrelatively weak. Hence, pulsed lasers with high peakfields are generally used. Since the sum frequencyintensity increases with the peak intensity of theincident beams, picosecond and femtosecond pulsesare optimal, although these shorter pulses result inlarger IR bandwidths. Nanosecond systems are gen-

erally much simpler to operate and have narrowerIR bandwidths but have the potential to contributeto significant heating of the interface unless care isexercised, such as employing an optical couplingscheme such as total internal reflection (TIR)54 thatcan enhance the VSF signal by several orders ofmagnitude, or other mechanisms such as samplerotation.55

Most VSFS studies to date have focused on thevibrational region around 3 µm because nanosecondand picosecond systems currently generate the high-est IR power densities there. Tunable IR light hasbeen produced by a number of optical parametricgeneration (OPG), oscillation (OPO), and amplifica-tion (OPA) systems as well as difference frequencymixing and stimulated Raman scattering. Figure 1ashows the geometry generally used for vapor/waterstudies where both nanosecond and picosecond lasersystems have been used. For liquid/liquid studies, theuse of nanosecond lasers has been facilitated byemploying a TIR geometry where the light is coupledto the interface through the organic (higher index)medium (see Figure 1b).56,57 The incident laser beamsstrike the interface near their respective criticalangles to generate sum frequency in reflection at itscritical angle. A TIR geometry has also been used instudies of solid/aqueous interfaces where the light ispassed through the backside of a transparent sur-face.58,59

Experiments employing femtosecond laser pulseshave recently been reported by several groups con-ducting VSFS studies.60-62 With the large band-width of femtosecond pulses (a few hundred wave-numbers), a significant portion of the vibrationalspectrum can be obtained without tuning. McGuireet al.60 demonstrated a Fourier transform spectro-scopic technique based on VSFS. This method re-sults in nearly unlimited spectral resolution and isbased on the Fourier transform of a SF-upconvertedinterferogram of an IR-induced polarization on thesurface of the sample. A second method61 capi-talizes on the variation of the SF exit angle as afunction of SF frequency, with the SF signal collectedwith a multielement detector. A third method dis-perses the broadband SF signal generated by fem-tosecond IR and narrow-band visible pulses with amonochromator and records the signal with a CCDcamera.62

Since the surface of an aqueous liquid, whether itis at a vapor/water interface or a liquid/liquid inter-face, provides an energetically attractive site formolecules that have both polar and apolar parts,obtaining a clean interface is a serious challenge inany VSF experiment. For example, our experimentshave shown remarkable sensitivity in the VSF waterspectrum to trace amounts of organic impurities thatmigrate to the interface.63 If these impurities arehighly surface active (this is generally the case), thespectrum of the impurity can dominate the interfacialspectrum. Therefore, samples and solvents must behighly purified, and all equipment that comes incontact with the samples must be extraordinarilyclean.

øzzz(2); øxxz

(2) ) øyyz(2); øxzx

(2) ) øyzy(2); øzxx

(2) ) øzyy(2)

(7)

IPPP ∝ |f̃zfzfzøzzz(2) + f̃zfifiøzii

(2) + f̃ifzfiøizi(2) +

f̃ififzøiiz(2)|2

ISSP ∝ |f̃ififzøiiz(2)|2

ISPS ∝ |f̃ifzfiøizi(2)|2

IPSS ∝ |f̃zfifiøzii(2)|2 (8)


III. Hydrogen Bonding and Structure of Water atAqueous Surfaces

A. Vapor/Water InterfaceThe VSF spectrum of the vapor/water interface was

first reported by Shen and co-workers.64,65 In morerecent years, it has been examined in additionalstudies as investigators have sought to unravel thecomplexity of the spectral response with extendedexperimental studies, improved analysis procedures,and advances in laser instrumentation.50,66-69 Figure3 shows representative spectra of the vapor/water

interface recently published by Wei and Shen70 forthree different polarization combinations. With SSPpolarization (Figure 3a), transition dipoles perpen-dicular to the surface plane are being sampledwhereas with SPS (Figure 3c) transition dipolesparallel to the interface are probed. The spectralregion corresponds to the OH stretching modes ofsurface water molecules. The vibrational spectrumof water in this region is of particular interestbecause the OH stretch modes are highly sensitiveto the local molecular environment.71-75 The watervibrational spectrum therefore provides a sensitiveprobe of the structure and energetics of the hydrogen-bond network at the interface. The sensitivity ofVSFS to surface water vibrational modes is ac-companied by a complexity in spectral interpretationthat is only beginning to be explored by researchersin depth. This complexity is due to the variety ofdifferent environments of the interfacial water mol-ecules and the broad nature of the peaks correspond-ing to water molecules that extensively hydrogenbond to other water molecules. This difficulty ininterpretation is not unique to surface vibrational

spectroscopy but has been a major point of contro-versy over the past decades in the interpretation ofbulk water spectra.71,73,74

Assignment of spectral features in this spectrumrely heavily on IR and Raman assignments of OHstretching modes taken from bulk water measure-ments.71,76-79 In most of the VSF studies, broadfeatures in the spectrum have been assigned tovibrational modes that encompass a wide distributionof hydrogen-bonded strengths for any particularassignment.16,43,50,64,65,67-70 The bands from approxi-mately 3000 to 3600 cm-1 have been discussed interms of symmetric stretch modes for water mol-ecules in a broad distribution of tetrahedral bondingenvironments, both symmetric and asymmetric. Theliterature has generally discussed the broad band inthis region in terms of two subbands that arecentered around 3200 and 3450 cm-1.16,43,50,64,65,67-70

The former, which is comparable in frequency to theIR and Raman spectra of ice,80 has generally beenattributed to strong intermolecular in-phase hydro-gen bonds of water molecules that give rise to ahighly correlated hydrogen-bonding network. Thespectrum represents a continuum of OH symmetricstretches (SS), ν1 of water molecules in a symmetricenvironment (SS-S). For simplicity, this subbandhas been referred to as the “ice-like” region becauseof its similarity in energy to OH bonds of watermolecules in bulk ice. The higher energy broad bandregion of ∼3250-3500 cm-1 is assigned to moreweakly correlated hydrogen-bonded stretching modesof molecular water that encompass both ν1 (OHsymmetric stretch) and to a lesser extent ν3 (OHasymmetric stretch) vibrational modes. Intensity inthis region corresponds with OH stretch intensitiesof bulk liquid water and hence has been referred toas “liquidlike” hydrogen-bonding modes. These watermolecules reside in a more asymmetrically (AS)bonded water environment, and hence, the intensityin this region has been labeled as SS-AS. Somedifferences in the spectral features in the tetrahe-drally coordinated region of the spectrum are ob-served depending upon whether a nanosecond69,81,82

or picosecond laser system is used,64,66,67,70,83,84 but ingeneral, the spectra all show similar trends. Theconsiderable intensity in the lower subband has ledShen and co-workers64 to conclude that the hydrogenbonding at this interface has an “ice-like” character.

A sharp spectral feature located near 3700 cm-1 isalso observed in the vapor/H2O spectrum. This regionis characteristic of vapor-phase water molecules, andthe peak corresponds to the free OH bond of surfacewater molecules.64 The relatively high energy of thismode reflects its lack of donor bonding with otherwater molecules. This free OH bond, which is directedinto the vapor phase, is energetically uncoupled fromthe adjacent intramolecular OH bond that doesparticipate in H-bonding with neighboring watermolecules (referred to here as the donor bond ordonor mode). This situation leads to two distinctintramolecularly uncoupled OH stretching modeswhich, analogous to gas-phase vibrational spectra ofwater dimers, give rise to a free and bonded OHstretch of the H-bond donating water molecule.85,86

Figure 3. SFG spectra of the water surface at 20 °C withthree different polarization combinations SSP, PPP, andSPS. (Reprinted with permission from ref 70. Copyright2001 American Physical Society.)


This intramolecularly uncoupled and unperturbedOH oscillator at 3700 cm-1 (Figure 3a) is locateddirectly between the coupled ν1 (3657 cm-1) and ν3

(3756 cm-1) modes of monomeric vapor water, asgoverned by the linear superposition of states. Thelocation of the uncoupled, bonded OH stretchingmode (donor mode) has been measured recently asdiscussed below.66

Deriving a quantitative relationship between spec-tral intensity and the number of water molecules thathave different degrees of hydrogen bonding is chal-lenging for several reasons. First, deconvolution ofthe broad spectral intensity in the 3100-3500 cm-1

spectral region into several peaks by use of variouscurve-fitting routines has long been controversial forbulk water studies. The nature of the nonlinearresponse of VSF makes this even more complicatedwithout employing additional experiments that canidentify particular water species as has been donein other VSF studies.63 Second, the broad spectralfeature in the tetrahedrally coordinated region rep-resents a broad range of water-water interactionsand assigning peaks to specific types of interactionsas one would do for more isolated water molecules isof course suspect. Third, the OH oscillator strengthof a water molecule can vary significantly with thedegree of hydrogen-bonding,71,73,87 with increasedbonding leading to increased oscillator strength.Thus, quantitative comparisons or even relativecomparisons between intensities of OH modes thatare strongly hydrogen bonded with those that areweakly hydrogen bonded require an understandingof how this bonding affects both the Raman and IRtransition probabilities for each assigned mode mea-sured by VSFS. Recall that the VSF resonant mac-roscopic susceptibility contains both factors (eq 3). Asan alternative, an early study of the number of watermolecules corresponding to the free OH mode, moni-tored the decrease of the free OH intensity asmethanol adsorbed at the surface, as a means ofquantifying the proportion of water molecules thatstraddle the interface. These studies suggest that atleast 20% of the water molecules fall into thiscategory.64

Recently there have been several studies that areproviding more detailed information about interfacialwater species contributing to the vapor/water inter-face.50,66 The first has involved development andapplication of a spectral analysis procedure for fittingthe VSF vapor/water spectrum that takes into ac-count interferences that arise between nearby vibra-tional modes due to the different symmetries of themodes and the relative orientation of the transitiondipole.50 The second has involved isotopic dilutionstudies that determine the spectral contribution fromthe OH donor mode of the water molecules thatstraddle the interface.66 Both of these studies havebenefited from related studies of the CCl4/H2O in-terface (to be described later) where weaker interfa-cial hydrogen bonding has allowed for specific waterspecies at this interface to be identified throughcomplementary IR and isotopic dilution experi-ments.63,88

The most rigorous spectral analysis proceduredeveloped and applied to the VSF spectra of watersurfaces that takes into account symmetry, phase,and relative orientation has appeared in a series ofrecent papers.50,63,88 Since VSFS is a coherent non-linear spectroscopic technique, each resonant vibra-tional mode has an inherent phase for a fixedorientation. Resonant modes that change phase withorientation of the molecule have the capability ofinterfering either constructively or destructivelywhen overlapped in frequency. The work considersa range of water species present at a water surfaceand the possible interference between these contrib-uting modes and takes into account the phase of theSF response from contributing vibrational modes.This phase is useful in obtaining an average orienta-tion of molecules at the surface by relating themacroscopic second-order susceptibility, ø(2), of thesystem to the molecular hyperpolarizabilities, âv, ofthe individual molecules at the interface.48,49 For thevapor/water spectrum, the signs of the øIJK,ν

(2) termscan be determined through a comprehensive fit of theobserved sum frequency spectra to eqs 1 and 3 andthe signs of the âlmn,ν components can be determinedthrough ab initio calculations.51,52 Application of thisanalysis to the weakly hydrogen-bonded region of thespectrum suggests the possible presence of weaklyhydrogen-bonded water monomer-like molecules inthe interfacial region. These types of water monomerswere first observed in CCl4/H2O VSF spectrum.63,88

The recent isotopic dilution experiments of HODat the vapor/water interface provide further informa-tion about contributing spectral features in the broadband between 3100 and 3500 cm-1.66 Isotopic dilutionstudies have been invaluable in previous studies ofIR and Raman water spectra, particularly the mea-surement of the HOD spectrum.41,89-93 Because of theintramolecular uncoupling of the OH and OD modes,the VSF spectrum is much simpler and allowsspectral-fitting procedures to be used that measureand enable separation of the nonresonant backgroundsignal from the resonant signal. The measurementof the VSF spectrum of HOD in D2O at the vapor/water interface provides a means of identifying thespectral region for contributing hydrogen-bondedmodes such as the donor mode of the water moleculesthat straddle the interface and the double donor modeof water molecules with their oxygen atoms pointingout of the interfacial region.

Figure 4(a-d) shows the VSF spectra in the OHstretch region of the vapor/water interface withdiffering concentrations of HOD/D2O/H2O.66 The topspectrum corresponds to the vapor/H2O interface,analogous to that shown in Figure 3a. The differencein the relative height of the free OH band and thebroad bonded OH band for these two spectra fromdifferent laboratories is due to a difference in cor-rection for the linear-IR Fresnel coefficient. ForFigure 4a the linear Fresnel coefficient used has beencalculated as a function of frequency (to correct forthe dispersion in the index of refraction),66 whereasfor Figure 3a a constant index of refraction has beenemployed.70 The experimental data from these twolaboratories prior to this correction are otherwise


identical including the relative height of the free OHand bonded OH modes. With increased D2O in theaqueous solution (Figure 4b-d), the number ofcontributing OH oscillators decreases as manifestedin the decrease in the resonant contribution from theOH oscillators. The VSF response from the pure D2Ospectrum (Figure 4d) provides an important mea-surement of the nonresonant VSF background thatis necessary to include in spectral analysis andfitting. Because of the complex nature of the VSFresponse, this background cannot merely be sub-tracted but must be deconvolved from the resonantresponse. This is easiest to achieve for the spectro-scopically simplest system where HOD and D2O arethe dominant species at the water surface (Figure 4c).The resonant response after removal of the nonreso-nant background is shown as a solid line (Figure 5a).The contributing peaks (dotted lines) derived for thisspectrum from the best fit to the experimental dataare also shown in Figure 5a. The overall best spectralfit to the data which includes the nonresonantresponse and the contributing resonant peaks isshown as a solid line in Figure 4c. The HOD surfacespectrum (Figure 5a) is dominated by the free OHmode at 3694 cm-1 and a broader peak near 3420cm-1 that the authors attribute to the OH bond ofHOD that is directed into the aqueous phase, thedonor mode. The frequency of the donor mode issimilar to that measured for the donor mode of HODat the CCl4/HOD/D2O interface.88 Also contributingin this spectral region are water molecules that onaverage have both their OH and OD bonds directedinto the aqueous phase with the OH bond having acomponent perpendicular to the interfacial plane(double donors). There are also two weak broad peaks

in the 3200, 3310 cm-1 region that the authorsattribute to more strongly bonded H2O molecules thatare present in low concentrations. In addition toestablishing the positions of the spectral peaks, thefree OH and donor OH peaks are found to be 180°out of phase, consistent with the opposite orientationof these modes relative to the interfacial plane. Asimilar analysis has been applied to the other isotopicmixtures of Figures 4 and 5. The peak areas of thecontributing OH oscillators scale appropriately withthe number of OH oscillators at the interface as theH2O/D2O concentration is varied, providing furtherevidence for the validity of the spectral analysis.

The derived resonant spectral response for thevapor/water interface allows a refinement of theearlier picture of contributing species in the vapor/water spectrum. The solid line in Figure 5c corre-sponds to the resonant VSF spectrum of the vapor/H2O interface after removal of the nonresonantcomponent and the correction for the dispersion inthe index of refraction. A large portion of the inten-sity in the vapor/water spectrum near 3450 cm-1 canbe attributed to the donor mode of water moleculesthat on average straddle the interface and otherpartially hydrogen-bonded interfacial water mol-ecules (double donor) near 3350 cm-1. Intensity from

Figure 4. VSF isotopic dilution measurements of the OHstretch region for vapor/water interface with differing molefractions (mf) of HOD, H2O, and D2O. Solid lines are thebest fits to the resonant + nonresonant response. SSPpolarization. (Reprinted with permission from ref 66.Copyright 2002 American Chemical Society.)

Figure 5. Resonant contributions (solid line) to the VSFspectra of Figure 4 for (a) 0.25/0.02/0.73 mf of HOD/H2O/D2O; (b) 0.41/0.51/0.08 mf HOD/H2O/D2O, and (c) 1.0 mfH2O. Spectral peaks (dotted lines) contributing to theresonant response as determined by the described spectralfitting procedure. SSP polarization. (Reprinted with per-mission from ref 66. Copyright 2002 American ChemicalSociety.)


these bonds and the free OH bond contribute to asignificant fraction of the overall intensity in thespectrum, not surprising since the polarization com-bination being used samples modes perpendicular tothe interface and all of these have a considerablecontribution in this direction. Also, these studiessuggest less intensity in the 3200-3300 cm-1 regionof the spectrum, less evidence for symmetricallytetrahedral coordinated water molecules than sug-gested in previous studies.16,43,50,64,65,67-70 Furtherevidence of a more “liquid-like” surface, rather thanone that has a structure more analogous to ice, comesfrom the frequency and bandwidth of the HODspectrum of Figure 5a that are similar to what isobtained for HOD in D2O at room temperature. Inthese studies the OH from HOD in D2O is near the3420 cm-1 measurement made in the surface VSFstudies. The OH spectrum of HOD in this region isknown to be highly temperature dependent andconsequently a good indicator of the bonding environ-ment of water molecules.71,94 HOD in D2O near 0 °Cis in the region of 3250-3300 cm-1.93 One mightconclude from these results that, given the fact thatthe transition probability of hydrogen-bonded speciesincreases with hydrogen-bond strength, the surfaceis largely comprised of water molecules that onaverage straddle the interface or act as doubledonors. However, the additional factor that must betaken into account is that these spectra largelyrepresent water molecules with nonisotropic direc-tional character, to which VSFS is highly sensitive.For tetrahedrally coordinated water molecules, watermolecules that are directionally isotropic in thesurface plane, cancellation should lead to minimalintensity from these molecules. Evidence for thiscomes from the in-plane response measured by Weiand Shen that shows virtually no intensity in the3200-3300 cm-1 region of the spectrum.70

A theoretical analysis of the VSF spectrum of thevapor/water interface in the OH stretch region basedon ab initio molecular orbital theory and moleculardynamics simulation has been performed by Moritaand Hynes.95 The essential features of the spectra ofthe early work of Du et al.84 and Shultz68,82,96 and co-workers has been reproduced and interpreted byusing an approximate modeling of the molecularvibrations and thus circumventing their explicitsimulation. The free OH bond band has been foundto be sensitive exclusively to the surface top mono-layer structure, while the hydrogen-bonded bandreflects the structure of the few top monolayers. Inagreement with the above-described studies66 andthose of Brown et al.,50 they also find that thedangling and hydrogen-bonded bands have oppositesigns of imaginary susceptibility, which indicatedifferent OH orientations at the surface. They de-termine that the population ratio between the OHbonds in the dangling bond frequency region andthose at lower frequencies can be estimated to be as35%:65%, respectively, but conclude that this per-centage is probably closer to 21% when they removethose dangling OH bonds that point inward towardthe bulk. They find that the vibrational modes below3600 cm-1 have overall symmetric stretching char-

acter, whereas those above 3600 cm-1 have overallantisymmetric character. The transition region around3600 cm-1 is attributed to an inhomogeneous mixtureof symmetric and antisymmetric modes, while in thetail frequency regions, the modes appear to approachlocalized OH stretching character.

B. Organic/Water InterfacesThere has long been an interest in the structure

of water adjacent to a hydrophobic medium. Theinterest arises since understanding molecular inter-actions of water next to a hydrocarbon or othernonpolar phase has direct relevance to areas such asmicellar formation, protein folding, chemical separa-tion, and oil extraction. There have been manytheoretical efforts directed toward understandingthese interfacial interactions on a molecular level,1-15

but few experimental studies have been possible.Recently, the vibrational spectroscopy of water at anorganic/water interface has been obtained.63,88 For thestudies reported, TIR VSFS has been used to mea-sure the hydrogen bonding and orientation of inter-facial water molecules at organic/water interfacesand to compare the results with the vapor/waterinterface. The organic liquids examined have prima-rily been CCl4, hexane, and hexene. Figure 6 shows

the spectra obtained from the CCl4/H2O (Figure 6a)and hexane/H2O (Figure 6b) interfaces using SSPpolarization. The spectra are dominated by inten-sity at higher energies (3400-3800 cm-1) character-istic of weakly or non-hydrogen-bonded water mol-ecules.71,97,98 They both show a sharp peak at higherenergies, at ∼3669 and 3680 ( 4 cm-1 for the CCl4/water and hexane/water interfaces, respectively. Forthese TIR experiments for which a nanosecond lasersystem is used, the nonresonant signal as measuredby D2O experiments is negligible. Consequently, thespectra shown represent the resonant OH response.A comparison of these spectra of Figure 6 with theresonant response from the vapor/water interface(Figure 5c) indicates that the hydrogen-bonding

Figure 6. VSF spectrum of the (a) CCl4/H2O and (b)hexane/H2O interfaces in the OH stretch region. SSPpolarization. (Reprinted with permission from ref 63.Copyright 2001 American Association for the Advancementof Science.)


interactions at the liquid/liquid interface as probedby SSP polarization are weaker than that of thevapor/water interface, as indicated by the overallspectral intensity of the former being at higherenergies than the latter. This weakening of interac-tion between water molecules is attributed to areduction in the coordination number of interfacialwater molecules and/or a weakening in the strengthof individual hydrogen bonds between interfacialwater molecules as they interact with the organicphase.

The weaker nature of the hydrogen bonding at thisinterface has facilitated assignment of spectral fea-tures in the CCl4/H2O spectrum. The “free OH” modefor water is clearly apparent at the CCl4/H2O inter-face, but its energy is red shifted (3669 ( 4 cm-1)relative to the vapor/water interface (3702 cm-1).64,65,67

From studies of HOD monomers in CCl4, the studyhas determined that this red shift is due to anattractive interaction between the dangling OH bondand the surrounding CCl4 molecules.63 The bindingenergy for the H2O-CCl4 complex is reported to be-1.4 kcal/mol.99 The results are consistent withsimulations of this interface that suggest a locallysharp transition between phases.11,100,101 VSF studiesof HOD at this interface provide further assignments.In addition to the free OH bond and free OD bond ofHOD at the interface (measured at 3664 and 2712cm-1, respectively), a broader peak is observed near3450 cm-1 that is attributed to the donor OH modeof the water molecules that straddle the interface orwater molecules with both bonds directed into theaqueous phase (double donor). This is similar to whatis observed in the follow-up studies described abovefor the vapor/water interface,66 but in the CCl4/H2Ostudies, the negligible nonresonant background sim-plifies the analysis. The remaining portion of the H2Ospectrum of Figure 6a between 3450 and 3700 cm-1

is attributed to weakly interacting water moleculesthat are either surrounded by CCl4 or bonded to otherwater molecules as electron donors only.63 Since thetrue monomers and the electron donor water mol-ecules are spectroscopically indistinguishable in theseexperiments, the authors refer to them as “monomer-like”. The OH stretch of water monomers in bulk CCl4show two characteristic peaks in the IR spectra, oneassociated with the symmetric OH stretch (SS or ν1)of water monomers (3616 cm-1) and the other (3708cm-1) associated with the asymmetric OH stretch (ASor ν3). These two peaks energetically bracket thedangling bond stretch mode in the VSF spectrum.Careful analysis of the data shows that both the SSand AS modes of monomer-like water molecules arepresent in the VSF spectrum, which respectivelyconstructively and destructively interfere with theneighboring dangling OH bond mode.50 The authorsconclude from the fits and the derived sign of thephases that the SS and AS observed intensitiesrepresent water molecules at the interface that havea net orientation with their hydrogens pointed intothe CCl4. The peak energies and widths determinedfrom the fits agree well with FTIR data of watermonomers in bulk CCl4. The orientation of thesewater molecules into the CCl4 phase is attributed to

the interfacial electrostatics which includes the non-negligible H2O-CCl4 interaction. The authors dem-onstrate how the orientation of these monomer-likewater molecules can be flipped 180° by variation ofthe pH in the aqueous phase.63,88

Overall, the results suggest that, relative to thevapor/water interface, the hydrogen bonding at theCCl4/H2O and hydrocarbon/H2O interface is veryweak. Intensity is observed in the strongly hydrogen-bonded region (3200-3400 cm-1), but overall thespectra are dominated by water species that haveonly weak interactions with other water moleculesand CCl4 molecules. This is the first spectroscopicevidence for the existence of monomeric and monomer-like water molecules at a water/hydrophobic inter-face. Figure 7 shows the fit (top) and contributing

spectral peaks (bottom) derived from the data de-scribed above for CCl4/H2O. Figure 8 provides a sche-matic of the different water species corresponding tothe spectral peaks in Figure 7 derived from thecombination of experiments and fitting. The relativephases derived from the fit to the data provide thenecessary orientation of the water molecules that aredepicted. For example, under SSP polarizations, theSS and AS of monomeric H2O (labeled as 3) haveopposite sign conventions (+ and - respectively),meaning they are near 180° out of phase.50 Given thatthe OH dangling bond (+ sign convention and labeledas 4) has a significant contribution perpendicular tothe interface, if the water monomers are orientedwith their dipoles in the same direction as thedangling bond, the SS(+) and AS(-) modes shouldconstructively and destructively interfere, respec-tively, with the dangling bond mode. This is what isobserved. Spectral fits place the AS and SS of thesemonomer and monomer-like waters with their hy-

Figure 7. (a) VSF spectrum of the CCl4/H2O spectrumwith the overall fit (upper). (b) Contributing spectral peaksfor the various OH modes derived from the fit to the datain part a and shown schematically in Figure 8. Peakslabeled 1, 2, and 3 are determined from the fit to be 180°out of phase from 3 and 4. See text for details. (Reprintedwith permission from ref 63. Copyright 2001 AmericanAssociation for the Advancement of Science.)


drogens oriented into the CCl4 phase at 3616 ( 2 and3706 ( 2 cm-1 respectively, compared with the FTIRmeasured peaks at 3616 and 3708 cm-1.

These results for the CCl4/H2O and hexane/H2O63

and other alkane/water101 systems described aredistinctly different than previous related studies ofthe CCl4/H2O interface by Gragson et al.38 and thehexane/H2O interface by Du et al.,64 which bothexhibited little spectral intensity in the 3500-3600cm-1 region and large intensities in the 3200 cm-1

region. This has been interpreted as an enhancementof the water structure at these surfaces. However, ithas been discovered in these later studies andconfirmed by surfactant addition studies that theVSF spectrum of interfacial water is highly sensi-tive to trace amounts of impurities (10-100 nM) thattend to concentrate at the interface. As the impuritiesare progressively removed, spectral features in the3400-3600 cm-1 region grow in and spectral inten-sity in the 3200 cm-1 region diminishes resulting inthe VSF spectrum in Figure 6a and b. These findingshave therefore modified previous interpretations andcan readily account for the differences between theseobservations and previous CCl4/H2O 102 and hexane/H2O65 VSF studies.

Recent molecular dynamics simulation studies ofthe organic/water interface have shed further lighton water structure and hydrogen bonding at theseinterfaces as well as the interpretation of the VSFresults of these systems.101 MD simulations of theCCl4, hexane, and vapor/water interfaces have beenconducted using the AMBER suite of MD programs.The SPC/E model was used to describe all watermolecule interactions. The method of Morita andHynes95 was used to calculate VSF spectra for theseinterfaces to allow for comparison with the experi-mental data. For the CCl4/water interface, the freeOH is calculated to be oriented on average roughly30° from the surface normal, resulting in a strongSF response under SSP polarization. The donor OHbonds are pointing into the bulk water (110° fromnormal) and consequently have a smaller and oppo-sitely signed contribution to øSSP

(2) than the free OHbonds. Interfacial forces also orient a large number

of water molecules in the interfacial region with bothhydrogens participating in hydrogen bonding withother water molecules. The interfacial molecules withthis type of stronger hydrogen bonding are found, onaverage, to have both of their hydrogens near theplane of the interface (∼95° from normal) toward thebulk water. The interfacial depth probed by the VSFexperiments can be estimated from the simulation.The simulated hexane/water interface, for example,displays a net orientation of water molecules for atotal of 9 Å. The full width half-maximum of øSSP

(2),an alternative depth, was determined to be 4.5 Å. TheVSF spectra calculated from these MD simulationsfor hexane/water and CCl4/water show remarkableagreement given the three-point model used in thesimulation. In both cases a sharp free OH peak isobserved at ∼3700 cm-1 as well as a broad hydrogen-bonded feature centered at ∼3400 cm-1. The calcu-lated relative intensities of the vibrational featuresfor these two systems also compare well with theexperimental results.

C. Water at Solid SurfacesThe molecular interaction of water with solid

surfaces forms the basis for a wide range of importantchemical and physical processes including wetting,corrosion, friction, adhesion, and erosion. For hydro-philic surfaces, bonding interactions between thesubstrate and the water molecules can lead tocomplete wetting of a surface and the formation oftwo-dimensional films. For hydrophobic solid surfaceswhere molecular interactions between a largely non-polar surface and water can lead to partial wetting,droplets are often formed on the surface. Withsurfaces of an ionic nature, water can form solubilizedfilms that erode or dissolve the surface. Aqueoussolutions in contact with metal and other reactivesurfaces can lead to oxidation and reactivity of thesurface that forms the basis of corrosive processesthat diminish the structure and function of manymaterials. It is only recently that a detailed knowl-edge of the molecular interactions between waterlayers and various solid surfaces has begun toemerge. Vibrational sum frequency spectroscopy isplaying a role in these advances as indicated by theoverview given below.

1. Surface Melting of Ice

The melting of ice and the molecular structure ofa liquidlike layer on a bulk ice surface has intriguedscientists for many decades. Considerable theoreticaland experimental effort has been expended over thepast decades in understanding the molecular proper-ties of a melting water surface because of its impor-tance in a wide range of processes including glacialflows, atmospheric aerosol processes, and frictionaleffects on ice surfaces. Experimentally, studies of icesurfaces pose challenges that are just beginning tobe overcome with techniques such as IR reflec-tance,78,94,103 X-ray diffraction,104,105 proton backscat-tering,106 ellipsometry,107,108 and interference micros-copy.109 Recently VSFS has been applied to thisissue.110 In these studies, the surface melting of the(0001) face of hexagonal ice (Ih) has been studied in

Figure 8. Schematic of the water molecules at the CCl4/H2O interface as determined by the VSF spectroscopy ofthis interface. The numbers correspond to the spectralpeaks shown in Figure 7b. (Reprinted with permission fromref 63. Copyright 2001 American Association for theAdvancement of Science.)


the range of 175-300 K. The free OH mode has beenused to monitor the degree of orientational disorder.Figure 9 shows the VSF spectrum of the (0001)

surface of ice at 232 K along with a spectrum of thewater surface for comparison. As shown, the spec-trum of ice is dominated by a strong but relativelybroad peak at 3150 cm-1 that resembles the mainO-H stretch peak observed in the Raman spectrumof Ih.111 The strength of this mode relative to the freeOH mode in the ice spectrum, particularly whencompared to the liquid water spectrum, demonstratesthe strong hydrogen-bonding interactions present atthe Ih surface. Using polarization studies of the freeOH bonds to measure the angular spread of the freeOH bond orientation, an orientational parameter isderived that is used to describe the degree of order-ing. The authors conclude that the onset of surfacemelting of ice is around 200 K, which is lower thanthat measured by X-ray scattering.105 They also findthat the degree of disorder of the water moleculesincreases with temperature. Another important con-clusion drawn from the studies is that near the bulkmelting temperature of 273 K, the order parameterof the ice surface is even lower than that of thesupercooled water surface, suggesting that the quasi-liquid layer on ice is different from the normal surfacelayer of water.

2. Water Adsorbed at Solid Substrate SurfacesThe first studies of water on a solid substrate were

performed at a quartz/water interface.83 As withprevious studies discussed in this review, the spectralregion corresponds to OH stretch modes of water. Theclean quartz surface is generally terminated bysilanol groups (SiOH) that can hydrogen bond withsurface water. This is certainly the case found forquartz/water studied near neutral pH as indicatedby no intensity in the free OH bond region due tothe bonding of water to this surface feature. Thedominance of a broad peak near 3200 cm-1 suggeststhat the hydrogen bonding at this surface is highly

coordinated and almost “ice-like” in nature. This andsimilar studies have shown that there is a strong pHdependence in the response of the OH stretch modesof water.16,83 At both low and high pH, stronghydrogen bonding is observed as evidenced by adominance of the OH-SS-S peak. There is a strongorientation of surface water molecules due to theeffect of increased hydrogen bonding or surface fieldestablished by the different electrostatics created bythe ions in solution. At intermediate pH, morebonding disorder is apparent in the spectrum. Mea-surements of the phase of the VSF response showthat the water dipoles have flipped by 180° as thepH is adjusted from 1.5 to 12.3.16

Later studies by Shen and co-workers16 examinedthe quartz/ice interface at various temperatures.Figure 10 shows these results. In these studies

conducted from -40 to 0.0 °C the intensity in the OHstretch region (3600-3100 cm-1) shows a shift around-0.5 °C to lower frequencies. A strong peak around3150 cm-1 dominates the spectrum. The spectrumlooks very similar to the ice/water spectrum describedabove. There is no evidence for the quasi-liquid layerat the quartz/ice interface at least up to -0.5 °C. Theauthors speculate the decrease in intensity withtemperatures near 0 °C could be due to deprotonationof the quartz surface such that it effectively increasesthe number of surface defects that would disrupt thepolar ordering.

Water adsorbed on a mica surface at room tem-perature is also found to have a more orderedhydrogen-bonding structure than bulk water.112 Thesestudies involved a combination of VSFS and scanningpolarization force microscopy (SPFM).113 The ODstretch modes of D2O on mica have been explored inthese studies to avoid complications arising from

Figure 9. VSF spectrum of the (0001) surface of ice at232 K (solid symbols). The inset shows a comparison of thespectrum with that of the water surface at 293 K (opensymbols). The polarization combination is SSP. (Reprintedwith permission from ref 110. Copyright 2001 AmericanPhysical Society.)

Figure 10. VSF spectrum (SSP) from a quartz/ice inter-face at various temperatures. (Reprinted with permissionfrom ref 16. Copyright 1998 Elsevier Science.)


the OH stretch modes that originate from the mica.The authors find that as the concentration ofwater on the surface increases in the submonolayerregime, the water evolves into a more ordered hy-drogen-bonded network as indicated by a down-shift in the OD stretch modes. At full monolayer, theVSF spectrum is indicative of an ice-like film. Nodangling OD bonds are found at the mica/waterinterface.

The combination of VSF and thermal desorp-tion has been used to monitor ice films grown onPt(111).114 In these studies the experimental re-sults suggest that ice films grown on Pt(111) attemperatures between 120 and 137 K are ferroelec-tric, where “ferroelectricity” is used here loosely todescribe the existence of a net polar ordering of watermolecules in ice films. They observe a strong en-hancement in the OH stretch resonances with filmthickness that they attribute to polar ordering.Figure 11 shows the VSF results that they obtain in

the OH stretch region as the films grow on thesurface. All incident beams and outgoing beams wereP polarized. They attribute the increased ordering ofwater with film thickness to the polar anchoring ofthe first ice monolayer on the platinum. This surface-induced ordering is estimated to have a decay lengthof ∼30 monolayers.

Yeganeh and co-workers115 examined the isoelectricpoint of the Al2O3/water interface by VSFS. Theisoelectric point (IEPS) of a solid surface as examinedin this paper is the point at which the interfacialcharge changes sign as the pH of the aqueous phaseis varied. This crossing point affects many interfacialprocesses because as water dipoles orient differentlyfor differing surface charges, competitive adsorption

can vary on either side of this point, and the dis-tribution of ions in the solution phase changes. Figure12 shows the results of their studies on Al2O3 wherethe total water VSF signal is plotted as a function ofthe pH of the solution. The signal is found to rise oneither side of the IEPS. The data were taken withP-polarized input and P-polarized VSF light detected.They show that as the pH of the aqueous phase isvaried across the isoelectric point, the water dipolesflip by 180°. They also find that the signal intensitydepends strongly on the hydroxyl number density atthe surface. A new methodology based on surfacecharge density or the determination of the IEPS of anonconductive, low surface area material is alsopresented.

Ionic surfaces in contact with water are challengingbecause of the more dynamic nature and reactivenature of these interfaces. Fluorite (CaF2) surfacesin contact with varied aqueous solutions have beenstudied by Becraft and Richmond.58 Fluorite is theprimary mineral used in the production of hydro-fluoric acid. Its acid/base behavior influences theadsorption of complexing agents and surfactants usedto separate CaF2 from other minerals often associatedwith it in the natural state. In these experiments itis shown,58 as with quartz/water and sapphire/waterinterfaces, that the bonding of water molecules in theinterfacial region is strongly dependent on pH. Underacidic conditions, a highly structured water phase isobserved at the surface as evidenced for the strongsignal near 3160 cm-1 (Figure 13a). As the pH israised from 2.9 to 5.1, the VSF response in thisregion is found to decrease with almost a zeroresponse at neutral pH (Figure 13a-c). The authorsattribute this to the zero point of charge for thissurface near neutral pH which leads to minimalwater alignment. As the pH is increased, the stronghydrogen-bonding network returns but with an ad-ditional feature growing in at 3657 cm-1 (Figure 13f).At pH 13.7, this feature dominates the spectrum.With the aid of fluoride addition studies, the authorsattribute this to calcium hydroxide species at thesurface.

Figure 11. VSF spectra in the OH stretch region for aset of ice films of different thicknesses on Pt(111). (Re-printed with permission from ref 114. Copyright 1999American Physical Society.)

Figure 12. Variation in total water VSF signal with pHof the solution. The marked minimum in the signalstrengths indicates that the IEPS of the sapphire surfaceis ∼8. (Reprinted with permission from ref 115. Copyright2001 American Physical Society.)


D. Effect of Adsorbates and Ions on SurfaceWater Structure and Bonding

1. Alkyl Surfactants Adsorbed at the Vapor/WaterInterface

When surfactants adsorb at a water surface, thesurface properties are significantly altered. In par-ticular, the lowered surface tension accompanyingsurfactant adsorption is one of the main reasons forthe effectiveness of many surfactants in commercialproducts including soaps, lubricants, and detergents.The pertinent question that has been asked forseveral decades is how the presence of the surfactantalters the hydrogen bonding of water at these sur-faces. This section reviews studies of surfactant/waterinterfaces that focus on the water structure. In a latersection, the VSF studies that examine the structureand conformation of surfactants adsorbed at aqueoussurfaces will be summarized.

A number of VSF studies have been conducted toaddress this particular question of what the waterstructure of an aqueous surface looks like in thepresence of varying quantities of surfactants. Thesehave been conducted at both the vapor/water39,43,116,117

and the organic/water interface.40,42,118 Most of these

studies have involved examining how the OH stretch-ing region of water (as discussed in the previoussection) is altered by the presence of alkyl surfac-tants. In many cases deuterated surfactants are usedto minimize the overlap between the OH vibrationalmodes of water and the C-H stretching modes of thesurfactants. With these studies, a coherent pictureis emerging on how the hydrogen bonding of wateris affected by the presence of a surfactant at thesurface. Some of the surfactants used in these studiesinclude pentadecanoic acid (PDA) which is not chargedat neutral pH and charged surfactants such asdodecyl trimethylammonium chloride (DTAC), so-dium dodecyl sulfate (SDS), or dodecylammoniumchloride (DAC). Earlier studies in this area havefocused on surface concentrations from approximately10-3 monolayer coverages to a complete monolayercoverage as indicated by a maximum in the adsorp-tion isotherm derived from surface tension measure-ments.39,40,42,43,117,118 Later studies conducted at theorganic/water interface have focused on trace con-centrations of surfactants at the interface (describedin the next section) where headgroup areas are amore appropriate unit for the interfacial concentra-tion.119 In these studies, the surfactants are under

Figure 13. VSF spectrum of the CaF2/H2O interface at pH (a) 2.9, (b) 5.1, (c) 6.4, (d) 9.3, (e) 12.3, and (f) 13.7. SSPpolarization. (Reprinted with permission from ref 58. Copyright 2001 American Chemical Society.)


isolated conditions with headgroup areas of >103

Å2/surfactant.The first studies to examine both the issue of

surfactant conformation and water structure simul-taneously for surfactants at the vapor/water interfacewere conducted by Gragson et al.43 Later studiesshowed a similar behavior for water at the surfactant/organic/water interface.42 Figures 14 and 15 provide

a comparison of the VSF spectrum of surfactantspentadecanoic acid (PDA), dodecylammonium chlo-ride (DAC), and sodium dodecyl sulfate (SDS) ad-sorbed at the vapor/water interface at approximatelymonolayer coverages.39,43 For simplicity, the broadpeak representing tetrahedrally coordinated watermolecules is separated into two peaks, one corre-sponding to the OH stretch modes of symmetricallybonded water molecules (OH SS-S) and at higherenergies the OH stretch of asymmetrically bondedwater molecules (OH SS-A). The CH stretch modesof the surfactants correspond primarily to the meth-ylene symmetric and asymmetric stretches (SS andAS), the methyl symmetric stretch, and a methylFermi resonance (FR). These modes will be discussedin more detail in section IV of this review. Both H2O(solid circles) and D2O (open circles) have been usedas the aqueous phase to allow separation of the CHmodes of the surfactant and OH modes of interfacialwater. Upon comparison of the uncharged PDA(Figure 14) and the charged surfactants DAC andSDS (Figure 15), it is immediately clear that thereis a large enhancement in the OH peaks of H2O inthe 3100-3500 cm-1 spectral region in the presenceof a charged surfactant. The uncharged surfactant(PDA) at the surface causes minimal change in theOH stretching region relative to a neat water spec-trum. When equal concentrations of mixed surfac-tants (DAC and SDS) adsorb at the interface, thesurface water spectrum also shows minimal change.43

The authors attribute the observed enhancement inthe OH stretch modes at lower energy in the pres-ence of charged surfactants to an increased align-ment of the interfacial water dipoles induced by thelarge electrostatic field present at these charged

interfaces. The interfacial water molecules are foundto attain their highest degree of alignment in thedouble layer region at surface surfactant concentra-tions well below maximum surface coverage. A secondfactor is the increased number of water moleculesbeing sampled due to the effect of the double layerfield established at the surface. As discussed insection II, the field effect leads to an additionalcontribution to the VSFS response at the interface,a third-order polarization term P(3), which containsthe electrostatic field dependence of the nonlinearpolarization induced at the interface. This field-induced third-order polarization has been examinedin previous SHG studies.120-122 In the absence of alarge electrostatic field, one would expect the inter-facial water molecules to be randomly orientedbeyond the top few water layers and thus not tocontribute to the VSF response. However, the pres-ence of a large electrostatic field aligns the interfacialwater molecules beyond the first few water layersand thus removes the centrosymmetry over thisregion, allowing more water molecules to contributeto the nonlinear polarization.39,43 Ionic strength stud-ies provide further confirmation for these conclusions.These studies show a decrease in enhancement of OHSS modes with increased ionic strength. As the ionicstrength is increased, there is an increased screeningof the surface charge or, alternatively, a change in

Figure 14. VSF spectra under SSP polarization conditionsfrom the air/H2O (filled circles) and air/D2O (open circles)interfaces with a monolayer (∼25 Å2/molecules) of PDA.(Reprinted with permission from ref 39. Copyright 1996American Chemical Society.)

Figure 15. VSF spectra under SSP polarization conditionsfrom the air/H2O (filled circles) and air/D2O (open circles)interfaces with (a) 14 mM DAC solution and (b) 8.1 mMSDS solution. (Reprinted with permission from ref 39.Copyright 1996 American Chemical Society.)


the Debye-Huckel screening length, consistent withthe observations. Similar observations of enhancedOH signal were made in pH-dependent studies offatty acid hexacosanoic acid studied at the vapor/water interface as a function of pH.117,123 As the fattyacid becomes ionized at higher pH values, the signalin the OH stretching region increases.

Another important observation apparent in Figure15 is the interference observed between OH stretchand CH stretch modes of the water and surfactant,respectively.43 The DAC spectrum shows destructiveinterference near 2970 cm-1 between the OH-SS-Smode and the CH3-FR mode (Figure 15a). The SDSspectrum shows a constructive interference betweenthese two modes (Figure 15b). The authors attributethis difference in the interference for cationic andanionic surfactants to differing orientations of theinterfacial water molecules, which results from op-posite electrostatic fields created by surfactants ofopposing charge. From this they conclude that theorientation of interfacial water molecules in thepresence of adsorbed cationic surfactants is with theoxygen atom pointed toward the air and for theanionic surfactant with the oxygen atom pointedtoward the bulk solution.

The two contributing factors to the increased OHintensity in the presence of surfactant, the increasedsampling volume due to the change in Debye-Huckelscreening length and field-induced orientation ofwater molecules, have been studied extensively inwork with ionic strength studies and deuteratedsurfactant studies.42 The interfacial potential wasvaried by changing the surface charge density (sur-factant concentration), the ionic strength, and thetemperature while monitoring the OH stretchingmodes of water. As the potential is increased atthe vapor/water interface, there is a progressiontoward more tightly bound interfacial water mol-ecules as manifested in an increase in intensity inthe OH-SS-S stretch region relative to the SS-ASregion. The authors conclude that this progressionis a direct consequence of the induced alignment ofthe interfacial water molecules resulting from thelarge electrostatic field produced in the interfacialregion due to the charged surfactant. Temperature-dependence measurements support this conclusion.

2. Alkyl Surfactants Adsorbed at Organic/Water Interfaces

The effect of surfactant on the ordering of waterat the organic/water is important in understandingthe role of water in the structure and formation ofmicrostructures such as micelles, vesicles, and otherthree-dimensional structures. These all involve acharged interface between water and a hydrophobicmedium such as the interior of a micelle or vesicle.There are a number of studies that have examinedthe effect of surfactant adsorption on the hydrogenbonding between water molecules at the organic/water interface. These studies, which have beenexclusively conducted in this laboratory at the CCl4/water interface, can be divided into two concentrationregimes. The earlier studies have focused on inter-facial concentrations in the 10-3 to 1 monolayerrange.41,42 The second, more recent, set of experi-

ments has been conducted at trace interfacial con-centrations.119 The higher concentration experimentsinvolved the types of measurements described abovefor the vapor/water interface where the interfacialpotential was varied and the water hydrogen bondingexamined.42 In these studies conducted at the CCl4/water interface, similar trends are observed as in thevapor/water studies in that an increased concentra-tion of surfactant results in a strong growth in OHstretch contribution corresponding to tetrahedrallycoordinated water molecules. The spectra of a mono-layer of SDS at the vapor/water and CCl4/waterinterfaces are remarkably similar, (Figures 15b andFigure 16). Both spectra show strong intensity in theOH-SS-S region of the spectrum, indicative ofstrong hydrogen bonding interactions. Isotopic dilu-tion studies of this interface in the presence of amonolayer of surfactant have also been conducted.42

These studies show a progressive shift in the OHstretching region to the blue as H2O molecules arereplaced by HOD molecules in the interfacial spec-trum. The frequency of the bonded OH bond of HODin the presence of the surfactants is found to becentered at 3460 cm-1.

The more recent studies of the effect of adsorbedsurfactant on the water hydrogen-bonding structurehave been conducted over a broader concentrationrange from trace interfacial concentrations to frac-tional monolayer coverage, with the majority of thefocus of these studies on the low concentrationregime.119 As mentioned in section III.B, the VSFspectrum of interfacial water at the organic/waterinterface is highly sensitive to trace amounts ofcharge at the interface. This observation has led toa series of studies that examine how the waterstructure at the organic/water interface changes asthe surfactant aqueous-phase concentration is variedfrom nanomolar aqueous-phase concentrations, wherethe interfacial concentration is so low that the systemcan be viewed as composed of isolated surfactants,to micromolar concentrations where electrostaticinteractions between surfactants and their solvatinginterfacial water molecules occur and interfacialpotentials become more homogeneous across theinterface. Figure 16 provides a view of the largechanges that occur in the VSF spectrum at anorganic/water interface as the SDS interfacial con-centration is varied over a wide concentration range.As the concentration is increased toward monolayercoverage (5 mM), spectra similar to the earlier CCl4/water studies42 and the vapor/water studies describedabove are progressively observed.43 The intensity inthe tetrahedral bonding region grows in and eventu-ally dominates the 5 mM spectrum with the ac-companying appearance of the CH modes of thesurfactant. As discussed in detail in the vapor/waterstudies, this growth in intensity from the OH modesof water in the 3100-3300 cm-1 region of thespectrum at these coverages is due to a combinationof an increased contribution from strongly hydrogen-bonding water species, an increased probe depth dueto the electrostatic field at the interface (an additionalP(3) effect), and an increased orientation of interfacialwater molecules induced by this field. Also observed


is a progressive loss in free OH intensity with highersurfactant concentration, indicative of a surface thatis increasingly covered by surfactant which bonds tothe free OH. Unlike the previous CCl4/H2O studiesof water,42 these were conducted with a nanosecondlaser system (TIR geometry) with IR tunability in thefree OH region.

3. Solutes, Acids, and Salts in the Aqueous Phase

There have been a number of VSF studies of thevapor/water interface that have focused on how thesurface water spectrum changes with acidity andionic concentration. These studies are relevant tounderstanding atmospheric reactions where aerosolscan have very high concentrations of acids and highionic strength. Some of the first and most compre-hensive studies in this area have examined thestructure of surface water for solutions of sulfuricacid and related salts.68,69,124-127 Figure 17 providesan example of how the VSF spectrum of waterchanges for different concentrations of aqueous solu-tions of sulfuric acid.123,124 The overall trends in thespectra have been seen in the other studies of thissystem by Schultz and co-workers.69,125-127 As shownin Figure 17, at low concentrations of acid, theintensity at lower frequencies below 3400 cm-1 isfound to increase. Both sets of studies conclude thatthere is significant orientation of surface watermolecules at lower acid concentrations. Second, withincreased acid concentration, the free OH modedecreases and eventually disappears. At the highestconcentrations, the OH spectra decrease and eventu-ally disappear. Raduge et al.124 observe a low-

frequency shoulder at ∼3060 cm-1 that they assignto the acid OH stretch vibration, suggesting that theacid concentration is appreciable. This peak is notobserved in the studies of Baldelli et al.126,127 Beyondthis point there is also considerable disagreementwith regard to the interpretation of the data. Baldelliet al.126,127 attribute the increased intensity at lowerfrequencies in the low acid region to an electricdouble layer at the interface, as negative ions pref-erentially adsorb at the liquid surface, and thedrop-off in intensity at higher concentrations todisordering of the surface water. At higher concen-trations, Raduge et al.124 conclude that the surfacetakes on a more crystalline structure, resemblingthat of a crystalline sulfuric acid hydrate with themolecules arranged in a centrosymmetric layeredstructure with the hydrogen bonds in the surfaceplane leading to vibrational modes that are SFinactive. Studies of supercooled sulfuric acid solutionsby Schnitzer et al.69 indicate that the water surfaceof liquid sulfuric acid solutions does not vary withtemperature.

Several other acid and salt solutions have also beenstudied by Schultz and co-workers.68,128,129 The struc-ture of water on HCl solutions has been examinedby VSF with the primary conclusions being that ionsin solution cause water on the surface to be reori-ented relative to pure water with the hydrogen atomsdirected toward the bulk solution. No signal due tomolecular HCl was observed, suggesting to theauthors that oriented water molecules and not mo-lecular HCl dominate the surface.128 For HNO3 solu-tions and liquid HNO3,129 the VSF studies indicatethat the surface consists of an electric field doublelayer comprised of subsurface anions and cations atlow mole fractions (0.005 and 0.01 mole fraction ofHNO3). These solutions are found to show more VSFsignal from OH stretch modes in the hydrogen-bonded region than simple salt solutions (NaCl,NaNO3, KHSO4), which the authors attribute to thesubsurface electric field that aligns water moleculesalong the surface normal. At higher concentrationswhere the intensity is reduced, the authors concludethat ionic complexes or molecules approach thesurface and disrupt the hydrogen-bonding networkat the water surface.

IV. Adsorbate Structure and Bonding at AqueousSurfaces

A. Surfactants at Vapor/Water Interfaces

Whereas many of the above-described studies havefocused on the effect of adsorbates on the interfacialwater structure, in the following sections the molec-ular structure of the adsorbed species has beenprobed. The most extensive studies to date have beenon surfactant structure where the surfactants consistof a polar or ionizable headgroup and a nonpolarhydrocarbon chain. Such molecules readily migrateto the interface where their long alkyl chains extendinto the hydrophobic portion of the interface (i.e., theair or organic phase) and the polar headgroup prefersthe more aqueous portion of the interfacial region. A

Figure 16. VSF spectra of the CCl4/H2O interface over arange of bulk concentrations of SDS. SSP polarization. Thehighest concentration corresponds to approximately amonolayer of SDS at the interface.


wide variety of surfactants of this nature have beenstudied, with the majority of the studies focused onmeasuring the conformation of the alkyl chain as afunction of surface concentration, surface pressure,and aqueous-phase composition.

The first studies to demonstrate that VSFS couldbe used to measure the molecular conformation ofsurfactants adsorbed at an vapor/water surface in-volved PDA. In these and follow-up studies, the CHstretch modes of the alkyl chain were monitored withvarious polarization combinations as the surfaceconcentration has been varied in a Langmuir troughfrom the liquid-condensed (LC) to liquid-extended(LE) phase.130,131 In the spectra presented as well asin later studies of PDA, the CH spectral region of thisinsoluble monolayer consists mainly of CH modes ofthe terminal methyl group.39 Figure 14a shows anexample of a spectrum of PDA at the vapor/waterinterface using SSP polarization which is similar towhat has been obtained in the earlier work.130,131 Thisfigure however also includes the OH stretch modesof surface water for reference to the earlier discussionin this review of how the surface water spectrum isaltered by the presence of a surfactant. As shown,two vibrational modes dominate the spectrum ofadsorbed PDA, the methyl symmetric stretch modenear 2875 cm-1 and the methyl Fermi resonancemode near 2935 cm-1. At a monolayer of coverage,

no contribution is seen from the methylene modes ofthe alkyl chain similar to what has been observed inother insoluble surfactant systems.39,54 For am-phiphilic molecules that form well-ordered monolay-ers, in a predominantly trans conformation, the CH2bonds have their dipoles directed toward opposingsides of the carbon backbone. This orientation pro-duces a cancellation of the CH2 stretching vibrationalmodes, and thus, a monolayer with no or few gauchedefects will exhibit only CH3 vibrational modes in theSF spectrum. In the earlier work of Shen,131 theseCH2 modes near 2850 cm-1 and in the 2930-2880cm-1 region begin to appear as the surface pressureof the monolayer changed. They concluded that theseobservations confirm the existence of highly orderedalkyl chains with nearly normal orientation when themonolayer is in the liquid-crystal (LC) phase. In theliquid-expanded (LE) phase, the chains were foundto have significant gauche defects as seen by theappearance of CH2 modes.

A number of studies of different alcohols adsorbedat the vapor/water interface have been conducted. Forthe simplest alcohol, methanol, information about themolecular orientation of methanol at the watersurface has been obtained by monitoring the CHstretch modes under different polarization combina-tions.132 The authors find that there is an excessnumber density of methanol at the interface as

Figure 17. VSF spectra of pure water and sulfuric acid-water mixtures (with different acid concentrations as indicated)at the liquid-vapor interface at 20 °C. SSP polarizations. (Reprinted with permission from ref 124. Copyright 1997 ElsevierScience.)


compared to the bulk concentration in agreementwith surface tension measurements. They find thatin a narrow range of bulk concentrations of methanol,the SF spectrum of the surface molecules changesdrastically. They attribute this to a special structureof CH3OH:H2O surface network. Another study ex-amined the VSF spectra of the liquid-vapor interfacefor n-alcohols of (C1-C8).133 In these experiments, allalcohols studied were found to be polar oriented withthe alkyl chains pointing away from the liquid. Forthe longer chains of hexanol, heptanol, and octanol,significant contributions from CH methylene modessuggest the presence of gauche defects in the alkylchains. In the water OH stretch region, the shift inspectra to lower frequencies indicates a well-orderedhydrogen-bonding network at the interface.133 Moredetailed studies of medium chain alcohols of C9-C14have been conducted with VSF and ellipsometry byBain and co-workers.134,135 These studies determinedthat gauche defects are evident in the solid mono-layer phase and increase only slightly in the mono-layer just above the phase transition. Using a modelof the chains as rigid rods, they calculated the areaper molecule and the chain tilt in the liquid phasefrom the VSF and ellipsometry data. They find thatthe area per molecule and tilt angle increase mono-tonically with chain length. The density of thehydrocarbon chains in the liquid monolayer phase isfound to be less dense than that in the solid mono-layer phase but significantly higher than in the bulkliquid alkane. Additionally, the area per molecule,chain tilt, and volume per methylene group in theliquid phase all increase with increasing chain length.Studies of the absolute orientation of crystallinemonolayers of amphiphilic R-hydroxy ω-bromo alco-hols (BrCnH2nOH, n ) 21,22) and the alkyl hydroxyesters (CmH2m+1 COO(CH2)nOH, m ) 14,15, n ) 10)at the vapor/water interface have been reported usinggrazing incidence X-ray diffraction (GIXD) for theformer and VSFS for the latter.136 Crystalline mono-layers of related alcohols are very efficient ice nuclea-tors of supercooled water drops because of the latticematch between the unit cell of the crystalline mono-layer cell and the ab lattice of hexagonal ice. Thestudies show the absolute orientation of the alcoholC-OH bonds at the water surface, which in turn canbe correlated with the ice-nucleating behavior of themonolayers on supercooled water. Glycerol/watermixtures have been studied to learn how the glyceroland water structure changes as the composition isvaried. In these studies, glycerol is found to partitionto the surface at all concentrations with the surfaceorientation of glycerol found to be constant throughmost of the concentration range.96

The most detailed spectroscopic study and spectralanalysis of alcohols has been conducted by Wolfrumand Laubereau.46 In this study of hexadecanol at thevapor/water interface, three vibrations of the termi-nal methyl group have been studied. Figure 18 showsthe spectrum obtained for this molecule under threedifferent polarizations. For SSP polarizations (Figure18a), the spectrum is dominated by the symmetricCH3 stretching vibration at 2875 cm-1. The maximumat 2936 cm-1 is attributed to an overtone of the

methyl bending vibration which gains intensity byFermi resonance with the symmetric stretchingvibration. A minimum in the spectrum is observedat 2958 cm-1 where the degenerate CH3 stretchingvibration is expected. Neither the symmetric norantisymmetric CH2 stretching vibration at 2850 or2939 cm-1 appear in the spectrum, indicating astraightened molecular chain in an all-trans config-uration. A theoretical analysis is presented thatpredicts opposite phases of adjacent vibrational modesleading to destructive interference in the SF spec-trum. This interference effect, similar to what wasdiscussed above for the OH modes of water, allowsthe determination of the line amplitudes of thedegenerate stretching mode. Figure 19 provides anexample of the calculated curves used to fit the VSFSSP spectrum of Figure 18a and 19a and the indi-vidual contributions of the three modes to the imagi-nary part (Figure 19b) and real part (Figure 19c) ofthe contributing tensors. The relative phases of thecontributions lead to the destructive and constructiveinterferences that contribute to the observed lineshapes of the spectral peaks in the VSF spectrum. Itis found, for example, that the stretching vibrationand the symmetric overtone of the bending vibration

Figure 18. Measured sum frequency signal of the ad-sorbed monolayer of hexadecanol on water in the CHstretching region versus frequency setting of the inputinfrared pulse for (a) SSP, (b) SPS, and (c) PPP polarizationcombinations. The solid lines are calculated curves pro-portional to |ø(2)

yzy|2(SSP), to |ø(2)yzy|2 (SPS), and a super-

position of all independent components of to |ø(2)ijk|2 (PPP).

(Reprinted with permission from ref 46. Copyright 1994Elsevier Science.)


have the same phase, whereas the contribution of thedegenerate vibration has the opposite phase. De-structive interference can occur in the high-frequencywing of the overtone, leading to the distinct minimumat 2958 cm-1. The tilt angle of the molecular chainwith respect to the surface normal in a spontaneouslyformed monolayer has been determined to be 8.2° (1.8°.

Several studies have appeared that have exploredboth the tail and headgroup regions of the surfac-tant. Using the insoluble surfactant CD3(CH2)19CN,Eisenthal and co-workers explored the orientation ofthe terminal deuterated methyl group and the polarnitrile group as a function of interfacial concentra-tion.137 For this Langmuir monolayer, the resultsindicate that the orientation of these two entities varywith surface concentration but in very different ways.It has been found that in the phase transition fromthe gas/liquid coexistence region to the liquid region,the spectral changes in the CN region are attributedto the breaking of hydrogen bonds with the CN groupwith the water squeezed out of the monolayer andsubsequent reorientation of the headgroup. Theorientation of the tail is found to be sensitive to the

monolayer density even in the gas-liquid coexistenceregion with the tail becoming continuously uprightupon compression. In another study of the tail andheadgroup of surfactants, Watry and Richmond138

examined the behavior of alkylbenzenesulfonate, asoluble surfactant, at both the vapor/water and CCl4/water interfaces. The results are compared withrelated alkylsulfonates to learn how the presence ofthe benzene ring next to the sulfonate headgroupaffects the behavior of the surfactant. The benzene-sulfonate molecule studied, dodecylbenzenesulfonate(DBS), is a commonly used commercial surfactantand has been known to be several orders of magni-tude more effective in reducing the surface tensionof water relative to alkysulfonate surfactants. Figure20 displays the spectrum of DBS at the vapor/water

interface. The spectral peaks between 2800 and 2950cm-1 correspond to the alkyl chain of DBS, whereasthe modes above 3000 cm-1 correspond to the CHvibrational modes of the benzene ring. By monitoringthe ratio of the intensity of the methylene and methylCH stretch symmetric modes, they conclude that thealkyl chains of the DDS are highly disordered at thevapor/water interface as a function of concentration.Increased surface concentration from fractions of amonolayer to a full monolayer does not result in anyincreased ordering. In contrast, as the interfacialconcentration of linear dodecylsulfonate increases atthe vapor/water interface, the chain ordering in-creases indicative of increased chain-chain interac-tions. The authors attribute this difference in order-ing to the presence of the benzene ring in DBS. Thelimiting surface area measurements and the orienta-tion of the benzene ring, measured by monitoring thephenyl C-H modes as a function of concentration,support a picture that DBS exists at the interface ina staggered headgroup geometry. This picture isconsistent with the benzene rings disrupting chain-

Figure 19. Measured sum frequency signal for SSPpolarization with least-squares fit proportional to to |ø(2)

yyz|2as in Figure 22a. The dashed line in part a is calculatedfor positive amplitudes Ayyz, (b) imaginary part to ø(2)

yyz,vand (c) real part to ø(2)

yyz,v for the individual vibrationalmodes; (- - -) symmetric stretching vibration at 2875 cm-1,(...) symmetric overtone at 2936 cm-1 enhanced by Fermiresonance (-‚-‚-) degenerate stretching vibration at 2958cm-1. (Reprinted with permission from ref 46. Copyright1994 Elsevier Science.)

Figure 20. VSF spectrum of DBS at the air/D2O interface.SSP polarization. The solid line is a fit to the dataassuming a Voigt functional form for the peaks. (Reprintedwith permission from ref 138. Copyright 2000 AmericanChemical Society.)


chain interactions for the first few methylene groupsadjacent to the benzene ring in DBS. This staggeredarrangement appears to significantly disorder thealkyl chains for DBS.

A study that combined second harmonic generationand VSF has demonstrated how molecular orienta-tion and conformation of a surfactant chromophorecan be measured at the vapor/water interface.139

These studies have involved pentyl-cyanoterphenylmolecules for which three parts of the molecule havebeen examined upon adsorption. The cyano head-group and pentyl chain has been examined by VSF,and the terphenyl ring has been examined by SHG.These molecules form a Langmuir monolayer on thewater surface. The results give a quantitativelyconsistent picture of the molecular configuration ifthe appropriate refractive indices for the monolayerare used. They find that in order for their results tobe physically reasonable and the deduced molecularconformation to be consistent with the commonlyaccepted one, a value of the index of refraction usedmust be different than that of a bulk phase value forthe molecule and intermediate between the index ofrefraction of air and water. Using a value of n′ ) 1.18( 0.04, they find that the molecules adsorb at theinterface with a tilt angle of 51.5° ( 1.5° from thesurface normal. Simple model calculations are usedto justify the experimentally determined value of n′.

Bain and co-workers conducted the most extensivestudies of a range of surfactants adsorbed at thevapor/water interface. One of their first studiesinvolved measuring the vibrational spectrum ofseven soluble surfactants adsorbed at this interface.23

These included three nonionic surfactants (dodecanol,C12-maltoside and CH3(CH2)n-1 (OCH2CH2)mOH, orCnEm), the anionic surfactant sodium dodecyl sulfate,one zwitterionic surfactant (C12-betaine), and twocationic surfactants, (C14 trimethylammonium bro-mide and didodecylammonium bromide). From theVSF spectra of the C-H stretch modes of the alkylchains, the degree of conformational disorder and theangle of the terminal methyl group are inferred. Theyfind that, in general, the number of gauche confor-mations increases as the area per chain increases.They also find that the angle of the methyl group,which is an indicator of the tilt of the hydrocarbonchains, is not simply related to the area per chain. Acomparison of surfactants with the same chain lengthand area per molecule shows that the structure ofthe chain region of the monolayer is sensitive to thenature of the headgroup and not just to the packingdensity. A series of cationic surfactants has been thefocus of another set of studies.55,59,140-143 In a studythat combines VSFS and neutron reflectivity, hexa-decyltrimethylammonium p-tosylate has been exam-ined. Single-chain cationic surfactants such as thisform small spherical micelles in dilute solutions, butin the presence of aromatic counterions, they cantransform into long, threadlike aggregates withdramatic effects on the optical properties and rheol-ogy of the solution.144 In these vapor/water studies,they show that the aromatic anions cause major andunexpected changes to the structure of the monolay-ers.141 They find that in order for the monolayer to

generate space for the bulky tosylate ions, the areaper surfactant molecule increases by about 25%. Ina related study, the behavior of this cationic surfac-tant in the presence of a series of halide counterionshas been examined by a combination of surfacetension, ellipsometry, and VSFS. Neither the VSFSnor ellipsometry data provided any firm evidence forspecific effects of the halide ions on the structure. Theprincipal effect of the counterion is to change theefficiency and effectiveness of the surfactant (bothdecreasing in the order of Br- > Cl- > F-). In anotherstudy, a series of alkyltrimethylammonium bromides(CH3(CH2)n-1N+(CH3)3Br-), CnTAB (n ) 12, 14, 16,18), has been examined at a constant area permolecule of 44 Å2 to understand the effect of chainlength on the molecular orientation. The data sug-gests that all four surfactants behave very similarlywith a chain tilt of 58° near the methyl terminus.Complimentary ellipsometric data suggests that thedensity of the chain region in the monolayers is closeto a liquid hydrocarbon. This and related neutronreflection data are consistent with the VSFS data.143

A study by Goates et al.44 examined the conforma-tional ordering of nonionic surfactants at the vapor/water interface. These nonionic surfactants that arein the category of alkyl poly(ethylene glycol) ethers,CnEm, are widely used as detergents, emulsifiers, anddispersants.145 In these studies, the goal is to deter-mine how the length of the poly(ethylene glycol) chain(EO) affects the structure of the monolayers whenthe length of the hydrocarbon chain is held constant.They find that for a constant headgroup area of 62Å2, the hydrophobic region of the monolayer has adensity close to that of a liquid hydrocarbon and astructure that varies little with the length of the EOchain.

Mixed monolayers of surfactants and hydrocarbonshave been the focus of several studies. A study ofn-eicosane and a monolayer film of dodecanol at thevapor/water interface has been examined in the C-Hstretch region by Sefler et al.146 McKenna et al.59 findthat mixed monolayers of hexadecyltrimethylammo-nium bromide and tetradecane formed at the air-water interface exhibit a first-order phase transitionfrom a conformationally disordered to a conforma-tionally ordered state as the temperature is lowered.The phase transition occurs at approximately 11 °Cabove the melting point of tetradecane. A two-dimensional phase transition has been studied in amixed monolayer of sodium dodecyl sulfate anddodecanol at the vapor/water interface.147 They findwith VSF and ellipsometry that at low temperatures,a monolayer at the surface of a solution containing99.9% SDS and 0.1% dodecanol is conformationallyordered and has a surface coverage comparable tothat of a monolayer of pure dodecanol at the sametemperature. At 16 °C, a first-order phase transitionto a phase that is less dense and more conformation-ally disordered is found. The high-temperature mono-layer phase is more disordered than the correspond-ing liquid phase in pure dodecanol. The solid phaseis found to contain equimolar amounts of SDS anddodecanol, whereas the liquid phase contains SDSand dodecanol in a ratio of 3:2. They attribute the


change in composition of the mixed monolayer to areduction in the interaction parameter of the mono-layer in the phase transition.

B. Water-Soluble Solutes Adsorbed at theVapor/Water Interface

In recent years, a number of studies have appearedthat examine the molecular structure of small water-soluble molecules at the vapor/water interface. Theimpetus behind most of these studies is to under-stand how atmospherically important molecules ad-sorb at this interface. Whereas atmospheric mol-ecules have been studied extensively in the gasphase, very few spectroscopic studies have investi-gated these molecules on liquid surfaces. Thesestudies provide new challenges for the VSF field dueto the relatively low signal levels of these smallmolecules, the complexity in spectral interpretationdue to interferences that can occur between variousvibrational modes of the solute molecules and thewater background, and the assignment of vibrationalmodes. Nevertheless, the surface sensitivity of themethod and the applicability to the study of impor-tant heterogeneous reactions at a liquid surfacemakes the future bright for this area of research.

Air/acetonitrile-water interfaces have been thefocus of two sets of studies. In the first set byEisenthal and co-workers,148,149 the VSFS measure-ments of this interface suggest a phase transition asthe solution composition is varied. This is manifestedin abrupt shifts in the CN vibrational frequency andorientation of acetonitrile molecules at the interfacewhen the bulk acetonitrile (CH3CN) concentrationreaches 0.07 mole fraction. At lower concentrationsit is found that the CN stretching vibrational fre-quency of acetonitrile at the interface is at a higherfrequency than that of neat bulk acetonitrile (Figure21a). At concentrations greater than 0.07 mole frac-tion acetonitrile in water, the frequency for surfaceacetonitrile molecules red shifts to a value that isnear that of neat bulk acetonitrile. The spectralchanges in the CN region are attributed to thebreaking of hydrogen bonds with the CN group asthe water is squeezed out of the monolayer withsubsequent reorientation of the headgroup. Themethyl group (measured with CD3CN) does notexhibit any frequency change as a function of aceto-nitrile concentration as shown in Figure 21b. Todevelop a broader picture of this behavior the studieswere followed by examination of air/proprionitrile-water and air/butyronitrile-water interfaces.150 Thesemolecules allow the study of the effect of the alkylchain on the surface interactions which, for theacetonitrile case, are dominated by the hydrogenbonding of the CN with water and by the dipole-dipole interactions among CN units. For the buty-ronitrile system, no phase transition is observed withinterfacial concentration. However, for proprionitrile,a phase transition as manifested by changes in CNfrequency and CN molecular orientation is observed.The chain length dependence of the behavior isattributed to a lower interfacial packing density formolecules as the chain length increases. Huang andWu examined the air/acetonitrile-water interface

and compared the behavior with the air/methanol-water interface.151 These studies involved a combina-tion of VSFS and third harmonic generation (THG).THG was used to obtain information about themicroscopic structure of these liquids in the bulkmedium. They found very different results for thesetwo systems. For acetonitrile, the studies suggest theexistence of microheterogeneity in the liquid mixtureswith acetonitrile mole fractions higher than 0.3. Thiscritical concentration for the bulk phase separationis larger than that appearing at the surface. Formethanol, the polar distribution of the methanolmolecules at the surface has been found to beenhanced by the interfacial water molecules.

Studies have been conducted on several sulfur-containing molecules examined at the vapor/waterinterface, specifically dimethyl sulfoxide (DMSO)152,153

and methane sulfonic acid (MSA).67 These moleculesare present as trace constituents in the atmosphere.Recently, DMSO has been proposed as the hetero-geneous precursor to atmospheric condensed phaseMSA through an atmospheric cycle originating withdimethyl sulfide, a phytoplankton degradationproduct.154-156 Aerosol particles containing MSA arethought to contribute to the class of aerosols whicheffectively scatter radiation out of the atmosphere.In the DMSO studies, the combination of surface

Figure 21. (a) Upper trace is a transmission spectrum ofa neat bulk CH3CN sample. The two bottom traces are sumfrequency spectra of CN vibration in the air/solutioninterface at bulk mole fractions of X ) 0.03 and 0.26. (b)The upper trace is a transmission spectrum of a neatCD3CN sample. The two bottom traces are sum frequencyspectra of the CD3 symmetric stretch in the air/solutioninterface at bulk mole fractions of X ) 0.02 and 0.26.(Reprinted with permission from ref 137. Copyright 1993American Institute of Physics.)


tension and VSF measurements show that for aque-ous mixtures of DMSO and water, the DMSO mol-ecules partition to the interface with a higher con-centration at the surface than in the bulk. Figure 22a

shows a spectrum of an aqueous solution of DMSOat 0.1 mole fraction. The data were taken with SSPpolarization. The sharp peak observed at 2920 cm-1

is assigned to the methyl symmetric stretch (SS) ofDMSO. This peak is found to be shifted slightly tohigher energy relative to the pure DMSO. The shiftto higher energy has been attributed to a decreasedinteraction of the DMSO sulfur lone pair with thetrans-CH of the methyl groups of DMSO. Thisdecreased interaction is indicative of the changingorientation of DMSO with surface concentration asaggregation increases. This observation combined

with changes in VSF intensity with DMSO surfaceconcentration led to the conclusion that, as thesurface concentration of DMSO increases, the DMSOmolecules are aggregating and reorienting such thatthe two methyl groups are becoming more perpendic-ular to the solution surface. The methyl asymmetricstretch of DMSO appears as a small peak at 2990cm-1. This mode destructively interferes with theCH3-SS48,49 as determined when appropriate fittingprocedures that take into account the phase relation-ship between these two modes are applied. This isvisually manifested in spectrum Figure 22a as anasymmetry in the CH3-SS mode and the dip in thespectrum near 2975 cm-1. The large intensity of theSS mode relative to the AS mode indicates that themethyl groups are oriented predominately out of theinterface.

Similar studies have been conducted with methanesulfonic acid.67 The VSF spectrum of this moleculeat the vapor/water interface is shown in Figure22b. The CH3-SS appears at 2940 cm-1 and theCH3-AS appears near 3030 cm-1. Once again, thestrong asymmetry in the CH3-SS peak is due to thedestructive interference between the two methylstretching modes. The VSF intensity centered around3100 cm-1 is attributed to the cooperative intermo-lecular hydrogen-bonding modes of surface water (seeearlier discussion). Unlike DMSO, as the MSA sur-face concentration is increased from low bulk con-centrations to pure MSA, the frequencies of both CH3modes do not shift. Additionally, the SF intensitytracks with the surface number density obtainedfrom surface tension measurements, leading to theconclusion that unlike DMSO, the methyl group ofsurface MSA does not reorient as a function of surfaceconcentration. It is interesting to compare the VSFspectrum of these molecules with water/acetonestudies.152 Acetone, ubiquitous in many regions of theatmosphere, has recently been shown to be thesecond highest concentration organic trace-gas con-stituent next to methane in regions of the NorthernHemisphere. It is believed that it may play a role inthe growth and surface chemistry of atmosphericaerosols.156 Figure 22c shows the spectrum of acetoneat an aqueous 0.1 mole fraction solution and atemperature of 15 °C.152 The CH3-SS mode appearsfor acetone at 2926 cm-1, which is slightly shiftedfrom the bulk Raman and IR studies which placeit at 2922 and 2924 cm-1, respectively.157 TheCH3-AS is less distinct, primarily due to the narrowseparation between these two peaks for acetone (∼43cm-1) relative to DMSO (∼84 cm-1) and MSA (89cm-1). The surface methyl groups of acetone prefer-entially orient away from the bulk liquid. VSF studiesof pure acetone surfaces have been reported157 withresults similar to the aqueous acetone studies of Allenet al.152 These studies were conducted at three polari-zations, SSP, PPP, and SPS. These studies, whichcombine VSF measurements and molecular dynamicssimulations, suggest that one of the methyl groupspoints away from the liquid surface and the other isembedded in the liquid. The assignment of the CH3-SS mode and its energy agrees well with the earlierwork of Allen et al.,152 although the destructive

Figure 22. VSF spectra using SSP polarization of thesurface of aqueous solutions containing (a) 0.1 mole fractionDMSO, (b) 0.1 mole fraction of MSA, and (c) 0.1 molefraction of acetone. (Reprinted with permission from ref152. Copyright 2000 Elsevier Science.)


interference and assignment of the CH3-AS is notspecifically addressed. The acetone surface is con-cluded to be more ordered than that of the bulk.

Ammonia-water complexes have been the focus ofwork by Shultz and co-workers.81,158 Understandingthe interaction of ammonia with different surfacesarises from its importance in heterogeneous catalysisand its relevance to various industrial processes.With ammonia as the most abundant alkaline com-pound in the atmosphere, it is the principal speciesthat neutralizes strong inorganic acids and hence isimportant in atmospheric chemistry.159 Figure 23

shows a VSF spectrum of concentrated ammonia. Themajor feature in the spectrum at 3312 cm-1 is anintense peak assigned to the symmetric N-H stretch,confirming the presence of NH3 at the interface. Thispeak is red shifted by ∼20 cm-1 from the fundamen-tal NH3 infrared gas-phase absorption. A weakerdeformation mode is observed at 3200 cm-1. Thedangling (free) OH peak is suppressed due to watermolecules complexing with ammonia at the interface.In increasingly dilute ammonia solutions, the N-HSS mode is less intense and the free OH peak ofwater appears. Comparison of the calculated andobserved VSF intensities for different polarizationcombinations has been used to determine the tiltangles for surface ammonia molecules, further char-acterizing the structure of the ammonia complex.Figure 24 is a schematic of the NH3-H2O complex

orientation derived from these studies. The C3 axisis concluded to be tilted between 25° and 38° relativeto the surface normal and a twist angle of g10°. Theorientation analysis yields an average configuration.The authors examine the influence of the localdielectric constant and the contributions of the Ra-man transition polarizability tensors on the orienta-tion analysis but find that these do not affect theconclusions about orientation.

C. Surfactants Adsorbed at Organic/WaterInterfaces

1. Charged Alkyl SurfactantsUnderstanding the molecular structure of adsor-

bates at the interface between two immiscible liquidsprovides new challenges for both experimentalistsand theorists. Few molecular studies to date havebeen conducted at the interface between an aqueousphase and an organic or hydrophobic liquid phasethat provide information about how surfactants oradsorbates orient and structure as they adsorb. Thefirst and the majority of these studies to date havebeen conducted using VSFS.160,161 Whereas the stud-ies of liquid/liquid interfaces described above focuson water structure and how it is affected by theadsorbates, in this section, the focus is on theadsorbate structure and in some cases how thismolecular structuring compares with similar adsorp-tion at a vapor/water interface.

Figure 23. (a) VSF spectrum of concentrated ammoniausing SSP polarization and solutions of decreasing am-monia concentration (b and c). (Reprinted with permissionfrom ref 81. Copyright 1998 Elsevier Science.)

Figure 24. Schematic of the NH3-H2O complex orienta-tion with respect to the surface XYZ axis. The C3 molecularaxis is approximately 38° from the surface normal Z.(Reprinted with permission from ref 158. Copyright 2000American Institute of Physics.)


The first successful measurements of the vibra-tional spectroscopy of alkyl charged surfactants ad-sorbed at a liquid/liquid interface have been reportedby Messmer et al.57 The goal of these VSFS studieshas been to measure the vibrational spectrum of asimple surfactant, SDS, at this interface to under-stand how such surfactants orient and assemble,similar to the above-described vapor/water studies.As mentioned earlier, the success of these otherwiselow signal experiments has come from the design ofappropriate cells to allow the experiments to beconduced by TIR. Carbon tetrachloride and D2O havebeen used in these experiments as well as later onesbecause of their transparency to the infrared lightaround 3 µm where the CH stretch modes of thesurfactants appear. These studies demonstrated thatspectra of SDS could be readily obtained at thisinterface. Analysis of the acquired spectra show thatfor a monolayer of SDS there is considerable disorderin the alkyl chains at all interfacial concentrations.Later more detailed studies examined both hydroge-nated and deuterated SDS to allow more accuratespectral assignments of CH stretch modes in theseand previous vapor/water studies.54 This study alsoexamined the chain ordering and the terminal methylgroup orientation of SDS as a function of interfacialconcentration. As in previous studies, the ratio of themethyl/methylene intensity has been used to deter-mine molecular conformation. The studies were ac-companied by surface tension measurements to allowcorrelation of the VSF signal and interfacial concen-tration. The studies show that at all interfacialconcentrations the surfactants display considerableconformational disordering. On average, the chainsare oriented normal to the interface. As the concen-tration increases, the chains show increased ordering.However, the persistence of the methylene signalindicates that the chains never reach the all-transconfiguration that is seen for monolayers at vapor/water or air/solid interfaces. This disorder has beenattributed to the presence of the CCl4 penetratinginto the chains and disrupting the van der Waalsinteractions between the chains.

In a related series of studies surfactants of differingcharged headgroups54,162 and chain lengths have beenexamined at the CCl4/D2O interface. In the former,SDS, sodium dodecylsulfonate (DDS), dodecylammo-nium chloride (DAC), and dodecyltrimethylammo-nium chloride (DTAC) have been studied as a func-tion of interfacial concentration and optical polari-zation. As with the previous SDS study,57 all indicatethe presence of gauche defects in the hydrocarbonchain as determined from the intensity ratio of themethyl to methylene symmetric stretch vibrationalmodes. An increase in the surface concentrationresults in a reduction of gauche defects in thehydrocarbon chain. The spectra suggest that theconformational ordering is different for the cationicversus anionic surfactants. The alkyl chains of thecationic surfactants possess the fewest gauche de-fects, whereas the anionic surfactants display moredisorder in the hydrocarbon chains at similar surfaceconcentrations. In studies of chain length, the con-formational ordering of three alkanesulfonates have

been examined, sodium hexanesulfonate (HS), so-dium undecanesulfonate (UDS), and sodiumdode-canesulfonate (DDS) adsorbed at the D2O/CCl4 in-terface.162 For all, an increase in interfacial concen-tration leads to a reduction of gauche defects in thehydrocarbon chains. The alkyl chain of HS displaysthe fewest gauche defects, while DDS and UDSdisplay more disorder in their hydrocarbon chains atsimilar surface concentrations. This is interpreted asa reduction in the possible number of gauche confor-mations in the shorter alkyl chains.

Both the headgroup and the chains of a chargedsurfactant at the CCl4/H2O interface have beenexamined in a study of sodium dodecylbenzene sul-fonate (DBS) where the orientation of the headgrouphas been examined by monitoring changes in phenylCH mode intensities.138 DBS is an important indus-trial and commercial surfactant used in cleansers anddetergents. VSFS studies have been conducted atboth the vapor/water and CCl4/D2O interfaces withthe former studies described in section IV.A. Theresults have been compared with sodium dodecylsul-fonate (DDS), which has the same alkyl chain length.DDS exhibits typical simple surfactant behaviorfound in the previous VSF studies.163 Figure 25 showsa plot of the ratio of methyl/methylene SS for thethese systems. For the DDS (Figure 25a), there is arise in the ratio with surface concentration indicativeof increased order with concentration until monolayercoverage results in a more constant value. In con-trast, at the vapor/water and CCl4/water interface(Figure 25b and c, respectively), DBS does notundergo any ordering of the chains as monolayercoverage is approached. The chains are highly dis-ordered at all surface concentrations at both inter-faces. Even in the presence of excess salt, whichscreens the charged headgroups and allows the DBSmolecules to pack more tightly, there is no significantchange in the order of the alkyl chains (Figure 25d).The authors attribute this to the disruptive natureof the bulky phenyl group in DBS that does not allowthe alkyl chains to order as well as DDS. Phenylgroup orientation for DBS has been examined byfollowing the strongest phenyl mode (analogous to ν2in benzene) which has its IR transition momentpointing from the alkyl chain to the sulfonate. Figure26 shows the SF intensity dependence on surfaceconcentration for the phenyl CH modes at the twointerfaces. For the vapor/water interface, the minimalchange in intensity at low concentrations followed bya sharp rise as a monolayer is achieved suggests thatthere is an abrupt change in orientation of the phenylgroup as the interface reaches a high level of packing.For the liquid/liquid interface the linear relationshipbetween the square root of the intensity and thesurface concentration indicates minimal reorientationwith concentration. Polarization studies indicate thatthe phenyl groups are oriented perpendicular to theinterface throughout this concentration range. Thecombined observations described and the results fromsurface tension measurements showing that theheadgroup areas at monolayer coverage for DDS andDBS are nearly the same, suggest that the phenylrings are in a staggered arrangement at the liquid/


liquid interface and that this staggering of theheadgroups disrupts the ability of the chains to orderwhile allowing the phenyl rings to maintain theirperpendicular orientation.

2. Biomolecules

Many important biomolecules in biological organ-isms and in the environment are classified as sur-factants because of their tendency to adsorb athydrophobic/hydrophilic interfaces. Phospholipids areone example of such biological surfactants. Thesebiomolecules constitute the major component of mostcell membranes and consist of a charged headgroupconnected to a pair of long acyl chains by means of athree-carbon glycerol backbone. At interfaces, theseamphiphilic molecules form Langmuir films whichexhibit a host of different phases and morpholo-gies.164,165 The interest in studying their behavior atliquid surfaces arises from the ability to use anyinsight gained from these studies to understand thenature of more complex phase behavior which takesplace in bilayer systems. Most previous studies ofphospholipid assembly using other methods havebeen conducted at vapor/water interfaces.165,166

The molecular structure of phospholipids as-sembled at the vapor/water and organic/water inter-faces has been studied in a number of VSFS studies.

The first VSF studies of phospholipids at an organic/water interface were reported by Walker et al.167 Inthese experiments, spectra of the CH stretchingregion of several different phosphocholines adsorbedto the D2O/CCl4 interface were measured. As withother surfactants, the relative intensity of the CH-SS modes have been used as a means of determiningchain conformation. The phospholipids used belongto a class of saturated, symmetric, dialkylphospho-cholines (PCs) having alkyl chain lengths of 12carbon atoms (dilauroyl-PC or DLPC), 14 carbonatoms (dimyristoyl-PC or DMPC), 16 carbon atoms(dipalmitoyl-PC or DPPC), and 18 carbon atoms(distearoyl-PC or DSPC) (Figure 27). These mono-layers formed from the breakup at the interface ofaqueous phase phosphocholine vesicles with theiradsorption measured by VSFS in conjunction withinterfacial tension measurements. The experimentswere conducted at a series of concentrations aboveand below the bilayer gel to liquid crystalline phasetransition temperature. It has been found that chainorder is dependent on both alkyl chain length andinterfacial phospholipid concentration. The interfa-cial pressure studies show that the forces within thebilayers of the aqueous vesicles control the interfacialconcentration at the liquid/liquid interface and henceof the subsequent conformation of the adsorbed PCs.

Figure 25. Ratio of the methyl to methylene symmetric stretch intensities as a function of surface concentration for (a)DDS at the air/water interface, (b) DBS at the air/water interface, (c) DBS at the air/water interface with 0.1 M NaCl, and(d) DBS at the CCl4/water interface with 0.1 M NaCl. (Reprinted with permission from ref 138. Copyright 2000 AmericanChemical Society.)


For PCs originating from the liquid crystalline stateof vesicle bilayers, the reduced barrier to vesiclebreakup at the interfaces results in a higher degreeof order among the alkyl chains. In contrast, strongercohesional forces in the vesicles in the gel stateinhibit PC deposition at the interface and the alkylchains of these expanded monolayers show greaterdisorder. At equivalent interfacial concentration, the

alkyl chain structure of the PCs show a dependenceon chain length. At the vapor/water interface, mono-layers composed of longer chain phospholipids showgreater alkyl chain ordering than the shorter chainPCs. For the same PCs studied at the water-CCl4interface, there is much less of a difference inordering with chain length and, in fact, there is aslight trend toward decreased chain ordering withchain length. The authors interpret this differenceas evidence that under ambient conditions the CCl4solvates the alkyl chains of monolayers adsorbed atthe water-CCl4 interface. Monolayers consisting oflonger chain species (i.e., DPPC and DSPS) are morereadily solvated than the shorter chains, and thereduction in interchain attractive forces leads togreater propensity for torsional distortion and gauchedefects. Similar chain length studies have beenconducted for comparison at the vapor/water inter-face.168,169 For the vapor/water interface, the longerchains are more conformationally ordered as theyexperience stronger interchain van der Waals forcesand form monolayers corresponding to more orderedmonolayers than those of their shorter chain coun-terparts. A later study in this laboratory extendedthese studies of symmetric PCs to those with longeralkyl chains up to C22.170 Different methods of mono-layer preparation have been necessary in thesestudies due to the insolubility of the larger PCs inthe aqueous phase. Samples were prepared by dis-solving the PC in chloroform and spreading it at theinterface by gently expelling small drops of thechloroform solution from a syringe tip placed under-neath the vapor/water interface above the liquid/liquid interface and allowing them to fall by gravityto the liquid/liquid interface. Several sample spread-ings were necessary in some cases to produce thedesired close-packed interfacial monolayer at theliquid/liquid interface. With this means of prepara-tion it has been found that the C18 and longer chainPCs form extremely well-ordered interfacial layerswith chains in a predominately all-trans conforma-tion while C16 and C15 PCs formed layers withdisordered chains. The C17 PCs produced layers withintermediate degree of order.

In another set of VSF studies of PCs, Smiley andRichmond examined the molecular level organizationof saturated symmetric and asymmetric chain PCs.171

The importance of this study is that a large majorityof biological phospholipids contain two dissimilarhydrocarbon chains per molecule. Little is knownabout asymmetric chain phospholipids in bilayers,and even less is known about their monolayer prop-erties. In these studies the PC monolayers have beenformed at the CCl4/H2O interface by injection of PCsinto the bulk aqueous phase. The PCs had variedchain combinations of CnCm where n ) 18, 16, 14,and 12 and m ) 18, 16, 14, 12, and 10. It has beenfound that three of the PCs studied, specifically C18/C16, C18/C18, and C16/C18, formed extremely well-ordered layers with primarily all-trans chain confor-mations. Highly asymmetric PCs showed relativelydisordered chains as might be expected from thereduced chain-chain interactions among the mis-matched portions of the longer chains. The shorter

Figure 26. Square root of the VSF intensity of the ν2 modeas a function of concentration for (a) DBS at the air/waterinterface with 0.1 M NaCl and (b) DBS at the CCl4/waterinterface with 0.1 M NaCl. The solid data points correspondto concentrations where surface pressure measurementsindicate maximum surface coverage. (Reprinted with per-mission from ref 138. Copyright 2000 American ChemicalSociety.)

Figure 27. Phospholipids used belong to a class ofsaturated, symmetric, dialkylphosphocholines (PCs) havingalkyl chain lengths of 12 carbon atoms (dilauroyl-PC orDLPC), 14 carbon atoms (dimyristoyl-PC or DMPC), 16carbon atoms (dipalmitoyl-PC or DPPC), and 18 carbonatoms (distearoyl-PC or DSPC). (Reprinted with permissionfrom ref 118. Copyright 1999 Elsevier Science.)


chain PCs of <16 carbons/chain are also relativelydisordered. The authors attribute the greater disor-der seen in the shorter chain PCs, irrespective ofchain mismatch, as a consequence of reduced chain-chain interactions relative to the longer chain PCs.Interestingly, the main differences measured in thesefilms could not be predicted based on the respectivegel-to-liquid crystalline phase transition tempera-tures. Disordered chains were found for PCs bothabove and below their respective transition temper-ature for these room-temperature studies. The au-thors speculate that this may result from interfaciallayer subtransitions.

A series of studies of biological systems has beenconducted by Cremer and colleagues that examinedbiomolecules, including proteins, as they adsorb atquartz surfaces. These recent studies shed light onhow water structures near these biological macro-molecules under various pH conditions.172-175

D. Surfactants and Adsorbates at Solid/AqueousInterfaces

Studies of molecular structure of solid/liquid in-terfaces provide further application for VSFS.52 Sev-eral studies have appeared that explore the structureof monolayers and surfactants at a solid/water in-terface. Bain and co-workers examined the coadsorp-tion of SDS and dodecanol a model hydrophobicsurface that was prepared by self-assembly of octa-decanethiol (ODT) on gold.176,177 In the first study,176

they show that VSFS could discriminate between thecoadsorbed SDS and dodecanol and demonstrate thesensitivity of VSFS to the packing density andconformational order in the adsorbed monolayer.Selective deuteration has been employed to distin-guish between the two adsorbed surfactants. In thesecond study,177 with quantitative measurements ofmixed monolayers, it has been found that dodecanolpartitioned to the interface at a higher level thanpresent in the bulk liquid. For example, a mixtureof 6 mM SDS + 10 µM dodcanol showed a monolayercomprised of 63% dodecanol and 37% SDS. It had apacking density comparable to that of a pure mono-layer of dodecanol. The composition of the monolayerhas also been found to be isotope dependent. It issuggested that the dependence arises from a highlysurface active impurity in the SDS or quantitativelyby a lattice model of the monolayer within regularsolution theory. Briggs et al.178 measured the adsorp-tion of a series of dichain sugar surfactants (di-(Cn-Glu)) from an aqueous solution onto a similardeuterated ODT surface chemisorbed on a gold-coated, chromium-primed silicon wafer. By monitor-ing the ratio of methyl and methylene SS modes theydetermined the relative conformational ordering inthe di-(Cn) chains. The SF spectra show that for di-(C6-Glu), the effectiveness and efficiency of adsorptionis only marginally affected by temperature up to 95°C. Using partially deuterated d30-di-(C6-Glu) mono-layers, they show that methylene resonances arisesolely from the tail groups. The methylene modeamplitudes are found to generally increase with di-(Cn-Glu) tail length as a result of more gauchedefects. The methyl mode amplitudes remain nearly

constant. Conformational disorder in the tail groupis found to increase with decreasing solution concen-tration and solution concentration below the criticalmicelle concentration. The monolayers show excep-tional thermal stabilities on hydrophobic surfaces.This makes them attractive candidates for formingtemperature-insensitive microemulsions with ap-plication in enhanced oil recovery. Zolk and co-workers179 investigated the solvation of olio(ethyleneglycol)-terminated self-assembled monolayers on goldin the C-H stretching region. Comparison of themonolayers in ambient atmosphere, in contact withwater, and in contact with carbon tetrachloride showthat the film structure is strongly disturbed by theinteraction of the liquid with the monolayer. Theseresults are consistent with the earlier conclusions ofmonolayers at the CCl4/H2O interface.54,161,167,180 Inrelated studies of dioctadecyl dimethylammoniumchloride (DOAC) adsorbed at quartz/liquid interfaces,it has been found that the chains can assume manydifferent conformations depending upon whether thesolvent is CCl4 or CDCl3, short-chain alkanes, alcohol,or water.180 Duffy and co-workers examined theadsorption of potassium oleate and sodium octanoateat the iron-water interface.181 These studies areimportant because of the use of lubricating films toreduce friction at metal surfaces. The moleculesstudied are believed to act as boundary layer lubri-cants at iron surfaces in aqueous solution. Thestrength and phase of the resonances in the VSFspectrum of oleate indicate that a bilayer was ad-sorbed at the iron surface. In comparison, adsorbedsodium octanoate did not contain any resonances atany applied potential implying that the ordered filmsformed by oleate are more effective at lubricatingthan the disordered films.

E. Electrochemical Interfaces

VSF measurements metal/aqueous electrode sur-faces have thus far been conducted on Au, Ag, andPt electrode surfaces. The most extensive studieshave been conducted with platinum single-crystallineand polycrystalline surfaces. Using a free electronlaser, Guyot-Sionnest and Tadjeddine demonstratedthe first use of VSFS to study ionic adsorption at anelectrified metal/aqueous electrolyte interface.182,183

These studies involved measurements of cyanide(CN-), thiocyanate (SCN-), and carbon monoxide(CO) adsorbed on polycrystalline platinum. The stud-ies indicate that the CN- and SCN- adsorb at thesurface but that the nitrogen atom can take twodifferent orientations, either facing or pointing awayfrom the surface. Daum and co-workers using abroadly tunable IR laser system conducted similarexperiments on Pt(111) and found evidence for co-valently bound CN-.184-186 The adsorption of CN-,SCN-, and related OCN- ions has also been exam-ined on Ag and Au electrodes where the adsorptionis found to vary with potential in characteristic waysfor each of the ions.187-190 Hydrogen adsorption onpolycrystalline and single-crystal Pt electrodes hasbeen examined in several studies by Peremans andTadjeddine.187,191-194 In these studies a significantpotential dependence in the VSF spectrum of the


Pt-H stretch is found. More than one peak for thismode appears, which the authors attribute to thebonding. The same investigators also examine theelectrochemically induced decomposition of methanolon Pt electrodes using VSFS by monitoring the COadsorption after decomposition.187,195-197 The CO hasbeen found to be present in both a single and multiplebonded state. The lifetime of a CO stretch vibrationon a Pt surface has also been examined and found tobe a few picoseconds.198-200 This lifetime is found tobe relatively independent of the solvent environmentor the application of electrochemical potential. Theorientation of acetonitrile at the Pt(111) electrode hasbeen examined as a function of concentration andelectrode potential.201 Acetonitrile is found to orientwith applied potential with the C-C bond perpen-dicular to the surface. Between 200 and 600 mV (vsthe normal hydrogen electrode) its orientation ispredominately with the methyl group directed towardthe metal while the CN group is toward the metalabove 800 mV. As water is added to the solution, theorientation is found to be disrupted.

F. Polymer Surfaces

Several studies have appeared that investigate theproperties of water at polymer surfaces.202-204 Kimand Shen204 studied the treatment of polyimidesurfaces treated with NaOH solution to improve theadhesion with metals. Both ultraviolet absorptionand VSFS have been employed in the studies. Theyfound that the conversion and subsequent etching ofthe polyimide film by the solution is more effectivein the amorphous part of the film. Drying of the filmconverts the surface amides back to imides. In studiesof the water/poly(ethylene glycol) interface, Dressenet al.202 show that poly(ethylene glycol), whose mo-lecular arrangement is originally relatively ordered,becomes disordered in the presence of water. Ad-ditionally, they find a new OH band that they identifyto water molecules that are strongly interacting withthe polymer. Three polymer blends immersed inwater have been examined in studies by Gracias etal.203 using both VSFS and scanning force microscopy.These studies involved polyethylene and polypropy-lenes of different molecular weights and structures.In aqueous solutions, polymer blends were found tosegregate with the more hydrophilic polymer goingto the surface. These studies have relevance tounderstanding the hydrated state of biopolymers thatcan be used as implants in the body.

V. Summary and Conclusions

There has long been a desire to understand themolecular structure and bonding that occurs ataqueous surfaces and interfaces. It is only recentlythat experimental techniques have become availableto provide this type of information. Vibrational sumfrequency spectroscopy is increasingly becoming atool of choice for such studies because of its abilityto measure the vibrational spectroscopy of moleculeswith inherent surface specificity, and the ability tomake such measurements in a conventional labora-tory setting. The future is particularly promising in

this area as studies proceed to longer wavelengthswhere different adsorbate modes can be measured,and to shorter time scales where interfacial dynamicscan be probed. This review has provided an overviewof VSFS studies that have been conducted on aqueoussurfaces and the contributions that the results havemade to the field.

As this review describes, the molecular structureand bonding of water at a vapor/water interface hasbeen the most extensively studied system since thefirst measurement was made by VSFS nearly adecade ago. Recent advances in laser technology anddetection methods, inclusion of appropriate normal-ization procedures for power and anomalous disper-sion, and improved sample preparation are resultingin a consensus from various laboratories as to themost accurate spectrum of OH stretching modes fromthe vapor/water interface. The current challenge isobtaining an interpretation of this spectrum thatinvolves identification of different types of waterbonding species at the interface. This requires at-tention to appropriate analysis procedures and spec-tral fitting routines that take into account the phaserelationships between contributing vibrational modesand interferences between adjacent resonant modesand orientational effects. Important input in decon-volution of broad spectral features, such as polariza-tion and isotopic dilution experiments, are providingvaluable new information. Equally important aretheoretical treatments of this and other aqueousinterfaces that assist in interpretation of the data viathe molecular simulations and simulation of VSFspectra. Studies of the adsorption of small solutemolecules that have relevance to atmospheric processare also being studied at these surfaces, and impor-tant information about the partitioning and structur-ing at the vapor/water interface is emerging. Thereare a plethora of molecules and systems that will bevaluable to study in this area in order to understandreactions and the molecular structure of the surfacesof aerosols and other aqueous surfaces. Beyond theroom-temperature vapor/water interface, VSF offersexciting opportunities for understanding ice surfacesand ice/water interfaces as a function of temperature,different ice crystal faces, and thin films of ice grownon different solids. The VSF studies will provideimportant surface-specific input to add to informationfrom a growing number of techniques that areproviding insight about molecular structure andinteractions at ice surfaces.

The structure and bonding of water in contact witha hydrophobic liquid is an area that has been largelydominated by theoretical efforts due to the paucityof experimental methods for probing this complexinterface. This review demonstrates the type ofinformation that can be gained from VSFS studiesin this area. VSFS has recently provided the firstdetailed molecular insight into contributing waterspecies at this type of interface, first at the CCl4/H2Ointerface and now with the extension to other hydro-carbon/water interfaces. The studies show weakerbonding interactions between interfacial water mol-ecules than at the vapor/water interface and alsoshow evidence for the importance of the interfacial


potential created by attractive interactions betweenwater dipoles and the polarizable organic fluid.Isotopic dilution studies and analysis procedures thatappropriately take into account the molecular orien-tation and the phase relationships between contrib-uting OH modes and the weaker bonding nature ofthis interface have allowed spectral assignments andspectral fitting to be done at a level beyond what hasbeen previously possible in VSF spectra of water. TheCCl4/H2O studies provide the framework for futurestudies of organic/water interfaces that hold muchpromise in future years. These experimental effortswill benefit from theoretical efforts in this area thatare just beginning to appear. As an extension of theseneat liquid/liquid studies, related VSFS studies ofthese interfaces in the presence of trace chargedspecies at the interface demonstrate the sensitivityof this technique to water molecules solvating chargeat this interface. Future studies will be important inthis area, particularly as it relates to an improvedunderstanding of charged species at hydrophobicsurfaces.

The review has also described the growing numberof studies that have examined the adsorption ofmolecules and surfactants at aqueous interfacesincluding vapor/water, organic/water, and solid/waterinterfaces. These studies conducted largely at bothvapor/water and liquid/liquid interfaces have largelyfocused on the conformation and ordering of alkylgroups on the adsorbates. As lasers with broaderwavelength capabilities in the infrared become moreprevalent and studies move beyond the C-H andN-H stretch modes of most current studies, valuableinformation will be gained about other parts of thesemolecules, particularly regions that reside at theaqueous surface. There are numerous biologicallyimportant interfacial processes involving surfactantsand solutes that would benefit from VSF studies thathave important relevance to respiration, anesthetics,ion transport, and macromolecular assembly. Thesame is true for the study of surfactants involved inremediation, oil extraction, lubrication, and com-mercially important products. At solid/liquid inter-faces, including electrochemically relevant systems,there are many important issues to examine aboutthe solvating layer at these solid surfaces and alltypes of adsorption including potential-induced ad-sorption.

VI. AcknowledgmentThe author acknowledges the financial support for

the range of studies from her laboratory describedin this review. These agencies include the NationalScience Foundation (CHE-9725751) for the studiesof water at various interfaces, Basic Energy Sciencesof the Department of Energy for the atmosphericstudies, the Office of Naval Research for the biomo-lecular studies, and the Petroleum Research Fundof the American Chemical Society for the surfactantstudies.

References(1) Stillinger, F. H.; Ben-Naim, A. J. Chem. Phys. 1967, 47, 4431.(2) Stillinger, F. H. Science 1980, 209, 451.

(3) Hummer, G.; Garde, S.; Garcı́a, A. E.; Paulaitis, M. E.; Pratt,L. R. J. Phys. Chem. B 1988, 102, 10469.

(4) Hummer, G.; Garde, S.; Garcia, A. E.; Pratt, L. R. Chem. Phys.2000, 258, 349.

(5) Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999,103, 4570.

(6) Wilson, M. A.; Pohorille, A.; Pratt, L. R. J. Phys. Chem. 1987,91, 4873.

(7) Sokhan, V. P.; Tildesley, D. J. Mol. Phys. 1997, 92, 625.(8) Weeks, J. D. J. Chem. Phys. 1977, 67, 3106.(9) Townsend, R. M.; Rice, S. A. J. Chem. Phys. 1991, 94, 2207.

(10) Croxton, C. A. Fluid interfacial phenomena; Wiley: New York,1986.

(11) Benjamin, I. J. Chem. Phys. 1992, 97, 1432.(12) Benjamin, I. Science 1993, 261, 1558.(13) Lee, S. H.; Rossky, P. J. J. Chem. Phys. 1994, 100, 3334.(14) Silverstein, K. A. T.; Haymet, A. D. J.; Dill, K. A. J. Am. Chem.

Soc. 1998, 120, 3166.(15) Schweighofer, K. J.; Benjamin, I. Chem. Phys. Lett. 1993, 202,

379.(16) Shen, Y. R. Solid State Commun. 1998, 108, 399.(17) Richmond, G. L. Annu. Rev. Phys. Chem. 2001, 292, 257.(18) Ren, Y.; Meuse, C. W.; Hsu, S. L.; Stidham, H. D. J. Phys. Chem.

1994, 98, 8424.(19) Willard, D. M.; Riter, R. E.; Levinger, N. E. J. Am. Chem. Soc.

1998, 120, 4151.(20) Chamberlain, J.; Pemberton, J. E. Langmuir 1997, 13, 3074.(21) Dluhy, R. A. J. Phys. Chem. 1986, 98, 1371.(22) Buontempo, J. T.; Rice, S. A. J. Chem. Phys. 1993, 98, 5835.(23) Bell, G. R.; Bain, C. D.; Ward, R. N. J. Chem. Soc., Faraday

Trans. 1996, 92, 515.(24) Knobler, C. M. Advances in Chemical Physics; Wiley: New York,

1990; Vol. 77.(25) Fiehrer, K. M.; Nathanson, G. M. J. Am. Chem. Soc. 1997, 119,

251.(26) Mitrinovic, D. M.; Zhang, Z.; Williams, S. M.; Huang, Z.;

Schlossman, J. L. J. Phys. Chem. B 1999, 103, 1779.(27) Brezesinski, G.; Thoma, M.; Struth, B.; Mohwald, H. J. Phys.

Chem. 1996, 100, 3126.(28) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Penfold, J. J. Chem. Soc.,

Faraday Trans. 1996, 92, 403.(29) Weinbach, S. P.; Kjaer, K.; Bouwman, W.; Als-Nielsen, J.;

Leiserowitz, L. J. Phys. Chem. 1996, 100, 8356.(30) Israelachvili, J. N. Intermolecular and Surface Forces; Academic

Press: New York, 1996; Vol. 2.(31) Eisenthal, K. B. Chem. Rev. 1996, 96, 1343.(32) Wirth, M. J.; Burbage, J. D. J. Phys. Chem. 1992, 96, 9022.(33) Shen, Y. R. Nature 1989, 337, 519.(34) Bloembergen, N. Opt. Acta 1966, 13, 311.(35) Corn, R. M.; Higgins, D. A. Chem. Rev. 1994, 94, 107.(36) Richmond, G. L.; Robinson, J. M.; Shannon, V. L. Prog. Surf.

Sci. 1988, 28, 1.(37) Bloembergen, N.; Simmon, H. J.; Lee, C. H. Phys. Rev. 1969,

181, 1261.(38) Gragson, D. E.; Richmond, G. L. J. Chem. Phys. 1997, 107, 9687.(39) Gragson, D. E.; McCarty, B. M.; Richmond, G. L. J. Phys. Chem.

1996, 100, 14272.(40) Gragson, D. E.; Richmond, G. L. J. Phys. Chem. B 1998, 102,

3847.(41) Gragson, D. E.; Richmond, G. L. J. Phys. Chem. B 1998, 102,

569.(42) Gragson, D. E.; Richmond, G. L. J. Am. Chem. Soc. 1998, 120,

366.(43) Gragson, D. E.; McCarty, B. M.; Richmond, G. L. J. Am. Chem.

Soc. 1997, 119, 6144.(44) Goates, S. R.; Schofield, D. A.; Bain, C. D. Langmuir 1999, 15,

1400.(45) Schreier, F. J. Quantum Spectrosc. Radiat. Transfer 1992, 48,

743.(46) Wolfrum, K.; Laubereau, A. Chem. Phys. Lett. 1994, 228, 83.(47) Lobau, J.; Wolfrum, K. J. Opt. Soc. Am. B 1997, 14, 2505.(48) Hirose, C.; Akamatsu, N.; Domen, K. Appl. Spectrosc. 1992, 46,

1051.(49) Hirose, C.; Akamatus, N.; Domen, K. J. Chem. Phys. 1992, 96,

997.(50) Brown, M. G.; Raymond, E. A.; Allen, H. C.; Scatena, L. F.;

Richmond, G. L. J. Phys. Chem. A 2000, 104, 10220.(51) Fredkin, D. R.; Komornicki, A.; White, S. R.; Wilson, K. R. J.

Chem. Phys. 1983, 78, 7077.(52) Bain, C. D. J. Chem. Soc., Faraday Trans. 1995, 91, 1281.(53) Yang, Y. J.; Pizzolatto, R. L.; Messmer, M. C. J. Opt. Soc. Am.

B 2000, 17, 638.(54) Conboy, J. C.; Messmer, M. C.; Richmond, G. L. J. Phys. Chem.

1996, 100, 7617.(55) Bell, G. R.; Li, Z. X.; Bain, C. D.; Fischer, P.; Duffy, D. C. J.

Phys. Chem. B 1998, 102, 9461.(56) Messmer, M. C.; Conboy, J. C.; Richmond, G. L. SPIE Proc. 1995,

2547, 135.


(57) Messmer, M.; Conboy, J. C.; Richmond, G. L. J. Am. Chem. Soc.1995, 117, 8039.

(58) Becraft, K.; Richmond, G. L. Langmuir 2001, 17, 7721.(59) McKenna, C. E.; Knock, M. M.; Bain, C. D. Langmuir 2000, 16,

5853.(60) McGuire, J. A.; Beck, W.; Wei, X.; Shen, Y. R. Opt. Lett. 1999,

24, 1877.(61) van der Ham, E. W. M.; Vrehen, Q. H. R.; Eliel, E. R. Opt. Lett.

1996, 21, 1448.(62) Richter, L. J.; Petralli-Mallo, T. P.; Stephenson, J. C. Opt. Lett.

1998, 23, 1594.(63) Scatena, L. F.; Richmond, G. L. Science 2001, 292, 908.(64) Du, Q.; Superfine, R.; Freysz, E.; Shen, Y. R. Phys. Rev. Lett.

1993, 70, 2313.(65) Du, Q.; Freysz, E.; Shen, Y. R. Science 1994, 264, 826.(66) Raymond, E. A.; Tarbuck, T.; Richmond, G. L. J. Phys. Chem. B

2002, 106, 2817.(67) Allen, H. C.; Raymond, E. A.; Richmond, G. L. J. Phys. Chem. A

2001, 105, 1649.(68) Shultz, M. J.; Schnitzer, C.; Simonelli, D.; Baldelli, S. Int. Rev.

Phys. Chem 2000, 19, 123.(69) Schnitzer, C.; Baldelli, S.; Shultz, M. J. Chem. Phys. Lett. 1999,

313, 416.(70) Wei, X.; Shen, Y. R. Phys. Rev. Lett. 2001, 86, 4799.(71) Scherer, J. R. In Advances in Infrared and Raman Spectroscopy;

Clark, R. J. H., Hester, R. E., Eds.; Heyden: Philadelphia, 1978;Vol. 5, p 149.

(72) Schuster, P.; Zundel, G.; Sandorfy, C. The Hydrogen Bond.Recent Developments in Theory and Experiment; North-Holland,1976; Vols. I-III.

(73) Eisenberg, D.; Kauzmann, W. The Structure and Properties ofWater; Oxford University Press: New York, 1969.

(74) Franks, F. Water: A Comprehensive Treatise; Plenum Press:New York, 1972; Vol. 1.

(75) Tanford, C. The Hydrophobic effect: Formation of Micelles andBiological Membranes; Wiley-Interscience Publications: NewYork, 1973.

(76) Walrafen, G. E.; Yang, W. H.; Chu, Y. C. ACS Symp. Ser. 1997,676, 287.

(77) Hare, D. E.; Sorenson, C. M. J. Chem. Phys. 1990, 84, 25.(78) Buch, V.; Devlin, J. P. J. Chem. Phys. 1999, 110, 3437.(79) Whalley, E.; Klug, D. D. J. Chem. Phys. 1986, 84, 78.(80) Bertie, J. E.; Whalley, E. J. Chem. Phys. 1964, 40, 1637.(81) Simonelli, D.; Baldelli, S.; Shultz, M. J. Chem. Phys. Lett. 1998,

298, 400.(82) Schnitzer, C.; Baldelli, S.; Shultz, J. J. J. Phys. Chem. B 2000,

104, 585.(83) Du, Q.; Freysz, R.; Shen, Y. R. Phys. Rev. Lett. 1994, 72, 238.(84) Du, Q.; Freysz, E.; Shen, Y. R. Am. Inst. Phys. 1994, 157.(85) Huang, Z. S.; Miller, R. E. J. Chem. Phys. 1989, 91, 6613.(86) Watts, R. O.; Reimers, J. R. J. Chem. Phys. 1984, 85, 83.(87) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford

University Press: Oxford, 1997.(88) Scatena, L. F.; Richmond, G. L. J. Phys. Chem. B 2001, 105,

11240.(89) Green, J. L.; Lacey, A. R.; Sceats, M. G. Chem. Phys. Lett. 1986,

130, 67.(90) Wall, T. T.; Hornig, D. F. J. Chem. Phys. 1965, 43, 2079.(91) Zoidis, E.; Yarwood, J.; Tassaing, T.; Danten, Y.; Besnard, M.

J. Mol. Liq. 1995, 64, 197.(92) Kint, S.; Scherer, J. R. J. Chem. Phys. 1978, 69, 1429.(93) Devlin, P. J. Chem. Phys. 1986, 90, 1322.(94) Wojcik, M. J.; Buch, V.; Devlin, J. P. J. Chem. Phys. 1993, 99,

2332.(95) Morita, A.; Hynes, J. T. Chem. Phys. 2000, 258, 371.(96) Baldelli, S.; Schnitzer, C.; Shultz, M. J. J. Phys. Chem. B 1997,

101, 4607.(97) Coker, D. F.; Miller, R. E.; Watts, R. O. J. Chem. Phys. 1985,

82, 3554.(98) Conrad, M. P.; Strauss, H. L. J. Phys. Chem. 1987, 91, 1668.(99) Chang, T.-M.; Dang, L. X. J. Chem. Phys. 1996, 104, 6772.

(100) Linse, P. J. Chem. Phys. 1987, 86, 4177.(101) Brown, M. G.; Walker, D. S.; Richmond, G. L. J. Phys. Chem.

2002, submitted.(102) Gragson, D. E.; Richmond, G. L. Langmuir 1997, 13, 4804.(103) Buch, V.; Devlin, J. P. J. Chem. Phys. 1991, 94, 4091.(104) Lied, A.; Dosch, H.; Bilgram, J. H. Phys. Rev. Lett. 1994, 72,

3554.(105) Dosch, H.; Lied, A.; Bilgram, J. H. Surf. Sci. 1995, 327, 145.(106) Golecki, I.; Jaccard, C. J. Phys. C 1978, 11, 4229.(107) Furukawa, Y.; Yamamoto, M.; Jurado, T. J. Cryst. Growth 1987,

82, 655.(108) Beaglehole, D.; Nason, D. Surf. Sci. 1980, 96, 357.(109) Elbaum, M.; Lipson, S. G.; Dash, J. G. J. Cryst. Growth 1987,

129, 491.(110) Wei, X.; Miranda, P. B.; Shen, Y. R. Phys. Rev. Lett. 2001, 86,

1554.(111) Scherer, J. R.; Snyder, R. G. J. Chem. Phys. 1977, 67, 4794.

(112) Miranda, P. B.; Xu, L.; Shen, Y. R.; Salmeron, M. Phys. Rev.Lett. 1998, 81, 5876.

(113) Salmeron, M.; Bluhm, H. Surf. Rev. Lett. 1999, 6, 1275.(114) Su, X.; Lianos, L.; Shen, Y. R.; Somorjai, G. A. Phys. Rev. Lett.

1998, 80, 1533.(115) Yeganeh, M. S.; Dougal, S. M.; Pink, H. S. Phys. Rev. Lett. 1998,

83, 1179.(116) Vogel, V.; Shen, Y. R. Annu. Rev. Mater. Sci. 1991, 21, 515.(117) Miranda, P. B.; Du, Q.; Shen, Y. R. Chem. Phys. Lett. 1998, 286,

1.(118) Walker, R. A.; Richmond, G. L. Colloids Surf. A 1999, 154, 175.(119) Scatena, L. F.; Richmond, G. L. Submitted for publication.(120) Zhao, X.; Ong, S.; Eisenthal, K. B. Chem. Phys. Lett. 1993, 202,

513.(121) Richmond, G. L. Langmuir 1986, 2, 132.(122) Corn, R. Anal. Chem. 1991, 63, 285A.(123) Miranda, P. B.; Shen, Y. R. J. Phys. Chem. B 1999, 103, 3292.(124) Raduge, C.; Pflumio, V.; Shen, Y. R. Chem. Phys. Lett. 1997,

274, 140.(125) Baldelli, S.; Schnitzer, C.; Shultz, M. J.; Campbell, D. J. J. Phys.

Chem. B 1997, 49, 10435.(126) Baldelli, S.; Schnitzer, C.; Campbell, D. J.; Shultz, M. J. J. Phys.

Chem. B 1999, 103-, 2789.(127) Baldelli, S.; Schnitzer, C.; Shultz, M. J.; Campbell, D. J. Chem.

Phys. Lett. 1998, 287, 143.(128) Baldelli, S.; Schnitzer, C.; Shultz, M. J. Chem. Phys. Lett. 1999,

302, 157.(129) Schnitzer, C.; Baldelli, S.; Campbell, D. J.; Shultz, M. J. J. Phys.

Chem. A 1999, 103, 6383.(130) Hunt, J. H.; Guyot-Sionnest, P.; Shen, Y. R. Chem. Phys. Lett.

1987, 133, 189.(131) Guyot-Sionnest, P.; Hunt, J. H.; Shen, Y. R. Phys. Rev. Lett. 1987,

59, 1597.(132) Wolfrum, K.; Graener, H.; Laubereau, A. Chem. Phys. Lett. 1993,

214, 41.(133) Stanners, C. D.; Du, Q.; Chin, R. P.; Cremer, P.; Somorjai, G.

A.; Shen, Y.-R. Chem. Phys. Lett. 1995, 232, 407.(134) Braun, R.; Casson, B. D.; Bain, C. D. Chem. Phys. Lett. 1995,

245, 326.(135) Casson, B. D.; Braun, R.; Bain, C. D. Faraday Discuss. 1996,

104, 209.(136) Edgar, R.; Huang, J. Y.; Popovitz-Biro, R.; Kjaer, K.; Bouwman,

W. G.; Howes, P. B.; Als-Nielsen, J.; Shen, Y. R.; Lahav, M.;Leiserowitz, L. J. Phys. Chem. B 2000, 104.

(137) Zhang, D.; Gutow, J.; Eisenthal, K. B. J. Phys. Chem. 1994, 98,13729.

(138) Watry, M.; Richmond, G. L. J. Am. Chem. Soc. 2000, 122, 875.(139) Zhuang, X.; Miranda, P. B.; Kim, D.; Shen, Y. R. Phys. Rev. B

1999, 59, 12632.(140) Knock, M. M.; Bain, C. D. Langmuir 1999, 16, 2857.(141) Bell, G. R.; Manning-Benson, S.; Bain, C. D. J. Phys. Chem. B

1998, 102, 218.(142) Ward, R. N.; Duffy, D. C.; Bell, G. R.; Bain, C. D. Mol. Phys.

1996, 88, 269.(143) Bell, G. R.; Bain, C. D.; Li, Z. X.; Thomas, R. K.; Duffy, D. C.;

Penfold, J. J. Am. Chem. Soc. 1997, 119, 10227.(144) Hoffmann, H.; Ebert, G. Angew. Chem., Int. Ed. Engl. 1988, 27,

902.(145) Porter, M. R. Handbook of Surfactants; Chapman and Hall:

London, 1994; Vol. 2.(146) Sefler, G. A.; Du, Q.; Miranda, P. B.; Shen, Y. R. Chem. Phys.

Lett. 1995, 235, 347.(147) Casson, B. D.; Bain, C. D. J. Phys. Chem. B 1998, 102, 7434.(148) Zhang, D.; Gutow, J. H.; Eisenthal, K. B. J. Chem. Phys. 1993,

98, 5099.(149) Eisenthal, K. B. Acc. Chem. Res. 1993, 26, 636.(150) Zhang, D.; Gutow, J. H.; Eisenthal, K. B. J. Chem. Soc., Faraday

Trans. 1996, 92, 539.(151) Huang, J. Y.; Wu, M. H. Phys. Rev. E 1994, 50, 3737.(152) Allen, H. C.; Raymond, E. A.; Richmond, G. L. Curr. Opin.

Colloids Surf. 2000, 5, 74.(153) Allen, H. C.; Gragson, D. E.; Richmond, G. L. J. Phys. Chem. B

1999, 103, 660.(154) Davis, D.; Chen, G.; Kasibhatla, P.; Jefferson, A.; Tanner, D.;

Eisele, F.; Lenschow, D.; Neff, W.; Berresheim, H. J. Geophys.Res. 1998, 103, 1657.

(155) Jefferson, A.; Tanner, D. J.; Eisels, F. L.; David, D. D.; Chen,G.; Crawford, J.; Huey, J. W.; Torres, A. L.; Berresheim, G. J.Geophys. Res. 1998, 103, 1647.

(156) Singh, H. B.; O’Hara, D.; Herlth, D. J. Geophys. Res. 1994, 99,1805.

(157) Yeh, Y. L.; Zhang, C.; Held, H.; Mebel, A. M.; Wei, X.; Lin, S.H.; Shen, Y. R. J. Chem. Phys. 2001, 114, 1837.

(158) Simonelli, D.; Shultz, M. J. J. Chem. Phys. 2000, 112, 6804.(159) Harrison, R. M.; Gieken, R. v. Atmospheric Particles; John Wiley

Sons: Chichester; New York, 1998; Vol. 5.(160) Richmond, G. L. Anal. Chem. News Views 1997, 69, 536A.(161) Conboy, J. C.; Messmer, M. C.; Walker, R.; Richmond, G. L. Prog.

Colloid Polym. Sci. 1997, 103, 10.


(162) Conboy, J. C.; Messmer, M. C.; Richmond, G. L. Langmuir 1998,14, 6722.

(163) Conboy, J. C.; Messmer, M. C.; Richmond, G. L. J. Phys. Chem.B 1997, 101, 6724.

(164) Mohwald, H. Annu. Rev. Phys. Chem. 1990, 41, 441.(165) McConnell, H. M. Annu. Rev. Phys. Chem. 1991, 171.(166) Knobler, C. M.; Desai, R. C. Annu. Rev. Phys. Chem. 1992, 43,

207.(167) Walker, R. A.; Conboy, J. C.; Richmond, G. L. Langmuir 1997,

13, 3070.(168) Walker, R. A.; Gruetzmacher, J. A.; Richmond, G. L. J. Am.

Chem. Soc. 1998, 120, 6991.(169) Walker, R. A.; Smiley, B. L.; Richmond, G. L. Spectroscopy 1999,

14, 18.(170) Smiley, B.; Richmond, G. L. Biopolymers (Biospectroscopy) 2000,

57, 111.(171) Smiley, B.; Richmond, G. L. J. Phys. Chem. B 1999, 103, 653.(172) Kim, J.; Cremer, P. S. J. Am. Chem. Soc 2000, 122, 12371.(173) Kim, J.; Kim, G.; Cremer, P. S. Langmuir 2001, 17, 7255.(174) Kim, J.; Cremer, P. S. ChemPhysChem 2001, 8/9, 543.(175) Kim, G.; Gurau, M.; Kim, J.; Cremer, P. Langmuir 2002, 18,

2807.(176) Bain, C. D.; Davies, P. B.; Ward, R. N. Langmuir 1994, 10, 2060.(177) Ward, R. N.; Davies, P. B.; Bain, C. D. J. Phys. Chem. B 1997,

101, 1594.(178) Briggs, A. M.; Johal, M. S.; Davies, P. B. Langmuir 1999, 15,

1817.(179) Zolk, M.; Eisert, M.; Pipper, J.; Herrwerth, S.; Eck, W.; Buck,

M.; Grunze, M. Langmuir 2000, 16, 5849.(180) Miranda, P. B.; Pflumio, V.; Saijo, H.; Shen, Y. R. Chem. Phys.

Lett. 1997, 264, 387.(181) Duffy, D. C.; Friedmann, A.; Boggis, S. A.; Klenerman, D.

Langmuir 1998, 14, 6518.(182) Guyot-Sionnest, P.; Tadjeddine, A. Chem. Phys. Lett. 1990, 172,

341.(183) Tadjeddine, A.; Guyot-Sionnest, P. Electrochim. Acta 1991, 36,

1849.(184) Daum, W.; Friedrich, K. A.; Klunker, C.; Knabben, D.; Stimming,

U.; Ibach, H. Appl. Phys. 1994, 59, 553.

(185) Friedrich, K. A.; Daum, W.; Klunker, C.; Knabben, D.; Stimming,U.; Ibach, H. Surf. Sci. 1995, 335, 315.

(186) Daum, W.; Dederichs, F.; Muller, J. E. Phys. Rev. Lett. 1998,80, 766.

(187) Tadjeddine, A.; Peremans, A.; Guyot-Sionnest, P. Surf. Sci. 1995,335, 210.

(188) Bowmaker, G. A.; Leger, J. M.; LeRille, A.; Melendres, C. A.;Tadjeddine, A. Soc. Faraday Trans. 1998, 94, 1309.

(189) LeRille, A.; Tadjeddine, A.; Zheng, W. Q.; Peremans, A. Chem.Phys. Lett. 1997, 271, 95.

(190) Ong, T. H.; Davies, P. B.; Bain, C. D. J. Phys. Chem. 1993, 97,12047.

(191) Peremans, A.; Tadjeddine, A. J. Chem. Phys. 1995, 103, 7197.(192) Peremans, A.; Tadjeddine, A. Phys. Rev. Lett. 1994, 73, 3010.(193) Tadjeddine, A.; Peremans, A. J. Chim. Phys. 1996, 93, 662.(194) Tadjeddine, A.; Peremans, A. J. Electroanal. Chem. 1996, 409,

115.(195) Peremans, A.; Tadjeddine, A. Chem. Phys. Lett. 1994, 220, 481.(196) Peremans, A.; Tadjeddine, A. J. Electroanal. Chem. 1995, 395,

313.(197) Tadjeddine, A.; Peremans, A. Surf. Sci. 1996, 368, 377.(198) Peremans, A.; Tadjeddine, A.; Guyot-Sionnest, P. Chem. Phys.

Lett. 1995, 247, 243.(199) Peremans, A.; Tadjeddine, A.; Zheng, W. Q.; LeRille, A.; Guyot-

Sionnest, P.; Thiry, P. Surf. Sci. 1996, 368, 384.(200) Schmidt, M. E.; Guyot-Sionnest, P. J. Chem. Phys. 1996, 104,

2438.(201) Baldelli, S.; Mailhot, G.; Ross, P.; Shen, Y.-R.; Somorjai, G. A.

J. Phys. Chem. B 2001, 105, 654.(202) Dressen, L.; Humbert, C.; Hollander, P.; Mani, A. A.; Ataka, K.;

Thiry, P. A.; Peremans, A. Chem. Phys. Lett. 2001, 333, 327.(203) Gracias, D. H.; Chen, Z.; Shen, Y. R.; Somorjai, G. A. Acc. Chem.

Res. 1999, 32, 930.(204) Kim, D.; Shen, Y. R. Appl. Phys. Lett. 1999, 74, 3314.

CR0006876


Date post:	18-Jun-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Molecular Bonding and Interactions at Aqueous Surfaces as...

Documents