Sum-Frequency Spectroscopy Analysis of Two-Component

Igor Prudovsky , Thallapuranam Kumar , Sarah Sterling , David Neivandt ... Dan Lis , Julien Guthmuller , Benoît Champagne , Christophe Humbert , Bert...
0 downloads 0 Views 284KB Size
Download PDF
15506

J. Phys. Chem. B 2006, 110, 15506-15513

Sum-Frequency Spectroscopy Analysis of Two-Component Langmuir Monolayers and the Associated Interfacial Water Structure Zhorro S. Nickolov,*,† David W. Britt,*,‡ and Jan D. Miller§ Department of Materials Science and Engineering and A. J. Drexel Nanotechnology Institute, Drexel UniVersity, Philadelphia, PennsylVania 19104, Department of Biological Engineering, Utah State UniVersity, Logan, Utah 84322, and Department of Metallurgical Engineering, UniVersity of Utah, Salt Lake City, Utah 84112 ReceiVed: May 23, 2006; In Final Form: June 19, 2006

Sum-frequency spectroscopy (SFS) in the CH and OH stretching regions was employed to obtain structural information about Langmuir monolayers on the H2O subphase of the model lipid dioctadecyldimethylammonium bromide (DOMA) and of the neutral surfactant methyl stearate (SME) and their mixtures and about the interfacial water structure underneath the films. These results were compared with the sum-frequency spectra of the interface between Langmuir monolayers of stearic acid and stearic acid-DOMA monolayers and water to prove that the uncompensated headgroup charge of DOMA at the interface is the reason for structuring of interfacial water close to the studied monomolecular films. Sum-frequency spectra on D2O subphase were also studied to account for the interference between the CH and OH spectral signatures because of the coherent nature of the SFS signals. Interfacial water structure proved to be a determining factor in the behavior of the mixed lipid monolayers. A mixing induced amplification in the surface potential ∆V observed in our previous work was explained with total increase of the dipole moment for the mixed films, bigger than the arithmetic average for DOMA and SME monolayers alone. The increase is due to the better packing of the molecules in the mixed films and to the decrease in the interfacial water dipole moment arising from a more disordered water structure underneath the mixed monolayers.

1. Introduction The Langmuir monolayer is widely used as a model interface that allows precise control of film chemistry, through choice of amphiphilic compounds, and film structure, through film compression or change in subphase composition, pH, temperature, and so forth. Floating at the air/water interface, the amphiphilic compounds (lipids, proteins, polymers, and surfactants) comprising the Langmuir monolayer orient their hydrophobic component toward air and their polar component toward the water in a manner that favors their self-assembly (assisted through moveable barriers) into well-organized monolayer films. When employing the Langmuir monolayer as a model interface to study interfacial phenomena such as protein adsorption, it is important to characterize the packing arrangement, phase behavior, and miscibility of the monolayer components. In its most basic form, the Langmuir trough consists of a tensiometer (Wilhelmy balance) to measure the change in subphase surface tension (i.e., the surface pressure) as the insoluble amphiphilic compounds are spread and compressed at the air/water interface. The surface pressure is related to the area occupied by the film according to a two-dimensional ideal gas law:

πA ) kT

(1)

where π is the surface pressure, A is the molecular area available * Authors to whom correspondence should be addressed. (Z.N.) Tel.: +1-215-895-1293; fax: +1-215-895-6760; e-mail: [email protected]; (D.B.) Tel.: +1-435-797-2158; Fax: +1-435-797-1248; e-mail: [email protected]. † Drexel University. ‡ Utah State University. § University of Utah.

at this surface pressure, and kT is the thermal energy of the system. Compressing the insoluble film from a low density (gaseous) state to a high density (gel or crystalline) state results in an increase in the surface pressure, and the resulting surface pressure versus molecular area isotherm is a signature of the particular amphiphile. Additional information regarding the packing and orientation of the monolayer components is assessed from the surface potential (∆V), which is readily measured using a vibrating Kelvin probe.1 From the surface potential, one can calculate the effective dipole moment of the monolayer through the Helmholtz relation:

µ ) o∆VA

(2)

where o is the permittivity of vacuum and A is the molecular area at the surface pressure at which ∆V is measured. Calculating the effective dipole moment from the measured surface potential provides an indication of lipid dipole orientation (alkyltail tilt angle), lateral density, and charge state. In the Demchak and Fort (DF) three-layer capacitor model for monolayers on a pure water subphase, it is assumed that ∆V is composed of three independent contributions arising from (1) orientation of interfacial water dipoles, (2) headgroup charges and dipoles, (3) tail-group dipoles.2 Only the orthogonal component of dipoles contributes to the measured surface potential, hence one observes an increase in the effective dipole moment upon compressing a monolayer as the alkyl tails are forced to assume a more orthogonal orientation as the available area in the plane of the air/water interface is restricted. Recently, we have used surface potential measurements to investigate mixing induced changes in lipid and interfacial water dipole moment orientation as factors contributing to amplifica-

10.1021/jp0631578 CCC: $33.50 © 2006 American Chemical Society Published on Web 07/18/2006

Sum-Frequency Spectroscopy Analysis

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15507

tion of surface potential for mixed cationic/nonionic lipid films that in turn results in enhanced protein adsorption.3 The fact that the surface potential value is not a simple average of the surface potential values of the pure and mixed films prompted us to seek an interface specific method to structurally characterize the films at the molecular level. As the structure of a Langmuir monolayer is inherently different from that of the bulk subphase and the thickness of the interfacial region is about several nanometers, advanced interface specific techniques are needed to study the interface at the microscopic level. A powerful array of analytical methods, such as FTIR spectroscopy,4,5 X-ray diffraction6 and reflection,7 and neutron scattering8 have been adapted to the air/water interface providing details into monolayer packing, orientation, and structure, which can be analyzed (minus the contribution of the interfacial water) according to the DF-model. Although these methods can be very useful for studying the monolayers, the nonlinear optical spectroscopic technique of sum-frequency generation spectroscopy (SFS) holds great promise and has emerged in the past decade as very informative and is often the only one which can be successful for investigating both the monolayer and the interfacial water structure.9-11 The capability of SFS to monitor nonintrusively the structure and packing of Langmuir, Langmuir-Blodgett, and selfassembled monolayers is especially fruitful.12-17 It has been shown that the CH regions of the vibrational spectra of model lipid monolayers, surfactants, and adsorbed proteins can be recorded with ease and can be used to investigate their structure and conformation. In the case of monolayers floating on water, the water structure immediately below the monolayers can be studied.12,17,18 This unique feature of SFS to access interfacial water structure, in addition to revealing the structure and packing of the monolayer, is particularly relevant for surface potential studies of Langmuir monolayers where the contribution of water to the surface potential is generally unknown. The aim of this study is to employ SFS to obtain structural information about Langmuir monolayers of the model lipid dioctadecyldimethylammonium bromide (DOMA) and of the neutral surfactant methyl stearate (SME) and their mixtures and about the interfacial water structure below the films. These results are compared with sum-frequency (SF) spectra of Langmuir monolayers of stearic acid and with SF spectra on D2O subphase to account for the interference between the CH and OH spectral signatures because of the coherent nature of SF signals.19 2. Sum-Frequency Spectroscopy SFS principles have been described in detail elsewhere.20-22 In brief, SFS is a surface-specific method based on secondorder nonlinear optical mixing in which two pulsed laser beams, one visible and the second infrared (IR), at frequencies ωvis and ωIR overlap temporarily and spatially on an interface to generate an output at the sum frequency ωSF ) ωIR + ωvis. The underlying principle for the surface specificity of SFS is that second-order nonlinear optical processes are forbidden in bulk media with inversion symmetry, but at a surface or interface the inversion symmetry is broken and sum-frequency generation becomes possible. If the IR frequency is tunable, the technique becomes sensitive to the molecular structure of the interface by detecting vibrational resonances of interfacial molecules. The intensity of the SF signal is proportional to the square of the surface nonlinear susceptibility |χ(2)|2:

I(ωSF) ∝ |F|2 |χ(2)|2I(ωIR)I(ωvis)

(3)

where ωSF ) ωIR + ωvis, I(ωIR), and I(ωvis) are the intensities of the tunable IR and visible beams, and |F|2 is a Fresnel factor accounting for the Fresnel coefficients for each of the fields at ωSF, ωIR, and ωvis. The second-order susceptibility can be written as the sum of a nonresonant and a resonant term:

χ(2) ) χNR(2)eiδ +

∑V |χV(2)|eiΦν

(4)

Here, the resonant term is summed over all vibrational resonances of the interfacial molecules, Φν are the phase angles of the different resonance modes (phase change relative to the incident beam), and δ is the phase of the nonresonant susceptibility. Since the SF signal is proportional to |χ(2)|2, the spectral line shape in SFS is often obtained as a result of interference between the different terms in eq 4. For a vibrational mode, the resonant term is expressed as follows:

χ(2) )

Aν ωIR - ων - iΓν

(5)

where Aν, ων, and Γν are the oscillator strength, resonant frequency, and homogeneous line width of the νth vibrational mode, respectively. Aν is a product of the infrared and Raman transition dipole moments and the number density of the interfacial molecules contributing to the SF signal. For a vibrational mode to be sum-frequency active, both the Raman and infrared transition moments must be nonzero. By choosing polarization conditions for the IR and visible beams and by selecting the polarization of the SF signal, vibrational modes having a particular orientation with respect to the plane of the interface can be studied. 3. Experimental Details Materials. Milli-Q water from a Millipore filtration system with a resisitivity of 18 MΩ‚cm (pH ) 5.5) was used as a subphase in the preparation of the Langmuir films. Dioctadecyldimethylammonium bromide ((CH3(CH2)17)2N(CH3)2Br, DOMA, or DODAB) of 99% purity was obtained from Sigma, methyl stearate (CH3(CH2)17COOCH3, SME) was a GC standard from Merck, and stearic acid (CH3(CH2)17COOH, STA), 99% pure, was from Aldrich. Chloroform (Sigma, HPLC grade) solutions of these lipids with concentration of 1 mM were prepared and stored at 4 °C until use. SME-DOMA and STADOMA mixtures with different mole ratios were prepared volumetrically from the stock solutions. Langmuir Trough and Monolayer Preparation. A Langmuir trough (MicroTrough S, Kibron, Inc., Finland) was used for the preparation of the monolayers. Because of its compact size (33 × 14 cm), it was easily integrated in the SF spectrometer detection scheme without compromising the alignment of the beams. Surface pressure and surface potential isotherms on water subphases (20 °C) were obtained by spreading the studied films into their “gaseous” (uncompressed) phases, waiting 10 min, and then compressing at 5 Å2 molecule-1 min-1 until collapse. Prior to the spreading, the surface area was reduced and surface active contaminants were suctioned off. An initial clean surface was indicated by zero surface pressures at low trough areas. Surface pressures, at which the SF spectra were taken, corresponded to the condensed phases of the films and were between 30 and 35 mN/m. The surface pressures were kept constant during spectra collection with the help of the integrated feedback of the trough.

15508 J. Phys. Chem. B, Vol. 110, No. 31, 2006

Nickolov et al.

Sum-Frequency Generation Spectrometer. In our experiments, we have used an EKSPLA, Ltd. sum-frequency generation spectrometer.23,24 The visible beam at 532 nm, pulse duration 20 ps, was generated after frequency doubling of the fundamental beam of a passive-active mode-locked picosecond Nd:YAG laser. The tunable IR beam was generated in an optical parametric generator (OPG) by first mixing the fundamental (1064 nm) and the third harmonic (365 nm) in an LBO nonlinear crystal and then by mixing the output from the OPG with the fundamental beam in a AgGaS2 crystal of a difference frequency generator (DFG). This tunable OPG/DFG system can cover the range 2.3-10 µm (4350-1000 cm-1 in vibrational frequencies). The visible and IR beams were overlapped spatially at the liquid/ air interface over an area of ca. 500 µm in diameter. Both pulses also overlapped temporally by using an optical delay line in the path of the visible beam. The incident angles of the visible and IR beams at the interface were 60° and 55° degrees, respectively, and the SF signal was collected in reflection at an angle of ca. 60°, was passed through collimating optics, holographic notch filter (532 nm, bandwidth < 20 nm, optical density > 6, Notch-Plus, Kaiser Optical Systems, Inc.), Glan polarizer, and was focused on the entrance slit of a monochromator. A photomultiplier and a gated electronics system were used for detection. The pulse energies of the visible and the IR beams at the sample were ca. 900 and 280 µJ, respectively. Two photodiodes were employed to monitor the output energy of the visible and IR pulses, and all SF spectra are normalized by their reference signals at each spectral point. The spectra have been collected in ssp polarization combination of the three beams (s, sum frequency; s, visible; p, infrared). Data were collected at 2 cm-1 increments, and each point is the average of 100 laser shots. The spectral resolution was determined by the spectral width of the IR beam which is e6 cm-1. All spectra were recorded at constant temperature of 24 °C. 4. Results and Discussion SF Spectra of Mixed Lipid (SME-DOMA) Monolayers. The SF spectra of pure DOMA monolayers; SME:DOMA monolayers at mixing ratios 2:1, 6:1, and 10:1; and of pure SME monolayers are shown in Figure 1, a-e. The O-H and C-H stretching regions of interfacial water and the lipid molecules were analyzed, respectively. Interfacial Water Structure. Significant O-H stretching band signals (3100-3700 cm-1) from the DOMA-water interface, Figure 1a, were recorded allowing us to infer that a highly ordered layer of water molecules forms beneath the DOMA monolayer. Normally, the interfacial water structure studied by SFS is characterized with two features, one centered at ca. 3240 cm-1 and a second at ca. 3440 cm-1.11,18 The first one is assigned to the OH vibrations of tetrahedrally coordinated water molecules and is often referred to as “icelike” structure component to account for the fact that the vibrational spectrum of ice is characterized with strong intensity in this region.25 The second is attributed to water molecules that are involved in not so well ordered structures, connected to other water molecules with distorted, weaker H-bonds. Either of these two assignments is, however, characteristic of water molecules with more disordered hydrogen bonding than in the solid (ice), and the term “water structure” should be accepted as describing a timeaveraged picture of the hydrogen bonding between water molecules in the liquid state. In addition, because of its inversion symmetry, bulk water does not generate SF signal. It is believed that the orientation and ordering of water molecules at an interface is most often a result of the alignment

Figure 1. Sum-frequency spectra at the air-water interface of model mixed lipid monolayers on H2O subphase in the C-H and O-H stretching regions: (a) Pure DOMA; (b) SME:DOMA ) 2:1; (c) SME: DOMA ) 6:1; (d) SME:DOMA ) 10:1; and (e) pure SME. Dashed lines indicate wavenumbers of the vibrational modes discussed in the text. From left: 2876, 2940, 3130, 3240, and 3440 cm-1.

of their molecular dipoles with the electric field present in the interfacial region. In the case of the DOMA-water interface, this field is due to the presence of N+(CH3)2 ions in the headgroups of the DOMA molecules arranged in the twodimensional ordered lattice of the Langmuir film. Thus, we will

Sum-Frequency Spectroscopy Analysis refer to the interfacial water structure promoted by the charged monolayer interface as “electric field induced”. When SME-DOMA mixed monolayers are studied, one more component in the OH band of interfacial water appears, located at ca. 3130 cm-1, Figure 1a-d, in addition to the two OH band components normally observed at water/air interfaces at ca. 3240 and 3440 cm-1. The lower wavenumber of this component characterizes it as corresponding to OH oscillators taking part in hydrogen bonding with increased strength. We assign this component to clusters of water molecules, which are directly bonded to the headgroups of the DOMA molecules. Although this component can be noticed in the OH band of the pure DOMA-water interface, Figure 1a, it is not well resolved. With the increase of the SME fraction in the mixed films, the structured interfacial water layer becomes thinner as seen by the decrease in intensity of the total OH band signal, Figure 1b-d. Interestingly, the intensity in the low-frequency part of the OH band (3130 cm-1) increases its contribution relative to the components at ca. 3240 and 3440 cm-1 with SME increase, Figure 1c and 1d. Finally, there is no OH signal from the pure SME-water interface, Figure 1e. The decrease in the total intensity of the OH band can be explained by the reduced DOMA headgroup dipole concentration which results in decreased thickness of the electric field induced ordered water layers. We believe that at concentrations of SME:DOMA of 10:1 most ordered water molecules at the interface are in clusters bonded to the headgroups. The absence of the OH band signal in the spectrum of the pure SME-water interface can be explained by the nonpolar nature of the SME molecules. The SME headgroups, in addition to being neutral in water, are somewhat hydrophobic because of the presence of a methyl group. Apparently, this also contributes to a more disordered water structure close to the mixed films and to a completely disordered (bulklike) water structure beneath the pure SME film. To support the hypothesis that the structuring of interfacial water close to the studied monomolecular Langmuir films is mostly due to the uncompensated headgroup charge of DOMA, we have examined the SF spectra of a mixed stearic acid (STA)-DOMA monolayer and have compared it to the spectra of the pure STA monolayer-water interface, Figure 2. In Figure 2a, there is no signal in the OH stretching band region of the mixed monolayer suggesting that the water molecules at the interface are completely disordered and thus, having inversion symmetry, are not capable of producing an SF signal. In contrast, the pure STA-water interface exhibits a weak but noticeable OH band signal characteristic of water structuring, Figure 2b. The structuring of water at the STA air interface is believed to be due to the partial deprotonation of the STA headgroups which results in partial negative charge of the film. The negatively charged STA film orients the water dipoles with their hydrogen atoms pointing at the interface. In a study of the effect of pH on the interfacial water structure at the air-STA-water interface by SFS, we have shown that the STA films are neutral below ca. pH ) 4.26 When pH increases, some of the headgroups’ protons are lost and the headgroups become negatively charged. Therefore, structuring of the interfacial water is detected in the OH stretching region. At the natural pH ) 5.6 for the experiments reported here, we believe that a significant number of the STA headgroups are deprotonated. This explains the OH spectra in Figure 2b. In the case of the mixed STA-DOMA monolayer, there are negative sites belonging to some of the deprotonated STA headgroups and positive ones that are due to the DOMA headgroups. As a whole, the effect on interfacial water is of a monolayer, which is believed to be overall

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15509

Figure 2. Sum-frequency spectra at the air-water interface of model lipid monolayers on H2O subphase in the C-H and O-H stretching regions: (a) STA:DOMA ) 2:1 and (b) pure STA. For more information, see text.

electrically neutral. Consequently, the water molecules close to the Langmuir film cannot assume ordered structure and produce no SF signal. Monolayer Structure from Analysis of the CH Stretching Band. In a closely packed surfactant monolayer at the air/water interface, the hydrocarbon tails are in all-trans conformation and the SF spectra consist of only two bands at ca. 2876 cm-1 and ca. 2940 cm-1 belonging to the end CH3 groups and corresponding to the C-H symmetric stretching vibration and to the C-H symmetric stretching in Fermi resonance with a bending overtone, respectively.27-29 The methylene group vibrations of the alkyl chain will only appear in the SF spectra in ssp polarization because of a break of the inversion symmetry existing around each of the C-C bond in all-trans conformation. This happens when there are some gauche defects in the acyl chains, and thus SFS is able to detect conformational changes in the lipid monolayers. The fully compressed state of DOMA Langmuir films in the C-H stretching region is characterized with only two peaks, Figure 1a. A shoulder at ca. 2840 cm-1 and the asymmetry at the low-frequency side of the 2940 cm-1 peak are an indication that there are also unresolved peaks belonging to the methylene group vibrations suggesting that there are some gauche defects in the alkyl chains.27-29 Following previous studies,30,31 we can assign the shoulder at ca. 2840 cm-1 to the symmetric methylene vibration and the asymmetry at ca. 2920 cm-1 to its Fermi resonance with a CH bending overtone. This appearance of the spectra makes us believe that the compressed DOMA monolayer is highly ordered, with alkyl chains in close to all-trans conformation with some gauche defects. In the analysis, so far we have neglected the contribution of the headgroup methyls of DOMA to the spectra. We believe that these two CH3 groups are symmetrically positioned with respect to the surface normal. Their dipole moments corresponding to the symmetric CH stretches are almost perpendicular to the interface, pointing in the opposite direction to the end

15510 J. Phys. Chem. B, Vol. 110, No. 31, 2006 methyl group dipoles and are thus reducing the SF intensity of the latter in ssp polarization combination. Therefore, the shape and appearance of the SF CH stretching spectra can be attributed to the alkyl chain structure only. Because of symmetry considerations, we also believe that the CH3 asymmetric stretching vibrations at ca. 2958 cm-1 do not appear in the SF spectra as their dipole moments have mutually opposite directions and thus their contributions cancel each other. The most interesting change in the CH stretching region of the mixed SME:DOMA monolayers (Figure 1b-d) in comparison with the spectrum of the pure DOMA monolayer (Figure 1a) is that the shoulders around 2840 and 2920 cm-1 disappear and the spectra are dominated by the more symmetrical now methyl bands at 2876 and 2940 cm-1. This behavior is in agreement with what we suggested earlier in our studies about more efficient packing structure in the SME-DOMA mixtures.2 This better film packing seems to be a result of the smaller SME molecules (∼20 Å2/molec) filling in the interstitial spaces between the larger DOMA molecules (∼60 Å2/molec) and confirms our previous compressibility data.2 With the increase in the ratio SME:DOMA, a new component in the CH stretching spectra appears around 2955 cm-1. This component can be assigned to the methyl asymmetric CH3 stretching vibration in the SME headgroup, but it has a lower frequency than the normally observed one at 2958-2962 cm-1.30 With the increase in the SME concentration, its intensity increases, becoming almost equal to the CH3 symmetrical Fermi resonance at 2940 cm-1 at mole ratio 10:1, Figure 1d. Finally, it dominates the SF spectrum of the pure SME monolayer, Figure 1e. Unfortunately, the appearance of the spectra in this region can be greatly influenced by an interference effect arising from the coherent nature of the SF generation process, and this seems to be the case in the range 2925-2975 cm-1, Figure 1. We observe an interesting “derivative-like” appearance around 2950 cm-1, and it is clearly present when the monolayers contain DOMA and have well-expressed water OH stretching band (see Figure 1). As the different resonant and nonresonant features in SFS have in general different phases (eq 4), when they overlap or are close enough, they can interfere constructively or destructively and give rise to a more complicated spectral pattern.19 Most probably, the reason for the spectral shape around 2950 cm-1 in Figure 1 is interference between the C-H stretching modes and the wing of the O-H stretching mode at ca. 3200-3400 cm-1. To prove this, we performed the same experiments with a subphase of D2O. Figure 3a-e shows the CH stretching SF spectra of pure STA, pure DOMA, mixed SME:DOMA, and pure SME monolayers, respectively, on a D2O subphase. Fitting of the spectra using eqs 3 and 5 was performed to analyze more precisely the observed components. The symmetric stretch methyl group components at 2876 and 2940 cm-1 have dominant presence in the spectra. The fit in the region between these two major bands can be perfect as in Figure 3b and 3c only when a negative component (180° out-of-phase) at ca. 2920 cm-1 is added. This fact and the wavenumber proximity of this component to the Fermi resonance CH2 symmetric stretch is under review and will be discussed in a separate study. If the fit is done without this component, as in Figure 3a, 3d, and 3e, the middle region is not very well approximated. Overall, both fits serve very well the purposes of the present analysis, and we continue discussing the main features of the pure monolayer spectra on D2O. For the pure STA and DOMA monolayers, there is one more weak component at ca. 2840 cm-1, assigned to the methylene symmetric stretching vibration, but it was not included in the

Nickolov et al. fit for STA because it is too weak. The presence of this component proves the existence of some small gauche imperfections in the alkyl chains of these monolayers on the D2O subphase. Another very strong component at 2958 cm-1 arises in SME containing monolayers, Figure 3c-e. It belongs to the asymmetric CH3 vibration, and its intensity grows significantly with the increase of the SME fraction in the mixed SME:DOMA monolayer, dominating the spectra for pure SME, Figure 3e. Comparison of the spectra in Figure 1 with those in Figure 3 clearly demonstrates that the shape of the CH spectra of the mixed monolayers on H2O subphase is strongly influenced by interference with the wing of the water OH band. When this influence is eliminated, as in the case of mixed and pure monolayers on D2O subphase (Figure 3), the CH stretching bands become well-defined and can be easily fit using the theoretical equations for SFS. The strong presence of the methyl asymmetric CH stretch in the SF spectra of SME:DOMA mixed monolayers and of pure SME monolayers needs an explanation from the point of view of SFS. When only the end methyl group vibrations in a highly ordered monolayer such as STA, Figure 3a, are recorded in ssp geometry, the asymmetric methyl stretch around 2958 cm-1 is absent from the spectrum in agreement with SFS selection rules because its dipole moment is oriented perpendicular to the IR laser polarization. Its presence in the spectra of SME containing monolayers, Figure 3c-e, can be explained with the fact that there is also an uncompensated dipole moment of the methyl group in the headgroup of SME. If the methyl dipole moment orientation changes from almost perpendicular to the monolayer surface to parallel to the interface with the increase in the SME fraction in the mixed films and transition to pure SME monolayer, the asymmetric methyl stretch will appear in the spectra. This is exactly what we observe in the spectra in Figure 3c-e. It was calculated from our Langmuir isotherms that the headgroup area of this SME configuration does not change too much and still remains around 20 Å2/molec, which does not contradict the assumption of a better packed structure of the mixed films compared to that of pure DOMA. Correlation of SFS Results with Surface Pressure and Surface Potential Measurements. The degree of lipid miscibility (dictating size scale of heterogeneity) in multicomponent lipid films is generally assessed from surface pressure (π) and surface potential (∆V) isotherms as well as from visual analysis of the films using some form of microscopy. In Langmuir monolayers, lipid miscibility is commonly evaluated in terms of deviation of film properties (“excess properties”) from an ideal (additive) mixing scenario.32 The most common parameter evaluated is ∆Amix, the difference between the measured average area per lipid molecule in the mixture and the ideal average of the pure components:

∆Amix(π) ) A12 (π) - (χ1A1(π) + χ2A2(π))

(6)

In this expression, A12 is the average lipid molecular area measured for the mixture at a given surface pressure (π), and A1 and A2 are the molecular areas of the pure components at the same surface pressure. Multiplying A1 and A2 by their respective mole fractions in the mixture, χ1 and χ2, yields their expected contributions to the net area on the basis of ideal mixing. Additional parameters such as surface potential (∆V) and effective dipole moment µ ) o∆VA, where o is the permittivity of the vacuum, can also be evaluated in the same manner to yield ∆(∆Vmix) and ∆µmix.33 An advantage of the effective dipole moment analysis is that it effectively normalizes

Sum-Frequency Spectroscopy Analysis

Figure 3. Sum-frequency spectra at the air-water interface of model lipid monolayers on D2O subphase in the C-H stretching region: (a) Pure STA; (b) pure DOMA; (c) SME:DOMA ) 2:1; (d) SME:DOMA ) 10:1; and (e) pure SME. Solid line indicates nonlinear curve fit to the experimental spectral points (hollow circles). Dashed lines indicate wavenumbers of the vibrational modes discussed in the text. From left: 2876, 2940, and 2958 cm-1.

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15511 the monolayer with respect to lipid dipoles per unit area accounting for mixing induced area changes. The π-A and corresponding ∆V-A isotherms for SME, DOMA, and various mixtures are presented in Figure 4.34 From the isotherms, it was concluded that the mixtures collapse at higher surface pressures (stronger films) and surface potentials (amplified charge) than the pure components. These enhanced properties may arise in part from a more efficient packing structure in the mixture, where smaller SME molecules (∼20 Å2/molec) fill in the interstitial spaces between the larger DOMA molecules (∼60 Å2/molec). SF spectra of DOMA, SME, and their mixtures in the CH stretching region, Figures 1a-1e and 3b-3e, support this conclusion. Perfectly ordered alkyl chains in all-trans conformation display only two methyl vibrational bands at 2876 and 2940 cm-1 in ssp geometry29 which is observed for the spectra in Figure 3. In addition, a band at 2860 cm-1 assigned to the asymmetric stretching vibration of the SME headgroup methyl is observed for the mixed monolayers and for SME. While in the pure DOMA spectrum, there are some gauche defects in the alkyl chains as proved by the presence of a weak methylene symmetric stretch component at ca. 2840 cm-1; the spectra of the mixed monolayers and SME display only the three bands mentioned above, and this is an indication that the alkyl chains are almost perfectly aligned parallel to each other in all-trans conformation and are closely packed, being more orthogonal to the interface than pure DOMA. This is a confirmation of the conclusion from our π-A and ∆V-A isotherm study, that the smaller size of the headgroup of SME is responsible for the formation of a well-packed, more orthogonal monolayer in the mixtures. The miscibility diagrams constructed from the isotherms for the mixtures studied here, Figure 5, clearly illustrate a mixing induced amplification in ∆V and film contraction.34 Moreover, ∆µmix follows the same trend as ∆(∆Vmix), indicating that an increase in lipid dipole density (∆Amix < 0) cannot explain the excess positive character of the mixtures since µ normalizes ∆V in terms of area. The analysis in these studies suggested that it appears that changes in headgroup or interfacial water dipoles must be the dominating factors behind the surface potential amplification and corresponding enhanced protein adsorption observed in Figure 5. The interfacial water structure deduced from the SF spectra in the OH region, Figure 1, helps us give a preference to the water dipole factor in the ∆V amplification. Indeed, with the increase in SME fraction in the mixtures, the water OH band in the SF spectra decreases in intensity and completely disappears for the pure SME monolayer/water interface. This demonstrates a transition from ordered, more structured water, with a particular orientation of the water molecules (dipoles) and high value of µ, to a disordered, bulklike water. As the net orientation of the interfacial water dipole moment is opposite to the end methyl group dipole moment for DOMA and mixed monolayers, the total dipole moment is a difference between these two. As the interfacial water below the mixed monolayers becomes more disordered with the SME mole fraction increase, the water dipole moment decreases. This, in combination with the increase in the end methyl dipole moment because of better packing and more orthogonal chain arrangement, leads to a total increase of the dipole moment for the mixed films, bigger than the arithmetic average for DOMA and SME monolayers alone. Thus, interfacial water structure proves to be a determining factor in the behavior of the mixed lipid monolayers.

15512 J. Phys. Chem. B, Vol. 110, No. 31, 2006

Nickolov et al.

Figure 4. Surface pressure and surface potential isotherms for the indicated mixtures of SME:DOMA on water.

The results from SFS analysis correlate very well with surface pressure-area and surface potential-area measurements and help obtain a better understanding of the interactions between the molecules of the mixed lipids and of the conditions necessary to design mixed monolayers with enhanced properties for effective protein adsorption. The high sensitivity of SFS to interfaces of lipid monomolecular films with water and the capabilities to perform in-situ nondestructive measurements offer significant, often unique, advantages over other surface-sensitive techniques used in similar studies before. More SFS studies at the lipid (protein) monolayer-water interface are needed to explore the full potential of the method for bionanotechnology and medical applications. Acknowledgment. This research was partially supported by the NSF-IMR Program (Grant DMR-0216904) and by the NSFNER Program (Grant BES 0404262 and Grant CTS 0210879). References and Notes

Figure 5. Mixing diagrams illustrating the surface potential and effective dipole moment amplification (top, bottom) and film contraction (middle) as SME is mixed with DOMA. The relative adsorption trend (measured by reflectivity at normal incidence) of ferritin (0.01 mg/mL in subphase) to the monolayers at 30 mN/m is presented for comparison (crosses, top).

5. Conclusions Sum-frequency spectroscopy proves to be a very successful method to study the structure and packing of mixed lipid monolayers on water as well as the structure of interfacial water.

(1) Goodman, T.; Bussmann, E.; Williams, C.; Taveras, M.; Britt, D. Langmuir 2005, 20, 3684. (2) Demchak, R. J.; Fort, T. J. J. Colloid Interface Sci. 1975, 46, 191. (3) Britt, D. W.; Mo¨bius, D.; Hlady, V. Phys. Chem. Chem. Phys. 2000, 2, 4594. (4) Binder, H. Vib. Spectrosc. 1999, 21, 151. (5) Elmore, D. L.; Shanmukh, S.; Dluhy, R. A. J. Phys. Chem. A 2002, 106, 3420. (6) Als-Nielsen, J.; Jacquemain, D.; Kjaer, K.; Leveiller, F.; Lahav, M.; Leiserowitz, L. Phys. Rep. 1994, 246, 251. (7) Thoma, M.; Schwendler, M.; Baltes, H.; Helm, C. A.; Pfohl, T.; Riegler, H.; Mo¨hwald, H. Langmuir 1996, 12, 1722. (8) Losche, M. Curr. Top. Membr. 2002, 52, 117. (9) Miranda, P. B.; Shen, Y. R. J. Phys. Chem. B 1999, 103, 3292. (10) Richmond, G. L. Chem. ReV. 2002, 102, 2693. (11) Gopalakrishnan, S.; Liu, D. F.; Allen, H. C.; Kuo, M.; Shultz, M. J. Chem. ReV. 2006, 106, 1155. (12) Gurau, M. C.; Kim, G.; Lim, S. M.; Albertorio, F.; Fleisher, H. C.; Cremer, P. S. Phys. Chem. Chem. Phys. 2003, 4, 1231. (13) Conboy, J. C.; Messmer, M. C.; Richmond, G. L. J. Phys. Chem. B 1997, 101, 6724. (14) Wolfrum, K.; Lobau, J.; Birkholzer, W.; Laubereau, A. Quantum Semiclassical Opt. 1997, 9, 257. (15) McKenna, C. E.; Knock, M. M.; Bain, C. D. Curr. Opin. Colloid Interface Sci. 2001, 6, 313.

Sum-Frequency Spectroscopy Analysis (16) Gibum, K.; Gurau, M. C.; Lim, S. M.; Cremer, P. S. J. Phys. Chem. B 2003, 107, 1403. (17) Bell, G. R.; Li, Z. X.; Bain, C. D.; Fischer, P.; Duffy, D. C. J. Phys. Chem. B 1998, 102, 9461. (18) Watry, M. R.; Brown, M. G.; Richmond, G. L. Appl. Spectrosc. 2001, 55, 321. (19) Moore, F. G.; Becraft, K. A.; Richmond, G. L. Appl. Spectrosc. 2002, 56, 1575. (20) Shen, Y. R. Annu. ReV. Phys. Chem. 1989, 40, 327. (21) Shen, Y. R. The Principles of Nonlinear Optics; Wiley-Interscience: New York, 1984. (22) Shen, Y. R. Surf. Sci. 1994, 299/300, 551. (23) Nickolov, Z. S.; Wang, X.; Miller, J. D. Spectrochim. Acta, Part A 2004, 60, 2711. (24) Nickolov, Z. S.; Miller, J. D. Sum Frequency Spectroscopy ReVeals New Possibilities to Study Interfaces, SME Annual Meeting Preprints, Feb 23-25, 2004, Denver, CO; pp 699-704.

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15513 (25) Buch, V.; Devlin, J. P. J. Chem. Phys. 1999, 110, 3437. (26) Nickolov, Z. S.; Miller, J. D.; Britt, D., in preparation. (27) Shen, Y. R. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 12104. (28) Richmond, G. L. Anal. Chem. 1997, 69, 536A. (29) Bain, C. D. Curr. Opin. Colloid Interface Sci. 1998, 3, 287. (30) MacPhail, R. A.; Strauss, H. L.; Snyder, R. G.; Elliger, C. A. J. Phys. Chem. 1984, 88, 334. (31) Guyot-Sionnest, P.; Hunt, J. H.; Shen, Y. R. Phys. ReV. Lett. 1987, 59, 1597. (32) Gaines, G. L. Insoluble Monolayers at Liquid-Gas Interfaces; Wiley-Interscience: New York, 1966. (33) Britt, D. W.; Jogikalmath, G.; Hlady, V. In Protein Interactions with Monolayers at the Air/Water Interface, Biopolymers at Interfaces, 2nd ed.; Malmsten, M., Ed.; Marcel Dekker: New York, 2003; pp 415-434. (34) Britt, D. W.; Goodman, T.; Selle, C. Materialwiss. Werkstofftech. 2003, 34, 1133.