J. Phys. Chem. B 2001, 105, 2197-2204
2197
Total Internal Reflection Fluorescence and Electrocapillary Investigations of Adsorption at the Water-Dichloroethane Electrochemical Interface. 2. Fluorescence-Detected Linear Dichroism Investigation of Adsorption-Driven Reorientation of Di-N-butylaminonaphthylethenylpyridiniumpropylsulfonate M. A. Jones and P. W. Bohn* Department of Chemistry, Materials Research Laboratory, and Beckman Institute for AdVanced Science and Technology, UniVersity of Illinois at Urbana-Champaign, 600 South Mathews AVenue, Urbana, Illinois 61801 ReceiVed: July 1, 2000; In Final Form: January 5, 2001
The potential-dependent adsorption and orientation of the zwitterionic amphiphile, di-N-butylaminonaphthylethenylpyridiniumpropylsulfonate, I, in the presence of dilauroylphosphatidylcholine (DLPC) at the H2O1,2-dichloroethane (DCE) interface was investigated using a combination of steady-state fluorescence, fluorescence-detected linear dichroism, and electrocapillary measurements. From electrocapillary measurements, DLPC was found to dominate the interfacial composition at all potentials, when DLPC and I were present in bulk DCE and H2O at ca. 2 and 1 µM concentrations, respectively. At potentials Ew - Eo > 0.32 V, the affinity of DLPC for the interface is diminished, and I becomes a more effective competitor for interfacial (o,w) is sites. Over the potential range 0.32 V e Ew - Eo e 0.47 V, the total interfacial excess of species ΓDLPC+I (o,w) (o,w) reduced, but the ratio ΓI /ΓDLPC is enhanced. The DC fluorescence signal increased in response to the enhanced interfacial population of the unprotonated monomeric I at positive potentials. Concurrent fluorescencedetected linear dichroism (FDLD) measurements found that I reorients toward the interface normal on the positive scan. Because the total interfacial excess decreases at these potentials, this behavior cannot be ascribed to a simple compression effect. Rather, it reflects the favored geometry at the compositions obtained at positive potentials. Potential step experiments showed phenomena occurring on three distinct time scales: ion reorganization on the millisecond time scale, an initial excursion of the DC fluorescence intensity over a few tens of seconds, and then a much longer evolution of the DC fluorescence (opposing the initial change) and the FDLD signal. The initial DC response can be explained by an interfacial reorganization of the chromophore that occurs in response to the applied potential before significant mass-transport occurs, while the slow time responses of both the IDC and the FDLD signal are attributed to a mass-transport-limited partitioning of the lipid species, into and out of the interfacial region.
Introduction The understanding of thin-film systems whose order parameters can be tuned promises a class of materials with welldefined and highly useful mechanical, optical, and electrical properties.1 Alignment of liquid-crystalline materials in external electric fields provides a classic example of how such systems can find useful applications as optical filters, as displays, and for data storage.2 In many instances, however, a stand-alone material that retains its molecular order in the absence of external perturbation is desired. Toward this end, processing strategies for liquid crystalline materials have been developed that allow the electric-field-induced alignment to be locked into materials by encapsulating the liquid crystal in a polymerizable matrix that holds it in the desired orientation once the electric field has been removed.2 A complementary approach for the development of such materials is to design order into a system from the bottom-up; this requires understanding the nature of the molecular interactions that dictate structure. Much of the work along these lines has focused on the self-organization of long-chain aliphatic * To whom correspondence should be addressed. E-mail: bohn@scs. uiuc.edu.
molecules into complex systems possessing order on the supramolecular length scale.3-8 The fundamental issues, key to exploiting the structure-function relationships in order to produce mesoscale ordering in these systems, were identified by Buscher as (1) surface coverage, monolayers vs aggregates; (2) mesoscale structure, ordered vs random; and (3) the intrinsic relationship between molecular orientation and surface coverage.9 The relationship between interfacial coverage, or number density Γ, and molecular orientation, as expressed by one of several order parameters, has been explored for a variety of systems. Buscher9 et al. investigated the assembly of polar selfassembled monolayers at the quartz-air interface. Using the second harmonic generation (SHG), they found that the average molecular orientation was a complex function of coverage as it progressed from submonolayer (large tilt angle, θ, of the molecular dipole relative to the surface normal), through the monolayer regime (factor of 2 decrease in tilt angle), to randomized aggregate structures. Neivandt10 et al. performed an attenuated total reflection (ATR) IR study to determine the orientation of a cationic surfactant adsorbed to the SiO2-H2O interface. They found a reduction in the orientation angle at higher coverage, and through desorption studies, they found that
10.1021/jp002370f CCC: $20.00 © 2001 American Chemical Society Published on Web 02/15/2001
2198 J. Phys. Chem. B, Vol. 105, No. 11, 2001 the change in surface coverage outran the molecular reorientation, an observation they attributed to the persistence of small ordered patches whose orientational order decreased over 30 h. Looking at amphiphilic organic ions on silica, Kung and Hayes11 studied wetting-dewetting transitions; they highlighted the fact that wet and dry films may not produce the same coverageorientation relationship due to environmentally driven aggregation phenomena. A much smaller number of studies have addressed the coverage-orientation relationship at liquid-liquid interfaces. Morrison and Weber12 were among the first to report total internal reflection fluorescence (TIRF) measurements of dye adsorption onto the interface between two immiscible solutions. These authors used a combination of fluorescence intensity, emission polarization, and emission spectra to probe the state of adsorbed dyes vs. coverage at the decalin/water interface. Takenaka13-16 et al. made a series of resonance Raman measurements to probe changes in the orientation of charged surfactants at the CCl4-H2O interface as a function of interfacial pressure. These systems were found to be almost freely oriented in the monolayer at low interfacial pressure but to subtend a definite angle with the interface at higher interfacial pressure. Higgins and Naujok17,18 used SHG to probe the molecular orientation of pure and mixed films of two structurally similar chromophores whose dipole moments are oriented ∼180° relative to one another when adsorbed to the H2O-DCE interface. Measurements of the molecular orientation in singlecomponent monolayers were reported, and an ∼180° angle between the dipole moments was demonstrated by observation of an SHG interference effect in mixed films. Wirth and Burbage19 studied the in-plane and out-of-plane reorientation of acridine orange at several liquid alkane-water interfaces. They found measurable differences in the out-of-plane orientation distribution among the several environments probed, the differences were attributed to surface roughness. In the present study, fluorescence-detected linear dichroism (FDLD) was employed to probe adsorption-driven reorientation of di-N-butylaminonaphthylethenylpyridiniumpropylsulfonate, I, in dilauroylphosphatidylcholine (DLPC) films at the H2O1,2-dichloroethane (DCE) interface. Previously we reported on the competitive adsorption of I and DLPC at this interface.20
An imposed potential difference between the aqueous and organic phases in the range 0.32 e Ew - Eo e 0.47 V influences the interfacial-bulk equilibrium, changing both the total number density of interfacial amphiphilic species, i.e., ΓDLPC+I, and the composition of the two-component film in a deterministic fashion. Based on the literature cited above, these changes in composition might be expected to influence the orientational order within the interfacial film and, thus, connect adsorption and orientational processes yielding new insights into the physicochemical dynamics of these important interfaces. Experimental Section Materials. Deionized (Millipore) H2O and 1,2-dichloroethane (Fisher, spectral grade, 99.9%, passed through a basic alumina column) were used as the ITIES solvents, with 100 mM LiCl and 1 mM tetra-n-butylammonium tetraphenylborate (TBATPB) (both Fluka puriss grade) supporting electrolytes, respectively.
Jones and Bohn The aqueous and organic phases were shaken together in a separatory funnel and allowed to equilibrate for at least 1 h prior to beginning an experiment. A reference interface between the aqueous 25 mM tetra-n-butylammonium chloride (TBACl) (Fluka, puriss) and the organic phase was used to sense the organic phase potential. For measurements involving added surfactant, dye I (Molecular Probes) was present in the aqueous phase, at ca. 1 µM. Based on the adsorption isotherm at open circuit potential, an excess of DLPC (Sigma) was added to the DCE phase at ca. 2 µM. Once assembled, the surfactant films were allowed to equilibrate for at least 2 h prior to beginning an experiment. Potentiometry. A Princeton Applied Research 273 potentiostat was configured, as described by Vanysek,21 for fourelectrode operation.22 Pt counter electrodes and Ag/AgCl reference electrodes were used in the aqueous and organic phases, respectively. The organic phase reference electrode was placed in a Luggin capillary which contained the TBACl (H2O)-DCE reference interface and was positioned within 1 mm of the H2O-DCE interface to minimize the potential drop between the interface and the organic reference sensing point. The electrochemical cell was as follows: Ag|AgCl|25 mM TBACl (H2O)|1 mM TBATPB (DCE)|100 mM LiCl (H2O)|AgCl|Ag. The interfacial potential is reported as the measured difference between the aqueous and organic phases, Ew - Eo. Tensiometry. The interfacial pressure of the H2O-DCE interface was monitored using the Wilhelmy plate method, employing a NIMA pressure sensor and a PTFE plate.23 Changes in interfacial tension were recorded as a function of potential, which was then referenced to an absolute scale by making a detachment measurement at a known potential. Electrocapillary measurements were made using a Plexiglas trough. Data collection was accomplished with Labview software. All experiments were performed on a pneumatic optical table in a Faraday cage at room temperature. All glassware was thoroughly cleaned in freshly prepared piranha:4:1 (v:v) H2SO4 (concentrated):H2O2 (30%) (Caution: Pirhana is extremely hazardous, handle with great care!), and was rinsed with copious amounts of deionized H2O prior to use. Fluorescence. For fluorescence measurements, the H2O/DCE interface was contained in a fused silica cuvette (25 × 25 mm, Spectrocell). The unfocused 457.9 nm line of an Ar+ laser (Coherent, Innova 300) was polarized with a Glan-laser prism prior to impinging on a photoelastic modulator (PEM-80, Hinds) with fast axis oriented at 45° to the incident polarization. The PEM was set for half-wavelength retardation at maximum displacement.24 The modulated radiation was attenuated (PI ∼ 60 µW) and totally internally reflected from the H2O/DCE interface at an incident angle of 73.6° (θc ∼ 67.4°). Fluorescence from the interface was gathered above the aqueous phase at f/2 efficiency, passed through a 500 nm long-pass filter, and then directed onto a cooled (-25 °C) photomultiplier tube (RCA C31034A). The excitation radiation was mechanically chopped (Stanford Research, SR540) at 1 kHz, and the resulting modulated PMT current was monitored by a lock-in amplifier (Stanford Research, SR530) and reported as IDC, the steadystate fluorescence level. A second lock-in amplifier (Stanford Research, SR530) was used to monitor the polarization signal, I| - I⊥, at 100 kHz. Data collection was accomplished on a microcomputer with Labview (National Instruments) software. Dichroism. Linear dichroism measurements are widely used to determine the average orientation of chromophores. In the total internal reflection (TIR) configuration, the magnitude of the evanescent field decays exponentially with the distance from
Investigations at the H2O-DCE Electrochemical Interface
J. Phys. Chem. B, Vol. 105, No. 11, 2001 2199
I| - I⊥ kpI0(e-A1Φ- e-A2Φ) ) IDC k I e-AΦ
(4)
p 0
where kp is a photomultiplier tube (PMT) constant, I0 is the incident light intensity, φ is the fluorescence quantum efficiency, A is the mean absorbance, and A1 and A2 are the absorbances at the extreme polarization states. Results
Figure 1. Normalized linear dichroic response as a function of orientation, θ, calculated from eq 3. The dashed line represents the I⊥ ) I| condition, which corresponds to θ* ) 29.10ο under these particular experimental conditions. The proportionality constant C ) 13.31 was determined by comparing the FDLD response and an independent dichroic ratio measurement.19
the DCE-H2O interface, and thus, samples adsorbed chromophores present at the interface. The evanescent field amplitudes are determined by the incident angle, β, the ratio of refractive indices, n21, and the dielectric properties of the adsorbed film. Approximating the interface as a two-layer system with the film dielectric properties set equal to that of the water phase, these field amplitudes were calculated along the three Cartesian coordinates according to25
Ex )
2(sin2 β - n221)1/2cos β (1 - n221)1/2[(1 + n221)sin2 β - n221]1/2 2cos β (1 - n221)1/2
(1b)
2sin β cos β (1 - n221)1/2[(1 + n221)sin2 β - n221]1/2
(1c)
Ey ) Ez )
(1a)
The fluorescence signal is given by
If ) κ|〈m|µ‚E|k〉|2
(2)
where 〈m| and 〈k| are the states involved in the transition, µ is the absorption transition moment, E is the electric field vector, and κ is a constant that includes the fluorescence quantum efficiency. Employing the Euler angle conventions, the intensity of fluorescence excited by the absorption of light polarized along the p and s directions was determined26 and an expression for the normalized linear dichroism was found
I| - I⊥ 1 ) C E2z Q|(θ)cos2 (θ) + (E2x - E2y)Q⊥(θ)sin2 (θ) IDC 2 (3)
[
]
where θ represents the angle between the transition moment vector (assumed to be parallel to the molecular long axis) and the z-axis (interface normal), Q⊥,|(θ) are collection efficiency terms27 that account for the f/2 collection angle in the emission optical path, and C is a constant that collects the proportionality and fluorescence quantum efficiency factors. The calculated signal response when θ was incremented through 90° is shown in Figure 1. Experimentally, the FDLD signal was obtained by taking the ratio of the polarization modulated signal I| - I⊥ to the steady-state fluorescence level IDC,28
Potential-Dependent Adsorption. The adsorption of surfaceactive species to the H2O-1,2-DCE interface was probed by monitoring the interfacial tension γ. To the zeroth order, the interfacial tension is determined by the identities of the two liquids and by the accumulation of species in the interfacial region. An applied potential redistributes these species, thus influencing γ. The potential dependence of the surface excess of species i, Γ(o,w) , was determined by measuring the electroi capillary response (γ vs E) for a surfactant concentration series and by applying the Gibbs equation.20,29 Figure 2 shows how the surface excess of DLPC and I at the H2O-1,2-DCE interface changes as a function of potential when DLPC is present at ca. 2 µM in 1,2-DCE and I is present at ca. 1 µM in H2O. DLPC was found to dominate the interfacial composition; at maximum coverage the DLPC occupied a mean molecular area of 80 Å2/ molecule, which is comparable to the 73 Å2/molecule reported at the nitrobenzene-H2O interface where the phospholipid was reported to be in the liquid expanded phase.30 During the 0.1 mV/s positive potential scan, there is an exchange of species over the range 0.32 e Ew - Eo e 0.47 V. Figure 3 shows how this exchange influences the total interfacial excess of lipid (o,w) (o,w) species, ΓDLPC+I , and the ratio of species, Γ(o,w) /ΓDLPC , as a I function of potential. The total number of lipid species at the interface decreases roughly by a factor of 2 during the positive scan; this response is dominated by the desorption of protonated H-DLPC+ from the interface. Alternatively, if the lipid composition rather than the total number density of species is considered, the ratio of I to DLPC present at the interface was found to increase by a factor of 26. Low-Frequency Fluorescence Response. Figure 4 shows the FDLD response to a low-frequency potential sweep (trace 4) at 0.1 mV/s applied to the H2O-1,2-DCE interface, in the presence of both DLPC at ca. 2 µM in DCE and I at 1 µM in H2O. The interfacial population of unprotonated monomeric I is reported by the fluorescence signal shown as trace 1 in Figure 4.31 It is worth noting that given the concentration of I in H2O and the sampling depth of the evanescent field at the H2ODCE interface, the ratio of interfacial dye molecules sampled to solution molecules sampled is ca. 103. This signal is averaged over all polarization states of the excitation light and corresponds to the steady-state fluorescence level, IDC. The increase in steady-state fluorescence at applied potentials Ew - Eo > 0.325 V suggests that unprotonated monomer I is partitioning to the interface in this potential regime. The offset between the apex of the potential scan and the maximum fluorescence response supports the idea that this is a mass-transport-limited phenomenon, in which I is continually transported to the interface at all times during which Ew - Eo > 0.325 V. Electrocapillary measurements reported previously20 support the notion that I-derived species are partitioning to the interface at the positive end of the potential range; however, simple mechanical measurements cannot distinguish between unprotonated monomeric I and other forms of I, e.g., aggregates or protonated species, HI+. However, since both protonation and aggregation are
2200 J. Phys. Chem. B, Vol. 105, No. 11, 2001
Figure 2. Comparison of the potential-dependent interfacial excess of a surfactant, I or DLPC, when the other surfactant is also present in bulk solution at nominal concentrations of 1 µM I in H2O and 2 µM DLPC in 1,2-DCE. The curves were calculated by measuring the electrocapillary response for a concentration series of the surfactant of interest, while holding the other surfactant composition constant and applying the Gibbs equation. The electrocapillary data were obtained from a positive-going potential sweep at 0.1 mV/s.
Figure 3. Comparison of the change in total interfacial excess of lipid (o,w) species, ΓDLPC+I , and the relative interfacial composition of lipid (o,w) (o,w) species, ΓI /ΓDLPC, as a function of potential.
known to shift the electronic spectroscopy of stilbazolium amphiphiles,32 the fluorescence measurements used here are a much better reporter for the interfacial population of unprotonated monomeric I. Trace 2 in Figure 4 shows the absolute value of the fluorescence-detected linear dichroism (FDLD) signal obtained by measuring the difference in fluorescence intensity when the sample is excited by light polarized along two orthogonal linear polarization states, || and ⊥, where the directions are referenced to the incident plane of the excitation radiation. In addition, the || direction is also parallel to the applied potential (perpendicular to the interface). As indicated in eq 2, this signal has been ratioed to the steady-state fluorescence level, IDC, to normalize for changes in the interfacial population of I. Clearly the dichroic difference represented by trace 2 undergoes substantial and repeatable changes as a function of applied interfacial potential. The direction of reorientation was extracted by monitoring the phase angle, R, between the FDLD signal and the drive voltage of the PEM (corrected for autophasing offsets in the lock-in amplifier). The PEM drive voltage and FDLD signals both report time-dependent events, the former giving the polarization state of the excitation radiation and the latter the intensity of fluorescence excited. In Figure 4, the phase angle between these two signals, R (trace 3), shifts by ∼180° at a specific repeatable point during the potential sweep, indicating a reversal in the sign of the dichroic difference signal displayed in trace 2, i.e., the molecular transition moment has rotated
Jones and Bohn through the angle where I⊥ ) I|; θ* ) 29.10° for the experimental parameters employed. A polarizer (set to || or ⊥) was inserted between the PEM and the sample to ascertain the phase angles corresponding to || and ⊥ predominant responses. On the positive sweep the FDLD signal initially was mostly determined by I⊥, then passed through I⊥ ) I| (minimum FDLD signal), and ended up being mostly determined by I|. Extracting more specific information about the degree of probe orientation requires calibrating the dichroic difference signal. The FDLD response (dichroic difference) was first calibrated against a dichroic ratio measurement to determine the value of C in eq 3. To determine molecular orientation from a linear dichroism experiment, the direction of the transition moment must be related to the molecular frame. For a good approximation, the transition moment of I corresponding to its lowest excited singlet is parallel to the long axis of the π-system.33 Assuming that the excitation and emission transition moments are parallel,33 and that there was little movement of the chromophore during its excited-state lifetime (