Heterogeneous Translational Dynamics of Rhodamine B in

pubs.acs.org/Langmuir © 2009 American Chemical Society

Heterogeneous Translational Dynamics of Rhodamine B in Polyelectrolyte Multilayer Thin Films Observed by Single Molecule Microscopy Joshua K. Carr, Ryan D. Himes, Connie H. Keung, Daniel L. Burden, and Peter K. Walhout* Department of Chemistry, Wheaton College, Wheaton, Illinois 60187 Received February 16, 2009. Revised Manuscript Received March 26, 2009 The lateral diffusion dynamics of rhodamine B (RB) in polyelectrolyte multilayer (PEM) thin films has been studied with single-molecule confocal fluorescence microscopy. The films were made with sodium poly(sodium 4-styrenesulfonate) (PSS) and poly(diallydimethlyammonium chloride) (PDDA). Analysis of the real-time emission intensity traces reveals three diverse components of translational motion: (1) fast diffusion of RB through the confocal detection volume; (2) reversible tracer adsorption processes; and (3) nanoconfined diffusion. These processes cover a wide range of time scales. Analysis via fluorescence correlation spectroscopy (FCS) involves multicomponent fitting of the autocorrelated emission data. The model includes a free Brownian diffusion parameter, D, and two rate constants of desorption, k-1 and k-2. For RB in a PSS/PDDA thin film made with 0.01 M NaCl in the polyelectrolyte buildup solutions, D=1.7
10-7 cm2/s, k-1=30 s-1, and k-2=0.1 s-1. FCS was also performed on RB/PEM samples made with NaCl concentrations of the buildup solutions ranging from 0.01 to 0.7 M. A weak dependence of D and k-1 on NaCl concentration was observed while k-2 increased linearly with [NaCl].

1. Introduction Polyelectrolyte multilayer (PEM) thin films and the more general layer-by-layer method of constructing thin films and microcapsules have attracted a great deal of interest from a wide variety of scientific and engineering disciplines in the past decade.1,2 Much focus lately has been on the use of these nanoengineered materials for drug delivery3-6 and the controlled release of other entrapped molecules.7-14 Other applications have focused on the ability to coat electrodes with PEM layers which either on their own or by virtue of doping or other functionalization favorably modify the electrochemistry of interest.15-18 In both of these areas of application, it is important to understand the interaction of smaller molecules with the PEM. Specifically, the diffusion of smaller molecules through the PEM is central to the functionality of many controlled release and electrochemical applications. *To whom correspondence should be addressed. E-mail: peter.k. [email protected]. (1) Decher, G.; Schlenoff, J. T. Multilayer Thin Films-Sequential Assembly of Nanocomposite Materials; Wiley-VCH: Weinheim, 2002. (2) Hammond, P. Adv. Mater. 2004, 16, 1271–1293. (3) Qiu, X.; Leporatti, S.; Donath, E.; M€ohwald, H. Langmuir 2001, 17, 5375– 5380. (4) Wood, K. C.; Chuang, H. F.; Batten, R. D.; Lynn, D. M.; Hammond, P. T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10207–10212. (5) Macdonald, M.; Rodriguez, N. M.; Smith, R.; Hammond, P. T. J. Controlled Release 2008, 131, 228–234. (6) Wang, Z.; Qian, L.; Wang, X.; Yang, F.; Yang, X. Colloids Surf., A 2008, 326, 29–36. (7) Sukhorukov, G. B.; Fery, A.; Brumen, M.; M€ohwald, H. Phys. Chem. Chem. Phys. 2004, 6, 4078. (8) Hiller, J. A.; Rubner, M. F. Macromolecules 2003, 36, 4078–4083. (9) Itoh, Y.; Matsusaki, M.; Kida, T.; Akashi, M. Chem. Lett. 2008, 37, 238–239. (10) Guyomard, A.; Nysten, B.; Muller, G.; Glinel, K. Langmuir 2006, 22, 2281– 2287. (11) Tedeschi, C.; Caruso, F.; M€ohwald, H.; Kirstein, S. J. Am. Chem. Soc. 2000, 122, 5841–5848. (12) Chung, A. J.; Rubner, M. F. Langmuir 2002, 18, 176–183. (13) Burke, S. E.; Barrett, C. J. Macromolecules 2004, 37, 5375–5384. (14) Wang, B.; Gao, C.; Liu, L. J. Phys. Chem. B 2005, 109, 4887–4892. (15) Farhat, T. R.; Schlenoff, J. B. J. Am. Chem. Soc. 2003, 125, 4627–4636. (16) Lutkenhaus, J. L.; Hammond, P. T. Soft Matter 2007, 3, 804–816. (17) Zhao, J.; Bradbury, C. R.; Fermin, D. J. J. Phys. Chem. C 2008, 112, 6832– 6841. (18) Song, M.; Ge, L.; Wang, X. J. Electroanal. Chem. 2008, 617, 149–156.

8330 DOI: 10.1021/la900573w

The engineering of PEM nanofiltration membranes also relies on understanding small molecule diffusion and permeation.19-21 Additionally, many other applications utilize the incorporation of small molecules and would benefit from a thorough understanding of their subsequent diffusive dynamics within a PEM thin film. Freely incorporated dye molecules have been explored for potential use in various optical applications owing to their high extinction coefficients and polarizability,22-27 and various sensors noncovalently incorporate key functional components such as dyes,28 enzymes,29 and nanocrystals.30,31 There are two significant issues related to the use of freely incorporated (noncovalently attached) molecules within PEM structures: the general characterization of their bulk uptake and release, and the specific nature of their immobilization and/or diffusion once incorporated into the PEM. Several studies have explored the former issue by studying dye uptake into a variety of PEM thin films and its subsequent release back into solution.10-14,32 This paper studies the latter issue of small molecule dynamics within a PEM by utilizing single-molecule (19) Tieke, B.; van Ackern, F.; Krasemann, L.; Toutianoush, A. Eur. Phys. J. E 2001, 5, 29–39. (20) Miller, M. D.; Bruening, M. L. Chem. Mater. 2005, 17, 5375–5381. (21) Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351–357. (22) Kometani, N.; Nakajima, H.; Asami, K.; Yonezawa, Y.; Kajimoto, O. J. Phys. Chem. B 2000, 104, 9630–9637. (23) He, J.-A.; Bian, S.; Li, L.; Kumar, J.; Tripathy, S. K.; Samuelson, L. A. J. Phys. Chem. B 2000, 104, 10513–10521. (24) Kaschak, D. M.; Lean, J. T.; Waraksa, C. C.; Saupe, G. B.; Usami, H.; Mallouk, T. E. J. Am. Chem. Soc. 1999, 121, 3435–3445. (25) Dai, Z.; D€ahne, L.; Donath, E.; M€ohwald, H. J. Phys. Chem. B 2002, 106, 11501–11508. (26) Tedeschi, C.; Li, L.; M€ohwald, H.; Spitz, C.; von Seggern, D.; Menzel, R.; Kirstein, S. J. Am. Chem. Soc. 2004, 126, 3218–3227. (27) Heflin, J. R.; Guzy, M. T.; Neyman, P. J.; Gaskins, K. J.; Brands, C.; Wang, Z.; Gibson, H. W.; Davis, R. M.; Van Cott, K. E. Langmuir 2006, 22, 5723–5727. (28) Dubas, S. T.; Iamsamai, C.; Potiyaraj, P. Sens. Actuators, B. 2006, 113, 370–375. (29) Davis, F.; Higson, S. P. J. Biosens. Bioelectron. 2005, 21, 1–20. (30) Komarala, V. K.; Rakovich, Y. P.; Bradley, A. L.; Byrne, S. J.; Corr, S. A.; Gun’ko, Y. K. Nanotechnology 2006, 17, 4117–4122. (31) Maheshwari, V.; Saraf, R. F. Science 2006, 312, 1501–1504. (32) Linford, M. R.; Auch, M.; M€ohwald, H. J. Am. Chem. Soc. 1998, 120, 178– 182.

Published on Web 06/08/2009

Langmuir 2009, 25(14), 8330–8339

Carr et al.

real-time fluorescence tracking and fluorescence correlation spectroscopy (FCS) to gain new insight into dye diffusion and immobilization dynamics. Several previous diffusion studies have been made of various small molecules in a variety of PEM structures, but a comprehensive understanding of diffusion has yet to emerge. Klitzing and M€ohwald used a F€ orster resonance energy transfer (FRET) technique to measure a diffusion coefficient for rhodamine B (RB) in a PEM thin film made from poly(allylammine hydrogen chloride) (PAH) and poly(sodium 4-styrenesulfonate) (PSS).33 Two different values were obtained: 1
10-14 cm2/s in the upper layers of the thin film and 2
10-16 cm2/s in the bulk of the film. Faster diffusion in the upper layers is due to their less dense, more mobile, and more hydrated state relative to the middle bulk layers.34-36 Other reported diffusion coefficients for small molecules, however, are strongly dependent on the polyelectrolyte used in the PEM, the degree of hydration, and the nature of the PEM structure. Estimates of the diffusion coefficients of similarly sized molecules at room temperature range from 10-16 to 10-8 cm2/s. The values reported herein for RB are even larger, on the order of 10-7 cm2/s. The wide range of possible polyelectrolyte mobilities and levels of hydration likely leads to a wide range of diffusive mechanisms for small molecules. Whether the diffusion of small molecules within a given PEM proceeds through water pockets and channels, or via hopping from one charge-compensating polyelectrolyte locale to another, or through plasticized polyelectrolyte segmental motion, or some combination of the three, remains a rather open question. One of the most studied PEM systems is the poly(diallyldimethylammonium chloride) (PDDA) polycation layered with the PSS polyanion. This system is insensitive to pH as both “strong” polyelectrolytes remain fully charged (except in highly acidic conditions when the sulfonate group is protonated). This is in sharp contrast to a PEM made of PAH and poly(acrylic acid) (PAA) which are both “weak”; the internal architecture of a PAA/ PAH PEM is extraordinarily sensitive to the pH of the polyelectrolyte buildup solutions and the pH of any subsequent solution to which the PEM is exposed.37 While the charge density of the polyelectrolytes constituting strong PEM thin films cannot be varied by pH, the presence of additional salt in the buildup solutions38,39 or in subsequent salt “doping” solutions can have a profound impact on the PEM properties. Schlenoff and coworkers have shown for PSS/PDDA thin films that by immersing them into various salt solutions many of their properties can be tuned, including the amount of hydration,40,41 the viscoelastic response,42 and the flux of an analyte to an underlying electrode.15,43 The primary impact of salt in the buildup solutions, on the other hand, is on the thickness increment added to the PEM with each additional bilayer.39 Salt increases the amount of polyelectrolyte added per layer during buildup by increasing the amount of “extrinsic” charge compensation (charge compensation by (33) Klitzing, R. v.; M€ohwald, H. Macromolecules 1996, 29, 6901–6906. (34) Ladam, G.; Schaad, P.; Voegel, J. C.; Schaaf, P.; Decher, G.; Cuisinier, F. Langmuir 2000, 16, 1249–1255. (35) McCormick, M.; Smith, R. N.; Graf, R.; Barrett, C. J.; Reven, L.; Spiess, H. W. Macromolecules 2003, 36, 3616–3625. (36) Smith, R. N.; Reven, L.; Barrett, C. J. Macromolecules 2003, 36, 1876–1881. (37) Shiratori, S. S.; Rubner, M. F. Macromolecules 2000, 33, 4213–4219. (38) Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153–8160. (39) Schlenoff, J. B.; Dubas, S. T. Macromolecules 2001, 34, 592–598. (40) Jaber, J. A.; Schlenoff, J. B. Langmuir 2007, 23, 896–901. (41) Schlenoff, J. B.; Rmaile, A. H.; Bucur, C. B. J. Am. Chem. Soc. 2008, 130, 13589–13597. (42) Jaber, J. A.; Schlenoff, J. B. J. Am. Chem. Soc. 2006, 128, 2940–2947. (43) Rmaile, H. H.; Farhat, T. R.; Schlenoff, J. B. J. Phys. Chem. B 2003, 107, 14401–14406.

Langmuir 2009, 25(14), 8330–8339

Article

small counterions) of the surface layer. Upon addition of a new layer, the higher extrinsic compensation of the surface requires more added polyelectrolyte to displace the salt and provide “intrinsic” (polyanion-polycation) charge compensation. Upon completion of the PEM, only the surface layer retains evidence of the salt concentration used during buildup, as the bulk layers are nearly completely intrinsically compensated, while the final surface layer retains the higher level of extrinsic compensation due to the salt in the buildup solutions.39 This paper reports on the translational self-diffusion of RB within various PSS/PDDA PEM thin films and is believed to be the first significant report measuring lateral diffusion in PEM films rather than diffusion normal to the plane of the film. PSS/ PDDA films were chosen because they have been well-characterized and their indifference to pH removes variations in the polyelectrolyte charge densities as a complicating parameter. The RB dynamics are measured with FCS and real-time emission intensity traces using a confocal fluorescence microscope. To our knowledge, this is the first characterization of PEM transport dynamics at the level of individual molecules.44 These singlemolecule techniques provide unprecedented insight into the dynamic heterogeneity that accompanies small molecule diffusion within a PEM. Very large diffusion coefficients are measured for RB, consistent with a prior preliminary report.45 The real-time analysis of the intensity data reveals discrete adsorption events whereby an RB molecule is apparently immobilized, or at least confined to a region smaller than the spatial sensitivity of the confocal technique. The diffusion and adsorption events were also studied as a function of salt concentration in polyelectrolyte buildup solutions. While RB diffusion has only a weak dependence on salt concentration, the rate constants for desorption increase significantly. This result is consistent with the notion that the observed dye resides solely in the surface layers of the PEM, but it does not rule out the possibility that the observed dynamics arise from dye sequestered near the substrate/PEM interface or from dye diffusing freely in water pockets or defects present throughout the film.

2. Experimental Section A. Sample Preparation. Poly(sodium 4-styrenesulfonate)

(PSS, Mw ≈ 70 kDa), poly(diallyldimethylammonium chloride) (PDDA, Mw ≈ 100-200 kDa), rhodamine B (RB) (99% pure), and rhodamine 6G (99% pure) were obtained from SigmaAldrich and used as received (Figure 1). Polyethylenimine (PEI) was obtained from Wako. NaCl (reagent grade) was obtained from Fluka. Dialysis of PSS and PDDA was not required because the identity of the counterions present with the polymers and of those added to the buildup solutions (see below) were the same. Prior to film deposition, precleaned glass microslides (VWR) were immersed for 1 h in “piranha” solution (3:1 concentrated sulfuric acid to 30% hydrogen peroxide) followed by extensive rinsing with deionized water. All slides had an initial layer of PEI deposited by immersing them in 10 mM PEI for 15 min, followed by two 1 min rinses in reagent grade 18 MΩ water (NERL Diagnostics). All polyelectrolyte concentrations of buildup solutions are given in terms of the repeat unit concentration. Polyelectrolyte multilayer (PEM) films were prepared via layerby-layer deposition using an automated slide stainer (Zeiss HMS). The polymer buildup solutions contained 2 mM PSS and PDDA, with a variable amount of added NaCl. For any sample, the (44) There has been at least one other single molecule study of dyes in PEM films, but this was strictly an examination of rotational diffusion in a 1-2 layer film: Li, Y.; Yip, W. T. Langmuir 2004, 20, 11039–11045. (45) Burden, D. L.; Walhout, P. K.; Elliott, J. T.; Chandler, E. L.; Scharf, R. G.; Culbertson, M. J.; Stults, D. A.; Rupp, E. L.; Poppen, S. D. Spectrosc. Lett. 2004, 37, 129–149.

DOI: 10.1021/la900573w

8331

Article

Carr et al.

Figure 1. Structures of polyelectrolyes, rhodamine B zwitterion, and rhodamine 6G. amounts of NaCl added to the PSS and PDDA solutions were the same. Each dip cycle consisted of a 5 min dip in PSS, followed by a series of four salt-free water rinses (15, 15, 60, and 15 s.), followed by a 5 min PDDA rinse, followed by four water rinses as before. For all samples studied, the dip cycle was repeated 20 times, yielding a 20-bilayer film on top of the initial one layer of PEI. PDDA was always the outermost layer. The slides typically were not annealed in a high concentration salt solution,46 but a brief study was made of one annealed slide (see the Supporting Information). RB was added to the PEMs by immersing them for 1 min in aqueous solutions of 70-200 nM RB, followed by two 1 min water rinses. RB is predominantly a zwitterion in ambient water with no net charge. It has a complex series of equilibria related to acid-base chemistry and aggregation, but in dilute neutral solution the situation is greatly simplified. The carboxylic acid group on RB has a pKa of 3.22, which means that, at a pH of 5.2 which is typical for ambient water, 99% of the RB is in the zwitterionic form.47,48 Dipping times were kept constant, but films made with more salt had a greater tendency to load the RB. In order to keep confocal fluorescence signals within a single-molecule window, more strongly loading films were immersed in more dilute RB solutions. Dipping the slides for several minutes longer than 1 min did not cause a noticeable increase in the average fluorescence intensity of the films. Generally, completed PEMs were clear, though some cloudiness was evident at higher NaCl concentrations (>0.5 M). The thickness of several PEM thin films with different buildup salt concentrations were measured with a phase-modulated ellipsometer (Beaglehole) by fitting reflectivity data to standard models. A 20-bilayer PSS/PDDA film assembled with 0.03 M NaCl in the polyelectrolyte buildup solutions on a silicon wafer (Virginia Semiconductor) with one PEI precursor layer had a thickness of 46 nm. The increase in PEM thickness with salt concentration was roughly linear, consistent with previous results.38

B. Confocal Microscope and Fluorescence Correlation Spectroscopy. Data was collected using a home-built confocal microscope. The microscope employs the chassis of an Axiotech Vario microscope (Carl Zeiss) with a 150
, 1.25 NA epiplanapochromat water-immersion objective. Water immersion is accomplished by putting a droplet of water directly on the PEM sample and immersing the microscope objective in the droplet. This water droplet then necessarily hydrates the underlying PEM thin film. The 543 nm line of an unpolarized 2 mW helium-neon laser (HGR020, Thorlabs) is directed through an optical density filter (held at OD=0.8) into a fiber optic cable via a fiber optic launch (HFB003, Thorlabs). The fiber output is directed through the objective and onto the sample, which rests on a 3D piezoelectric stage (17MAX301, Melles Griot) to allow for precise positioning of the sample. For typical trials, the laser light had a power of 26.0-29.8 μW after exiting the objective. The laser beam is focused on the sample so as to create a diffraction-limited Gaussian detection volume. The fluorescence from the sample that reenters the objective passes through a dichroic filter to (46) Dubas, S. T.; Schlenoff, J. B. Langmuir 2001, 17, 7725–7727. (47) Ramette, R. W.; Sandell, E. B. J. Am. Chem. Soc. 1956, 78, 4872–4878. (48) Mchedlov-Petrosyan, N. O.; Kholin, Y. V. Russ. J. Appl. Chem. 2004, 77, 414–422. While up to 1% of the RB is present in the cationic form, there is no reason to believe that the cation is preferentially loaded into the PEM (see ref 43).

8332 DOI: 10.1021/la900573w

remove backscattered laser light and then is focused through a 100 μm (diameter) pinhole onto an avalanche photodiode (APD) (SPCM-AQR-15, Perkin-Elmer). The signal from the APD is then passed to a hardware autocorrelator (5000/60
0 Multiple Tau Digital Correlator, ALV), which passes both an intensity trace (count rate versus time) and its autocorrelation function to the computer. The autocorrelator generates the autocorrelation function from the intensity trace via the following function: Z GðτÞ ¼ lim

T f¥

T -T

yðtÞ yðt -τÞ dt

ð1Þ

where t is time (from the intensity trace), τ is correlation time, y(t) is the number of counts at time t, and the limits of (T indicate integration over the entire trace. For this experiment, τ was varied discretely between 10-5 and 31 s. Data was typically collected for 120 s. Real-time analysis was performed by sending the output from the APD to a multichannel scalar (Perkin-Elmer, MCSPlus). Integration times (bin widths) were varied to facilitate direct analysis of the real-time data which revealed several event types, the duration of which varied over 4 orders of magnitude. Replicate measurements were performed at six separate sample locations which were randomly spaced by ∼7 μm. The physical dimensions of the focused laser were determined by calibration using RB and its published diffusion constant in free solution. From the observed free solution crossing time, the standard 3D analytical autocorrelation model was employed to extract values for radius and axial dimensions of the detection profile.49 For 2D surface measurements, the parameter of greatest importance is the detection volume radius (ω), which was determined to be 265 nm. C. Power Dependence. To determine the effects of optical excitation power on the measured kinetic parameters, autocorrelations were collected at four optical powers from a single PEM sample made with 0.1 M NaCl. Lower powers required longer collection times to acquire reasonable signal-to-noise for the autocorrelations. For routine FCS measurements, the chosen optical powers of 26.0-29.8 μW (12-14 kW/cm2) represent a compromise between the desire to minimize RB photobleaching and the need to achieve adequate signal-to-noise for a reason ably short data collection time. Longer collection times were not desirable due to the introduction of additional error from occasional slow drift in the mechanical stages and optics. D. PEM Hydration Dependence. To determine the effects of film hydration, data was also collected using a 100
/1.25 N.A. oil-immersion objective (Edmund Industrial Optics). In this setup, the PEM is not directly hydrated with a water droplet; instead, a coverslip is suspended above the film on a thin platform (see Figure 2). This creates a film-air interface which greatly reduces the amount of water in the PEM. Use of a coverslip with the oilimmersion objective is typical in FCS to avoid the sample coming into contact with the oil. Data was collected using an optical power of 114 μW (measured exiting the objective) and a collection time of 240 s. A higher optical power and longer collection times were needed due to a less tightly focused beam profile at the sample. The size of the detection volume was measured by raster (49) Rigler, R.; Mets, U; Widengren, J.; Kask, P. Eur. Biophys. J. 1993, 22, 169– 175.

Langmuir 2009, 25(14), 8330–8339

Carr et al.

Article

Figure 2. Schematic drawing of the experimental setups using the water-immersion and oil-immersion microscope objectives. Relative sizes are not drawn to scale. The height of the detection volume is ∼1 μm. The air gap was ∼0.5 mm. scanning using a 170 nm diameter fluorescent bead and resulted in a Gaussian profile with a 1/e2 radius of 825 nm.

3. Results A. Real-Time Monitoring Experiments. Figure 3 shows individual events acquired in real-time monitoring experiments with a variety of integration times ((A) 1 ms/bin; (B) 5 ms/bin; (C) 20 ms/bin). (An “event” refers to a period of RB fluorescence, the time scale of which indicates RB diffusion or immobilization within the confocal detection volume). Generally, events begin upon entry of a dye into the detection region and are terminated upon exit. Events that span 3-4 orders of magnitude in time can be identified. Background emission from the slide surface is relatively high; however, individual events are statistically distinguishable above the background noise. The red line demarks the 3 standard deviation (SD) threshold above the average background count rate. Count rates exceeding this threshold are distinguishable from background fluctuations at the 99% confidence level and can be attributed to RB emission. Shot noise during the events arises from standard photon statistics. In addition, brief photophysical dark states of RB and alterations in quantum efficiency, similar to that exhibited by other rhodamine dyes,50 potentially add further fluctuation to events arising from individual dye molecules. Single-molecule detection events from various data sets have been artificially spliced together in Figure 3 to enable comparison and do not represent the actual acquisition frequency. The actual event acquisition frequency varied depending upon the sample evaluated but generally ranged from 1 to 10 events/s. While the type of event observed is independent of the integration time, Figure 3 groups similar events with the integration time that is most suitable for their acquisition and display. Four categories of events are observed in the real-time data: (a) brief transitory events, ranging in time from 1 to 10 ms (A1-A4); (b) adsorption events of short (20 ms) duration (A5); (c) adsorption events of long (100-500 ms) duration (B1-B3, C1); and (d) long (20-1500 ms) events that are indicative of diffusive microconfinement (B4, C2). The events displayed here represent a range of observations and are not intended to imply the actual relative occurrence frequency of the different event types. Samples constructed under higher added NaCl tended to produce a greater number of events that shared features indicative of diffusive confinement. Adsorption events are characterized by a rectangular appearance. (The term “adsorption” is used here to refer to a transitory immobilization of the dye within the PEM and should be distinguished from the general loading of the dye into the PEM from solution during synthesis of the sample). This rectangular appearance can be caused by the rapid movement of a dye molecule into the confocal observation zone, followed by reversible immobilization. Single-molecule events with a similar (50) Nie, S.; Chiu, D. T.; Zare, R. N. Anal. Chem. 1995, 67, 2849–2857.

Langmuir 2009, 25(14), 8330–8339

Figure 3. Examples of real-time monitoring of RB fluorescence in 20-bilayer PSS/PDDA thin film. Photons detected within a given bin width are plotted versus the time of the data collection. Horizontal line demarks the 3 standard deviation level above the average background noise. Events have been compiled from different data sets and are not necessarily from the same spot on the sample.

appearance have been reported by other investigators in systems where adsorption is known to take place.51,52 Similar dye adsorption has also been reported in PEM thin films15 and polyelectrolye solutions53 on the macroscopic level. In order to characterize general trends in single-molecule behavior via FCS, an analytic model that is appropriate for describing simultaneous diffusion and adsorption processes is required. The general situation of a diffusing fluorophore (51) Wirth, M. J.; Swinton, D. J. Appl. Spectrosc. 2001, 55, 1013–1017. (52) Zhong, Z.; Lowry, M.; Wang, G.; Geng, L. Anal. Chem. 2005, 77, 2303– 2310. (53) Caruso, F.; Donath, E.; M€ohwald, H.; Georgieva, R. Macromolecules 1998, 31, 7365–7377.

DOI: 10.1021/la900573w

8333

Article

Carr et al.

involved in a dynamic equilibrium first described by Elsen and Magde has been adapted by Wirth et al.54 for the case of rare adsorption of the fluorophore: k1

A þBaC k -1

K ¼

k1 ½C ¼ k -1 ½A½B

ð2Þ

ð3Þ

where A is the adsorption site, which is nonfluorescing and nondiffusing, B is the diffusing fluorophore with diffusion constant D, and C is the adsorbed species, which fluoresces but does not diffuse. The model was developed under the following assumptions: (1) K[B],1; (2) D/ω2 .k1[A]; and (3) k-1 .k1[A]. Condition (1) is easily satisfied by the conditions of singlemolecule spectroscopy where [B] is inherently low. Condition (2) is satisfied when strong adsorption is much less likely than a step of time, which is shown to be true for our samples. Condition (3) is satisfied when the rate of adsorption is less than the rate of desorption. Evaluation of the real-time record for a range of PEM salt preparations indicates that k1[A]=0.1 s-1 and k1[A]=10 s-1 for low (0.01 M) and high (0.7 M) salt, respectively. Estimates of the off rate constants generally varied from k-1=0.2 s-1 at low salt (0.01 M) to k-1 = 50 s-1 at high salt (0.7 M). k1[A] was estimated by counting the number of adsorption events in a given time period and dividing by that period. k-1 and k-2 are derived from the real-time data by assessing typical duration times of the adsorption events. For each salt concentration, the assumption of condition (3) is generally met. Thus, all the kinetic assumptions implicit in the derivation of the strong adsorption model are reasonably approximated in our RB/PEM samples. However, the data in Figure 3 also suggest the presence of confined motion (B4, C2), which stands in distinct contrast to an immobilizing adsorption process. The duration of B4 and C2 is long, but the signal clearly changes in a continuous manner, suggesting that dye molecules are executing a slow transition into and out of the regions of more intense laser illumination. Shot noise and photophysical dark states (which typically occur on the order of microseconds) cannot account for the rolling intensity changes on this longer time scale. The pattern could potentially be produced by solvated dye that becomes temporarily corralled inside aqueous-filled cavities in the thin film structure. Such cavities containing confined water in thin 10 layer PSS/PAH films have recently been studied by NMR.55 It is also possible that the shorter peaks observed on top of event C2 arise from cooccupancy of the detection volume by a second or third dye molecule. In general, however, the features of events B4 and C2 are consistent with a dye molecule moving slowly in a nonBrownian, or confined, fashion across a significant portion of the Gaussian confocal detection profile, even with the possible presence of additional molecules in the detection volume. Events exhibiting these features occur more frequently in PEM films constructed with higher salt concentrations. The range and relative continuity of the intensity changes suggests that the size of the putative cavities is on the order of several hundred nanometers. (54) Wirth, M. J.; Ludes, M. D.; Swinton, D. J. Appl. Spectrosc. 2001, 55, 663– 669. (55) Chavez, F. V.; Sch€onhoff, M. J. Chem. Phys. 2007, 126, 104705. The cavities studied by NMR were determined to be only 1 nm in width for PSS/PAH, but the cryoporometry technique used is not as sensitive to cavity sizes on the order of 102 nm.

8334 DOI: 10.1021/la900573w

It is possible that confined motion could also be giving rise to the rectangularly shaped events that are ascribed to adsorption. Given the size and Gaussian profile of the confocal detection area (d=530 nm), along with the S/N levels obtained in our measurements, dyes need only be confined to a spatial region of less than 102 nm to maintain an emission intensity that stays constant to within the shot noise limit of the signal. Events A5, B1, B2, B3, and C1 maintain the appearance of physically adsorbed molecules. However, given the experimental S/N levels, the possibility that dyes remain mobile but are trapped within small (