From Exponential to Linear Growth in Polyelectrolyte Multilayers

Yi Zou , Lingxiao Xie , Sean Carroll , Maria Muniz , Heather Gibson , Wei-Zen Wei , Haipeng Liu , and ..... Mara L. Macdonald , Natalia M. Rodriguez ,...
0 downloads 0 Views 159KB Size
Download PDF
4376

Langmuir 2006, 22, 4376-4383

From Exponential to Linear Growth in Polyelectrolyte Multilayers Claudine Porcel,†,‡ Philippe Lavalle,‡ Vincent Ball,‡ Gero Decher,† Bernard Senger,‡ Jean-Claude Voegel,‡ and Pierre Schaaf*,† Institut Charles Sadron, Centre National de la Recherche Scientifique, UniVersite´ Louis Pasteur, 6 rue Boussingault, 67083 Strasbourg Cedex, France, and Institut National de la Sante´ et de la Recherche Me´ dicale, Unite´ 595, Faculte´ de Chirurgie Dentaire, UniVersite´ Louis Pasteur, 11 rue Humann, 67085 Strasbourg Cedex, France ReceiVed NoVember 29, 2005. In Final Form: January 10, 2006 There exist two types of polyelectrolyte multilayers: those whose thickness increases linearly with the number of deposition steps, which are nicely structured, and those whose thickness increases exponentially, which resembles hydrogels. This simple picture has recently slightly evolved with the finding that some exponentially growing films enter into a linear growth phase after a certain number of deposition steps. In this study, we investigate the buildup process of hyaluronic acid/poly(L-lysine) (HA/PLL) multilayers that constitute one of the best known exponentially growing systems. The films are built by using two deposition methods: the well-known dipping method and the more recent spraying method where the polyelectrolyte solutions are sprayed alternately onto a vertical substrate. The goal of this study is twofold. First, we investigate the influence of the main parameters (i.e., spraying rate and spraying time) of the spraying method on the film growth process. We find that, as for the dipping method, the film thickness first evolves exponentially with the number of deposition steps, and after a given number of deposition steps, it follows a linear evolution. We find that similar behavior is observed with the dipping method. Second, because the spraying method allows the very fine variation of the different parameters of the buildup, we use this method to investigate the exponential-to-linear transition. We find that this transition always takes place after about 12 deposition steps whatever the values of the parameters controlling the deposition process. We discuss our results in light of a model proposed by Hu¨bsch et al. (Hu¨bsch, E.; Ball, V.; Senger, B.; Decher, G.; Voegel, J. C.; Schaaf, P. Langmuir 2004, 20, 1980-1985) and later by Saloma¨ki et al. (Saloma¨ki, M.; Vinokurov, I. A.; Kankare, J. Langmuir 2005, 21, 11232-11240) in which it is assumed that the exponential-to-linear transition is due to a film restructuring that progressively forbids the diffusion of one of the polyelectrolytes constituting the film over part of the film. This “forbidden” zone then grows with the number of deposition steps so that the outer zone of the film that is still concerned with diffusion keeps a constant thickness and moves upward as the total film thickness increases.

Introduction The alternate deposition of polyanions and polycations on solid surfaces leads to the formation of films called polyelectrolyte multilayers (PEM).1 These films have received considerable attention during the past decade because of their widespread potential applications in fields as different as photodiodes,2,3 optical devices,4 filtration membranes,5,6 self-supported membranes with highly enhanced Young’s moduli,7,8 fuel cell membranes,9 biologically active membranes,10 drug release,11,12 or biologically active coatings.13-16 Until recently, two types of buildup processes have been reported: (i) the one in which the thickness and mass of the film increase linearly with the number of deposited pairs of layers1 and (ii) the process leading to films that grow exponentially with the number of deposition steps.17-20 Poly(styrene sulfonate)/ * Corresponding author. E-mail: [email protected]. Phone: +33(0)388-414012. Fax: +33-(0)388-414099. † Institut Charles Sadron, Universite ´ Louis Pasteur. ‡ Institut National de la Sante ´ et de la Recherche Me´dicale, Universite´ Louis Pasteur. (1) Decher, G. Science 1997, 277, 1232-1237. (2) Fou, A. C.; Onitsuka, O.; Ferreira, M.; Rubner, M. F.; Hsieh, B. R. J. Appl. Phys. 1996, 79, 7501-7509. (3) Eckle, M.; Decher, G. Nano Lett. 2001, 1, 45-49. (4) Hiller, J.; Mendelsohn, J. D.; Rubner, M. F. Nat. Mater. 2002, 1, 59-63. (5) Liu, X. Y.; Bruening, M. L. Chem. Mater. 2004, 16, 351-357. (6) Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003, 19, 7038-7042. (7) Mamedov, A.; Kotov, N. A.; Prato, M.; Guldi, D. M.; Wicksted, J. P.; Hirsch, A. Nat. Mater. 2002, 1, 190-194. (8) Tang, A.; Kotov, N. A.; Magonov, S.; Ozturk, B. Nat. Mater. 2003, 2, 413-418.

poly(allylamine hydrochloride)21 (i.e., PSS/PAH) and poly(styrene sulfonate)/poly(diallyldimethylammonium) (i.e., PSS/ PDADMA) architectures built at room temperature and at moderate salt concentration (e.g., 0.15 M NaCl) constitute two prominent examples of linearly growing films. During each layer deposition step, the polyelectrolytes (PE) from the solution are electrostatically attracted by the oppositely charged PE constituting the last layer of the film that has already been deposited on the solid substrate. This PE deposition leads to charge over(9) Farhat, T.; Hammond, P. T. AdV. Funct. Mater. 2005, 15, 945-954. (10) Onda, M.; Lvov, Y.; Ariga, K.; Kunitake, T. J. Ferment. Bioeng. 1996, 82, 502-506. (11) Izumrudov, V. A.; Kharlampieva, E.; Sukhishvili, S. A. Biomacromolecules 2005, 6, 1782-1788. (12) Vazquez, E.; Dewitt, D. M.; Hammond, P. T.; Lynn, D. M. J. Am. Chem. Soc. 2002, 124, 13992-13993. (13) Chluba, J.; Voegel, J. C.; Decher, G.; Erbacher, P.; Schaaf, P.; Ogier, J. Biomacromolecules 2001, 2, 800-805. (14) Jessel, N.; Atalar, F.; Lavalle, P.; Mutterer, J.; Decher, G.; Schaaf, P.; Voegel, J.-C.; Ogier, J. AdV. Mater. 2003, 15, 692-695. (15) Benkirane-Jessel, N.; Lavalle, P.; Hu¨bsch, E.; Holl, V.; Senger, B.; Haı¨kel, Y.; Voegel, J.-C.; Ogier, J.; Schaaf, P. AdV. Funct. Mater. 2005, 15, 648-654. (16) Thierry, B.; Kujawa, P.; Tkaczyk, C.; Winnik, F. M.; Bilodeau, L.; Tabrizian, M. J. Am. Chem. Soc. 2005, 127, 1626-1627. (17) Picart, C.; Lavalle, P.; Hubert, P.; Cuisinier, F. J. G.; Decher, G.; Schaaf, P.; Voegel, J.-C. Langmuir 2001, 17, 7414-7424. (18) Kujawa, P.; Moraille, P.; Sanchez, J.; Badia, A.; Winnik, F. M. J. Am. Chem. Soc. 2005, 127, 9224-9234. (19) Elbert, D. L.; Herbert, C. B.; Hubbell, J. A. Langmuir 1999, 15, 53555362. (20) Picart, C.; Mutterer, J.; Richert, L.; Luo, Y.; Prestwich, G. D.; Schaaf, P.; Voegel, J.-C.; Lavalle, P. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1253112535. (21) Ladam, G.; Schaad, P.; Voegel, J.-C.; Schaaf, P.; Decher, G.; Cuisinier, F. J. G. Langmuir 2000, 16, 1249-1255.

10.1021/la053218d CCC: $33.50 © 2006 American Chemical Society Published on Web 03/23/2006

Exponential/Linear Growth

Langmuir, Vol. 22, No. 9, 2006 4377

compensation at the interface, which in turn causes electrostatic repulsion and limits the PE adsorption to only one additional monolayer.21,22 It was shown that linearly growing PEM films exhibit a somewhat “fuzzy” but layered structure, with each PE layer interpenetrating only with the neighboring ones.23,24 The thickness per layer typically lies between 1 and 10 nm depending on parameters such as the ionic strength of the PE solutions, their pH (for weak PE), and, to some extent, the molecular weight of the PE. The second type of buildup process, in which the mass and thickness of the PEM grow exponentially with the number of deposition steps, was reported more recently. Hyaluronic acid/ poly(L-lysine)17,20 (HA/PLL) and hyaluronic acid/chitosan18,25 constitute two examples of exponentially growing films. At present, the explanation of their buildup process is based on the diffusion “in” and “out” of the whole film during each bilayer deposition step of at least one of the PEs constituting the film. In the case of HA/PLL, it was found that during each deposition step HA interacts only with the upper layer of the film, whereas PLL diffuses in and out of the entire film structure.20 More precisely, consider a film after the deposition of HA. The HA solution is then replaced by the buffer solution, and the film bears an outer negative excess charge. When this film is brought into contact with the PLL solution, the PLL chains first interact with the outer HA layer as would be the case for a linearly growing PEM film. In addition, PLL chains diffuse into the film, down to the substrate. The film then consists of HA and PLL chains interacting strongly with each other and of free PLL chains that are expected to interact only weakly with the film matrix. It must be noticed that an exchange process between free and bound PLL chains may take place in the film so that the free chains at a given time are not necessarily those that have diffused from the solution into the film at an earlier stage of the construction. When the PLL solution is then replaced by a pure buffer solution, some of the free PLL chains present in the film diffuse out of it with the consequence of decreasing the chemical potential inside the film and increasing the height of the electrostatic barrier to be overcome by the PLL chains remaining in the film.26 Therefore, part of the stored free PLL chains cannot escape from the film. When the film is then brought in contact with the HA solution again, this electrostatic barrier disappears because of neutralization by the adsorbing polyanions followed by the formation of a negative excess charge on top of the PEM film. All the remaining free PLL chains then diffuse out of the film. As soon as they reach the film/solution interface, they are complexed by HA chains and are expected to remain bound to the interface. These HA/PLL complexes form the new outer layer of the film. The HA deposition process stops either when the film contains no free PLL chains or when the exposure to the HA solution is interrupted by a new rinsing phase. The structure of these films is quite different from that of linearly growing ones. In particular, exponentially growing films behave as polyanion/polycation gels as was confirmed by Collin et al.27 Exponentially growing films received more and more attention because of their potential applications in various domains such as the field of biomaterials.

However, this mere classification of PEM films into two categories seems to be an oversimplification. Indeed, it has already been observed that even a PSS/PAH film may grow exponentially when the PE solutions are prepared in the presence of a high salt concentration.28 The same observation has been made for the PSS/PDADMA films.29 Furthermore, Michel et al. reported that the evolution of the thickness of the exponentially growing film poly(L-glutamic acid)/poly(allylamine) (PGA/PAH) becomes linear after the deposition of about 10 pairs of layers.30 This film was constructed by using the spraying method. Hu¨bsch et al.31 observed a similar behavior but by constructing films from a polyanion solution composed of mixtures of polyanions and a single polycation solution. More recently, Saloma¨ki et al.32 investigated the influence of the effect of temperature on the PEM film buildup. These authors found that while the thickness of PSS/PAH and PSS/PDADMA films increases linearly with the number of deposition steps at room temperature, their buildup process becomes exponential when the temperature is raised. This effect is also salt-dependent. In addition, they found that as the number of deposition steps becomes large the exponential growth turns into a linear increase, as observed by Michel et al.30 Following the assumption of Ladam et al.,21 Saloma¨ki et al.32 also proposed that the films may be subdivided in three main domains (Figure 1). Domain I corresponds to the initial buildup phase and is in contact with the substrate, domain III is in contact with the solution, and domain II corresponds to the internal part of the film that is surrounded by domains I and III. These authors assume that at least one of the PEs can diffuse into the entire domain III during each deposition step. They further assume that the composition of domain III is a function of the polycation and polyanion concentrations of the buildup solutions only. During the exponential growth phase, the thickness of domain III increases. However, when this zone becomes too thick, diffusion is no longer able to transfer material everywhere into this zone when the time allowed for diffusion is kept constant. From this point on, domain II appears and develops. This scenario is nothing else than the consequence of the diffusion process: in a given time lapse, the diffusing species can reach only a finite depth in the film. Conversely, in a finite time lapse (equal to the duration of the exposure to HA), the useful PLL chains are extracted from a finite zone located underneath the film/solution interface. A very similar model was proposed by Hu¨bsch et al. to explain the exponential-to-linear transition for blended polyelectrolyte multilayers.31 Another factor that may take place and that strengthens the above argument is the gradual rearrangement of the PE chains in the film causing the film to become progressively less and less penetrable. Of course, this process starts close to the substrate where rearrangements took place first. If this phenomenon occurs, then it contributes to reduce the thickness of the zone that acts as an effective PLL reservoir and thereby favors the appearance of domain II earlier than expected for a homogeneous film. As the buildup progresses, the thickness of domain II increases whereas that of domain III is constant and moves with the film/ solution interface. This process corresponds to the linear growth

(22) Caruso, F.; Lichtenfeld, H.; Donath, E.; Mo¨hwald, H. Macromolecules 1999, 32, 2317-2328. (23) Schmitt, J.; Gru¨newald, T.; Decher, G.; Pershan, P. S.; Kjaer, K.; Lo¨sche, M. Macromolecules 1993, 26, 7058-7063. (24) Jomaa, H. W.; Schlenoff, J. B. Macromolecules 2005, 38, 8473-8480. (25) Richert, L.; Lavalle, P.; Payan, E.; Shu, X. Z.; Prestwich, G. D.; Stoltz, J.-F.; Schaaf, P.; Voegel, J.-C.; Picart, C. Langmuir 2004, 20, 448-458. (26) Lavalle, P.; Picart, C.; Mutterer, J.; Gergely, C.; Reiss, H.; Voegel, J.-C.; Senger, B.; Schaaf, P. J. Phys. Chem. B 2004, 108, 635-648. (27) Collin, D.; Lavalle, P.; Garza, J. M.; Voegel, J.-C.; Schaaf, P.; Martinoty, P. Macromolecules 2004, 37, 10195-10198.

(28) Picart, C.; Gergely, C.; Arntz, Y.; Voegel, J.-C.; Schaaf, P.; Cuisinier, F. G. J.; Senger, B. Biosens. Bioelectron. 2004, 20, 553-561. (29) McAloney, R. A.; Sinyor, M.; Dudnik, V.; Goh, M. C. Langmuir 2001, 17, 6655-6663. (30) Michel, M.; Izquierdo, A.; Decher, G.; Voegel, J.-C.; Schaaf, P.; Ball, V. Langmuir 2005, 21, 7854-7859. (31) Hu¨bsch, E.; Ball, V.; Senger, B.; Decher, G.; Voegel, J. C.; Schaaf, P. Langmuir 2004, 20, 1980-1985. (32) Saloma¨ki, M.; Vinokurov, I. A.; Kankare, J. Langmuir 2005, 21, 1123211240.

4378 Langmuir, Vol. 22, No. 9, 2006

Porcel et al.

Figure 1. Schematic representation of the multilayer buildup mechanism by the model of the three zones. (a) At the beginning of the film buildup, the deposition of the first layers strongly depends on the properties of the substrate surface. This region is composed of only the first pairs of layers in the vicinity of the substrate surface and corresponds here to domain I. (b) The number of deposition steps increases, and the diffusion process takes place in domain III, leading to exponential growth of the film thickness. (c) As the construction goes on, the film undergoes a restructuring of the bottom layers of zone III, leading to the formation of a restructured zone denoted as zone II. Because this new zone is supposed to become impermeable for the diffusion process, domain III reaches a limited thickness, and the film grows linearly with the number of deposition step.

regime that follows the exponential phase. The linear growth process, usually observed at room temperature and moderate salt concentration (0.15 M NaCl) for PSS/PAH films, for example, corresponds to the case where domain III extends only over a very limited thickness, typically that of one or two PE layers. This explanation constitutes an interesting unifying view of the buildup process of the PEM films. However, even though this explanation sounds very appealing, Saloma¨ki et al. did not supply experimental evidence supporting it. To extend the study of the exponential-to-linear transition, after its observation by Michel et al.30 in our group, we used both the dipping and the spraying methods to buildup HA/PLL films, which are well known for their exponential behavior. The spraying method allows us to vary the contact times of the PE solutions with the film in a very precise way. This method, first introduced by the group of Schlenoff,33 was recently investigated in more detail by Izquierdo et al.34 and extended by Porcel et al.35 to construct films by simultaneously spraying both PEs. In the present article, we shall report our investigation of the HA/PLL system with the goal to gain as much information as possible on the exponential-to-linear transition. In particular, we shall critically discuss the explanation of Saloma¨ki et al.32 We also wish to investigate how the two main buildup parameters entering into the spraying process (i.e., spraying time and spraying rate) affect the film growth. Finally, we shall compare the spraying buildup process with the more conventional dipping method. Materials and Methods Polyelectrolyte Solutions. Poly(L-lysine) (PLL, P-2636, Mw ) 6.7 × 104 g mol-1) was purchased from Sigma (St. Quentin Fallavier, France), and hyaluronic acid (HA, Mw ) 4.0 × 105 g mol-1) was purchased from Bioiberica (Barcelona, Spain). Poly(ethylenimine) (PEI, Lupasol WF, Mw ) 2.5 × 104 g mol-1) was obtained from BASF (Germany). All solutions were prepared using ultrapure water (Milli-Q Ultrapure Water System, Millipore) with a resistivity of 18.2 MΩ cm. The PEI solution was obtained by dissolution in pure Milli-Q water of an amount leading to a concentration of 2 mg mL-1. HA and PLL solutions were prepared by the dissolution of adequate amounts of PE in a 0.15 M NaCl solution. The final PE concentrations were 3 × 10-3 M in monomer units. For the experiments concerning the determination of the influence of the (33) Schlenoff, J. B.; Dubas, S. T.; Farhat, T. Langmuir 2000, 16, 9968-9969. (34) Izquierdo, A.; Ono, S. S.; Voegel, J.-C.; Schaaf, P.; Decher, G. Langmuir 2005, 21, 7558-7567. (35) Porcel, C. H.; Izquierdo, A.; Ball, V.; Decher, G.; Voegel, J.-C.; Schaaf, P. Langmuir 2005, 21, 800-802.

spraying rate, these solutions were diluted by factors of 2 and 10 to obtain PE concentrations of 1.5 × 10-3 and 3 × 10-4 M, respectively. Multilayer Buildup by the Spraying Method. For the buildup of the PEM films by spraying, we used polypropylene air pump sprays of 200 mL functioning with an Airspray system (Fisher Bioblock Scientific, Illkirch, France). The pressure inside the spray bottles was fixed to 1 bar. All of the films were made on silicon wafers (100) (Wafernet, INC., San Jose, CA) cleaned with EtOH, 1:1 MeOH/HCl(aq) solution, and with a 98% w/w H2SO4 solution. Then the substrates were extensively rinsed with Milli-Q water and dried under nitrogen. The alternate spraying process consisted of alternately spraying the polyanion (HA) and the polycation (PLL) solutions perpendicularly to a vertically fixed substrate to allow for the drainage of the solution. To promote the adsorption of the first HA layer, a precursor layer of PEI was always deposited on the bare substrate by spraying for 3 s. The HA (or the PLL) solution was sprayed onto the substrate surface for x s (or for y s), and then we waited (30 x) s (or (30 - y) s) before rinsing by spraying the buffer solution for 20 s. Finally, we waited 10 s more before depositing the PLL polycation (or the HA polyanion) in the same manner. The obtained film can be denoted PEI3s-(HAxs/PLLys)n after the deposition of n pairs of HA/PLL layers. The spraying rate of the PE solutions per surface unit on the substrate surface was estimated to be 0.0040.005 mL s-1 cm-2. FITC-Labeled Poly(L-lysine). Poly(L-lysine) labeled with fluorescein isothiocyanate (FITC, Sigma, F-4274, lot no. 013K2528, Mw ) 389.4 g mol-1) was prepared according to the method described by Hermanson.40 The excess FITC not linked to the PLL was removed by dialysis (Spectra Por, dialysis membrane, molecular weight cutoff 10 × 103 g mol-1) against a 0.15 M NaCl solution at 4 °C and in the dark. This dialysis step was repeated five times until no FITC release could be detected in the dialysate by measuring its absorbance at 488 nm. The degree of substitution of the PLLFITC was calculated to be 0.024 fluorophore per monomer of L-lysine. Multilayer Buildup by the Dipping Process. For ellipsometer measurements, the films were constructed by hand dipping. The same silicon wafers with the same cleaning method as in the spraying method were used as substrates. They were first dipped into the PEI solution for 3 min (or 10 min) and then rinsed by dipping consecutively in two different 0.15 M NaCl baths for 2 min. The deposition of HA (or PLL) consisted of dipping the substrate the necessary time fixed for the experiment (3 or 10 min) in the PE solution before two rinsing steps of 2 min in two different buffer baths. The film thickness was measured by means of an ellipsometer as will be described in the next section. The concentrations of the PE solutions were equal to 3 × 10-3 M. For confocal microscopy the film buildup was carried out with an automated dipping robot (Riegler & Kirstein GmbH, Berlin,

Exponential/Linear Growth Germany). Glass slides (VWR, Fontenay sous Bois, France), cleaned in the same manner as the silicon slides used for the spray process were first dipped into the PEI solution for 10 min to adsorb a precursor layer. They were then rinsed by simple dipping in Milli-Q water for 10 min more. After this step, HA and PLL could be alternately adsorbed by using the same process except that the Milli-Q rinsing water bath was replaced by the buffer solution of 0.15 M NaCl. A pair of layers is achieved in 40 min. For n pairs of layers, the film is denoted PEI-(HA/PLL)n. Ellipsometry. Film thicknesses were determined in the dry state with a Plasmos SD 2100 ellipsometer (Jobin Yvon, France) at a fixed angle of 45° for the films constructed by the spraying and the dipping methods. The measurements were performed after drying by blowing N2 over the wet films for 30 s. Great care was taken to avoid NaCl crystallization over the substrates, which would significantly modify the film homogeneity. The ellipsometer works according to the rotating analyzer method. The light source is a 632.8 nm He-Ne laser. The real part of the refractive index of the silicon wafers and the PEM film are respectively fixed to 3.865 and 1.465 at the working wavelength. Their imaginary parts equal -0.02 and 0.0, respectively, because the PE does not absorb light at this wavelength. The refractive index value of the HA/PLL films has been chosen in a first approximation to be close to that of PGA/PAH films as measured by means of optical waveguide light-mode spectroscopy.36 One can point out that even if this value is not strictly exact it introduces a systematic but only slight error in the thickness. Moreover, as will be seen, the thickness values found by ellipsometry and those determined by AFM after scratching the film are comparable, validating the approximate value taken for the refractive index. The estimate of the thickness is always the average of 10 measurements performed over an area of 3 cm2. The error bars correspond to the standard error of the mean. Atomic Force Microscopy (AFM). Atomic force microscopy images were obtained in contact mode in air or in buffer solution with the Nanoscope IIIa microscope from Digital Instruments (Santa Barbara, CA). Cantilevers with spring constants of 0.01 and 0.03 N m-1 and with silicon nitride tips were used (model MLCT-AUHW, Park Scientific, Sunnyvale, CA). We always performed several scans over a given area. These scans had to produce reproducible images for us to ascertain that there is no sample damage induced by the tip. Deflection and height mode images were scanned simultaneously at a fixed scan rate (between 0.5 and 2.5 Hz) with a resolution of 512 × 512 pixels2. The scan sizes varied from 10 × 10 µm2 to 80 × 80 µm2. Profilometric section analyses of a scratched film allowed us to determine precisely the quality of the film and its thickness over the scanned area. The scratches were achieved with a syringe tip and were always imaged perpendicular to the fast scan axis. The profiles correspond to a cross section along this axis. Confocal Laser Scanning Microscopy (CLSM). Confocal laser scanning microscopy investigations of the films were performed in a liquid (0.15 M NaCl). CLSM observations were carried out with a Zeiss LSM microscope using a ×40/1.4 oil immersion objective and with 0.4 µm z-section intervals. The adsorption of the PLLFITC at 2.5 × 10-3 M in monomer units was performed by depositing 500 µL of the solution on top of the films prepared by dipping and were always kept in the wet state. After 2 min, the PLLFITC solution was removed by replacement with a buffer solution. FITC fluorescence was detected upon excitation at 488 nm through a cutoff dichroic mirror and an emission band-pass filter at 505 nm (green). Virtual vertical sections can be visualized, hence allowing the thickness of the film to be determined. The PEM films were never dried for these CLSM observations. Optical Waveguide Light-Mode Spectroscopy (OWLS). The construction of PEI-(HA/PLL) films was observed by OWLS. To summarize, OWLS is an optical technique that senses the film with an evanescent wave roughly over 300 nm. In this area, it allows NTE and NTM, the transverse electric and magnetic effective refractive (36) Boulmedais, F.; Ball, V.; Schwinte´, P.; Frisch, B.; Schaaf, P.; Voegel, J.-C. Langmuir 2003, 19, 440-445.

Langmuir, Vol. 22, No. 9, 2006 4379

Figure 2. Influence of the drying on the buildup of (PEI3s-(HA3s/ PLL3s)20) films made by spraying. (9) Thickness with a rehydration step before each buildup step and (O) without a rehydration step. (+) (see text). A film of 20 pairs of layers made without any intermediate drying steps. The inset shows the evolution of the film thickness on a logarithmic scale. The linear part (straight line) corresponds to the exponential growth zone, which extends from 4 to 12 pairs of layers. The thicknesses were measured by ellipsometry, and the films are in the dry state. indexes, respectively, to be determined. Then, these indexes allow us to obtain structural information about the film (optical thickness and refractive index). This technique has always been applied with success to (HA/PLL) PEM.17 The experimental setup and the deposition conditions can be found elsewhere.36

Results General Features of the HA/PLL Film Buildup by Spraying. Figure 2 represents the typical evolution of the film thickness with the number of bilayer deposition steps and the influence of the drying on the buildup for films obtained by the spraying method. The PE spraying time was fixed at 3 s for each PE deposition, and the concentration of the PE solutions was kept fixed at 3 × 10-3 M. After the deposition of four pairs of layers, the thickness increases exponentially up to the deposition of 10-12 PE pairs of layers (inset in Figure 2) before increasing linearly with the number of deposition steps. The film thickness at which this exponential-to-linear transition takes place is on the order of 100 nm for the film in the dry state, as measured by ellipsometry. We verified that differences between the thicknesses measured by ellipsometry and those measured by AFM when the films were partially scratched and a profile recorded never exceeded 10%. This confirms that we do not introduce a significant error by assuming that the film refractive index is equal to 1.465 for ellipsometer measurements. The results of three different experiments are summarized in Figure 2. In the first experiment, the film was dried after each polyanion/polycation deposition step. Then, it was rehydrated with the same 0.15 M NaCl aqueous solution before the next PE solution was sprayed on the substrate when continuing the buildup process. In the second experiment, the film was dried, and the thickness was measured only after the deposition of every three pairs of layers for the exponential growth part and every two pairs for the linear one. Eventually in the third experiment, 20 pairs of layers were deposited in one shot, avoiding the intermediate dryings. No differences in the film thicknesses were seen. This shows that the drying and rehydration processes do not affect the film buildup process significantly. Thus, in the following, we dried the film whenever needed.

4380 Langmuir, Vol. 22, No. 9, 2006

Figure 3. Spraying versus dipping: comparison of the evolution of the PEM film thickness as a function of the number of added pairs of layers between films made by dipping or spraying. (9) PEI3s(HA15s/PLL9s)17 film obtained by spraying, (O) PEI-(HA/PLL)20 obtained by dipping for 3 min per PE, and (2) PEI-(HA/PLL)18 obtained by dipping for 10 min per PE. All of the PE concentrations were equal to 3 × 10-3 M. In all experiments, the film thicknesses were measured by ellipsometry after drying.

To investigate further the influence of the drying process on the film thickness, we performed AFM thickness measurements of the same films in the dried and rehydrated states. We found that when constituted by more than six bilayers the films swell by a factor of 4 when rehydrated after a few minutes. We used this swelling factor throughout this study to get an idea of the thickness of the film when it is in contact with the aqueous solutions even though it must be taken only as an indication because we do not control the humidity of air in contact with the film in the dry state. A similar study was performed by Barrett et al.37 on hyaluronic acid/poly(allylamine) films, and the factor of 4 that we find typically lies in the range of their reported values. This confirms that (HA/PLL)n films are highly hydrated with a water content of at least 75%. Dipping versus Spraying. To verify that the transition from the exponential to the linear growth regime for the HA/PLL system is not a peculiarity of the spraying method, we performed several experiments with the conventional dipping method. Figure 3 represents two cases where the films were built up to 20 pairs of layers, the contact time of the PE solutions with the substrate being equal to 3 and 10 min, respectively. In both cases, the transition from the exponential to the linear growth regime sets in after the deposition of 10-12 bilayers. The evolution of the thickness of a film obtained by the spraying method is also represented in this Figure. A very similar evolution

Porcel et al.

of the film thickness for a PEM film constructed by dipping the surface for 3 min in the PE solutions and by spraying the PLL solution for 9 s and the HA solutions for 15 s is observed, with the concentrations of the PE solutions being equal to 3 × 10-3 M in monomer units. The buildup processes performed with the dipping and the spraying methods do not seem to be fundamentally different from each other even though the characteristic time scales involved in the two approaches are not the same. Moreover, as will be seen below, the exponential-to-linear transition always takes place at about n ) 12 for films constructed by the spraying method over a very large range of buildup conditions (i.e., different spraying times and spraying rates). The first studies performed on the HA/PLL system where the film was grown by the dipping method showed that during the initial stages the substrate was not uniformly covered by a film but rather by islands.17 These islands grew and finally coalesced as the number of deposition steps increased. Could this morphological transition be responsible for the exponential-tolinear transition in the growth regime? As observed in Figure 2, the drying of a film after the deposition steps has no influence on the evolution of its thickness with n. Yet, after 6 HA/PLL deposition steps (n g 6) we observe by AFM, after drying, a continuous film (Figure 4) that remains continuous after rehydration. Now, because the exponential-to-linear transition takes place for n ) 10-12, it cannot be ascribed to the islandsto-continuous film transition. Linear Growth Regime. Previous studies clearly established that exponential growth is due to the diffusion of at least one PE in and out of the whole film at each bilayer deposition step.20 In the case of a HA/PLL film, it was shown that PLL constitutes the diffusing species, whereas HA does not diffuse through the film. One can now ask whether PLL still diffuses in and out of the whole film when the linear growth regime is reached. To answer this question, CLSM experiments were performed on a film consisting of 80 HA/PLL pairs of layers (i.e., far beyond the onset of the linear growth regime). The film ended in an HA layer. On top of the PEM film we added PLLFITC, and the film was imaged less than 2 min after the PLLFITC deposition. Figure 5 shows that the film section became homogeneously green over its full height. This proves that labeled PLL can penetrate down to the bottom of the PEM architecture even when the linear growth regime is reached. To analyze further the film buildup process, we also use a complementary technique (OWLS). When the film becomes thicker than the penetration depth of the evanescent wave, the signal becomes sensitive only to changes in the refractive index of the film near the deposition substrate. Such changes were

Figure 4. AFM height image in air of a scratched PEI-(HA15s/PLL6s)8 film and a scratched section. The image size is 50 × 50 µm2, and the z scale ranges from 0 to 300 nm. The PE solution concentrations are 3 × 10-3 M. The scratched zone is located on the left of the image.

Exponential/Linear Growth

Langmuir, Vol. 22, No. 9, 2006 4381

Figure 5. CLSM observations of vertical sections of PEI-(HA/ PLL)80-HA-PLLFITC films made by dipping, the last buildup step corresponding to the deposition of a PLLFITC solution on top of the PEI-(HA/PLL)80-HA film and another buffer rinse (image size: 76.8 × 29.8 µm2). The white line indicates the silica slide. The thickness of the film was on the order of 18 µm. All of the PE solution concentrations used to build up the films were 3 × 10-3 M in monomer units.

Figure 6. OWLS observations of the evolution of NTM during the buildup of a PEI-(HA/PLL)24 film made by dipping. The PE solution concentrations were 3 × 10-3 M. The arrow at the 12th pair of layers represents the position at which the exponential-to-linear transition is observed for films characterized in the dry state by means of ellipsometry measurements.

interpreted in previous studies to be the consequence of the diffusion of a PE in and out of the film.17,20 Other optical techniques based on evanescent waves such as surface plasmon resonance18 were used and interpreted similarly. Figure 6 shows a typical evolution of the OWLS signal with time. Only the evolution of the effective refractive index of the transverse magnetic polarization is shown, but the evolution of the transverse electric polarization is similar. Figure 6 reveals that the cyclic evolution of the signal begins during the exponential growth phase and extends far into the linear growth regime. Thus, even in the linear regime, the refractive index of the film changes in the penetration zone of the evanescent wave during each deposition step. However, the amplitude of the signal variations decreases slowly during the cyclic evolution as the buildup process goes on. This indicates an attenuation of the process leading to the refractive index change at the bottom of the film. We shall now concentrate on the influence of the two main parameters that govern the spraying buildup process in the linear regime. First, we focus on the influence of the spraying time. Figure 7A represents the evolution of the film thickness, d, as a function of the number of polyanion/polycation deposition steps for different spraying times of the PLL solution (at 3 × 10-4 M). The spraying time of the HA solution (3 × 10-3 M) was kept constant (15 s). One observes that the thickness

Figure 7. Evolution of the film thickness as a function of the number of added layer pairs for PEI3s-(HA/PLL)n films for different PLL spraying times. (A) Results of PEI3s-(HA15s/PLLys)20 where y corresponds to (9) 3, (O) 6, (2) 9, ()) 12, and (left-facing solid triangle) 15 s. The PLL concentration was reduced by a factor of 10 in comparison to the HA concentration and was equal to 3 × 10-4 M. (B) Saturation state is obtained for PEI3s-(HA3s/PLLys)20 for y equal to (9) 3, (O) 6, (2) 9, ()) 12, and (left-facing solid triangle) 15 s. Here, the PE concentrations are both equal to 3 × 10-3 M.

increment, ∆d, per layer is strongly dependent on the PLL spraying time. These experiments were repeated for two larger PLL concentrations (i.e., spraying rates). Figure 8A summarizes the evolution of the thickness increment per polyanion/polycation deposition step in the linear regime with the spraying time, PLLτs, of the PLL solution for concentrations equal to 3 × 10-4, 1.5 × 10-3, and 3 × 10-3 M. As can be seen in Figure 8A, at low spraying rates, the thickness increment increases linearly with the spraying time over the investigated time range. As the spraying rate (or PLL concentration) increases, the thickness increment also increases with the spraying time but less rapidly than PLLτs and might even saturate for the largest concentration and the longest spraying time. Figure 7B shows the evolution of d with n for four different PLL spraying times at high spraying rates (PLL concentration equal to 3 × 10-3 M) but with a small, constant HA spraying time, HAτs, equal to 3 s. The thickness increment per polyanion/ polycation pair of layers becomes almost independent of the PLL spraying time. Similar experiments were conducted by keeping the spraying time of PLL constant and varying that of HA. The results are summarized in Figure 8B. Here too, at low HA concentration, the thickness increment per bilayer increases linearly with the spraying time, HAτs. In contrast, the increment (37) Burke, S. E.; Barrett, C. J. Biomacromolecules 2005, 6, 1419-1428.

4382 Langmuir, Vol. 22, No. 9, 2006

Figure 8. Evolution of the thickness increment per pair of layers, ∆d, (measured in the linear growth regimes) as a function of the spraying time of one of the PEs (i.e., PLLτs in the upper panel and HAτ in the lower panel). (A) Results for PEI -(HA /PLL ) films s 3s 15s ys 20 for different PLL concentrations: (9) 3 × 10-4, (O) 1.5 × 10-3, -3 and (2) 3 × 10 M, with the HA concentration equal to 3 × 10-3 M. (B) Results for PEI3s-(HAxs/PLL9s)20 films for different HA concentrations: (9) 3 × 10-4, (O) 1.5 × 10-3, and (2) 3 × 10-3 M, with the PLL concentration equal to 3 × 10-3 M.

practically reaches a plateau value at the highest HA concentration (3 × 10-3 M) for a spraying time HAτs that is long enough. Finally, it must be pointed out that under all of the investigated deposition conditions the exponential-to-linear transition took place at about n ) 12, indicating the independence of the location of this transition from the spraying parameters.

Discussion PEM films are usually constructed by bringing the substrate into contact with the PE solutions for a given time during each deposition step. We observe that after the deposition of a few pairs of layers, which are sensitive to the presence of the substrate, the film growth increases exponentially before becoming linear. The crossover from the exponential to the linear growth regime is expected once the film reaches a critical thickness for which the diffusing PE is no longer able to diffuse in and out of the whole film during the finite contact time.31 This is one of the explanations advanced by Saloma¨ki et al.32 for the exponentialto-linear transition. In addition, these authors assume that a restructuration of the film occurs during each deposition step, leading to a densification of the PEM film. This densification starts, of course, at the bottom of the film, where the number of restructuration steps is largest if one assumes that the deposition of each pair of layers is accompanied by such a restructuring. It results in the progressive reduction of the mobility of the PE

Porcel et al.

inside the film until diffusion becomes blocked in the deepest region of the film (i.e., where restructuring has been active the most frequently). In this step, a whole zone of the PEM film can no longer play its role of PLL reservoir, whether the PLL diffuses in or out of the film. This process is summarized in Figure 1. Do our experimental results support these assumptions? First, assume that the exponential-to-linear transition is solely due to the limited diffusion zone accessible during a finite time lapse in a homogeneous film. This would imply that by increasing the contact time of the film with the solution of diffusing PE (i.e., PLL in the present work) the occurrence of the exponentialto-linear transition should be shifted to a larger number of deposition steps. Experimental results do not suggest this trend. Indeed, when films are built by the spraying process, the characteristic contact time between a film and the PE solutions is on the order of 10-20 s, whereas with the dipping method this time is typically on the order of a few minutes. However, Figure 3 shows that, in both cases, the exponential-to-linear transition takes place in about 10-12 deposition steps. Moreover, the exponential-tolinear transition seems fairly independent of the characteristic parameters governing the spraying buildup process (spraying time and spraying rate) as can be seen in Figure 7A and B. The independence of the exponential-to-linear transition with the number of deposition steps, n, seems to invalidate this assumption. Now, assume that during each deposition step there is a film restructuring that preferentially reduces the diffusion of the PE through the most restructured parts of the film (i.e., those close to the adsorption substrate). Indeed, this assumption could explain why the exponential-to-linear transition always takes place after the same number of deposition steps. The reason for this might be that, during the buildup process, layers existing for more than this number of steps have been restructured so often that they have been rendered practically impenetrable to the diffusing PE. This could be due to a gradual densification of the layers or a gradual change in the Donnan potential after each restructuring. However, is this mechanism compatible with the CLSM and OWLS experiments displayed in Figures 5 and 6, respectively? At first sight, this seems not to be the case. Indeed, when a PLLFITC solution is deposited on top of a thick, HA-terminated film, namely, an (HA/PLL)80-HA film, the film becomes totally green, within less than 2 min, suggesting the diffusion of PLLFITC into the whole film. However, recent experiments have shown that PE from the solution can exchange PE from the film.38 In particular, we have shown that free PLL chains in the film can exchange with bound PLL chains involved in strong interactions with HA.20 Moreover, the exchange process between PLL chains from the film and PLLFITC chains from the solution must take place rapidly as indicated by the rapid coloration of the film. Such an exchange process certainly takes place in the zone of the film where free PLLFITC chains have diffused. If, after the exchange process, the bound PLLFITC chains diffused through the whole film, then the film could become entirely green even without diffusion of free PLL chains through the whole multilayer. Such a diffusion process of bound PLL chains is strongly suggested by fluorescence recovery after photobleaching (FRAP) experiments, and the diffusion coefficient of bound PLL was estimated to be on the order of 10-9-10-10 cm2 s-1.39 The diffusion of bound PLL chains could take place through a kind of reptation/exchange process where a bound PLL chain moves (38) Ball, V.; Hu¨bsch, E.; Schweiss, R.; Voegel, J.-C.; Schaaf, P.; Knoll, W. Langmuir 2005, 21, 8526-8531. (39) Picart, C.; Mutterer, J.; Arntz, Y.; Voegel, J.-C.; Schaaf, P.; Senger, B. Microsc. Res. Technol. 2005, 66, 43-57. (40) Hermanson, G. T. Bioconjugate Techniques; Academic Press: San Diego, CA, 1996; p 303.

Exponential/Linear Growth

and forms new electrostatic interactions with anionic sites that became vacant because of the diffusion of another bound PLL chain. The value of the diffusion coefficient is fully compatible with the fact that a film whose thickness is on the order of 1 µm becomes entirely green within a few seconds. This mechanism could thus reconcile the observation that the green coloration extends down to the substrate and the hypothesis that PLL chains do not diffuse into the region that has undergone the most restructuring. OWLS also suggests, at first sight, that PLL diffuses down to the substrate even in the linear growth regime. However, it must be mentioned that the optical signals vary rapidly (in less than 40 s) after each new contact with a PE solution. After 20 deposition steps, an HA/PLL film typically reaches a thickness on the order of 1 µm in the wet state. This indicates that the diffusion coefficient of the PLL chains should at least be on the order of 10-9-10-10 cm2 s-1, which is a typical value for PLL in HA/PLL films as already indicated. This interpretation would thus invalidate the assumption of a restructured zone, practically impenetrable to the PLL chains, that grows during the film buildup and where PLL diffusion would be hindered. However, as for the CLSM, the OWLS experiments can also be explained in another way: the OWLS signal might be due to the fact that PLL diffusion in the upper zone of the film also affects the counterion concentration in the whole film. This could induce the swelling of the film down to the deposition substrate and thus lead to a change in its refractive index. This process is also expected to be very rapid. We discuss now the evolution of the thickness increment in the linear regime under the hypothesis that it is a consequence of the hindrance of the diffusion of free PLL chains into the zone of the PEM film affected by a given critical number nc of successive restructurings. First, it is worth noting that HA and PLL do not play the same role. Whereas PLL diffuses in and out of at least part of the film, HA does not. When a film is in contact with a PLL solution and when equilibrium is reached between the film and the solution, the concentration cFPLL of free PLL chains inside the film increases with the concentration of PLL chains in the solution. After nc deposition steps, a zone starting at the bottom of the film becomes inaccessible to free PLL chains (zones I and II in Figure 1). For a given PLL concentration in solution, a maximum number of free PLL chains can then diffuse into a multilayer. This number per unit area of the film is equal to cFPLL × dIII where dIII represents the thickness of zone III. When the concentration of the PLL solution increases, it is expected that cFPLL reaches a plateau. This, in turn, implies that a maximum number of HA chains can be bound onto the PEM film during each HA deposition step in the linear growth regime. Because the thickness dIII is directly related to both the number of free PLL chains in the film and the number of HA chains that will be complexed by the free PLL chains that diffuse out of the film, dIII also reaches a plateau value. Thus, there must exist an upper limit for the thickness increment reached at high PLL and high HA concentrations and for long spraying times. This upper thickness increment limit seems to be on the order of 60-70 nm/deposition step for the film in the dry state (250-300 nm/ deposition step in the hydrated state) and is observed for HA and PLL concentrations exceeding 3 × 10-3 M, HA spraying times exceeding 15 s, and PLL spraying times exceeding roughly 10 s. For PLL solutions of concentration 3 × 10-3 M and for HA solutions of concentrations lower than 3 × 10-3 M (e.g., 3 × 10-4 and 1.5 × 10-3 M), for the same thickness increment to be observed, one should spray longer to spray the same amount

Langmuir, Vol. 22, No. 9, 2006 4383

of HA onto the surface. Such long spraying times are unfortunately out of reach with our experimental device. Within the range of spraying times investigated here, the amount of HA that is sprayed onto the surface is not sufficient to extract, during each deposition step, the total number of free PLL chains from the film. The thickness increment in the linear regime thus increases with the HA spraying time, the increase being even fairly linear with HAτ . s Finally, when the film is built by using an HA concentration of 3 × 10-3 M and an HA spraying time of 15 s, all of the free PLL chains in the film will be extracted during each deposition step whatever the concentration of the PLL solution. The evolution of the thickness increment, ∆d, with the PLL spraying time, PLLτ , at different PLL concentrations depends on the number of s free PLL chains that diffuse into the film during the PLL spraying step. Decreasing the concentration of the PLL solution will also decrease the equilibrium concentration of free PLL chains in the film. However, decreasing the PLL spraying time at constant PLL concentration in the solution will not have a similar effect because diffusion is not a process in which the flux varies linearly with time. It is expected that decreasing the PLL concentration of the solution by a factor X is not equivalent to decreasing the spraying time by the same factor X by keeping the PLL concentration of the solution fixed. A detailed, quantitative analysis is beyond the scope of this article. In any case, it is expected that by sufficiently decreasing both parameters the thickness increment also decreases.

Summary We have investigated the buildup process of hyaluronic acid/ poly(L-lysine) (HA/PLL) multilayers. The films were built by using the dipping method and the more recent spraying method where the polyelectrolyte solutions are sprayed alternately onto a vertical substrate. We find that, with both methods, after the deposition of the first zone that is still influenced by the presence of the substrate, the film thickness evolves exponentially with the number of deposition steps and after a given number of deposition steps its thickness evolution becomes linear. Second, because the spraying method allows us to vary the different parameters of the buildup very finely, we used this method to investigate the exponential-to-linear transition. We find that this transition always takes place after about 12 deposition steps whatever the values of the parameters controlling the deposition process. We discuss our results in light of a model proposed by Hu¨bsch et al.31 and later by Saloma¨ki et al.32 in which it is assumed that the exponential-to-linear transition is due to a film restructuring that progressively forbids the diffusion of one of the polyelectrolytes constituting the film over part of the film. This forbidden zone then grows with the number of deposition steps so that the outer zone of the film that is still concerned with diffusion keeps a constant thickness and moves upward as the total film thickness increases. Our analysis leads to the conclusion that even though our experimental facts do not prove that the crossover from the exponential to the linear growth regime is due to a restructuring (and hence a densification) of the deepest region of the film, they are fully compatible with this hypothesis. Further studies are needed to confirm the existence of such a restructuring and thus fully explain the exponential-to-linear transition in other PEM films whose growth starts exponentially. Acknowledgment. This work was supported by ACI Nanoscience (project no. NR204). LA053218D