Influence of the Polyelectrolyte Molecular Weight on Exponentially

Nov 8, 2006 - Institut Charles Sadron, 6 rue Boussingault, 67083 Strasbourg Cedex, Institut National de la Sante´ et de la. Recherche Me´dicale, Uni...
0 downloads 0 Views 238KB Size
Download PDF
1898

Langmuir 2007, 23, 1898-1904

Influence of the Polyelectrolyte Molecular Weight on Exponentially Growing Multilayer Films in the Linear Regime C. Porcel,†,‡ Ph. Lavalle,‡,§ G. Decher,† B. Senger,‡,§ J.-C. Voegel,‡,§ and P. Schaaf*,† Institut Charles Sadron, 6 rue Boussingault, 67083 Strasbourg Cedex, Institut National de la Sante´ et de la Recherche Me´ dicale, Unite´ 595, 11 rue Humann, 67085 Strasbourg Cedex, and Faculte´ de Chirurgie Dentaire, UniVersite´ Louis Pasteur, 1 place de l’Hoˆ pital, 67000 Strasbourg, France ReceiVed September 18, 2006. In Final Form: NoVember 8, 2006 Alternated deposition of polyanions and polycations on a charged solid substrate leads to the buildup of polyelectrolyte multilayer (PEM) films. Two types of PEM films were reported in the literature: films whose thickness increases linearly and films whose thickness increases exponentially with the number of deposition steps. However, it was recently found that, for exponentially growing films, the exponential increase of the film thickness takes place only during the initially deposited pairs of layers and is then followed by a linear increase. In this study, we investigate the growth process of hyaluronic acid/poly(L-lysine) (HA/PLL) and poly(L-glutamic acid)/poly(allylamine) (PGA/ PAH) films, two films whose growth is initially exponential, when the growth process enters the linear regime. We focus, in particular, on the influence of the molecular weight (Mw) of the polyelectrolytes. For both systems, we find that the film thickness increment per polyanion/polycation deposition step in the linear growth regime is fairly independent of the molecular weights of the polyelectrolytes. We also find that when the (HA/PLL)n films are constructed with low molecular weight PLL, these chains can diffuse into the entire film during each buildup cycle, even for very thick films, whereas the PLL diffusion of high molecular weight chains is restricted to the upper part of the film. Our results lead to refinement of the buildup mechanism model, introduced previously for the exponentially growing films, which is based on the existence of three zones over the entire film thickness. The mechanism no longer needs all the “in” and “out” diffusing polyanions or polycations to be involved in the buildup process to explain the linear growth regime but merely relies on the interaction between the polyelectrolytes with an upper zone of the film. This zone is constituted of polyanion/polycation complexes which are “loosely bound” and rich in the polyelectrolyte deposited during the former deposition step.

Introduction The alternate deposition of polyanions and polycations onto solid substrates leads to the formation of films called polyelectrolyte multilayer (PEM)1 films. These films have received considerable interest due to 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 Two types of multilayer growth processes were reported: films †

Institut Charles Sadron. 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. (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.

with thickness that increases linearly with the number of deposition steps1 and other systems whose thickness grows exponentially.17-19 Linear growth is due to the fact that during the buildup process the polyelectrolytes from the solution interact only with polyelectrolytes of opposite charge forming the outer layer of the film.1 For the exponential growth, on the other hand, it was stated that it results from the diffusion “in” and “out” of the whole film of at least one of the polyelectrolytes constituting the multilayer.19 However, this view appears oversimplified since the recent observation that the exponential growth regime in fact only holds over a limited number of deposition steps, before the crossover to a linear growth regime.20,21 One can, however, point out that the thickness increment is usually much smaller for linearly growing films than for exponentially growing ones in their linear growth regime (typically a few nanometers versus hundreds of nanometers). The exponential to linear growth transition is still not fully understood. It has been proposed that it results from a gradual restructuration of the multilayer consecutive to the in and out diffusion process, which progressively hinders the diffusion of polyelectrolytes into the restructured zone, which becomes (16) Thierry, B.; Kujawa, P.; Tkaczyk, C.; Winnik, F. M.; Bilodeau, L.; Tabrizian, M. J. Am. Chem. Soc. 2005, 127, 1626-1627. (17) Elbert, D. L.; Herbert, C. B.; Hubbell, J. A. Langmuir 1999, 15, 53555362. (18) Picart, C.; Lavalle, P.; Hubert, P.; Cuisinier, F. J. G.; Decher, G.; Schaaf, P.; Voegel, J.-C. Langmuir 2001, 17, 7414-7424. (19) 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. (20) Saloma¨ki, M.; Vinokurov, I. A.; Kankare, J. Langmuir 2005, 21, 1123211240. (21) Michel, M.; Izquierdo, A.; Decher, G.; Voegel, J.-C.; Schaaf, P.; Ball, V. Langmuir 2005, 21, 7854-7859.

10.1021/la062728k CCC: $37.00 © 2007 American Chemical Society Published on Web 01/03/2007

Polyelectrolyte Mw Influence on Multilayer Films

Langmuir, Vol. 23, No. 4, 2007 1899

Scheme 1. Schematic Representation for the Three-Zone Buildup Mechanism Model of an Exponentially Growing PEM Filma

a Key: (a) At the beginning, the deposition of the first layers mainly depends on the properties of the substrate surface. Only the first pairs of layers in the vicinity of the substrate surface belong to this case, and they represent zone I. (b) As the number of deposition steps increases, the diffusion process takes place in zone III, leading to an exponential growth of the film thickness. (c) The construction goes on, and the film undergoes a restructuration of the bottom layers of zone III, leading to the formation of a restructured zone denoted as zone II. This new zone is supposed to hinder the diffusion process, so zone III reaches a constant thickness. From this point on, the film grows linearly with the number of deposition steps, the thickness increment per polyanion/polycation deposition step being equal to ∆d as indicated when the number of deposition steps increases from n to n + 1. The further thickness increment of the film concerns exclusively zone II.

inaccessible by diffusion.22 This “forbidden zone” then grows with the number of deposition steps (and restructuration cycles), and the outer zone of the film, which is still concerned by diffusion, keeps a constant thickness and moves upward as the total film thickness increases. This process was first proposed by Hu¨bsch et al.23 and more recently by Saloma¨ki et al.20 It is represented schematically in Scheme 1. Even though it is appealing, this model still needs further experimental validation. In particular, up to now, no direct proof of film restructuration has been furnished. Moreover, the thickness of the “diffusion zone” was not determined (zone III in Figure 1). Also, only very few studies reported the effect of the molecular weight of the polyelectrolytes on the exponential growth regime.24 In particular, its influence on the linear growth regime of exponentially growing films has not yet been investigated. This study is aimed at answering some of these questions for exponentially growing films. In particular, we study the influence of the polyelectrolyte molecular weights on the linear growth regime and on the poly(L-lysine) (PLL) diffusion behavior in hyaluronic acid (HA)/PLL films. Materials and Methods Polyelectrolyte Solutions. Different poly(L-lysine), poly(Lglutamic acid), and poly(allylamine hydrochloride) polymers corresponding to various molecular weights have been used: PLL20 (P-7890, lot no. 044K5102, Mw(MALLS) ) 2.15 × 104), PLL55 (P-2636, lot no. 024K5118, Mw(MALLS) ) 5.57 × 104), PLL360 (P-1524, lot no. 093K5158, Mw(MALLS) ) 36.18 × 104), PGA44 (P-4761, lot no. 104K5110, Mw(vis) ) 4.44 × 104; PGA ) poly(L-glutamic acid)), PGA100 (P-4886, lot no. 083K5120, Mw(vis) ) 9.78 × 104), PAH15 (283215-5G, lot no. 13413HB, Mw(MALLS) ) 1.5 × 104; PAH ) poly(allylamine)), and PAH70 (28322-3, lot no. 05212MO-083, Mw(MALLS) ) 7.0 × 104). They were all purchased from Sigma (St. Quentin Fallavier, France). Two different hyaluronic acids of molecular weights 41.1 × 104 (HA400) from Bioiberica (Barcelona, Spain) and 13.23 × 104 (HA130, 80190, lot no. 002941) from Lifecore (Chaska, MN) were used. The molecular weights of HA, PLL, and PGA have been determined by size exclusion chromatography to check the overlap between the different (22) Porcel, C.; Lavalle, P.; Ball, V.; Decher, G.; Senger, B.; Voegel, J. C.; Schaaf, P. Langmuir 2006, 22, 4376-4383. (23) Hu¨bsch, E.; Ball, V.; Senger, B.; Decher, G.; Voegel, J. C.; Schaaf, P. Langmuir 2004, 20, 1980-1985. (24) Kujawa, P.; Moraille, P.; Sanchez, J.; Badia, A.; Winnik, F. M. J. Am. Chem. Soc. 2005, 127, 9224-9234.

Figure 1. Influence of the molecular weights of HA and PLL on the growth of (HA/PLL)n films: (a) buildup of (HA/PLL)n films with Mw ) 130000 HA and three PLLs with molecular weights of (9) 20000, (O) 55000, and (2) up to 360000; (b) buildup of (HA/ PLL)n films using a Mw ) 360000 PLL and two different HAs with molecular weights of (9) 130000 and (O) 400000. molecular weight distributions, which showed only negligible overlaps. Poly(ethylenimine) (PEI; Lupasol WF, Mw ) 2.5 × 104) 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

1900 Langmuir, Vol. 23, No. 4, 2007 Scheme 2. Standard Deposition Sequence for the Deposition of One Polyelectrolyte Layer

dissolution in pure Milli-Q water of the amount required to reach a concentration of 2 mg/mL. HA and PLL solutions were prepared by dissolving the adequate amounts of polyelectrolytes in 0.15 M NaCl solution. The final polyelectrolyte concentrations were 1.5 × 10-3 M for the films made for ellipsometric measurements and 3.0 × 10-3 M for films for confocal microscopy analysis. Multilayer Buildup by the Spray Process.25-27 For the buildup of the PEM films by the spray procedure, we used an automated spray device developed in our laboratory. The nebulization systems are the ElectroSpray nebulizers LC/MSD G-1946-60098 from Agilent Technologies. The flow of the polyelectrolyte or the rinsing solutions is provided by syringe pumps KDS210 from KD Scientific at a constant flow rate of 1200 µL/min. The gas supply is provided by electrical floodgates STNA 1016, 220 V/50 Hz, with a pressure of 1 bar. The gas used here is nitrogen. The syringe pumps as well as the floodgates are driven by software developed in our laboratory and aimed at programming different deposition sequences for the construction of the multilayers. All the films were built on silicon wafers (100) (Wafernet, Inc., San Jose, CA) cleaned with EtOH, 1:1 MeOH/HClaq solution, and pure H2SO4. Then the substrates were rinsed with Milli-Q water and dried under nitrogen. The alternate spray process consisted in spraying alternatively the polyanion (HA) and the polycation (PLL) solutions perpendicularly to a vertically fixed substrate which allowed the drainage of the solution. The standard deposition sequence for the construction of one polyelectrolyte layer is described in Scheme 2. It consists in the spraying of one polyelectrolyte (HA or PLL) onto the substrate surface for 20 s. Then we wait 20 s before rinsing by spraying for 30 s the rinsing solution of NaCl, 0.15 M. Finally, we wait 10 s more before the deposition of the next polyelectrolyte layer following the same sequence. The deposition of a pair of layers is achieved in 160 s. Finally, the film is denoted PEI-(HAMw(HA)/PLLMw(PLL))n (or PEI(PGAMw(PGA)/PAHMw(PAH))n) corresponding to a film constituted of n pairs of layers and built with the couple of HA and PLL (or PGA and PAH) of molecular weights Mw(HA) and Mw(PLL) (or Mw(PGA) and Mw(PAH)). Multilayer Buildup by the Dipping Process. For ellipsometry, films built by dipping have also been handmade with the following procedure: the silicon wafers were first dipped for 10 min into a PEI solution and then rinsed by two consecutive 2 min dippings in 0.15 M NaCl baths. Before any ellipsometer measurement, the films were dried. For confocal microscopy the film buildup was carried out with an automated dipping robot (Riegler & Kirstein GmbH, Berlin, Germany). Glass slides (VWR, Fontenay sous Bois, France), cleaned in the same manner as for the spray process, were first dipped in the PEI solution for 10 min to adsorb a precursor layer. Then they were rinsed by a simple dipping in Milli-Q water for a further 10 min. After this step, HA and PLL were alternately adsorbed by using the same process except replacing the Milli-Q water by a 0.15 M NaCl solution. A pair of layers was deposited in 40 min. For n pairs of layers, the film was denoted as those obtained by the spraying method. FITC- and Rhodamine-Labeled Poly(L-lysine). Poly(L-lysine) labeled with fluorescein isothiocyanate isomer I, FITC (Sigma, F-4274, Mw ) 389.4) or rhodamine B isothiocyanate, Rho (Sigma, R-1755, Mw ) 536.08) was prepared according to the method (25) Schlenoff, J. B.; Dubas, S. T.; Farhat, T. Langmuir 2000, 16, 9968-9969. (26) Izquierdo, A.; Ono, S. S.; Voegel, J.-C.; Schaaf, P.; Decher, G. Langmuir 2005, 21, 7558-7567. (27) Porcel, C. H.; Izquierdo, A.; Ball, V.; Decher, G.; Voegel, J.-C.; Schaaf, P. Langmuir 2005, 21, 800-802.

Porcel et al. described by Hermanson (Hermanson, G. T. Bioconjugate Techniques; Academic Press: New York, 1996; p 303). The excess of FITC or Rho not linked to PLL was removed by dialysis (Spectra Por, molecular weight cutoff 10 × 103) against a 0.15 M NaCl solution at 4 °C and in the dark. This dialysis step was repeated five times until no FITC or Rho releases could be detected in the dialysates by following the absorbance at 488 nm (FITC) or 543 nm (Rho). The degrees of substitution of PLLFITC and PLLRho were, respectively, 0.024 and 0.060 fluorophore per monomer of L-lysine. 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 a light wavelength of 632.8 nm (He-Ne laser). The real part of the refractive index of the silicon wafers was fixed to 3.865 and its imaginary part to -0.02. The real refractive index value of the PEM films has been chosen equal to 1.465, close to the value for PGA/PAH films determined by Boulmedais et al. by means of optical waveguide lightmode spectroscopy (OWLS).28 Since the PEM films did not absorb light, the imaginary contribution was taken equal to zero. Ten different measurements were performed at different locations, over an area of 1 cm2, to get an estimation of the standard deviation of the thickness. The technique and its use for measuring PEM film thicknesses have been discussed in detail previously.22 Confocal Laser Scanning Microscopy (CLSM). Confocal laser scanning microscopy investigations of the films were performed in 0.15 M liquid solution. CLSM observations were carried out with a Zeiss LSM microscope using a 40×/1.4 oil immersion objective and a 0.4 µm z-section interval. The adsorption of PLLFITC or PLLRho at 0.5 mg/mL was performed by dipping the film for a few minutes in a PLLFITC or PLLRho solution immediately before its observation. FITC or Rho fluorescences were detected upon excitation at 488 nm (FITC) or 543 nm (Rho), through a cutoff dichroic mirror and an emission band-pass filter of 505-530 nm (green) and 585 nm (red), respectively. Virtual vertical sections can be visualized, hence allowing the film thickness to be determined. The PEM films were never dried.

Results Film Thickness Evolution. We used mainly the spraying method to construct the films. To follow the evolution of the film thickness during the buildup, we dried the films every two polyanion/polycation deposition steps, measured the thickness by ellipsometry, rehydrated the film, and pursued the multilayer deposition process. Parts a and b of Figure 1 show the evolution of the film thickness as a function of the number, n, of deposited pairs of layers for five different (HA/PLL)n multilayer architectures. Figure 1a corresponds to (HA/PLL)n films obtained with the Mw ) 130000 kDa HA and with three different PLLs of Mw ) 20000, 55000, and 360000, whereas Figure 1b represents a similar evolution for the Mw ) 360000 PLL and the two HAs of Mw ) 130000 and 400000. For all the cases, one observes that the films first grow exponentially before the thickness increment per pair of layers deposited becomes constant (linear growth regime of an exponentially growing film). In a first approximation, the exponential to linear transition always takes place when the film thickness measured in the dry state is in the range of 150-250 nm. This transition is, however, reached for different numbers of polyanion/polycation deposition steps and lies, for the investigated HA/PLL systems, between 12 and 18 pairs of layers. Table 1 summarizes the thickness increments per pair of layers in the linear domain for the different experiments. It is clear that the thickness increment is fairly independent of the PLL molecular weight, whereas it decreases when the HA molecular weight increases. It is on the order of 30 nm/pair of (28) Boulmedais, F.; Ball, V.; Schwinte´, P.; Frisch, B.; Schaaf, P.; Voegel, J.-C. Langmuir 2003, 19, 440-445.

Polyelectrolyte Mw Influence on Multilayer Films

Langmuir, Vol. 23, No. 4, 2007 1901

Table 1. Film Thickness Increments (∆d, nm/pair of layers) in the Linear Regime Obtained for the Different HA and PLL Couples by Varying Their Molecular Weightsa ∆d

pure films HA130 HA400

films with HA change HA130-HA400 HA400-HA130 films with PLL change PLL360-PLL55

HA130/PLL20 HA130/PLL55 HA130/PLL360 HA400/PLL20 HA400/PLL55 HA400/PLL360

26.3 32.3 30.7 17.2 16.1 18.5 ∆d1

∆d2

(HA130/PLL360)20-(HA400/PLL360)10 32.7 19.13 (HA400/PLL360)20-(HA130/PLL360)10 20.39 34.66 ∆d1

∆d2

(HA130/PLL360)20-(HA130/PLL55)5 25.8 27.8

a The standard deviation is about 2.0 nm/pair of layers. The films built with a unique HA and PLL couple are named “pure films”, whereas those for which the HA molecular weight is changed during the construction are denoted as “films with HA change”. In the latter case, two thickness increments, ∆d1 and ∆d2, are given, corresponding respectively to the linear growth regime for the first and the second couples used.

layers for the Mw ) 130000 HA and 16-19 nm/pair of layers for the Mw ) 400000 HA. Richert et al. already observed a similar effect for PEM films built by dipping with the couple hyaluronic acid/chitosan for which an increase of the chitosan molecular weight reduced the thickness increment during the buildup process.29 In a former study,22 we found that the ratio between the thicknesses measured in the hydrated state and in the dry state was approximately 4 for (HA/PLL)n multilayers so that in the hydrated state the thickness increment should respectively be on the order of 120 nm/pair of layers for the Mw ) 130000 HA and 70-80 for the Mw ) 400000 HA. To investigate the influence of the already deposited multilayer on the thickness growth in the linear domain, we performed experiments where the (HA/PLL)n film construction began with a first couple of polyelectrolyte molecular weights and then, during the buildup, one of the polyelectrolytes was replaced by the same polyelectrolyte but of different molecular weight. Thus, two kinds of buildup experiments could be achieved by a change of either the PLL or the HA molecular weight. To investigate the influence of the PLL molecular weight change on the HA/ PLL film buildup by spraying, an (HA400/PLL360)20-(HA400/ PLL55)5 film has been built. Replacing the Mw ) 360000 PLL by a smaller one of Mw ) 55000 did not modify the film thickness increment (on the order of 25 nm/pair of layers) in the linear regime (see Table 1). This increment is the same as those determined for the “pure films” built with the different PLL molecular weights and the Mw ) 400000 HA. The experiments based on a change of the HA molecular weight had a different behavior. A film was first built with the Mw ) 130000 HA and then pursued with the Mw ) 400000 HA, and vice versa, for a fixed PLL molecular weight of 360000. Figure 2 represents typical results for such buildups. The thickness increments found before, ∆d1, and after, ∆d2, the transition are gathered in Table 1. One observes that the thickness increment for an (HA/PLL)n film is not influenced by the presence of the underlying (HA/PLL)n film built with a different HA molecular weight: the thickness increment is similar to the value found when the film is directly deposited on a bare silica substrate. Moreover, one can observe that the change of the thickness increment takes place as soon (29) 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.

Figure 2. Buildup of an HA/PLL film by spraying with a change of the HA molecular weight during the construction, (HA130/ PLL360)20-(HA400/PLL360)10. The slopes in the linear regime are respectively equal to ∆d1 ) 32.7 nm/pair of layers and ∆d2 ) 19.1 nm/pair of layers.

Figure 3. Buildup of an (HA130/PLL360)20-(HA400/PLL360)5 film by dipping with a change of the HA molecular weight during the construction. The slope does not change when HA130 is replaced by HA400. It is equal to 16.8 nm/pair of layers.

as the first HA is replaced by the second HA: there is no progressive smooth transition of the increment from ∆d1 to ∆d2 over several deposition steps. This is observed if one starts with a multilayer first constructed with the Mw ) 130000 HA and then pursued with the Mw ) 400000 HA; a similar observation is made if one uses the reverse buildup procedure. The above-reported results were obtained for films built with the spraying method. To check whether the buildup process is influenced by the deposition method, we also performed experiments with the more conventional dipping method. We built (HA/PLL)n multilayers by first using a Mw ) 130000 HA until the linear growth regime was reached. Then, we used a Mw ) 400000 HA to pursue the buildup process (Figure 3). The reverse experiment was also done (data not shown). In both cases we no longer observed a change of the thickness increment per pair of layers deposited. This indicates that the dependence of the thickness increment on the HA molecular weight for films built by spraying is in fact related to this deposition procedure. This can be explained by the presence of the drainage in the spraying deposition process, which does not exist in the conventional dipping method.22,26 The drainage induces a highvelocity gradient at the film/solution interface. Strong shear forces are applied on the polyelectrolytes which interact with the film.

1902 Langmuir, Vol. 23, No. 4, 2007

Porcel et al.

Table 2. Film Thickness Increments (∆d, nm/pair of layers) in the Linear Growth Domain for (PGA/PAH)n Films with Different Combinations of Polyelectrolyte Molecular Weightsa ∆d PGA44 PGA100

PGA44/PAH70 PGA100/PAH15 PGA100/PAH70

23.7 ( 0.6 24.3 ( 0.7 20.6 ( 0.5

a The standard deviations presented here correspond to the error in the determination of slopes calculated from a linear fit. The standard deviation corresponding to the reproducibility of the experiment is on the order of 3 nm/pair of layers.

Their strength increases with the size of the interacting polyelectrolytes and thus with their molecular weights. Since HA is a polymer with a large persistence length, its chains are expected to be more extended than those of PLL and more sensitive to the drainage effect. This explains the influence of the HA molecular weights on the linear growth regime. All these results were performed with the HA/PLL system. We performed similar experiments with (PGA/PAH)n multilayers, another strongly exponentially growing film.28 The films were, here too, constructed by the spraying method. The results are gathered in Table 2. For this system the thickness increments were almost independent of the molecular weights of both polyelectrolytes in the range investigated (44000 or 100000 for PGA and 15000 or 55000 for PAH). This observation was strengthened by an experiment in which a PGA/PAH multilayer was first built with Mw ) 15000 PGA and then pursued with a Mw ) 70000 PGA, the PAH molecular weight being kept constant (15000). No changes in the thickness increment could be observed after replacement of the first PGA molecular weight by another. From these experiments one can conclude that the thickness increment per deposited pair of layers in the linear regime for exponentially growing films is fairly independent of the polyelectrolyte molecular weight. In the case where a molecular weight effect was observed, it seemed to be related to the spraying procedure, and even for this method it was only observed when the molecular weight of HA was changed. Diffusion Behavior of PLL in HA/PLL Multilayers. Earlier studies devoted to (HA/PLL)n film buildup revealed that PLL chains, deposited on top of an HA/PLL multilayer, could penetrate into the whole architecture.19 This was observed by the use of fluorescently labeled PLL (the same molecular weight as that used to build up the films, namely a Mw ) 32000 PLL) deposited on top of (HA/PLL)n films, rendering them entirely and uniformly colored as found by CLSM observations. We performed here similar experiments using Mw ) 20000, 55000, and 360000 PLL and keeping the HA molecular weight fixed at 400000. Figure 4 shows that the film built with the Mw ) 20000 PLL becomes entirely and uniformly red once the PLL of the same molecular weight labeled with rhodamine, PLL20Rho, has been deposited on top of the film. The image was acquired 3 min after the deposition. This indicates that the Mw ) 20000 PLL chains can diffuse into the entire film during a dipping step of a few minutes. A similar result was obtained for the diffusion of a labeled Mw ) 55000 PLL in an HA/PLL film built with a PLL of similar molecular weight. On the other hand, in Figure 5A one observes that for a film constituted by 80 pairs of layers and built with the Mw ) 360000 PLL, only the upper 4 µm of the multilayer becomes green 3 min after the deposition of the Mw ) 360000 PLLFITC. PLL360FITC can no longer diffuse over the entire film thickness as demonstrated by the fluorescent profile determined 1 h after the deposition of the PLL360FITC layer (Figure 5A). The full film thickness could be estimated by depositing a PLL55Rho layer which could diffuse

over the entire film section in a few minutes. The thickness of the whole film was about 16 µm. This indicates that over a time scale on the order of several minutes (and up to 1 h), PLL360FITC diffuses only into an upper zone of the film. A further experiment concerning the diffusion of the Mw ) 360000 PLL has been performed. An additional (HA400/PLL360)60 film has been deposited on top of an (HA400/PLL360)60-HA400/PLL360FITC multilayer. In this case, mainly the central zone of the film appeared green as can be observed in Figure 5B, the image being taken 48 h after the deposition of the PLL360FITC layer and only a few hours after the last deposited layer of the film. The intensity profile indicates that, even if the diffusion of the embedded PLL360FITC is mainly restricted to a localized zone, the entire film becomes slightly green, indicating that the “none-diffusive” zone is not fully impermeable to PLL360FITC. This diffusion intensity profile remains unchanged even after one week. A similar fluorescent intensity profile is observed for a film on top of which PLL360FITC has been deposited for 1 day. The whole thickness of the film becomes slightly green, but a strong green strip of 4 µm thickness remains visible, indicating that, even if PLL360FITC can diffuse over the entire film, the diffusion process outside the upper 4 µm zone is slow. Thus, PLL360FITC diffuses rapidly only over a restricted upper zone of about 4 µm thickness. Finally, we also observe that when an (HA400/PLL360)n film is brought into contact with a Mw ) 20000 PLLFITC, the film becomes uniformly and entirely green after a few minutes, indicating that PLLFITC of low molecular weights penetrates rapidly into the whole film architecture built with the Mw ) 360000 PLL. In contrast, when the Mw ) 360000 PLLFITC is brought into contact with an (HA400/PLL)n film constructed with a lower molecular weight of PLL (20000 or 55000), only the upper part of the film becomes green, indicating that high molecular weight PLL does not penetrate into the whole film, even though it has been built with a PLL of lower molecular weight.

Discussion We investigate in this study the influence of the polyelectrolyte molecular weights on the buildup of “exponentially” growing PEM films. As already reported, the thickness of all the films first increases exponentially with the number of deposition steps before entering a linear growth regime.20,22 Kujawa et al. already investigated the influence of the polyelectrolyte molecular weights on the exponential growth phase for the hyaluronic acid/chitosan system.24 These authors found that polyelectrolytes of higher molecular weights led to thicker films after a given number of deposition steps. However, according to them, this is not due to the exponential growth rate, which is independent of the polyelectrolyte molecular weights, but rather to an earlier onset of the steep exponential growth phase when high polyelectrolyte molecular weights are used. In the present work, the analysis was not focused on the exponential growth phase but on the linear domain of the exponentially growing films. The exponential/linear transition was previously attributed to a film restructuration occurring after a given number of deposition steps which forbids the diffusion of the diffusing polyelectrolyte into a restructured zone at the bottom of the multilayer.20,22 We demonstrate here that the labeled PLL with the lowest molecular weights PLL (20000 and 55000) diffuse rapidly and uniformly into the whole thickness of (HA400/ PLLMw(PLL))n multilayer films, whereas the diffusion of a Mw ) 360000 PLL is always restricted to an upper zone about 4 µm in thickness. This result is observed whatever the PLL molecular weight used to build up the films and even when the labeled PLL

Polyelectrolyte Mw Influence on Multilayer Films

Langmuir, Vol. 23, No. 4, 2007 1903

Figure 4. CLSM observations (left panel) of a vertical section of a PEI-(HA400/PLL20)60-HA400/PLL20Rho film made by dipping and for polyelectrolyte solution concentrations of 3 × 10-3 M in monomer units (image size 76.8 µm × 15.6 µm). The white bar indicates the position of the glass slide. The fluorescent intensity profile of PLL20Rho is also presented (right panel).

Figure 5. CLSM observations of vertical sections of PEI-(HA400/PLL360)n-HA/PLLlabeled films made by dipping and for polyelectrolyte solution concentrations of 3 × 10-3 M in monomer units. The white bars indicate the position of the glass slide. (A) PEI-(HA400/PLL360)80HA400/PLL360FITC-HA400/PLL55Rho observed 1 h after the deposition of PLL55Rho (image size 76.8 µm × 27.9 µm). The fluorescent intensity profiles for both PLL360FITC and PLL55Rho are given on the right. (B) PEI-(HA400/PLL360)60-HA400/PLL360FITC-(HA400/PLL55)60 observed 48 h after the deposition of the PLL360FITC layer (image size 76.8 µm × 22.8 µm). The profile of the PLL360FITC fluorescent intensity has been followed over one week and is shown in the graph on the right.

layer is deposited on a film which has reached the linear regime. Finally, the thickness increment per pair of layers in the linear regime is independent of the PLL molecular weight. Considering also the observations made by Kujawa et al.24 for the exponential phase, one can conclude that, besides the buildup onset, the whole buildup process of exponentially growing films is, for the dipping method, independent of the polyelectrolyte molecular weights. The differences in the film thickness for a given number of deposition steps and for a given polyelectrolyte type but with different molecular weights are thus entirely due to a more or less rapid onset of the film buildup. With a low molecular weight labeled PLL, the film in the linear growth regime contains the labeled PLL in its whole thickness although we assumed that PLL does not diffuse in and out of zone II. This fact was previously explained by a “reptation/ exchange” mechanism of the PLL chains.22 This diffusion process of PLL chains involved in HA/PLL complexes which form the multilayer should take place at a constant local PLL concentration and thus at a constant local “fixed charge” density. The present data seem, however, to invalidate this statement. Indeed, the Mw

) 360000 PLL never diffuses over the entire thickness of an (HA400/PLL)n film for experimental time scales of a few minutes to a few hours, even through a reptation/exchange process and whatever the PLL molecular weight used to build the film. This implies that, if a PLL of low molecular weight is deposited on top of an HA/PLL film built with high molecular weight PLL and if it diffuses freely only into zone III, it cannot diffuse through a reptation/exchange process over the entire film (in particular zone II) since this would require simultaneously a reptation/ exchange process of high PLL molecular weight outside of this domain. Thus, the homogeneous diffusion of low molecular weight PLL over the whole (HA400/PLL360)n film previously observed must be due to a diffusion process of freely diffusing low molecular weight PLL during the contact time of the film with the PLL solution accompanied by the diffusion outside of this compartment of small charged ions to maintain the “fixed charge density” constant. The PLL of low molecular weights should thus diffuse in and out of the whole film during each deposition step and thus contribute to an exponential growth regime of the film. However, when PLL of Mw ) 360000 is

1904 Langmuir, Vol. 23, No. 4, 2007

replaced by PLL of Mw ) 55000 during the construction of an HA400/PLL film in the linear growth regime, the film thickness increment per pair of layers deposited remains unchanged despite the differences between their freely diffusing domains. This raises the problem of the origin of the linear growth regime of exponentially growing films. For all the constructions corresponding to the different couples of polyelectrolytes with different molecular weights, the (HA/ PLL)n films always enter the linear growth regime when their thickness (in the wet state) reaches a critical value in the range of 0.5-1 µm. This thickness is also much smaller than the 4 µm zone over which the high molecular weight PLL is supposed to freely diffuse into the films. So, even though the diffusion domain of the high molecular weight PLL is restricted to this part of the film, it is still too large to explain the position of the transition between the exponential and the linear growth regimes. The results concerning the diffusion behavior of PLL with different molecular weights in the HA/PLL films lead to the conclusion that the thickness increment per pair of layers seems to be independent of the thickness of the zone in the film in which the PLL freely diffuses once the diffusion zone has reached a critical thickness, which is on the order of 0.5-1 µm. This can then be explained in the following way: when an (HA/PLL)n film ending in PLL is brought into contact with an HA solution, the HA chains can interact with an upper zone of the film which is rich in PLL and whose HA/PLL complexes do not interact as strongly as those of the central part of the film. This upper “loose” zone would correspond to zone III in the three-zone model, and its extension would be on the order of 0.5-1 µm for the HA/PLL system. The free PLL chains of this upper zone could diffuse freely in and out of the multilayer and be complexed by the HA chains from the solution. However, not all these complexes interact strongly enough to remain attached to the film surface. Only a fraction of them participate in the construction of the new layer on top of the film. This could constitute a reason explaining why the PLL diffusion zone is decoupled from the thickness increment per pair of layers. As long as the film has not reached a thickness equal to the extension of zone III, the HA chains from the solution interact with the whole film and the amount of HA chains which participate in the mass increase of the film is proportional to this thickness. This leads to the initial exponential growth. A similar

Porcel et al.

argument holds for the PGA/PAH films whose thickness increments are also independent of the molecular weights of both polyelectrolytes. However, even though the in and out diffusion process no longer entirely drives the film growth process in the linear regime, it probably still exists. Moreover, the presence of the limited diffusion zone of the high molecular weight PLL, even when the film is built with a low molecular weight PLL, shows the inhomogeneity of the film structure. The constant in and out diffusion process can gradually induce a film restructuration which progressively renders PLL diffusion more and more difficult. We suggested previously two origins for this hindrance: (i) a gradual increase of the Donnan potential due to a progressive increase of positive fixed local charges or (ii) a gradual densification of the film.22 The experiments performed here allow the first possibility to be ruled out. Indeed, low molecular weight PLL can diffuse over the entire film, and the Donnan potential should apply equally on low and high molecular weight PLL. To summarize, the three-zone model previously presented appears to remain valid, but the buildup process is no longer based on the polyelectrolyte diffusion in and out of the film but on the interaction between HA and an entire zone (zone III) in the upper part of the film, this zone being constituted of PLL/HA complexes which are “loosely bound” and rich in PLL. Moreover, there are still free PLL chains present in the remaining part of the film, but all of these molecules do not participate “actively” in the film formation, possibly because when they diffuse out of the film, they form complexes with HA and these complexes do not remain anchored strongly enough on the top of the film. However, the intimate mechanism leading to the formation of the new additional layer remains still unknown. Thus, is it PLL which diffuses out of the upper zone and is complexed by HA, is it HA that, during the interaction with the excess PLL, leads to a surface remodeling and interaction with the upper film zone, or is it a combination of both processes or another mechanism? The present data can still not answer this question, which remains open. Acknowledgment. This work was supported by ACI “Nanoscience” (Project No. NR204). LA062728K