Equilibrium Aspects of Polycation Adsorption on Silica Surface: How

Lauriane AlisonAhmet F. DemirörsElena TervoortAlexandra TelekiJan ... Wacklin , Kell Mortensen , Lise Arleth , Robert Feidenhans'l , and Marité Cár...
0 downloads 0 Views 300KB Size
Download PDF
5872

Langmuir 2005, 21, 5872-5881

Equilibrium Aspects of Polycation Adsorption on Silica Surface: How the Adsorbed Layer Responds to Changes in Bulk Solution† Yulia Samoshina,‡ Tommy Nylander,*,‡ Victor Shubin,§ Rogert Bauer,⊥,† and Krister Eskilsson# Physical Chemistry 1, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden; Colloid Chemistry Department, Chemical Faculty, Saint-Petersburg State University, Universitetsky pr. 2, St. Petersburg, 198904 Russia; Institute for Mathematics and Physics, KVL, Thorvaldsensvej 40, DK-1871, Frederiksberg C, Denmark; and Kemira Kemi AB, Box 902, SE-251 09 Helsingborg, Sweden Received January 9, 2005. In Final Form: April 13, 2005 Adsorption of cationic high molecular weight polyacrylamides (CPAM) (Mw is about 800 kDa) with different fractions of cationic units τ ) 0.09 and τ ) 0.018 onto silica surface was studied over a wide range of pH (4-9) and KCl concentration (cs ) 10-3-10-1 M) by in-situ null ellipsometry. We discuss how the adsorbed layer depends on the bulk conditions as well as kinetically responds to changes in solution conditions. The adsorbed amount Γ of CPAM increases with pH for all studied electrolyte concentrations until a plateau Γ is reached at pH > 6. At low pH we observed an increase in adsorbed amount with electrolyte concentration. At high pH there is no remarkable influence of added salt on the values of the adsorbed amount. The thickness of adsorbed polymer layers, obtained by ellipsometry, increases with electrolyte concentration and decreases with pH. At low cs and high pH the polyelectrolyte adsorbs in a flat conformation. An overcompensation of the surface charge (charge reversal) by the adsorbed polyelectrolyte is observed at high cs and low pH. To reveal the reversibility of the polyelectrolyte adsorption with respect to the adsorbed amount and layer thickness, parameters such as polyelectrolyte concentration (cp), cs, and pH were changed during the experiment. Generally, similar adsorbed layer properties were obtained independent of whether adsorption was obtained directly to initially bare surface or by changing pH, cs, or the concentration of polyelectrolyte solution in the presence of a preadsorbed layer, provided that the coverage of the preadsorbed layer was low. Once a steady state of the measured parameters (Γ, d) was reached, experimental conditions were restored to the original values and corresponding changes in Γ and adsorbed layer thickness were recorded. For initially low surface coverage it was impossible to restore the layer properties, and in this case we always ended up with higher coverage than the initial values. For initial high surface coverage it was usually possible to restore the initial layer properties. Thus, we concluded that polyelectrolyte appears only partially reversible to changes in the solution conditions due the slow rearrangement process within the adsorbed layer.

Introduction Cationic polyelectrolytes are commonly used in applications such as water treatment as flocculants and papermaking as retention agents.1-3 Adsorption of polyelectrolytes on solids is determined by a number of factors, such as the nature and the charge of the surface, the charge density of the polyion, the molecular weight and concentration of the polymer, the salt concentration, and nonelectrostatic interactions of the macromolecules with the surface and with each other. The effects may have a complex interdependence on each other, but progress in both in theoretical modeling and phenomenological knowledge of polymer adsorption at solid/liquid interface allows † This paper is dedicated to the memory of Professor Rogert Bauer, who died in May 2004. We are grateful for the enlightening and inspiring discussions with him, being such a great scientist and human. ‡ Lund University. § Saint-Petersburg State University. ⊥ Institute for Mathematics and Physics, KVL. # Kemira Kemi AB.

(1) Schwoyer, W. L. K., Ed. Polyelectrolytes for Water and Wastewater Treatment; CRC Press: Boca Raton, FL, 1981. (2) Napper, D. H. Polymeric Stabilization of Colloidal Dispersions; Academic Press: New York, 1983. (3) Swerin, A.; Odberg, L. Fundamentals of Papermaking Materials, Transactions of the Fundamental Research Symposium, 11th, Cambridge, UK, Sept. 1997 1997, 1, 265-350.

us to understand or even predict the influence of pH, ionic strength, and molecular weight on the “equilibrium” adsorption state.4 Nonetheless, the question of reversibility is still open,5-7 and it is generally believed that a large fraction of remaining discrepancies between experimental and theoretical results are due to the nonequilibrium behavior of real systems. For adsorption of high molecular weight polyelectrolytes on oppositely charged surfaces it is accepted that equilibrium is seldom, if ever, achieved. Hence, kinetics of adsorption and the control of the adsorption by regulating the electrostatic interactions have received more attention lately.6 Indeed, complementary studies have indicated that the polymer adsorption process is in many cases of practical relevance, kinetically controlled with the polymer molecules arrested in a “nonequilibrium” adsorption state. These dynamical aspects of polymer adsorption are, however, not yet fully understood. Further assessment of the kinetic barriers for achieving equilibrium adsorption of polyelectrolytes at charged (4) Fleer, J. I.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman & Hall: London, 1993. (5) Shubin, V.; Linse, P. Macromolecules 1997, 30, 5944-5952. (6) Gobel, J. G.; Besseling, N. A. M.; Stuart, M. A. C.; Poncet, C. J. Colloid Interface Sci. 1999, 209, 129-135. (7) Geffroy, C.; Labeau, M. P.; Wong, K.; Cabane, B.; Stuart, M. A. C. Colloids Surf. A 2000, 172, 47-56.

10.1021/la050069q CCC: $30.25 © 2005 American Chemical Society Published on Web 05/24/2005

Polycation Adsorption on Silica Surface

surfaces from solution can be obtained by determining how the adsorbed layer structure responds to changes in bulk solution.8,9 In many applications, e.g., hair conditioners, these dynamic phenomena can be utilized to achieve a certain layer structure at the interface, but the full potential of the dynamic aspects of polymer adsorption has so far not been exploited. A variety of experimental techniques10 have been developed and applied to investigate the polymer adsorption phenomena at solid-liquid interfaces. Among these methods used for determination of adsorbed layer thickness are ellipsometry,11 neutron reflectivity,12 and optical waveguide light-mode spectroscopy,13,14 which is based on the confinement of light in a high refractive index layer. The latter technique allows independent determination of the refractive index and the thickness of the adsorbed layer. In ellipsometry, however, the problem of a separation between film thickness and film refractive index can be overcome by choosing the suitable substrate (30 nm silica oxide layer on the top of the silicon wafer) and by proper optical characterization of the substrate in two ambient media. This approach has previously been demonstrated in a number of studies.15,16 For adsorption of macromolecules, optical waveguide light-mode spectroscopy and ellipsometry were shown to provide comparable results.17 The output ellipsometry data ordinarily only permit the determination of mean layer thickness, which is obtained in the frame of a homogeneous slab model of the adsorbed layer. One of the advantages of ellipsometry is that it is effective in studying fast adsorption processes. The inhomogienity of the adsorbed layer in the normal direction can be revealed by, for instance, neutron reflectivity measurements, but in comparison with ellipsometry, neutron reflection in general does not allow studying fast dynamic processes. In this study the objective was to determine how the adsorbed layer is affected by the way we approach a certain solution condition, i.e., cycling and stepwise changes of pH and electrolyte concentration and stepwise increases in polymer concentration. We report the adsorption kinetics of a polyelectrolyte with 9% molar charge density onto silica surface under different solution conditions (pH, salt concentration cs, polymer concentration cp). Important characteristics of the adsorption processsadsorbed amount and layer thicknessswere accessed by in-situ null ellipsometry. This technique allows measurements with a time resolution of a few seconds, which is crucial for the purpose of our study. Since the silica surface charge is pH- and cs-dependent, the experiments were conducted under thoroughly controlled conditions. Cycling of pH, cs, and cp allowed us to estimate the degree of reversibility of the polyelectrolyte adsorption. (8) Terada, E.; Samoshina, Y.; Nylander, T.; Lindman, B. Langmuir 2004, 20, 1753-1762. (9) Notley, S. M.; Biggs, S.; Craig, V. S. J.; Wagberg, L. Phys. Chem. Chem. Phys. 2004, 6, 2379-2386. (10) Kawaguchi, M.; Takahashi, A. Adv. Colloid Interface Sci. 1992, 37, 219-317. (11) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light; North-Holland: Amsterdam, 1989. (12) Tiberg, F.; Nylander, T.; Su, T. J.; Lu, J. R.; Thomas, R. K. Biomacromolecules 2001, 2, 844-850. (13) Ramsden, J. J.; Lvov, Y. M.; Decher, G. Thin Solid Films 1995, 254, 246-251. (14) Ramsden, J. J.; Lvov, Y. M.; Decher, G. Thin Solid Films 1995, 261, 343-344. (15) Tiberg, F.; Landgren, M. Langmuir 1993, 9, 927-932. (16) Landgren, M.; Jonsson, B. J. Phys. Chem. 1993, 97, 1656-1660. (17) Hook, F.; Voros, J.; Rodahl, M.; Kurrat, R.; Boni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. Colloids Surf. B 2002, 24, 155-170.

Langmuir, Vol. 21, No. 13, 2005 5873

Experimental Section Materials and Chemicals. Polymers. Cationic polyacrylamides (CPAM) used in this study are copolymers of acrylamide (AM) and (3-(methacrylamido)propyl)trimethylammonium chloride (MAPTAC) with mean molecular weights Mw ) 800 kDa and different fractions of MAPTAC units τ ) 0.09 and τ ) 0.018. MAPTAC has a quaternary amine group; i.e., the charge density does not depend on pH. Furthermore, the polymer is stable against hydrolyses up to pH 9. The CPAM samples were synthesized at Laboratorie de Physico-chimie macromoleculare, Universite Pierre et Marie Curie, Paris. Polymers were used without further purification. The fractions of cationic MAPTAC units were determined by polyelectrolyte titration using an anionic polymer, potassium poly(vinyl sulfate), KPVS, and cationic o-toluidine blue, OTB, as indicator. Polymer stock solutions were prepared by dissolving of the polymer powder in water to a final polymer concentration of 1000 ppm. The solutions were stored no longer than a week at t ) 10 °C before use. The water used in all ellipsometry experiments was deionized and passed through a Milli-Q filtration system (Millipore) with a pH ≈ 5.6 ( 0.2. Analytical grade inorganic electrolytes from Merck were used without purification. Silica Particles. A sample of nonporous silica was kindly provided by Nissan Chemicals Ltd. (Japan). The sample was centrifugated to remove any smaller particles. The collected particles were redispersed in water, and the diameter of the now rather monodisperse silica beads was determined by transmission electron microscopy to 3000 Å, as specified by the manufacturer. The sample was then treated with 0.1 M HCl and washing with double-distilled water in order to further clean the sample and to protonate the -OH surface groups. The washing was done via consecutive centrifugation-decantation-resuspension cycles until neutral pH was reached. Finally, the sample was ultrasonicated and then kept as 5% suspension in water. Ellipsometry. The ellipsometry measurements were performed with a modified, automated Rudolph Research thin-film null ellipsometer, model 43603-200E, equipped with highprecision step motors and controlled by a personal computer. A detailed description of the instrument is given by Landgren and Jo¨nsson.15,16 In the ellipsometric measurements, polished Si/ SiO2 wafers were used as substrates. The wafers had a mean SiO2 thickness of 300 Å. They were cleaned in a two-step process as described elsewhere.10 To obtain accurate measurements of the thickness and surface excess, it is necessary to know the optical properties of the substrate. The optical properties of the oxidized silicon substrate were obtained by measuring in two ambient media, in air and electrolyte solution.15,16 Hereby, the complex refractive index (N2 ) n2 + jk2) of the bulk silicon and the thickness (d1) and the refractive index (N1 ) n1) of the silica layer were obtained. All measurements were conducted at a wavelength of 4015 Å and an angle of incidence on the surface set at 68°. The experiments were performed in situ using a 5 mL trapezoidal cuvette of optical glass, thermostated at 25 ( 0.1 °C. Agitation was achieved by means of a magnetic stirrer at 300 rpm. The solution in the cuvette could be changed under well-defined flow rate without emptying the cuvette, by means of a multichannel peristaltic pump. The cuvette was also equipped with a miniature glass electrode, and the pH was carefully adjusted by adding KOH or HCl before and, if necessary, during the experiments. A known volume of the concentrated polymer solution (cp ) 1000 ppm) was injected with a micropipet in order to obtain the desired final concentration of the polymer in the cell. After injecting the stock solution, the changes in Ψ and ∆ in one zone were recorded as a function of time. These values of Ψ and ∆ were corrected for optical imperfections using the data from the four zone measurements. The measurements were conducted until no further changes in ellipsometric angles were detected. In the cycling experiments, the electrolyte concentration, polymer concentration, or pH was changed by flushing the cuvette with the appropriate aqueous solution. Once a steady state of the measured parameters (Γ, d) was reached, the original experimental conditions were restored by again changing the bulk solution, and the response in Γ and d was recorded. In all cases the exchange to an appropriate polymer or polymer-free salt solution was performed at a continuous flow 20 mL/min.

5874

Langmuir, Vol. 21, No. 13, 2005

Samoshina et al.

The results of Ψ and ∆ measurements were interpreted within the framework of an optical four-layer model, assuming isotropic media and planar interfaces. The mean refractive index nf and the average thickness df of the adsorbed layer were calculated as described previously.15 From these values the adsorbed amount was calculated according to de Feijter et al.:18

Γ)

(nf - n0)df dn/dc

(1)

where Γ is the adsorbed amount, mg m-2, and n0 is the refractive index of the ambient bulk solution. Here we used the refractive index increment value dn/dc ) 0.169 for the polymer according to Shubin.19 Potentiometric Titration. Surface charge density of silica particles was obtained by conventional acid-base potentiometric titration. The details of the experimental setup and evaluation of the data are described previously.20,21 In the absence of polyelectrolyte, titration was carried out using a “dynamic” procedure where aliquots of acid/base were added every 2-3 min. In the presence of adsorbing polyelectrolyte, however, a “quasi-static” procedure was adopted, where pH was adjusted stepwise and was maintained at a desired level by continuous addition of acid/base for 2-3 min. The time lag between consecutive changes of pH in this case was about 10 min. Static Light Scattering. The stock solutions of CPAM were dissolved in water to a concentration of 10 g/L and equilibrated for at least 24 h. These were diluted to 1 g/L and were kept for 20 min at 22 °C before injection into the HPLC system. A LC10AD HPLC pump from Shimadzu and a Rheodyne injector was coupled to a Superdex 200 column (7.8 × 300 mm, from Amersham-Pharmacia). After the column, the liquid passed through a static light scattering instrument with a laser operating at 6900 Å (Dawn EOS equipped with a K2 flow cell, Wyatt Technology, Santa Barbara, CA) followed by a refractive index detector and a UV detector (RID-10A and MHC10A, Shimadzu, Japan). The elution buffer was 10 mM NaHPO4/H2PO4, pH 7, containing 0.15 M NaCl. A 500 µL injection loop was used. The flow rate was 0.5 mL/min. All column runs were performed at ambient temperature. Bovine serum albumin from Sigma, having a molar mass of 66.8 kg/mol for the monomer, was applied to the column as 1 g/L in order to calibrate the instrument. The concentration of the species appearing inthe elution profiles was calculated from the refractive index using an incremental index of refraction of 1.69 × 10-4 g/L. In order to estimate the molar mass and radius of gyration for the aggregates, a Berry plot, i.e., the inverse square root of the light scattering signal vs q2, was made. (The light intensities were measured at 16 angles, out of which 14 were used in a q range of 5-24 µm-1.) The radius of gyration was derived from the slope of this line and the molar mass from the intercept with the vertical axis.

Results Surface Charge Density of Silica and the Effects of Polyelectrolyte Adsorption. To get a picture of the charge stoichiometry upon polyelectrolyte adsorption, surface charge density measurements were performed. The surface charge densities (σ0) of the bare silica particles, measured over a wide range of pH and KCL concentrations, are presented in Figure 1. The obtained data are characteristic for silica, and the observed charge densities agree well with literature values for nonporous silica.22,23 A principal mechanism by which silica surface (amorphous nonporous silica beads or oxidized silicon wafers) acquires the charge in contact with KCl solution is the dissociation (18) De Feijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 1759-1772. (19) Shubin, V. Langmuir 1994, 10, 1093-1100. (20) Shubin, V.; Samoshina, Y.; Menshikova, A.; Evseeva, T. Colloid Polym. Sci. 1997, 275, 655-660. (21) Shubin, V. J. Colloid Interface Sci. 1997, 191, 372-377. (22) Bolt, G. H. J. Phys. Chem. 1957, 61, 1166-1169. (23) Iler, R. K. The Chemistry of Silica; Wiley: New York, 1979.

Figure 1. Surface charge density of the bare silica particles and in the presence of adsorbed CPAM with τ ) 0.09 and with τ ) 0.018 as a function of pH at different electrolyte concentrations: (a) at 0.001 M KCl; (b) at 0.01 M KCl; (c) at 0.1 M KCl.

of silanol groups and adsorption of charge-determining ions, such as OH- ions. The surface charge is measurable at around pH 5, and thereafter σ0 increases progressively with pH. The effect of the background electrolyte concentration on σ0 is due to screening of the surface charges, which in turn leads to a decrease in the surface potential term and a lowering of the effective pKa of the silanol groups. The addition of polyelectrolyte significantly increases the surface charge density of silica (Figure 1), and the effect is mostly pronounced at low salt concentration. The increase in charge density in the presence of CPAM is expected since adsorption of polyelectrolyte at the interface affects the dissociation of the silanol groups and hence the surface charge density. The magnitude of this effect is determined by the concentration of cationic species near the surface. Therefore, we observed that polyelectrolyte with τ ) 0.09 has a more pronounced effect on σ0 compared to the polyelectrolyte with τ ) 0.018 at the same pH and salt concentration.

Polycation Adsorption on Silica Surface

Langmuir, Vol. 21, No. 13, 2005 5875

Figure 3. Adsorption isotherm of CPAM (τ ) 9%). pH ) 5.6, cs ) 0.1 M, T ) 25 °C.

Figure 2. Concentration, radius of gyration (a), and molar mass (b) profiles as a function of the retention time for CPAM solution with τ ) 0.09 applied to the HPLC system at a concentration of 1 g/L. Lower curves represent relative concentration.

Moreover, the influence of polyelectrolyte charge density is stronger at low ionic strength and high pH. At low ionic strength the polymer adsorbs in a relatively flat interfacial conformation with a maximum number of segments in contact with the surface. As the electrolyte concentration increases, the screening of the electrostatic attraction between the charged segments and the surface as well as electrostatic repulsion within the adsorbed layer increases progressively. Moreover, competitive adsorption of the counterions increases. These factors lead to a looser attachment of polyelectrolyte molecules to the surface and, consequently, to a weaker influence on the number of titratable silanol groups. As observed in Figure 1, the effect of polyelectrolyte on σ0 decreases with increasing electrolyte concentration and decreasing polyelectrolyte charge density. Furthermore, nonelectrostatic interactions of polyelectrolyte segments with the surface may not be negligible even if they are weak, if the segment concentration near the surface is high enough. Polydispersity of the polymer sample In Figure 2 the results from static light scattering experiments are shown. The concentration, molar mass, and radius of gyration profiles as a function of retention time were similar for the two samples as expected due to the relatively high ionic strength (0.15 M NaCl). Therefore, only the results for the CPAM sample with τ ) 0.09 are shown in Figure 2a,b. The polymer is polydisperse with mass ranging from 500 to 1400 kg/mol and radius of gyration ranging from 500 to 800 Å. This might contribute to the slow adsorption kinetics observed under certain conditions in the ellipsometric experiments. We, however, always compare results from the same stage of the adsorption process. For instance, the change of solution conditions is achieved by flushing the cuvette with a fresh polymer solution adjusted to the new experimental conditions, e.g., pH and ionic strength. Note also that the

Figure 4. Adsorbed amount of CPAM with τ ) 0.09 on silica surface as a function of electrolyte concentration at different values of bulk equilibrium at pH ) 4 (O), 5 (1), 5.6 (4), 8 ([), and 9 (0) and with τ ) 0.018 at pH ) 5.6 (×).

surface-to-volume ratio is very small so any fractionation at the surface will hardly affect the composition in the solution. Adsorption Isotherm. Figure 3 shows the adsorption isotherm of the polymer with τ ) 0.09 on silica at cs ) 0.1 M and pH ) 5.6. The isotherm shows that the polymer has high affinity toward the oppositely charged surface. The isotherm reaches an adsorption plateau at 10 ppm bulk polymer concentration. The adsorption isotherm in Figure 3 was obtained from separate experiments. It is, however, important to note that a similar isotherm is obtained when the polymer concentration was increased sequentially. In all experiments discussed below, the polymer concentration was, unless stated otherwise, 50 ppm, which is well into the plateau region of the isotherm. The plateau values of the adsorbed amount and adsorbed layer thickness for the polymer with τ ) 0.09 are for a wide range of electrolyte concentration and pH summarized in Figures 4 and 5. For comparison, data for adsorption of the polymer with τ ) 0.018 at pH ) 5.6 are also presented. The presented values are the mean of results from at least four measurements with a standard deviation of 10%. Numerous experimental studies on the influence of salt concentration and pH on adsorption of polyelectrolytes onto various surfaces have given a good picture as to the adsorption mechanism (cf. refs 4 and 24). In general, the polyelectrolyte will attempt to neutralize any opposite charge present on the surface. Furthermore, the adsorp-

5876

Langmuir, Vol. 21, No. 13, 2005

Figure 5. Average layer thickness of CPAM with τ ) 0.09 on silica surface as a function of electrolyte concentration at different values of bulk equilibrium at pH ) 4 (O), 5.6 (4), and 9 (0) and with τ ) 0.018 at pH ) 5.6 (×).

tion of the polyelectrolyte at the interface can lead to an overcompensation of the surface charge and hence charge reversal. Adsorbed polyelectrolytes are generally oriented parallel to the surface in very thin layers, with a thickness often less than the radius of gyration of the polymer. In our case, the radii of gyration under the given experimental conditions are 500-800 Å as determined by static light scattering (see Figure 2), while the adsorbed layer thickness is about 100-300 Å depending on the solution conditions. The concentration of salt is important, and added electrolyte can either promote or hamper adsorption of the polyelectrolyte. This is illustrated in Figure 4, where the adsorbed amount is plotted as a function of salt concentration at different pH. At low pH (4-5.6), the adsorbed amount increases gradually with cs. At high pH (8-9) we observed that the adsorbed amount is rather insensitive to the electrolyte concentration, although the values at high ionic strength are slightly lower than those at low ionic strength. As discussed above, the charge density of the silica surface is high at these pH’s (cf. Figure 1), and this implies a stronger interaction of the polymer with the surface. For the whole studied pH range, the thickness of the layer increases with the KCl concentration (Figure 5). At low cs the polyelectrolyte molecule tends to adsorb at the surface in a flat conformation, forming numerous contacts with the surface. At higher salt content the electrostatic attraction between charged polymer segments and the surface is screened, and the cations compete for the surface sites. This results in a more extended conformation of the polymer at the surface with fewer points of contacts. The average layer thickness of adsorbed polyelectrolyte also decreases with increasing pH or surface charge density. This is a reflection of the fact that the surface charge compensation occurs in a thin layer adjacent to the surface. The high density of the polymer layer at pH 8-9 even at low salt concentration is obvious when comparing the amount adsorbed (Figure 4) and thickness of the adsorbed layer (Figure 5). This will in turn increase the electrostatic and steric barrier for adsorption, and therefore no further changes in adsorbed amount are observed when pH is increasing from 8 to 9 at constant salt concentration. Although the adsorbed amount gradually increases with pH, i.e., surface charge density, for the whole range of salt concentrations (Figure (24) Lyklema, J. Fundamentals of Interface and Colloid Science; Academic Press: London, 1995; Vol. II.

Samoshina et al.

4), it reaches a plateau at pH > 6. Further neutralization of surface charges when pH increases can be achieved by collapse of the polyion onto the surface since surface charge compensation occurs in a thin layer adjacent to the surface. This is indirectly confirmed by the fact that layer thickness at higher pH is the smallest one (Figure 5). High ionic strength also gives a more efficient screening of the surface charges, and one would therefore expect that the driving force for adsorption of an oppositely charged polyelectrolyte is reduced. The polyelectrolyte has to compete with a larger number of small ions, which implies that under certain conditions desorption of the polymer can occur.25 This screening-reduced regime was not observed for our system. In fact, the opposite was observed, and the adsorbed amount at pH 5.6 in the presence of 1 M KCl is higher (1.05 mg/m2, data not shown) than at 0.1 M salt. The effect of the polyelectrolyte charge density on the adsorbed amount and the layer thickness has also been studied. Some of the results for the polymers with τ ) 0.09 and τ ) 0.018 are given in Figures 2 and 3 for different electrolyte concentration at pH ) 5.6. The adsorbed amount increases only slightly with decreasing charge density at a given electrolyte concentration, while the corresponding increase in thickness is substantial at low ionic strength. Kinetics of Adsorption. The time-resolved measurements allows us to follow the kinetics of the adsorption process, both in terms of amount adsorbed, Γ, and adsorbed layer thickness, df, and typical data are shown in Figure 10. The first part of the kinetics curve is linear with a high rate of adsorption, followed by a reduction of the adsorption rate until steady state is reached. Polymer adsorption can generally be regarded as a three-step process, consisting of the transport from the bulk to the surface, attachment to the surface, and rearrangement in the surface layer. The attachment process can be limited by an electrostatic and steric barrier, build up by previously adsorbed macromolecules. The rearrangement process, in fact, involves optimizing polymer segment-surface contacts and takes place on a much longer time scale than the initial mass accumulation.26,27 Consequently, the adsorbed polymer is kinetically trapped at the interface, and therefore polymer adsorption often appears to be irreversible.5,7,28 For diffusion-controlled adsorption the initial rate of adsorption dΓ/dt can in the simplest case be modeled as29

dΓ D ) cp dt δ

(2)

Here D is the diffusion constant, and δ is the thickness of the unstirred layer, which for the used cuvette geometry and stirring rate has been estimated to 100 µm.30 The initial adsorption rate typically observed in our experiments for cp ) 50 ppm is about 0.005 ( 0.001 mg/(m2 s). For a polyacrylamide homopolymer of similar molecular weight the diffusion coefficient has been estimated to 9.4 × 10-12 m2/s,31 which together with δ ) 100 µm (25) Van de Steeg, H. G. M.; Cohen Stuart, M. A.; De Keizer, A.; Bijsterbosch, B. H. Langmuir 1992, 8, 2538-2546. (26) Stuart, M. A. C.; Hoogendam, C. W.; deKeizer, A. J. Phys.: Condens. Matter 1997, 9, 7767-7783. (27) Ka¨llrot, N. Personal communication, 2004. (28) van de Ven, T. G. M. Adv. Colloid Interface Sci. 1994, 48, 121140. (29) Cuypers, P. A.; Willems, G. M.; Kop, J. M. M.; Corsel, J. W.; Janssen, M. P.; Hermens, W. T. ACS Symp. Ser. 1987, 343, 208-221. (30) Tiberg, F.; Joensson, B.; Lindman, B. Langmuir 1994, 10, 37143722.

Polycation Adsorption on Silica Surface

corresponds to a diffusion-controlled adsorption rate of 0.0047 mg/(m2 s) calculated according to eq 2. This simplified analysis gives a value similar to the measured value and indicates that the initial adsorption process is diffusion-controlled. This has also been found for adsorption of several other polyelectrolytes on oppositely charged surfaces.32 Recent Brownian dynamics simulation of the adsorption of a single uncharged polymer molecule at the surface have indicated that the attachment to the surface occurs very fast compared to the “spreading” process, as determined by the radius of gyration in the plane parallel to the flat surface.27 As the polymer surface coverage increases, the rate of adsorption at a certain point also becomes dependent on the barrier created by already adsorbed polymers, and the rate of adsorption decreases to finally approach zero at surface saturation. It is worth noting that in many cases we observed that even if there is still further accumulation of mass on the surface, no increase in the thickness of the adsorbed layer is observed, and therefore the density of the adsorbed layer will increase. Reversibility of the Polyelectrolyte Adsorption upon Dilution. The reversibility of the polyelectrolyte adsorption upon dilution was investigated by exchanging the polymer solution with corresponding salt solution without polymer after steady state of adsorption had been reached. The extent of desorption was found to be negligible under all conditions investigated (data not shown), which is generally observed for polymer adsorption. Here we also note that the adsorption isotherm obtained by stepwise increase in the polymer concentration, where steady state appeared to be reached before the next aliquot of polymer solution, is similar to the one obtained from separate experiments for each concentration. Dijt showed that desorption is extremely low for polymers because they often adsorb according to a highaffinity adsorption isotherm.33 Because of the high affinity, the concentration of free polymer near the surface, in equilibrium with any adsorbed amount below the saturation, is extremely low, even when the adsorption is completely reversible. These free molecules must diffuse away from the surface before more polymer molecules can desorb. The rate of desorption is therefore limited by the transport step. Reversibility of the Polyelectrolyte Adsorption when Changing Ionic Strength and pH. We studied the reversibility of polyelectrolyte adsorption by changing the experimental conditions (cs- and pH-cycling). In this way we moved the system away from equilibrium, after which the system was restored to the initial conditions and the response in adsorbed amount and layer thickness was monitored. The kinetics and the extent to which the adsorbed layer relaxes back to the original state are a measure of the reversibility of adsorption. Only one parameter was changed at a time; i.e., upon changing pH, the ionic strength and the polymer concentration were kept constant. For change in pH at constant ionic strength, the conformation of the polyelectrolyte in bulk remains constant, while the surface charge density of the surface varies (Figure 1). However, changing the ionic strength (31) Ying, Q. C.; Wu, G. W.; Chu, B.; Farinato, R.; Jackson, L. Macromolecules 1996, 29, 4646-4654. (32) Hoogeveen, N. G.; Stuart, M. A. C.; Fleer, G. J. J. Colloid Interface Sci. 1996, 182, 133-145. (33) Dijt, J. C.; Stuart, M. A. C.; Fleer, G. J. Macromolecules 1992, 25, 5416-5423.

Langmuir, Vol. 21, No. 13, 2005 5877

Figure 6. Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at cs ) 0.1 M and cp ) 50 ppm. The black arrows on the time scale indicate the changing of bulk pH from 4 to 9 and from 9 back to 4.

Figure 7. Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at cs ) 0.1 M and cp ) 50 ppm. The black arrows on the time scale indicate the changing of bulk pH from 9 to 4 and from 4 back to 9.

will affect the polymer intra- and intermolecular interaction as well as the charge density of the surface (Figure 1). Changing the pH at High Ionic Strength. Figure 6 shows the adsorption from a polymer solution containing 0.1 M KCl at pH 4 and the response when increasing the pH from 4 to 9 and then restoring the pH back to 4. The adsorbed amount and the adsorbed layer thickness corresponding to the adsorption onto the bare silica at pH ) 4 and 0.1 M KCl are 0.42 mg/m2 and 320 Å, respectively (Figure 4). When pH increased (region II), the adsorbed amount increases to 1.3 mg/m2. This is accompanied by a shrinking of the adsorbed layer (to 250 Å). When pH is restored back to 4 (region III), the adsorbed layer respond by polymer desorption and swelling. We note that both the values of the adsorbed amount and the thickness are larger than for the adsorption to the bare surface (region I). Figure 7 shows the adsorption from a polymer solution containing 0.1 M KCl, where the changes in pH are the reverse from Figure 6. Initially, the pH is 9, and the response when decreasing the pH to 4 and then restoring the pH to 9 is shown. The adsorbed amount and the adsorbed layer thickness corresponding to the adsorption onto the bare silica at pH ) 9 are about 1.3 mg/m2 and 250 Å, respectively. This is similar values as obtained if the polymer had first been adsorbed from pH 4 and pH then was adjusted to 9 (region II, Figure 6). When the pH

5878

Langmuir, Vol. 21, No. 13, 2005

Figure 8. (a) Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at cs ) 0.001 M and cp ) 50 ppm. The black arrows on the time scale indicate the changing of bulk pH from 4 to 9 and from 9 back to 4. (b) A qualitative picture of changes in conformation of adsorbed layer upon changes in bulk solution, corresponding to the data in (a).

is lowered to 4, the polymer desorbs, to an adsorbed amount of 0.87 mg/m2, while the adsorbed layer thickness increases to 390 Å(region II). The adsorbed amount is significantly higher than adsorption onto the initially bare silica at pH ) 4 at the same ionic strength (region I, Figure 6). The adsorbed layer is also significantly thicker. It is interesting to note that the same values were obtained in Figure 6 when the pH is restored to 4 (region III), after it had been equilibrated at pH 9. Changing the pH at Low Ionic Strength. Figure 8 shows the adsorption from a polymer solution containing 0.001 M KCl at pH 4 and the response when increasing the pH from 4 to 9 and then restoring the pH to 4. The main features of the figure are similar to the corresponding data recorded at high ionic strength (Figure 6), and the results are schematically illustrated in Figure 8b. There are, however, some quantitative differences if we compare the first (region I) and the final state (region III) of Figures 6 and 8. Although the average densities of adsorbed layer in region I for both ionic strength are about the same (≈13 mg/mL), the adsorbed amount and the adsorbed layer thickness at low ionic strength are almost half of those at high ionic strength. Furthermore, when we sequentially restore the initial solution conditions to pH 4 after having increased pH to 9, average density is very low, about 6-7 mg/mL, at low salt content (Figure 8, region III). At high salt content the corresponding layer is much more compact with an average density of about 20-22 mg/mL. Figure 9 shows the adsorption from a polymer solution containing 0.001 M KCl where the changes in pH are the reverse from Figure 8. Initially, the pH is 9, and the response when decreasing the pH to 4 and then restoring the pH to 4 is shown. Part b of Figure 9 shows a schematic drawing of the structure of the layer as discussed below. Also in Figure 9 the features are similar to the corresponding data reported in Figure 7, which were recorded at high ionic strength. The adsorbed amount decreased to 0.75 mg/m2, and the layer swells up to 300 Å when pH is lowered after adsorption to the bare surface at pH 9 (Figure 9, region II). A substantial polymer desorption and swelling of the adsorbed layer was observed, but the

Samoshina et al.

Figure 9. (a) Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at cs ) 0.001 M and cp ) 50 ppm. The black arrows on the time scale indicate the changing of bulk pH from 9 to 4 and from 4 back to 9. (b) A qualitative picture of changes in conformation of adsorbed layer upon changes in bulk solution, corresponding to the data in (a).

Figure 10. Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at pH ) 5.6 and cp ) 50 ppm. The black arrows on the time scale indicate the changing of the ionic strength of bulk solution from 0.1 to 0.001 and from 0.001 back to 0.1.

density of the layer remained higher than when the polymer was adsorbed to the initially bare surface at pH 4 (region I, Figure 8). This observation is analogous to the region III of Figure 8, but here the desorption and swelling of the layer are more substantial. However, it is contrary to what was observed at high ionic strength, where the same values of Γ and d were observed when pH was decreased from 9 to 4 (Figures 6 and 7), irrespective of the history of the adsorption process. Changing the Ionic Strength. Figure 10 shows the adsorption at pH 5.6 from a polymer solution containing 0.1 M KCl and the response when changing the ionic strength to 0.001 M and then restoring the ionic strength to 0.1 M. Desorption occur fast when reducing the ionic strength of the solvent (region II). This can be attributed to the reduced screening of the charge, which in turn increases the electrostatic repulsion between adsorbed molecules, facilitating desorption from the surface. Although the adsorbed amount is reduced with the decreased ionic strength, the adsorbed layer thickness remains almost constant at 250 Å.

Polycation Adsorption on Silica Surface

Figure 11. Adsorbed amount (O) and adsorbed layer thickness (4) of CPAM on silica surface as a function of time at pH ) 5.6 and cp ) 50 ppm. The black arrows on the time scale indicate the changing of the ionic strength of bulk solution from 0.001 to 0.1 and from 0.1 back to 0.001.

There are, however, still some changes in layer thickness occurring on the long time scale. Even if the polymer desorbs, the adsorbed amount and adsorbed layer thickness are higher than the values obtained in adsorption to an initially bare surface at 0.001 M salt solution (Figures 4 and 5). The observed difference shows that even after partial desorption, the adsorbed molecules remaining on the surface retain their configuration attained during the initial adsorption stage. Region III in Figure 10 shows how the adsorbed layer response when restoring the ionic strength to 0.1 M. The adsorbed amount increases and reaches similar value as obtained during adsorption to the initially bare surface, 0.85 mg/m2 (region I). Layer thickness does not change significantly from region II to region III but increases slightly with time as was observed in region I. Together with the identical plateau values for the adsorbed amount and the layer thickness in regions I and III in Figure 10, this suggests that the adsorption is reversible at 0.1 M salt concentration. Figure 11 shows the adsorption at pH 5.6 from a polymer solution, where the changes in ionic strength are the reverse from Figure 10. Initially the ionic strength is 0.001 M and then increases to 0.1 M before being restored to 0.001 M. When the ionic strength was increased from 0.001 M (region I) to 0.1 M (region II), the adsorbed amount increases from 0.4 to 0.8 mg/m2 and the adsorbed layer thickness from 100 to 400 Å. In this case the layer is thicker but has similar density as if the polymer was adsorbed on an initially bare surface from a solution with 0.1 M KCl (region I, Figure 10). When the ionic strength is changed back from 0.1 M to the initial experimental conditions, 0.001 M KCl (region III), the adsorbed amount drops down to value initially obtained in 0.001 M KCl (region I). In contrast, the adsorbed layer thickness does not change but remains at about 400 Å. This is a much thicker and less dense adsorbed layer compared to if the polymer was adsorbed under identical conditions to the initially bare silica surface (100 Å). Discussion Charge Balance at the Interface. With the combination of the ellipsometric data for adsorption of polyelectrolyte and silica surface charge in the presence of polyelectrolyte, we can now quantitatively discuss the relation between polyelectrolyte adsorption and surface charge formation. The ratio between the charge of adsorbed polyelectrolyte and the surface charge, usually referred as a “charge ratio” r, is an important characteristic

Langmuir, Vol. 21, No. 13, 2005 5879

Figure 12. Ratio of adsorbed polyelectrolyte charge density to the silica surface charge density, r, as a function of electrolyte concentration for τ ) 0.09 at pH 5 (1), 5.6 (4), 8 ([), and 9 (0) and for τ ) 0.018 at pH ) 5.6 (×). The data were calculated from the surface charge density obtained from Figure 1 and the ellipsometry data in Figures 4 and 10, assuming that the charge densities of the silica wafers are similar to the one on the silica particles.

of the adsorption process. It defines to what extent the substrate charge is compensated (or neutralized) upon adsorption of polyelectrolyte. If electrostatics is the only driving force for adsorption, a range of situations are expected between an almost exact charge compensation (r ≈ 1) and a total displacement of polyelectrolyte by electrolyte ions.34 An undercompensation is typical for adsorption of polyelectrolytes with low τ and at high cs, i.e., when the small ions are more efficient in compensating the surface charge. When attractive nonelectrostatic interaction of a polymer with a substrate plays a role, charge reversal can take place (r > 1). Overcompensation can also be the result of mismatch between the charges at the interface and the polyelectrolyte charges35,36 or, to some extent, be the result of nonequilibrium features of adsorption (slow conformational rearrangements).32 Figure 12 shows some results for r calculation for different pH as a function of salt concentration for CPAM with τ ) 0.09, and results for r at pH ) 5.6 for polymer with τ ) 0.018 are also included. (Even if the exact number of ionizable silanol groups is different between two types of surfaces, we assume that the surface charge of the silica beads closely approximates the surface charge on silica wafers due to the same surface chemistry according to the methods of pretreatment.) First of all, at low pH, large surface charge overcompensation is observed. Since only a few negative charges are present on the surface at low pH, it is difficult to measure σ0. However, the amount of the adsorbed material is substantial and is much higher than needed for charge neutralization. Here we note that the charges on the polymer are connected, and the charges on the surface are discrete and far apart at low charge densities. Therefore, if only one of the charges on the polymer interact with the opposite charge on the surface, we very easily arrive at charge overcompensation at low surface charge densities. In addition, it is also well-known that neutral polyacrylamide can adsorb on silica from an aqueous solutions, although the nature of this interaction is not clearly understood.37 (34) Linse, P. Macromolecules 1996, 29, 326-336. (35) Eriksson, L.; Alm, B.; Stenius, P. Colloids Surf. A 1993, 70, 47-60. (36) Shin, Y.; Roberts, J. E.; Santore, M. M. J. Colloid Interface Sci. 2002, 247, 220-230.

5880

Langmuir, Vol. 21, No. 13, 2005

Figure 12 shows that at pH ) 5.6 significant overcompensation is observed for polyelectrolyte with a τ ) 0.09 (r ) 4-10 depending on ionic strength), while for the polyelectrolyte with τ ) 0.018 the degree of overcompensation is lower at all salt concentrations. This is due to the fact that, even if the polymer with τ ) 0.018 produces more extended adsorbed layers due to higher conformational entropy, the majority of segments in loops and tails is not charged. Other results in Figure 12 can also be related to the conformation of adsorbed polyelectrolyte. The r parameter increases with the ionic strength for moderate pH. This is in line with the observation from Figure 5, where adsorbed layer thickness increases at a given pH. Here sufficient screening favors formation of extended adsorbed layers, where charged segments tend to move away from the surface. These segments are mainly the loops/tails, which are responsible for the overcompensation of the surface charge in the adsorbed layer. The occupation area per molecule then decreases, which makes it possible for more polyelectrolyte to be accommodated at the interface (cf. Figure 4). At higher pH (8-9), the surface charge density in the presence of adsorbed polymer (Figure 1) is almost independent of the ionic strength, and at the same time, the adsorbed amount is approximately constant upon alteration of pH (Figure 4). This implies that the number of dissociated silanol groups per area unit occupied by one polyelectrolyte molecule is constant and ≈800 silanol groups/polymer molecule. The polylectrolyte adopts a rather flat conformation at the interface at high pH, and the accumulation of excess charge is highly unfavorable in close vicinity to the surface. Charge overcompensation is therefore not expected to occur under these conditions. Indeed, as can be seen from Figure 12, the r value is close to 1 for all salt concentrations. Adsorption Appears To Be Only Partly Reversible. If we compare the adsorbed layer properties obtained under certain conditions with results obtained when the these conditions have been recreated using other initial conditions, e.g., by changing either pH, cs, or the concentration of polyelectrolyte, similar results are obtained if the initial conditions give a low surface coverage. For example, the same value of adsorbed amount of 0.85 mg/ m2 could be obtained when the polymer stock solution is added directly to a polymer concentration 50 ppm or if the concentration was increased stepwise, where the first addition give a bulk concentration of 5 ppm and the plateau adsorbed amount of 0.65 mg/m2. This does not occur during the opposite process, where decreasing polyelectrolyte concentration only leads to slight desorption. This does not necessarily prove irreversibility.24 It has been shown that the subsurface volume fraction of polymer reaches values in the range of 10-10 even after several dilutions.38 Further solvent exchange has hardly any effect because of the high affinity shape of the adsorption isotherms and the slow transport rate through the stagnant layer outside the surface at these low polymer concentrations. The increase in ionic strength reduces the range of electrostatic repulsion within the adsorbed layer but also tends to weaken the interactions with the surface. This leads to an increase in adsorbed amount but also to a more expanded layer as the electrostatic interaction with the surface is weaker. The differences observed due to the (37) Samoshina, Y.; Diaz, A.; Becker, Y.; Nylander, T.; Lindman, B. Colloids Surf. A 2003, 231, 195-205. (38) Scheutjens, J. M. H. M.; Fleer, G. J. In The Effect of Polymers on Dispersion Properties; Tadros, T. F., Ed.; Academic Press: London, 1982; p 145.

Samoshina et al.

path chosen to reach at a certain conditions might be explained as follows. Starting at low ionic strength, the polymer adsorbs in a flat conformation due to a strong interaction between polymer segments and the surface as well as larger electrostatic repulsion between the polymer segments. Increasing the ionic strength then weakens the polymer-surface interaction, which leads to partial detachment of polymer segments and swelling of the layer. This decreases the barrier for adsorption of polymer and the adsorbed amount increases. When low ionic strength is restored, the polymer segments located far from the surface are unlikely to reattached to the surface even if some polymer molecules are desorbed due to the increasing range of polymer-polymer electrostatic repulsion. When starting at high ionic strength, the adsorbed layer is expanded, and a decrease of ionic strength leads to desorption of polymer molecules, while the thickness of the layer remains constant. When we changed the ionic strength at a constant pH, desorption occurred upon reducing the ionic strength of the solvent. However, the adsorbed layer thickness remains constant during the experiment. When the ionic strength of the bulk solution was increased, the adsorbed amount also increases. What is the driving mechanism for the additional adsorption? The surface charge density does not change from region II to region III in Figure 10 or from region I to region II in Figure 11, and the charge overcompensation is already substantial as discussed before. The lateral interactions within adsorbed layer become less repulsive, but rearrangement has been shown to be too slow to give the fast response of the adsorbed layer we observed. The possible scenario is patchwise desorption in a previous step (region II of Figure 10). The majority of the molecules in region II preserve their conformation attained in region I. Since almost one-third of them desorbs, there are bare patches at the interface available for adsorption, which can be occupied now, when ionic strength has been increased. When pH was changed at a given ionic strength, additional adsorption occurred when we switched pH from 4 to 9 and was accompanied by a shrinking of the layer.The driving force for additional adsorption is thought to be of electrostatic nature since the surface charge density increases sharply from pH ) 4 to pH ) 9. The electrostatic interaction between polyelectrolyte molecule and the surface is more extensive, and the macromolecule therefore collapses at the interface to compensate the surface charge. When the bulk conditions again were restored to pH 4, the layer responded by desorbing polymer and swelling the layer, but the adsorbed amount and the thickness are larger than for the adsorption via the direct route. When we switch pH from 9 to 4, the surface charge density decreases sharply (Figure 1). The overcompensation of the surface charge becomes very high, and the driving force for desorption is the high electrostatic potential caused by accumulation of excessive positive charges in the adsorbed layer. The process of desorption involves a loss of part of the contacts with the surface, and part of the molecules then detach and diffuse away from the surface, while the remaining ones build a looser layer than prior to desorption. The rate of desorption should depend on the number and strength of the bonds formed with the surface. It has been estimated that one molecule of such a high molecular weight polyacrylamide forms ≈800 bonds with the surface at pH ) 9.When we switch pH from 9 to 4, it is very unlikely that all these bonds will break at once. When pH or ionic strength was changed at constant polymer concentration from conditions of low surface

Polycation Adsorption on Silica Surface

coverage to those giving high surface coverage, a very slow and incomplete response of the adsorbed layer was observed. The adsorbed amount and the thickness of the layer never attained the levels observed when polymer was adsorbed directly at that particular condition to the bare surface. The slow response points on several barriers toward attaining the equilibrium layer structure. This involves both electrostatic and steric interactions as discussed above. Finally, in cases where the adsorbed layer appears to respond reversibly, it should be borne in mind that this does not necessarily mean that the layers are the same. The conformation of the polymer in the adsorbed layer could still be different, even though the adsorbed amount and the average layer thickness are the same. Although the adsorbed amount and the layer thickness (Figure 8, region II and Figure 9, region I) at pH 9 and low ionic strength appear very similar, the layer responds differently to a decrease in pH. Here we note that the evaluation of the thickness from the ellipsometry data assumes that the layer is uniform and homogeneous, and calculated thickness has to be regarded as an average, where the contribution to the dense regions are larger. Thus, differences in lateral and perpendicular density profiles will not be taken into account, and differences in the segment density profiles may give the same average thickness. Then the fraction of bound segments in region II, Figure 8, can be smaller than in that for the situation in Figure 9, region I, without being reflected as differences in d as the total adsorbed amount is the same. Then, after lowering the surface charge by decreasing the pH, the polymers with a lower number of contact points with the surface are more easily detached from the surface, even if they have the same adsorption energy per segment as when there are a larger number of contact points between the macromolecule and the surface. A vast amount of studies on various aspects of the reversibility of protein adsorption have been published. With respect to reversibility of adsorption/desorption of proteins, distinction should be made between reversibility toward dilution of the solution,39,40 changes in pH and ionic strength,41,42 changes in temperature,43 addition of other types of surface-active substances,44 and exchange against dissolved proteins.45,46 Proteins are very complex polyelectrolytes as they are polyampholytes. In addition, globular proteins adopt a (39) Welzel, P. B. Thermochim. Acta 2002, 382, 175-188. (40) Norde, W.; Anusiem, A. C. I. Colloids Surf. 1992, 66, 73-80. (41) Kondo, A.; Fukuda, H. J. Colloid Interface Sci. 1998, 198, 3441. (42) Galisteo, F.; Norde, W. J. Colloid Interface Sci. 1995, 172, 502509. (43) Ferna´ndez, A.; Ramsden, J. J. J. Biol. Phys. Chem. 2001, 1, 81-84. (44) Norde, W.; Favier, J. P. Colloids Surf. 1992, 64, 87-93. (45) Vermonden, T.; Giacomelli, C. E.; Norde, W. Langmuir 2001, 17, 3734-3740.

Langmuir, Vol. 21, No. 13, 2005 5881

compact three-dimensional structure in aqueous solution, and at an interface the protein conformation changes in most cases. An entropy gain caused by those conformational changes has in fact been considered to be one of the driving forces for protein adsorption.40,47 The question of reversibility of protein adsorption is therefore far more complex than for synthetic polyelectrolyte adsorption. Concluding Remarks Our measurements indicate that polyelectrolyte adsorption appears only partially reversible due to the high affinity to the surface, which slows down the rearrangement process. Thus, our general observation is that relaxation occurs more easily when the direction of the process is from low to the higher surface coverage. The mechanism of desorption is, to a large extent, controlled by electrostatic interactions, where both decreased surface charge and increased repulsion between segments (excluded volume effect) can lead to desorption. We have discussed how the formed structure of adsorbed layer depends not only on the bulk conditions but also on the preapplied conditions, i.e., the history of environmental changes. The obtained knowledge about underlying mechanisms can be exploited to tune the behavior of polyelectrolyte films on a nanometer scale by manipulating the electrostatic interactions. Many practical formulations such as pharmaceutical suspensions, inks, cosmetics, paints, etc., require a rational designing of interfacial arrangement of complex multicomponent systems containing polyelectrolytes. Sometimes the desired structure could not be obtained by direct film formation. However, by preapplying certain bulk conditions, enhanced (or reduced) surface concentration and desired conformation of polymer can be achieved. It is also important to remember that many applications require that the polymer films respond in a controllable manner when it is exposed to a changing environment such as changes in pH and ionic strength. Acknowledgment. We are grateful to Dr. Ilias Iliopoulos, Laboratorie de Physico-chimie macromoleculare, Universite Pierre et Marie Curie, Paris, for providing the cationic polyelectrolytes and for the stimulating discussions and suggestions by Bjo¨rn Lindman and Per Linse. The silica substrates were kindly provided by Dr. Stefan Welin- Klintsro¨m, Linkping University, Sweden. Yulia Samoshina thanks Svenska Insitutet, Visby Program, and the Swedish Foundation for Strategic Research (Stiftelsen fo¨r Strategisk Forskning, SSF) for financial support through the Colloid & Interface Technology program. LA050069Q (46) Giacomelli, C. E.; Norde, W. J. Colloid Interface Sci. 2001, 233, 234-240. (47) Norde, W. Adv. Colloid Interface Sci. 1986, 25, 267.