Journal of Colloid and Interface Science 413 (2014) 147–153
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Overshoots of adsorption kinetics during layer-by-layer polyelectrolyte film growth: Role of counterions Cédric C. Buron ⇑, Claudine Filiâtre Institut UTINAM, UMR CNRS 6213, Université de Franche-Comté, 16, Route de Gray, 25030 Besançon, France
a r t i c l e
i n f o
Article history: Received 26 June 2013 Accepted 19 September 2013 Available online 29 September 2013 Keywords: Overshoot Polyelectrolyte Multilayer Reflectometry Adsorption kinetics Counterions SiO2 particles
a b s t r a c t Layer-by-layer adsorption of polycation poly(trimethylammonium ethyl methacrylate chloride) (MADQUAT) and polyanion poly(acrylic acid, sodium salt) (PAA) on silicon oxide substrates was studied in order to understand non-regular multilayer buildup. Indeed during MADQUAT adsorption, a special variation in the signal, called overshoots, was monitored with stagnation point adsorption reflectometry (SPAR). These overshoots were observed under different experimental conditions (pH, polymer concentration, salt concentration and molecular weight variations). A method called ‘‘substrate thickness method’’ was applied to determine the thickness, the refractive index and the adsorbed amount variation during the overshoot. Results clearly showed a decrease in the adsorbed amounts, but which does not necessarily indicate dissolution of the adsorbed multilayer. Therefore we also investigated layer-by-layer adsorption on colloidal silica particles (SiO2) under the same conditions as those in reflectometry. Indeed the high specific surface area of particles allowed titration of chloride ions and unadsorbed polymer in the bulk during the adsorption process. An increase in chloride ions concentration with no increase in the polymer concentration was observed. Accordingly, overshoots in cases of MADQUAT/PAA multilayer are not due to desorption of cationic polyelectrolyte but to desorption of chloride ions from the adsorbed multilayer film. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Adsorption of polyelectrolyte is a major topic in academic research but also relevant to many industrial processes. Depending on the purpose, polyelectrolytes can either be deposited, forming either a monolayer or a multilayer film. In order to buildup a multilayer film, a simple, low-cost technique can be used. This method is based on a layer-by-layer (LbL) adsorption of oppositely charged polyelectrolytes driven by electrostatic attractions. Since Decher’s research, [1] other studies have also shown the possibility of building a multilayer film using hydrogen bonding [2]. The alternate sequential adsorption is now extended to a wide range of compounds such as proteins [3–5], DNA [6,7], or inorganic particles [8–10]. To understand the mechanism involved in the growth of a polyelectrolyte film, many experimental techniques have been used. Among them, in situ techniques such as optical reflectometry (also called stagnation point adsorption reflectometry, SPAR), ellipsometry, surface plasmon resonance, optical waveguide light microscopy or quartz crystal microbalance with dissipation have been used to monitor the evolution of the adsorbed amount of each ⇑ Corresponding author. E-mail address: [email protected] (C.C. Buron). 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.09.040
polymer, the refractive index or the thickness of the deposited film. These parameters varied in duration from a few seconds up to several hours. Therefore numerous investigations have been carried out on the kinetic of polymer adsorption [11,12]. Under certain experimental conditions, specific evolution of the monitoring output appears at the beginning of adsorption. For example, in the study of adsorption kinetic by optical reflectometry, a phenomenon called overshoot was monitored on the reflectometer signal during polyelectrolyte adsorption. Indeed overshoot corresponds to a rapid increase in the reflectivity signal followed by a slow decrease. It not only has been observed during the adsorption of grafted copolymers poly(acrylamide)/poly(ethyleneoxide) [13], proteins [14] or cationic starch [15] but it has also appeared in multilayer films composed of cationic/anionic polyelectrolytes [16–18], cationic polyelectrolytes and colloidal silica (18 nm) [19], protein and polyelectrolyte [20], polypeptides [21] or proteins [22]. There is ongoing debate in the literature over possible explanations for overshoots observed in the polyelectrolyte multilayer buildup. Indeed Cohen Stuart and co-researchers have suggested a combined adsorption/desorption process of the polyelectrolytes [18,23]. They studied the effect of KCl, NaNO3 and CaCl2 salt concentration on the occurrence of the overshoot during polyelectrolyte adsorption. Using a stability diagram, they showed the
148
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
possibility of the formation of a soluble complex composed of the introduced polyelectrolyte and the previously adsorbed polymer above a critical salt concentration. Another explanation is a conformational rearrangement of polymer molecules at the surface. Filippov and Filippova [24] have theoretically demonstrated that the overshoots are controlled by conformational kinetics of the adsorbed layer. Using experimental data, Wågberg and co-researchers have also proposed as an explanation for the overshoots, a rearrangement of the polyelectrolyte conformation in the adsorbed layer resulting in a change in the reflectometer signal [25]. However any conformational changes at a constant adsorbed amount (optical mass) cannot be detected using reflectometry [13]. In our previous studies, we also observed overshoots in the reflectometer signal during the growth of multilayer films with oppositely charged polyelectrolytes [26,27]. In view of these findings, the aim of the present work was to investigate the reason for the overshoots appearing during the buildup of a multilayer made of a strong cationic polyelectrolyte and a weak anionic polyelectrolyte with special attention given to the counterions of each polyelectrolyte. Using an optical reflectometer equipped with an impinging jet cell, several physico-chemical parameters (pH, ionic strength, monovalent and divalent ions, weak and strong polyelectrolytes with different molecular weights) were tested to observe their effect on the overshoots. The refractive index and thickness variations during an overshoot were determined using a method previously developed by our team (substrate thickness method) [28]. Adsorption of layer-by-layer films was also carried out onto colloidal silica particles which due to their large specific surface area allowed us to determine the amount of adsorbed polymer. Adsorption of each polyelectrolyte was recorded using the titration of the unadsorbed amount of polymer. Concentration of chloride ions was measured in situ during MADQUAT adsorption with an ion-selective electrode.
2. Materials and methods 2.1. Chemicals and solutions Quaternized polydimethylaminoethyl methacrylate chloride (MADQUAT) and poly(acrylic acid, sodium salt) (PAA) were synthesized by COATEX (Genay, France). MADQUAT (207 g per molar unit) with different molecular weights was used as a strong cationic polyelectrolyte. PAA (94 g per molar unit) has an average molecular weight of 10 kDa. PAA has frequently been used as a weak anionic polyelectrolyte for adsorption studies [29,30] and has an intrinsic ionization pK of 5 in 10 3 M NaCl solution and a half ionization pK around 6.8 [31]. Poly(4-styrenesulfonate, sodium salt) (PSS) with an average molecular weight of 70 kDa was purchased from Alfa Aesar. PSS is a strong anionic polyelectrolyte which was totally ionized across the whole range of pH used in this study [32]. The chemical structures of the polyelectrolytes are presented in Fig. 1. Polymers were dissolved in ultrapure water having a resistivity above 18 MX cm (Milli-Q Plus, Millipore). The solution pH was adjusted with HCl (Aldrich) or NaOH (Aldrich) and the appropriate amount of analytical grade NaCl (Aldrich) was added to set the ionic strength. Polyelectrolyte solutions were prepared just before each experiment. Colloidal silica particles were synthesized using the Stöber method in ethanol [33]. After several washing/centrifugation cycles, particle size was measured in water using dynamic light scattering (Zetasizer 3000HS, Malvern). Monodispersed particles of 258 nm in diameter with a polydispersity index of 0.082 were obtained.
Fig. 1. Chemical structures of the polyelectrolytes used in this study.
2.2. Optical reflectometry An impinging jet cell was mounted on the reflectometer, which created a stagnation point in the center of the cell where the solution was introduced through a cylindrical channel drilled into the glass prism [34]. A linearly polarized He–Ne laser beam entered the cell through one side of the prism. The beam was reflected on the stagnation point, then emerged from the cell, and was split into its parallel (p) and perpendicular (s) components (with respect to the plane of incidence) by using a beam splitter cube. Photodiodes were used to detect the beams. The resulting electric intensities, Ip and Is, were then converted into electric tension and finally recorded. A two-way valve was used to switch between solutions with or without polyelectrolyte. Signals recorded during the injection of the polyelectrolyte free solution and polyelectrolyte solution were respectively denoted by S0 and S. Finally (S S0)/S0 represents, at least semi-quantitatively, the changes in adsorbed amount [18]. Silicon wafers coated with a silicon oxide (SiO2) layer were used as the substrate. Prior to each experiment, wafers were cleaned with piranha solution (one part of hydrogen peroxide (30%) in three parts of sulfuric acid (98%)), then rinsed and finally stored in ultrapure water until use. All experiments followed the same experimental procedure. In order to obtain a stable baseline of the reflectometer signal, a polymer free solution of the same pH and salt concentration as the subsequent polymer solutions was introduced into the cell for 5 min. Since the silicon oxide carried a negative surface charge under experimental conditions of pH 5.5 [35], MADQUAT was then introduced into the cell. In most cases in this study, ten layers of alternate polycation and polyanion (five bilayers) were deposited by switching the injection of the solution every 10 min. If not mentioned in the text, no rinsing step was done between each injection. For each of the experiments the solution flow rate was maintained at a constant rate. The experiments were repeated successfully three times with reproducibility of the order of 5%. In order to determine the mean refractive index and thickness of the adsorbed layer during the overshoots, a method called substrate thickness method was used. Briefly, the principle of the analytical procedure is based on a comparison between the experimental and calculated reflectivities ratio Rp/Rs for the same adsorption experiment carried out with at least two substrates of different thicknesses. Reflectometer output S is linked to the reflectivities ratio by an apparatus transmission factor f. For a more complete description of the procedure, please refer to our previous paper [28]. In the present work, silicon wafers with a silicon oxide layer of 115 ± 2 nm or 298 ± 1 nm were purchased from ACM (France) and were used to apply the method. Thicknesses of SiO2
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
layers were measured using a spectroscopic phase-modulated ellipsometer (Jobin Yvon, Model UVISEL). 2.3. Adsorption onto silica colloidal particles (SiO2) Silica particles (1 g) were dispersed in 100 mL of 10 3 M or 10 1 M sodium salt solution. The adsorption of the cationic polyelectrolyte was first done by adding 1 mL of a concentrated MADQUAT solution (100 g L 1). After 10 min, polymer-coated particles were separated by centrifugation at 13,000 rpm for 8 min and then dispersed in a solution containing 2 g L 1 of PAA for 10 min. The same procedure was applied successively in order to form six adsorbed layers onto the particles. In addition, concentration of polymer solutions was progressively increased to take into account the exponential growth of the adsorbed amounts [26]. In order to ensure good particle dispersion between each adsorption step, a 5 min ultrasonication step was included in the procedure. This was achieved using a classic ultrasonic bath (Elma T420) at a constant ultrasonic frequency of 75 kHz and an ultrasonic power of 70 W. The pH of the suspensions was adjusted to pH 5.5 prior to adding each polyelectrolyte solution. During the multilayer buildup, concentration of chloride ions was measured in situ using a silver combination electrode with a Hg/Hg2SO4 reference electrode (Radiometer Analytical, XM 900). Polymer concentration in the solution, i.e. the unadsorbed macromolecules, was determined after centrifugation using the total organic carbon technique (Shimadzu, TOC 5050). 3. Results and discussion 3.1. Effect of polyelectrolyte concentration A multilayer film made of five MADQUAT-PAA bilayers (denoted (MADQUAT-PAA)5) was built-up using different polyelectrolyte concentrations. The results of the study by reflectometry are presented in Fig. 2 with reflectometer output variation as a function of time. Each step is related to the adsorption of a polyelectrolyte given that each experiment begins with cationic polyelectrolyte adsorption. Results clearly showed a strong increase in the reflectometer plot (i.e., adsorbed amount) when the polyelectrolyte concentration increased. However, the first adsorption of MADQUAT did not depend on the polymer concentration because DS/S0 had almost the same value. Indeed polymer adsorp-
149
tion is ruled by the formation of an electrokinetic barrier [36] while adsorption kinetic is governed by the transport of the polymer to the surface vicinity. During the adsorption process, a progressive accumulation of charged polymer onto the substrate leads to repulsive electrostatic interactions between the adsorbed polyelectrolyte and the incoming polyelectrolyte that progressively stop the adsorption. The adsorption of MADQUAT seems to be of the high affinity type due to the rapid initial increase in the reflectometer signal. As stated in the introduction, the intention in this study was to investigate the reason for a peculiar shape of the signal, called overshoot, which was recorded during adsorption of MADQUAT. For example, at the 9th adsorption step (see arrow in Fig. 2), reflectometer output quickly increased and then slowly decreased. Moreover the amplitude of overshoots was higher when the polymer concentration increased. The same results were observed by Bijsterbosch et al. for monolayer adsorption of grafted copolymer onto silica [13]. 3.2. Effect of the electrolyte type and concentration In our previous studies, the overshoots were observed during layer-by-layer MADQUAT/PAA film growth under different experimental conditions. They were only recorded during MADQUAT adsorption. We noticed that the overshoots were more pronounced when the concentration of sodium chloride salt was increased [27]. But this phenomenon was not linked to the adsorbed amount of polymer. Interestingly the adsorbed amount was higher for 10 2 M than 10 1 M in NaCl. The type of salt was then changed to observe the effect on the growth of the multilayer. For a series of monovalent cations Li+, Na+, K+, Cs+ having the same counterion (Cl ) and for different monovalent anions NO3-, Cl with Na+ counterion, an overshoot was again observed during polycation adsorption. The ionic size of tested cations and anions does not seem to affect the overshoot. Concentration of each type of salt was also modified in the range of 10 3 M to 10 1 M without any modification of the overshoots. Finally the nature of the counterion and more precisely the effect of divalent cations on multilayer adsorption were investigated [37]. Surprisingly the overshoot phenomenon disappeared when BaCl2 and ZnCl2 were used as electrolytes. It may result from the formation of insoluble complexes between divalent cation and PAA. 3.3. Effect of molecular weight of polyelectrolytes Layer-by-layer deposition was then carried out at pH 5.5 with two PAA molecular weights of 10000 Da and 170000 Da (results not presented here). Adsorbed amounts after 5 deposited bilayers were the same even if the length of PAA polymer chain was very different. Overshoots were also recorded during MADQUAT adsorption but they appeared to be independent of the PAA molecular weight. However, the effect was highlighted when MADQUAT with different molecular weights was used (Fig. 3). There was a difference in the reflectometer signal recorded after adsorption of two bilayers. These measurements are reported in terms of specific viscosity which depends on the molecular weight of polymer according to the Mark-Houwink equation. For a specific viscosity (gsp) of 3.45 and 6.55, there was still an overshoot for MADQUAT adsorption whatever the number of adsorbed layers. However, for gsp = 14, overshoot vanished from the first layer to the fifth layer of adsorbed MADQUAT. Therefore the length of the polycation chain plays an important role in the overshoot occurrence. 3.4. Effect of the polymer charge density
Fig. 2. Reflectometer output variation as a function of polyelectrolyte concentration during (MADQUAT/PAA)5 multilayer formation (pH 5.5, NaCl 10 3 M).
10
Since the overshoot was the highest at a NaCl concentration of 1 M, polyelectrolyte solutions were prepared at this concentra-
150
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
ditions such as divalent cations, high length polycation chain and pH variation caused the overshoot to disappear. Multilayer assembly was then done using two strong polyelectrolytes: MADQUAT and PSS under the same conditions (pH 5.5, NaCl 10 1 M). Fig. 5 presents a steady increase in DS/S0. The value of each increment and also the adsorbed amount of MADQUAT or PSS, are the same for each polyelectrolyte adsorption step. Each polyelectrolyte is fully charged and can adsorb in an extended conformation onto the surface. The construction is similar to a stacking of polyelectrolyte layers since the amount of charge brought by each polyelectrolyte cannot change during the adsorption. On the other hand, a variation in ionization degree between PAA in solution and PAA within several kinds of multilayers was measured by Choi and Rubner [38]. A careful adjustment of the polymer charge density for both polyelectrolytes could be a way of avoiding the formation of overshoot.
Fig. 3. MADQUAT molecular weight effect on (MADQUAT/PAA)5 multilayer formation. Overshoots are indicated by an arrow for the 9th adsorbed layer. Molecular weights are given in terms of specific viscosity, gsp (pH 5.5, NaCl 10 3 M, Cp 10 mg L 1).
3.5. Thickness, refractive index and adsorbed amount variations during the overshoot
tion and pH was adjusted at pH 4 or pH 5.5 leading to a variation in the anionic polyelectrolyte charge density. Reflectometer outputs started to change after adsorption of 3 layers, as previously observed (Fig. 4). At the end of the experiment, the adsorbed amount of LbL film was higher for pH 4 than pH 5.5. Overshoots were only observed at pH 4 during the adsorption of the first and the second layer of MADQUAT. The same observation was made for NaCl 10 3 M solution (unshown results). When pH is increased to pH 5.5, charge density along the backbone of PAA changes because carboxylic acid groups are weak acidic functions. The charge of a weak polyelectrolyte is tunable with respect to the pKa and pH values. The degree of ionization of PAA is equal to 5% at pH 4 and 30% at pH 5.5 which was shown by Choi and Rubner [38]. When the polyelectrolyte is less charged, more adsorbed polymer chains are needed to reach the electrokinetic barrier. Consequently, a decrease in PAA charge density led to an increase in the adsorbed amounts of polymers and to ordinary variation in the reflectometer output (i.e., DS/S0 reaches a plateau). At this stage, the adsorbed amount seems to be a key parameter in understanding the overshoots in adsorption kinetic. Indeed, an increase in the adsorbed amount using several different experimental con-
New experiments were then carried out by increasing the MADQUAT injection time of MADQUAT solution from 10 min to 2 h. In order to obtain significant and well-defined overshoots, three polyelectrolyte bilayers were first prepared by sequential adsorption of MADQUAT and PAA for 10 min on two substrates (115 and 298 nm silica thicknesses). MADQUAT was then injected for 2 h. The results are presented in Fig. 6a. When a silicon wafer with a SiO2 thickness of 298 nm was used the overall negative value of DS/S0 can be explained by using the sign of the sensitivity factor (As) defined by Dijt et al. [34]. Indeed it was negative for this value of SiO2 thickness and quite similar in absolute value to a SiO2 thickness equal to 115 nm. Therefore both curves showed the same trend. Interestingly the overshoot vanished after 45 min and the output then increased very slowly until reaching a plateau. The explanation of the overshoot is still a subject of discussion in the literature. Indeed, Kovacevic et al. [18] suggested that the initial increase in adsorption followed by a slow decrease in reflectometer output was due to an adsorption/desorption process. Under these conditions, polyelectrolyte complexes formed at the interface slowly desorbed. However Wågberg et al. [25] assumed that there were some rearrangements of the polyelectrolytes in the adsorbed multilayer without necessarily desorption. To demonstrate this, polyelectrolyte adsorption was followed by quartz
Fig. 4. Reflectometer output variation as a function of pH of the solution during (MADQUAT/PAA)5 multilayer formation (NaCl 10 1 M, Cp 10 mg L 1).
Fig. 5. Reflectometer output variation during (MADQUAT/PSS)5 multilayer formation (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
151
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
Fig. 6. (a) Effect of polyelectrolyte supply: injection of MADQUAT for 2 h for two silicon oxide thicknesses (b) Calculated adsorbed amount by substrate thickness method for the seventh layer (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
crystal microbalance with dissipation. A close correlation between energy dissipation into the multilayers and a decrease in the adsorption as measured by reflectometry was established. Consequently, they concluded that desorption of polyelectrolyte complexes was not necessarily the origin of reflectometer signal decrease. However as it was clearly stated by Cohen Stuart and co-researchers [13] and as it can be demonstrated using an optical model of the interface, reflectometry is a technique that determines the adsorbed amount of polymers. However, any conformational changes cannot be detected for a constant adsorbed amount. In order to determine whether overshoots were due to desorption, rearrangement or both, we applied the substrate thickness method during MADQUAT adsorption using two different silicon oxide thicknesses (115 nm and 298 nm). Final values for thickness and refractive index were compared with spectroscopic ellipsometry measurements after drying and are summarized in Table 1. Results obtained from the substrate thickness method led to results similar to those using ellipsometry. Nevertheless a lower refractive index value was obtained by reflectometry compared to that of ellipsometry. This was probably due to the state of the multilayer because the film was not dried in the reflectometry experiments. Water molecules, trapped in the multilayer, induced a decrease in the refractive index value. However values were consistent and adsorbed amount variations were determined. Evolution of the adsorbed amounts during the overshoot is shown in Fig. 6b. After injection of MADQUAT (see the arrow in Fig. 6b) the uptake quickly increased and then slowly decreased, just like the reflectometer signal. Therefore desorption occurs as suggested by Cohen Stuart and co-researchers however it was not necessarily polyelectrolyte desorption and this will be discussed in the next section. 3.6. Injection of PAA at different times of the overshoot Initially two bilayers (NaCl 10 1 M) were deposited onto the substrate. Then the two-way valve that allowed the injection of
polyelectrolyte solution was switched for PAA adsorption after 2, 5 and 15 min of MADQUAT adsorption. Surprisingly in Fig. 7 the adsorbed amount of PAA did not depend on when the PAA injection started. Indeed the height of each increment was the same as for the previous study, i.e. after 2 h of MADQUAT injection (Fig. 6a). If the desorption of polyelectrolyte occurred, a constant charge density of the adsorbed multilayers was unlikely. Consequently, the adsorbed PAA amount should not be the same. This point seems to be in favor of a rearrangement in the multilayer film. However, we observed desorption (see Fig. 6b). These two results were combined to suggest an explanation of overshoots other than those proposed by Kovacevic et al. [18] or Wågberg et al. [25] that is polyelectrolyte desorption or polymer layer rearrangement without desorption respectively. Considering that there was no desorption of polyelectrolyte, the only way to explain the evolution of the reflectometer signal during MADQUAT adsorption is the ejection of ions from the LbL film leading to a modification of the refractive index. Several studies have shown that ions are embedded in the film to ensure electroneutrality especially when a weak polyelectrolyte is used. Thus for a poly(allylamine hydrochloride) (PAH)/PSS multilayer film, 50–80% of the PAH charge was screened by small ions [39,40]. In fact when MADQUAT bound to the adsorbed PAA layer, a change in PAA ionization degree occurred. Choi and Rubner [38] estimated a 25% increase in the degree of ionization at pH 5.5 for a multilayer made of poly(diallyldimethylammonium chloride) (PDADMAC) and PAA (pKa 5.5). PDADMAC is a strong polyelectrolyte, like MADQUAT. Consequently unbonded charges of MADQUAT, which were not used to bind the surface, were screened by chloride ions. They can interact with newly created anionic PAA charges which induces desorption of Cl from multilayer films. A simple calculation showed that, if a MADQUAT monomer was adsorbed, the mass of chloride ions represented 17% of the total monomer mass. Therefore, if chloride ions were removed from the multilayer, the adsorbed amount will inevitably decrease. The observed overshoots were probably due to the ejection of ions from the LbL film. Due to the time scale, this movement
Table 1 Thickness and refractive index of (MADQUAT/PAA)4 deposited bilayers onto silicon oxide layer. Note that the third MADQUAT adsorption was carried out for 2 h (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
Substrate thickness method (115 and 298 nm) Ellipsometry onto 115 nm Ellipsometry onto 298 nm
in situ After drying After drying
Thickness (nm)
Refractive index
18.2 ± 3.0 16.4 ± 2.0 20 ± 2.5
1.380 1.431 1.444
152
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153 Table 2 Evolution of MADQUAT concentration (Cp) and chloride ion concentration ([Cl ]) over time during the 4th MADQUAT adsorption at pH 5.5. Substrate: Si/SiO2/(MADQUAT/ PAA)3. Time (min)
NaCl 10
0 2 12 40
1.83 1.62 1.59 1.58
Cp (g L ±0.002
3 1
)
M
1
NaCl 10 [Cl ] (mM) ±0.1
Cp (g L ±0.002
0.95 3.3 4.6 5.0
6.14 5.36 5.26 5.25
1
)
M [Cl ] (M) ±0.1 0.120 0.128 0.129 0.130
tially complexed with quaternized amino groups of MADQUAT and not involved in the linking process.
3.8. Adsorption of MADQUAT/PAA LbL onto SiO2 colloidal particles Fig. 7. Injection of PAA solutions (see arrows) at different times (2 min, 5 min and 15 min) of overshoot encountered during MADQUAT adsorption (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
of ions seems to be a process with low desorption kinetic. Schlenoff et al. [41] revealed that no small ions were detected in PDADMAC/ PSS films. Their experiment supports our explanation and observations because no overshoot was recorded at NaCl 10 1 M when MADQUAT/PSS were assembled together (see Section 2.4). 3.7. Effect of rinsing step Initially, no rinsing was done between each polyelectrolyte adsorption. So a rinsing step of 5 min using a solution of the same pH and ionic strength was added in the buildup process. Unfortunately this did not change the overshoot shape. Consequently, a rinsing step was applied during an overshoot as it is presented in Fig. 8. For the fifth polymer layer, MADQUAT solution was injected for 5 min, a rinsing with NaCl 10 1 M solution was then done for 5 min and finally followed by another injection of MADQUAT. Surprisingly, output stopped decreasing during the rinsing step giving a flat signal and then decreased when MADQUAT solution was again injected. Firstly, this suggests that the overshoot phenomenon was only due to MADQUAT adsorption. Secondly, when the multilayer was flushed with a MADQUAT free solution, no desorption was observed. Consequently, MADQUAT adsorption leads to an ejection of counterions which are probably chloride ions ini-
The use of a substrate with an area of few square millimeters avoids the measurement of parameter variations such as pH or ion concentration in the bulk during the adsorption of polyelectrolytes. To circumvent the problem of a substrate with a small surface area, experiments were done using silica particles with a specific surface area of 10.6 m2 g 1 (calculated using the following equation: Sp = 3/(Rq), with R = 248 nm and q = 2.2 g cm 3). As for the reflectometry experiments, three bilayers of MADQUAT/PAA were first deposited onto particles using the process described in the experimental section. Then, during the adsorption of the 7th MADQUAT layer, the variations in unadsorbed polymer and chloride ion concentrations were recorded. Multilayer assemblies were performed at 10 3 M and 10 1 M in NaCl. Results are summarized in Table 2. From the addition of MADQUAT solution until 40 min following this, concentration of unadsorbed polymer decreased for both salt concentrations. At the same time an increase in chloride ion concentration was recorded especially for NaCl 10 3 M concentration. For 10 1 M, the slight variation was probably due to the high initial concentration in Cl . When MADQUAT began to adsorb onto the LbL film, an ejection of chloride ions occurred. Chloride ions and the negative charge of PAA compete to bind with the MADQUAT positive charge. The interactions between negative and positive charges of each polyelectrolyte were stronger because an adsorption was recorded. Indeed the values described an isotherm of adsorption which is well-known in the field of polymer research and also demonstrated that, in the case of MADQUAT/ PAA multilayer, there was no desorption of a complex/coacervate of polyelectrolytes. Therefore we infer that ions were ejected from the adsorbed multilayer.
4. Conclusion
Fig. 8. Effect of rinsing during MADQUAT adsorption on two previously adsorbed MADQUAT/PAA bilayers (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
The growth of a layer-by-layer film made of a strong cationic polyelectrolyte (MADQUAT) and a weak anionic polyelectrolyte (PAA) was investigated using optical fixed-angle reflectometry. Our attention was focused on the special evolution of the reflectometer output called overshoot. Indeed during MADQUAT adsorption, the reflectometer signal exhibits a peak followed by a slow decrease. Several parameters such as polymer concentration and molecular weight, pH and type of salt and anionic polyelectrolyte (strong or weak) were changed in order to investigate their effects on the overshoots. Consequently, overshoot vanished when the pH was fixed at pH 4 during the self-assembly or when a strong polyelectrolyte (PSS) was used instead of a weak polyelectrolyte (PAA). This clearly indicates the role of the charge density of each polyelectro-
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
lyte on this phenomenon that is confirmed by the significant effect of divalent cations. Furthermore, the determination of adsorbed amount variation during MADQUAT adsorption has shown a decrease in the uptake, suggesting a polymer desorption. However, as other researchers have claimed, it is thought that there is no polymer desorption but some rearrangement in the multilayer. Indeed, an experiment based on the injection at different steps of the overshoot has shown that the adsorbed amount of PAA did not change which is not in accordance with the desorption of polymer. To be able to measure the variations in the unadsorbed polymer and chloride ions concentrations in bulk solution during adsorption, colloidal silica particles were used. Titration of unabsorbed amount of MADQUAT has shown a decrease in MADQUAT concentration in the working solution indicating an adsorption of the polyelectrolyte. So overshoot is not due to desorption of a polyelectrolyte complex in the case of MADQUAT/PAA multilayer. Finally, an ejection of ions from the multilayer film during the adsorption of the cationic polyelectrolyte highlighted by following the chloride ion concentration in the bulk solution. In our case, the observed overshoots in the reflectometer signal are due to a decrease in the adsorbed multilayer refractive index owing to the chloride ion ejection. Further experiments will be performed to follow the concentration of sodium ions which may also be ejected from the multilayer during the construction. Acknowledgment This article is dedicated to the memory of Professor Alain Foissy who died in June 2009. At the end of his career, his research work focused on the influence of small ions in layer-by-layer polyelectrolyte adsorption. References [1] [2] [3] [4]
G. Decher, Science 277 (1997) 1232. S.A. Sukhishvili, S. Granick, Macromolecules 35 (2002) 301. F. Caruso, K. Niikura, D.N. Furlong, Y. Okahata, Langmuir 13 (1997) 3427. F. Caruso, D.N. Furlong, K. Ariga, I. Ichinose, T. Kunitake, Langmuir 14 (1998) 4559. [5] E. Lojou, P. Bianco, Langmuir 20 (2004) 748. [6] J. Zhang, L.S. Chua, D.M. Lynn, Langmuir 20 (2004) 8015.
153
[7] K. Ren, J. Ji, J. Shen, Bioconjugate Chem. 17 (2005) 77. [8] T. Sennerfors, F. Tiberg, J. Colloid Interface Sci. 238 (2001) 129. [9] T. Sennerfors, D. Solberg, F. Tiberg, Adsorption of polyelectrolyte-nanoparticle systems on silica: influence of ionic strength 254 (2002) 222. [10] F. Caruso, H. Lichtenfeld, M. Giersig, H. Möhwald, J. Am. Chem. Soc. 120 (1998) 8523. [11] J.C. Dijt, M.A. Cohen Stuart, G.J. Fleer, Macromolecules 25 (1992) 5416. [12] J.C. Dijt, G.J. Fleer, M.A. Cohen Stuart, J.E. Hofman, Colloids Surf. 51 (1990) 141. [13] H.D. Bijsterbosch, M.A. Cohen Stuart, G.J. Fleer, Macromolecules 31 (1998) 8981. [14] P. Van Dulm, W. Norde, J. Colloid Interface Sci. 91 (1983) 248. [15] Y. Li, W.-G. Hou, W.-Q. Zhu, Colloids Surf., A 303 (2007) 166. [16] N.G. Hoogeveen, M.A. Cohen Stuart, G.J. Fleer, M.R. Bohmer, Langmuir 12 (1996) 3675. [17] D. Kovacevic, S. Glavanovic, N. Peran, Colloids Surf., A 277 (2006) 177. [18] D. Kovacevic, S. van der Burgh, A. de Keizer, M.A. Cohen Stuart, Langmuir 18 (2002) 5607. [19] L.-E. Enarsson, L. Wågberg, J. Colloid Interface Sci. 325 (2008) 84. [20] O. Svensson, L. Lindh, M. Cárdenas, T. Arnebrant, J. Colloid Interface Sci. 299 (2006) 608. [21] S. Grohmann, H. Rothe, S. Eisenhuth, C. Hoffmann, K. Liefeith, Biointerphases 6 (2011) 54. [22] H. Shirahama, J. Lyklema, W. Norde, J. Colloid Interface Sci. 139 (1990) 177. [23] D. Kovacevic, S. van der Burgh, A. de Keizer, M.A. Cohen Stuart, J. Phys. Chem. B 107 (2003) 7998. [24] L.K. Filippov, N.L. Filippova, J. Colloid Interface Sci. 178 (1996) 571. [25] L. Wågberg, G. Pettersson, S. Notley, J. Colloid Interface Sci. 274 (2004) 480. [26] CC. Buron, C. Filiâtre, F. Membrey, H. Perrot, A. Foissy, J. Colloid Interface Sci 296 (2006) 409. [27] C.C. Buron, C. Filiatre, F. Membrey, C. Bainier, D. Charraut, A. Foissy, J. Colloid Interface Sci. 314 (2007) 358. [28] C.C. Buron, F. Membrey, C. Filiâtre, A. Foissy, Colloids Surf., A 289 (2006) 163. [29] S.S. Shiratori, M.F. Rubner, Macromolecules 33 (2000) 4213. [30] S.T. Dubas, J.B. Schlenoff, Langmuir 17 (2001) 7725. [31] I. Pochard, P. Couchot, A. Foissy, Colloid Polym. Sci. 1088 (1998) 276. [32] M. Balastre, J. Persello, A. Foissy, J.F. Argillier, J. Colloid Interface Sci. 219 (1999) 155. [33] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62. [34] J.C. Dijt, M.A. Cohen Stuart, G.J. Fleer, Adv. Colloid Interface Sci. 50 (1994) 79. [35] J. Persello, in: E. Papirer (Ed.), Adsorption on Silica Surfaces, vol. 90, 1998, pp. 297–342. [36] G.J. Fleer, M.A. Cohen Stuart, J.M.H.M. Scheutjens, T. Cosgrove, B. Vincent, Polymers at Interfaces, Chapman and Hall, London, 1993. [37] C.C. Buron, C. Filiâtre, F. Membrey, C. Bainier, L. Buisson, D. Charraut, A. Foissy, Thin Solid Films 517 (2009) 2611. [38] J. Choi, M.F. Rubner, Macromolecules 38 (2005) 116. [39] P. Lavalle, C. Picart, J. Mutterer, C. Gergely, H. Reiss, J.-C. Voegel, B. Senger, P. Schaaf, J. Phys. Chem. B 108 (2004) 635. [40] J. Schmitt, T. Gruenewald, G. Decher, P.S. Pershan, K. Kjaer, M. Loesche, Macromolecules 26 (1993) 7058. [41] J.B. Schlenoff, H. Ly, M. Li, J. Am. Chem. Soc. 120 (1998) 7626.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Overshoots of adsorption kinetics during layer-by-layer polyelectrolyte film growth: Role of counterions Cédric C. Buron ⇑, Claudine Filiâtre Institut UTINAM, UMR CNRS 6213, Université de Franche-Comté, 16, Route de Gray, 25030 Besançon, France
a r t i c l e
i n f o
Article history: Received 26 June 2013 Accepted 19 September 2013 Available online 29 September 2013 Keywords: Overshoot Polyelectrolyte Multilayer Reflectometry Adsorption kinetics Counterions SiO2 particles
a b s t r a c t Layer-by-layer adsorption of polycation poly(trimethylammonium ethyl methacrylate chloride) (MADQUAT) and polyanion poly(acrylic acid, sodium salt) (PAA) on silicon oxide substrates was studied in order to understand non-regular multilayer buildup. Indeed during MADQUAT adsorption, a special variation in the signal, called overshoots, was monitored with stagnation point adsorption reflectometry (SPAR). These overshoots were observed under different experimental conditions (pH, polymer concentration, salt concentration and molecular weight variations). A method called ‘‘substrate thickness method’’ was applied to determine the thickness, the refractive index and the adsorbed amount variation during the overshoot. Results clearly showed a decrease in the adsorbed amounts, but which does not necessarily indicate dissolution of the adsorbed multilayer. Therefore we also investigated layer-by-layer adsorption on colloidal silica particles (SiO2) under the same conditions as those in reflectometry. Indeed the high specific surface area of particles allowed titration of chloride ions and unadsorbed polymer in the bulk during the adsorption process. An increase in chloride ions concentration with no increase in the polymer concentration was observed. Accordingly, overshoots in cases of MADQUAT/PAA multilayer are not due to desorption of cationic polyelectrolyte but to desorption of chloride ions from the adsorbed multilayer film. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Adsorption of polyelectrolyte is a major topic in academic research but also relevant to many industrial processes. Depending on the purpose, polyelectrolytes can either be deposited, forming either a monolayer or a multilayer film. In order to buildup a multilayer film, a simple, low-cost technique can be used. This method is based on a layer-by-layer (LbL) adsorption of oppositely charged polyelectrolytes driven by electrostatic attractions. Since Decher’s research, [1] other studies have also shown the possibility of building a multilayer film using hydrogen bonding [2]. The alternate sequential adsorption is now extended to a wide range of compounds such as proteins [3–5], DNA [6,7], or inorganic particles [8–10]. To understand the mechanism involved in the growth of a polyelectrolyte film, many experimental techniques have been used. Among them, in situ techniques such as optical reflectometry (also called stagnation point adsorption reflectometry, SPAR), ellipsometry, surface plasmon resonance, optical waveguide light microscopy or quartz crystal microbalance with dissipation have been used to monitor the evolution of the adsorbed amount of each ⇑ Corresponding author. E-mail address: [email protected] (C.C. Buron). 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.09.040
polymer, the refractive index or the thickness of the deposited film. These parameters varied in duration from a few seconds up to several hours. Therefore numerous investigations have been carried out on the kinetic of polymer adsorption [11,12]. Under certain experimental conditions, specific evolution of the monitoring output appears at the beginning of adsorption. For example, in the study of adsorption kinetic by optical reflectometry, a phenomenon called overshoot was monitored on the reflectometer signal during polyelectrolyte adsorption. Indeed overshoot corresponds to a rapid increase in the reflectivity signal followed by a slow decrease. It not only has been observed during the adsorption of grafted copolymers poly(acrylamide)/poly(ethyleneoxide) [13], proteins [14] or cationic starch [15] but it has also appeared in multilayer films composed of cationic/anionic polyelectrolytes [16–18], cationic polyelectrolytes and colloidal silica (18 nm) [19], protein and polyelectrolyte [20], polypeptides [21] or proteins [22]. There is ongoing debate in the literature over possible explanations for overshoots observed in the polyelectrolyte multilayer buildup. Indeed Cohen Stuart and co-researchers have suggested a combined adsorption/desorption process of the polyelectrolytes [18,23]. They studied the effect of KCl, NaNO3 and CaCl2 salt concentration on the occurrence of the overshoot during polyelectrolyte adsorption. Using a stability diagram, they showed the
148
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
possibility of the formation of a soluble complex composed of the introduced polyelectrolyte and the previously adsorbed polymer above a critical salt concentration. Another explanation is a conformational rearrangement of polymer molecules at the surface. Filippov and Filippova [24] have theoretically demonstrated that the overshoots are controlled by conformational kinetics of the adsorbed layer. Using experimental data, Wågberg and co-researchers have also proposed as an explanation for the overshoots, a rearrangement of the polyelectrolyte conformation in the adsorbed layer resulting in a change in the reflectometer signal [25]. However any conformational changes at a constant adsorbed amount (optical mass) cannot be detected using reflectometry [13]. In our previous studies, we also observed overshoots in the reflectometer signal during the growth of multilayer films with oppositely charged polyelectrolytes [26,27]. In view of these findings, the aim of the present work was to investigate the reason for the overshoots appearing during the buildup of a multilayer made of a strong cationic polyelectrolyte and a weak anionic polyelectrolyte with special attention given to the counterions of each polyelectrolyte. Using an optical reflectometer equipped with an impinging jet cell, several physico-chemical parameters (pH, ionic strength, monovalent and divalent ions, weak and strong polyelectrolytes with different molecular weights) were tested to observe their effect on the overshoots. The refractive index and thickness variations during an overshoot were determined using a method previously developed by our team (substrate thickness method) [28]. Adsorption of layer-by-layer films was also carried out onto colloidal silica particles which due to their large specific surface area allowed us to determine the amount of adsorbed polymer. Adsorption of each polyelectrolyte was recorded using the titration of the unadsorbed amount of polymer. Concentration of chloride ions was measured in situ during MADQUAT adsorption with an ion-selective electrode.
2. Materials and methods 2.1. Chemicals and solutions Quaternized polydimethylaminoethyl methacrylate chloride (MADQUAT) and poly(acrylic acid, sodium salt) (PAA) were synthesized by COATEX (Genay, France). MADQUAT (207 g per molar unit) with different molecular weights was used as a strong cationic polyelectrolyte. PAA (94 g per molar unit) has an average molecular weight of 10 kDa. PAA has frequently been used as a weak anionic polyelectrolyte for adsorption studies [29,30] and has an intrinsic ionization pK of 5 in 10 3 M NaCl solution and a half ionization pK around 6.8 [31]. Poly(4-styrenesulfonate, sodium salt) (PSS) with an average molecular weight of 70 kDa was purchased from Alfa Aesar. PSS is a strong anionic polyelectrolyte which was totally ionized across the whole range of pH used in this study [32]. The chemical structures of the polyelectrolytes are presented in Fig. 1. Polymers were dissolved in ultrapure water having a resistivity above 18 MX cm (Milli-Q Plus, Millipore). The solution pH was adjusted with HCl (Aldrich) or NaOH (Aldrich) and the appropriate amount of analytical grade NaCl (Aldrich) was added to set the ionic strength. Polyelectrolyte solutions were prepared just before each experiment. Colloidal silica particles were synthesized using the Stöber method in ethanol [33]. After several washing/centrifugation cycles, particle size was measured in water using dynamic light scattering (Zetasizer 3000HS, Malvern). Monodispersed particles of 258 nm in diameter with a polydispersity index of 0.082 were obtained.
Fig. 1. Chemical structures of the polyelectrolytes used in this study.
2.2. Optical reflectometry An impinging jet cell was mounted on the reflectometer, which created a stagnation point in the center of the cell where the solution was introduced through a cylindrical channel drilled into the glass prism [34]. A linearly polarized He–Ne laser beam entered the cell through one side of the prism. The beam was reflected on the stagnation point, then emerged from the cell, and was split into its parallel (p) and perpendicular (s) components (with respect to the plane of incidence) by using a beam splitter cube. Photodiodes were used to detect the beams. The resulting electric intensities, Ip and Is, were then converted into electric tension and finally recorded. A two-way valve was used to switch between solutions with or without polyelectrolyte. Signals recorded during the injection of the polyelectrolyte free solution and polyelectrolyte solution were respectively denoted by S0 and S. Finally (S S0)/S0 represents, at least semi-quantitatively, the changes in adsorbed amount [18]. Silicon wafers coated with a silicon oxide (SiO2) layer were used as the substrate. Prior to each experiment, wafers were cleaned with piranha solution (one part of hydrogen peroxide (30%) in three parts of sulfuric acid (98%)), then rinsed and finally stored in ultrapure water until use. All experiments followed the same experimental procedure. In order to obtain a stable baseline of the reflectometer signal, a polymer free solution of the same pH and salt concentration as the subsequent polymer solutions was introduced into the cell for 5 min. Since the silicon oxide carried a negative surface charge under experimental conditions of pH 5.5 [35], MADQUAT was then introduced into the cell. In most cases in this study, ten layers of alternate polycation and polyanion (five bilayers) were deposited by switching the injection of the solution every 10 min. If not mentioned in the text, no rinsing step was done between each injection. For each of the experiments the solution flow rate was maintained at a constant rate. The experiments were repeated successfully three times with reproducibility of the order of 5%. In order to determine the mean refractive index and thickness of the adsorbed layer during the overshoots, a method called substrate thickness method was used. Briefly, the principle of the analytical procedure is based on a comparison between the experimental and calculated reflectivities ratio Rp/Rs for the same adsorption experiment carried out with at least two substrates of different thicknesses. Reflectometer output S is linked to the reflectivities ratio by an apparatus transmission factor f. For a more complete description of the procedure, please refer to our previous paper [28]. In the present work, silicon wafers with a silicon oxide layer of 115 ± 2 nm or 298 ± 1 nm were purchased from ACM (France) and were used to apply the method. Thicknesses of SiO2
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
layers were measured using a spectroscopic phase-modulated ellipsometer (Jobin Yvon, Model UVISEL). 2.3. Adsorption onto silica colloidal particles (SiO2) Silica particles (1 g) were dispersed in 100 mL of 10 3 M or 10 1 M sodium salt solution. The adsorption of the cationic polyelectrolyte was first done by adding 1 mL of a concentrated MADQUAT solution (100 g L 1). After 10 min, polymer-coated particles were separated by centrifugation at 13,000 rpm for 8 min and then dispersed in a solution containing 2 g L 1 of PAA for 10 min. The same procedure was applied successively in order to form six adsorbed layers onto the particles. In addition, concentration of polymer solutions was progressively increased to take into account the exponential growth of the adsorbed amounts [26]. In order to ensure good particle dispersion between each adsorption step, a 5 min ultrasonication step was included in the procedure. This was achieved using a classic ultrasonic bath (Elma T420) at a constant ultrasonic frequency of 75 kHz and an ultrasonic power of 70 W. The pH of the suspensions was adjusted to pH 5.5 prior to adding each polyelectrolyte solution. During the multilayer buildup, concentration of chloride ions was measured in situ using a silver combination electrode with a Hg/Hg2SO4 reference electrode (Radiometer Analytical, XM 900). Polymer concentration in the solution, i.e. the unadsorbed macromolecules, was determined after centrifugation using the total organic carbon technique (Shimadzu, TOC 5050). 3. Results and discussion 3.1. Effect of polyelectrolyte concentration A multilayer film made of five MADQUAT-PAA bilayers (denoted (MADQUAT-PAA)5) was built-up using different polyelectrolyte concentrations. The results of the study by reflectometry are presented in Fig. 2 with reflectometer output variation as a function of time. Each step is related to the adsorption of a polyelectrolyte given that each experiment begins with cationic polyelectrolyte adsorption. Results clearly showed a strong increase in the reflectometer plot (i.e., adsorbed amount) when the polyelectrolyte concentration increased. However, the first adsorption of MADQUAT did not depend on the polymer concentration because DS/S0 had almost the same value. Indeed polymer adsorp-
149
tion is ruled by the formation of an electrokinetic barrier [36] while adsorption kinetic is governed by the transport of the polymer to the surface vicinity. During the adsorption process, a progressive accumulation of charged polymer onto the substrate leads to repulsive electrostatic interactions between the adsorbed polyelectrolyte and the incoming polyelectrolyte that progressively stop the adsorption. The adsorption of MADQUAT seems to be of the high affinity type due to the rapid initial increase in the reflectometer signal. As stated in the introduction, the intention in this study was to investigate the reason for a peculiar shape of the signal, called overshoot, which was recorded during adsorption of MADQUAT. For example, at the 9th adsorption step (see arrow in Fig. 2), reflectometer output quickly increased and then slowly decreased. Moreover the amplitude of overshoots was higher when the polymer concentration increased. The same results were observed by Bijsterbosch et al. for monolayer adsorption of grafted copolymer onto silica [13]. 3.2. Effect of the electrolyte type and concentration In our previous studies, the overshoots were observed during layer-by-layer MADQUAT/PAA film growth under different experimental conditions. They were only recorded during MADQUAT adsorption. We noticed that the overshoots were more pronounced when the concentration of sodium chloride salt was increased [27]. But this phenomenon was not linked to the adsorbed amount of polymer. Interestingly the adsorbed amount was higher for 10 2 M than 10 1 M in NaCl. The type of salt was then changed to observe the effect on the growth of the multilayer. For a series of monovalent cations Li+, Na+, K+, Cs+ having the same counterion (Cl ) and for different monovalent anions NO3-, Cl with Na+ counterion, an overshoot was again observed during polycation adsorption. The ionic size of tested cations and anions does not seem to affect the overshoot. Concentration of each type of salt was also modified in the range of 10 3 M to 10 1 M without any modification of the overshoots. Finally the nature of the counterion and more precisely the effect of divalent cations on multilayer adsorption were investigated [37]. Surprisingly the overshoot phenomenon disappeared when BaCl2 and ZnCl2 were used as electrolytes. It may result from the formation of insoluble complexes between divalent cation and PAA. 3.3. Effect of molecular weight of polyelectrolytes Layer-by-layer deposition was then carried out at pH 5.5 with two PAA molecular weights of 10000 Da and 170000 Da (results not presented here). Adsorbed amounts after 5 deposited bilayers were the same even if the length of PAA polymer chain was very different. Overshoots were also recorded during MADQUAT adsorption but they appeared to be independent of the PAA molecular weight. However, the effect was highlighted when MADQUAT with different molecular weights was used (Fig. 3). There was a difference in the reflectometer signal recorded after adsorption of two bilayers. These measurements are reported in terms of specific viscosity which depends on the molecular weight of polymer according to the Mark-Houwink equation. For a specific viscosity (gsp) of 3.45 and 6.55, there was still an overshoot for MADQUAT adsorption whatever the number of adsorbed layers. However, for gsp = 14, overshoot vanished from the first layer to the fifth layer of adsorbed MADQUAT. Therefore the length of the polycation chain plays an important role in the overshoot occurrence. 3.4. Effect of the polymer charge density
Fig. 2. Reflectometer output variation as a function of polyelectrolyte concentration during (MADQUAT/PAA)5 multilayer formation (pH 5.5, NaCl 10 3 M).
10
Since the overshoot was the highest at a NaCl concentration of 1 M, polyelectrolyte solutions were prepared at this concentra-
150
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
ditions such as divalent cations, high length polycation chain and pH variation caused the overshoot to disappear. Multilayer assembly was then done using two strong polyelectrolytes: MADQUAT and PSS under the same conditions (pH 5.5, NaCl 10 1 M). Fig. 5 presents a steady increase in DS/S0. The value of each increment and also the adsorbed amount of MADQUAT or PSS, are the same for each polyelectrolyte adsorption step. Each polyelectrolyte is fully charged and can adsorb in an extended conformation onto the surface. The construction is similar to a stacking of polyelectrolyte layers since the amount of charge brought by each polyelectrolyte cannot change during the adsorption. On the other hand, a variation in ionization degree between PAA in solution and PAA within several kinds of multilayers was measured by Choi and Rubner [38]. A careful adjustment of the polymer charge density for both polyelectrolytes could be a way of avoiding the formation of overshoot.
Fig. 3. MADQUAT molecular weight effect on (MADQUAT/PAA)5 multilayer formation. Overshoots are indicated by an arrow for the 9th adsorbed layer. Molecular weights are given in terms of specific viscosity, gsp (pH 5.5, NaCl 10 3 M, Cp 10 mg L 1).
3.5. Thickness, refractive index and adsorbed amount variations during the overshoot
tion and pH was adjusted at pH 4 or pH 5.5 leading to a variation in the anionic polyelectrolyte charge density. Reflectometer outputs started to change after adsorption of 3 layers, as previously observed (Fig. 4). At the end of the experiment, the adsorbed amount of LbL film was higher for pH 4 than pH 5.5. Overshoots were only observed at pH 4 during the adsorption of the first and the second layer of MADQUAT. The same observation was made for NaCl 10 3 M solution (unshown results). When pH is increased to pH 5.5, charge density along the backbone of PAA changes because carboxylic acid groups are weak acidic functions. The charge of a weak polyelectrolyte is tunable with respect to the pKa and pH values. The degree of ionization of PAA is equal to 5% at pH 4 and 30% at pH 5.5 which was shown by Choi and Rubner [38]. When the polyelectrolyte is less charged, more adsorbed polymer chains are needed to reach the electrokinetic barrier. Consequently, a decrease in PAA charge density led to an increase in the adsorbed amounts of polymers and to ordinary variation in the reflectometer output (i.e., DS/S0 reaches a plateau). At this stage, the adsorbed amount seems to be a key parameter in understanding the overshoots in adsorption kinetic. Indeed, an increase in the adsorbed amount using several different experimental con-
New experiments were then carried out by increasing the MADQUAT injection time of MADQUAT solution from 10 min to 2 h. In order to obtain significant and well-defined overshoots, three polyelectrolyte bilayers were first prepared by sequential adsorption of MADQUAT and PAA for 10 min on two substrates (115 and 298 nm silica thicknesses). MADQUAT was then injected for 2 h. The results are presented in Fig. 6a. When a silicon wafer with a SiO2 thickness of 298 nm was used the overall negative value of DS/S0 can be explained by using the sign of the sensitivity factor (As) defined by Dijt et al. [34]. Indeed it was negative for this value of SiO2 thickness and quite similar in absolute value to a SiO2 thickness equal to 115 nm. Therefore both curves showed the same trend. Interestingly the overshoot vanished after 45 min and the output then increased very slowly until reaching a plateau. The explanation of the overshoot is still a subject of discussion in the literature. Indeed, Kovacevic et al. [18] suggested that the initial increase in adsorption followed by a slow decrease in reflectometer output was due to an adsorption/desorption process. Under these conditions, polyelectrolyte complexes formed at the interface slowly desorbed. However Wågberg et al. [25] assumed that there were some rearrangements of the polyelectrolytes in the adsorbed multilayer without necessarily desorption. To demonstrate this, polyelectrolyte adsorption was followed by quartz
Fig. 4. Reflectometer output variation as a function of pH of the solution during (MADQUAT/PAA)5 multilayer formation (NaCl 10 1 M, Cp 10 mg L 1).
Fig. 5. Reflectometer output variation during (MADQUAT/PSS)5 multilayer formation (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
151
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
Fig. 6. (a) Effect of polyelectrolyte supply: injection of MADQUAT for 2 h for two silicon oxide thicknesses (b) Calculated adsorbed amount by substrate thickness method for the seventh layer (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
crystal microbalance with dissipation. A close correlation between energy dissipation into the multilayers and a decrease in the adsorption as measured by reflectometry was established. Consequently, they concluded that desorption of polyelectrolyte complexes was not necessarily the origin of reflectometer signal decrease. However as it was clearly stated by Cohen Stuart and co-researchers [13] and as it can be demonstrated using an optical model of the interface, reflectometry is a technique that determines the adsorbed amount of polymers. However, any conformational changes cannot be detected for a constant adsorbed amount. In order to determine whether overshoots were due to desorption, rearrangement or both, we applied the substrate thickness method during MADQUAT adsorption using two different silicon oxide thicknesses (115 nm and 298 nm). Final values for thickness and refractive index were compared with spectroscopic ellipsometry measurements after drying and are summarized in Table 1. Results obtained from the substrate thickness method led to results similar to those using ellipsometry. Nevertheless a lower refractive index value was obtained by reflectometry compared to that of ellipsometry. This was probably due to the state of the multilayer because the film was not dried in the reflectometry experiments. Water molecules, trapped in the multilayer, induced a decrease in the refractive index value. However values were consistent and adsorbed amount variations were determined. Evolution of the adsorbed amounts during the overshoot is shown in Fig. 6b. After injection of MADQUAT (see the arrow in Fig. 6b) the uptake quickly increased and then slowly decreased, just like the reflectometer signal. Therefore desorption occurs as suggested by Cohen Stuart and co-researchers however it was not necessarily polyelectrolyte desorption and this will be discussed in the next section. 3.6. Injection of PAA at different times of the overshoot Initially two bilayers (NaCl 10 1 M) were deposited onto the substrate. Then the two-way valve that allowed the injection of
polyelectrolyte solution was switched for PAA adsorption after 2, 5 and 15 min of MADQUAT adsorption. Surprisingly in Fig. 7 the adsorbed amount of PAA did not depend on when the PAA injection started. Indeed the height of each increment was the same as for the previous study, i.e. after 2 h of MADQUAT injection (Fig. 6a). If the desorption of polyelectrolyte occurred, a constant charge density of the adsorbed multilayers was unlikely. Consequently, the adsorbed PAA amount should not be the same. This point seems to be in favor of a rearrangement in the multilayer film. However, we observed desorption (see Fig. 6b). These two results were combined to suggest an explanation of overshoots other than those proposed by Kovacevic et al. [18] or Wågberg et al. [25] that is polyelectrolyte desorption or polymer layer rearrangement without desorption respectively. Considering that there was no desorption of polyelectrolyte, the only way to explain the evolution of the reflectometer signal during MADQUAT adsorption is the ejection of ions from the LbL film leading to a modification of the refractive index. Several studies have shown that ions are embedded in the film to ensure electroneutrality especially when a weak polyelectrolyte is used. Thus for a poly(allylamine hydrochloride) (PAH)/PSS multilayer film, 50–80% of the PAH charge was screened by small ions [39,40]. In fact when MADQUAT bound to the adsorbed PAA layer, a change in PAA ionization degree occurred. Choi and Rubner [38] estimated a 25% increase in the degree of ionization at pH 5.5 for a multilayer made of poly(diallyldimethylammonium chloride) (PDADMAC) and PAA (pKa 5.5). PDADMAC is a strong polyelectrolyte, like MADQUAT. Consequently unbonded charges of MADQUAT, which were not used to bind the surface, were screened by chloride ions. They can interact with newly created anionic PAA charges which induces desorption of Cl from multilayer films. A simple calculation showed that, if a MADQUAT monomer was adsorbed, the mass of chloride ions represented 17% of the total monomer mass. Therefore, if chloride ions were removed from the multilayer, the adsorbed amount will inevitably decrease. The observed overshoots were probably due to the ejection of ions from the LbL film. Due to the time scale, this movement
Table 1 Thickness and refractive index of (MADQUAT/PAA)4 deposited bilayers onto silicon oxide layer. Note that the third MADQUAT adsorption was carried out for 2 h (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
Substrate thickness method (115 and 298 nm) Ellipsometry onto 115 nm Ellipsometry onto 298 nm
in situ After drying After drying
Thickness (nm)
Refractive index
18.2 ± 3.0 16.4 ± 2.0 20 ± 2.5
1.380 1.431 1.444
152
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153 Table 2 Evolution of MADQUAT concentration (Cp) and chloride ion concentration ([Cl ]) over time during the 4th MADQUAT adsorption at pH 5.5. Substrate: Si/SiO2/(MADQUAT/ PAA)3. Time (min)
NaCl 10
0 2 12 40
1.83 1.62 1.59 1.58
Cp (g L ±0.002
3 1
)
M
1
NaCl 10 [Cl ] (mM) ±0.1
Cp (g L ±0.002
0.95 3.3 4.6 5.0
6.14 5.36 5.26 5.25
1
)
M [Cl ] (M) ±0.1 0.120 0.128 0.129 0.130
tially complexed with quaternized amino groups of MADQUAT and not involved in the linking process.
3.8. Adsorption of MADQUAT/PAA LbL onto SiO2 colloidal particles Fig. 7. Injection of PAA solutions (see arrows) at different times (2 min, 5 min and 15 min) of overshoot encountered during MADQUAT adsorption (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
of ions seems to be a process with low desorption kinetic. Schlenoff et al. [41] revealed that no small ions were detected in PDADMAC/ PSS films. Their experiment supports our explanation and observations because no overshoot was recorded at NaCl 10 1 M when MADQUAT/PSS were assembled together (see Section 2.4). 3.7. Effect of rinsing step Initially, no rinsing was done between each polyelectrolyte adsorption. So a rinsing step of 5 min using a solution of the same pH and ionic strength was added in the buildup process. Unfortunately this did not change the overshoot shape. Consequently, a rinsing step was applied during an overshoot as it is presented in Fig. 8. For the fifth polymer layer, MADQUAT solution was injected for 5 min, a rinsing with NaCl 10 1 M solution was then done for 5 min and finally followed by another injection of MADQUAT. Surprisingly, output stopped decreasing during the rinsing step giving a flat signal and then decreased when MADQUAT solution was again injected. Firstly, this suggests that the overshoot phenomenon was only due to MADQUAT adsorption. Secondly, when the multilayer was flushed with a MADQUAT free solution, no desorption was observed. Consequently, MADQUAT adsorption leads to an ejection of counterions which are probably chloride ions ini-
The use of a substrate with an area of few square millimeters avoids the measurement of parameter variations such as pH or ion concentration in the bulk during the adsorption of polyelectrolytes. To circumvent the problem of a substrate with a small surface area, experiments were done using silica particles with a specific surface area of 10.6 m2 g 1 (calculated using the following equation: Sp = 3/(Rq), with R = 248 nm and q = 2.2 g cm 3). As for the reflectometry experiments, three bilayers of MADQUAT/PAA were first deposited onto particles using the process described in the experimental section. Then, during the adsorption of the 7th MADQUAT layer, the variations in unadsorbed polymer and chloride ion concentrations were recorded. Multilayer assemblies were performed at 10 3 M and 10 1 M in NaCl. Results are summarized in Table 2. From the addition of MADQUAT solution until 40 min following this, concentration of unadsorbed polymer decreased for both salt concentrations. At the same time an increase in chloride ion concentration was recorded especially for NaCl 10 3 M concentration. For 10 1 M, the slight variation was probably due to the high initial concentration in Cl . When MADQUAT began to adsorb onto the LbL film, an ejection of chloride ions occurred. Chloride ions and the negative charge of PAA compete to bind with the MADQUAT positive charge. The interactions between negative and positive charges of each polyelectrolyte were stronger because an adsorption was recorded. Indeed the values described an isotherm of adsorption which is well-known in the field of polymer research and also demonstrated that, in the case of MADQUAT/ PAA multilayer, there was no desorption of a complex/coacervate of polyelectrolytes. Therefore we infer that ions were ejected from the adsorbed multilayer.
4. Conclusion
Fig. 8. Effect of rinsing during MADQUAT adsorption on two previously adsorbed MADQUAT/PAA bilayers (pH 5.5, NaCl 10 1 M, Cp 10 mg L 1).
The growth of a layer-by-layer film made of a strong cationic polyelectrolyte (MADQUAT) and a weak anionic polyelectrolyte (PAA) was investigated using optical fixed-angle reflectometry. Our attention was focused on the special evolution of the reflectometer output called overshoot. Indeed during MADQUAT adsorption, the reflectometer signal exhibits a peak followed by a slow decrease. Several parameters such as polymer concentration and molecular weight, pH and type of salt and anionic polyelectrolyte (strong or weak) were changed in order to investigate their effects on the overshoots. Consequently, overshoot vanished when the pH was fixed at pH 4 during the self-assembly or when a strong polyelectrolyte (PSS) was used instead of a weak polyelectrolyte (PAA). This clearly indicates the role of the charge density of each polyelectro-
C.C. Buron, C. Filiâtre / Journal of Colloid and Interface Science 413 (2014) 147–153
lyte on this phenomenon that is confirmed by the significant effect of divalent cations. Furthermore, the determination of adsorbed amount variation during MADQUAT adsorption has shown a decrease in the uptake, suggesting a polymer desorption. However, as other researchers have claimed, it is thought that there is no polymer desorption but some rearrangement in the multilayer. Indeed, an experiment based on the injection at different steps of the overshoot has shown that the adsorbed amount of PAA did not change which is not in accordance with the desorption of polymer. To be able to measure the variations in the unadsorbed polymer and chloride ions concentrations in bulk solution during adsorption, colloidal silica particles were used. Titration of unabsorbed amount of MADQUAT has shown a decrease in MADQUAT concentration in the working solution indicating an adsorption of the polyelectrolyte. So overshoot is not due to desorption of a polyelectrolyte complex in the case of MADQUAT/PAA multilayer. Finally, an ejection of ions from the multilayer film during the adsorption of the cationic polyelectrolyte highlighted by following the chloride ion concentration in the bulk solution. In our case, the observed overshoots in the reflectometer signal are due to a decrease in the adsorbed multilayer refractive index owing to the chloride ion ejection. Further experiments will be performed to follow the concentration of sodium ions which may also be ejected from the multilayer during the construction. Acknowledgment This article is dedicated to the memory of Professor Alain Foissy who died in June 2009. At the end of his career, his research work focused on the influence of small ions in layer-by-layer polyelectrolyte adsorption. References [1] [2] [3] [4]
G. Decher, Science 277 (1997) 1232. S.A. Sukhishvili, S. Granick, Macromolecules 35 (2002) 301. F. Caruso, K. Niikura, D.N. Furlong, Y. Okahata, Langmuir 13 (1997) 3427. F. Caruso, D.N. Furlong, K. Ariga, I. Ichinose, T. Kunitake, Langmuir 14 (1998) 4559. [5] E. Lojou, P. Bianco, Langmuir 20 (2004) 748. [6] J. Zhang, L.S. Chua, D.M. Lynn, Langmuir 20 (2004) 8015.
153
[7] K. Ren, J. Ji, J. Shen, Bioconjugate Chem. 17 (2005) 77. [8] T. Sennerfors, F. Tiberg, J. Colloid Interface Sci. 238 (2001) 129. [9] T. Sennerfors, D. Solberg, F. Tiberg, Adsorption of polyelectrolyte-nanoparticle systems on silica: influence of ionic strength 254 (2002) 222. [10] F. Caruso, H. Lichtenfeld, M. Giersig, H. Möhwald, J. Am. Chem. Soc. 120 (1998) 8523. [11] J.C. Dijt, M.A. Cohen Stuart, G.J. Fleer, Macromolecules 25 (1992) 5416. [12] J.C. Dijt, G.J. Fleer, M.A. Cohen Stuart, J.E. Hofman, Colloids Surf. 51 (1990) 141. [13] H.D. Bijsterbosch, M.A. Cohen Stuart, G.J. Fleer, Macromolecules 31 (1998) 8981. [14] P. Van Dulm, W. Norde, J. Colloid Interface Sci. 91 (1983) 248. [15] Y. Li, W.-G. Hou, W.-Q. Zhu, Colloids Surf., A 303 (2007) 166. [16] N.G. Hoogeveen, M.A. Cohen Stuart, G.J. Fleer, M.R. Bohmer, Langmuir 12 (1996) 3675. [17] D. Kovacevic, S. Glavanovic, N. Peran, Colloids Surf., A 277 (2006) 177. [18] D. Kovacevic, S. van der Burgh, A. de Keizer, M.A. Cohen Stuart, Langmuir 18 (2002) 5607. [19] L.-E. Enarsson, L. Wågberg, J. Colloid Interface Sci. 325 (2008) 84. [20] O. Svensson, L. Lindh, M. Cárdenas, T. Arnebrant, J. Colloid Interface Sci. 299 (2006) 608. [21] S. Grohmann, H. Rothe, S. Eisenhuth, C. Hoffmann, K. Liefeith, Biointerphases 6 (2011) 54. [22] H. Shirahama, J. Lyklema, W. Norde, J. Colloid Interface Sci. 139 (1990) 177. [23] D. Kovacevic, S. van der Burgh, A. de Keizer, M.A. Cohen Stuart, J. Phys. Chem. B 107 (2003) 7998. [24] L.K. Filippov, N.L. Filippova, J. Colloid Interface Sci. 178 (1996) 571. [25] L. Wågberg, G. Pettersson, S. Notley, J. Colloid Interface Sci. 274 (2004) 480. [26] CC. Buron, C. Filiâtre, F. Membrey, H. Perrot, A. Foissy, J. Colloid Interface Sci 296 (2006) 409. [27] C.C. Buron, C. Filiatre, F. Membrey, C. Bainier, D. Charraut, A. Foissy, J. Colloid Interface Sci. 314 (2007) 358. [28] C.C. Buron, F. Membrey, C. Filiâtre, A. Foissy, Colloids Surf., A 289 (2006) 163. [29] S.S. Shiratori, M.F. Rubner, Macromolecules 33 (2000) 4213. [30] S.T. Dubas, J.B. Schlenoff, Langmuir 17 (2001) 7725. [31] I. Pochard, P. Couchot, A. Foissy, Colloid Polym. Sci. 1088 (1998) 276. [32] M. Balastre, J. Persello, A. Foissy, J.F. Argillier, J. Colloid Interface Sci. 219 (1999) 155. [33] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62. [34] J.C. Dijt, M.A. Cohen Stuart, G.J. Fleer, Adv. Colloid Interface Sci. 50 (1994) 79. [35] J. Persello, in: E. Papirer (Ed.), Adsorption on Silica Surfaces, vol. 90, 1998, pp. 297–342. [36] G.J. Fleer, M.A. Cohen Stuart, J.M.H.M. Scheutjens, T. Cosgrove, B. Vincent, Polymers at Interfaces, Chapman and Hall, London, 1993. [37] C.C. Buron, C. Filiâtre, F. Membrey, C. Bainier, L. Buisson, D. Charraut, A. Foissy, Thin Solid Films 517 (2009) 2611. [38] J. Choi, M.F. Rubner, Macromolecules 38 (2005) 116. [39] P. Lavalle, C. Picart, J. Mutterer, C. Gergely, H. Reiss, J.-C. Voegel, B. Senger, P. Schaaf, J. Phys. Chem. B 108 (2004) 635. [40] J. Schmitt, T. Gruenewald, G. Decher, P.S. Pershan, K. Kjaer, M. Loesche, Macromolecules 26 (1993) 7058. [41] J.B. Schlenoff, H. Ly, M. Li, J. Am. Chem. Soc. 120 (1998) 7626.
