Effect of colloidal silica and electrolyte on the structure of an adsorbed cationic polyelectrolyte layer

Effect of colloidal silica and electrolyte on the structure of an adsorbed cationic polyelectrolyte layer

Colloids and Surfaces A: Physicochemical and Engineering Aspects 155 (1999) 145 – 154 www.elsevier.nl/locate/colsurfa Effect of colloidal silica and ...

164KB Sizes 2 Downloads 30 Views

Colloids and Surfaces A: Physicochemical and Engineering Aspects 155 (1999) 145 – 154 www.elsevier.nl/locate/colsurfa

Effect of colloidal silica and electrolyte on the structure of an adsorbed cationic polyelectrolyte layer Sara Stemme a, Lars O8 dberg a,*, Martin Malmsten b a

Swedish Pulp and Paper Research Institute-STFI, Box 5604, SE-114 86 Stockholm, Sweden b Institute for Surface Chemistry-YKI, Box 5607, SE-114 86 Stockholm, Sweden Received 21 October 1997; accepted 6 January 1999

Abstract The interactions between an adsorbed high molecular weight cationic polyacrylamide (C-PAM) and anionic components have been investigated by ellipsometry. The anionic components used were non-aggregated and microaggregated anionic colloidal silica particles (ACS) and an anionic polyacrylamide (A-PAM). The C-PAM was adsorbed to a silica surface and the different anionic components were then added to the system. The effect of adding an electrolyte (NaCl) to some of these systems was also investigated. The thickness of the adsorbed layer, the adsorbed amount and some kinetic aspects of the adsorption and desorption processes were studied. A three-fold expansion was observed for the C-PAM layer when non-aggregated and microaggregated ACS was added. When large amounts of NaCl are added there is a great decrease in layer thickness for a system with non-aggregated ACS. A system with microaggregated ACS particles is less affected by an increase in the electrolyte concentration. More ACS is adsorbed to a C-PAM layer in the presence of NaCl than in its absence. When A-PAM is added to the C-PAM layer, the thickness and adsorbed amount is much less influenced than if ACS is added. A small increase in adsorption and in layer thickness is first observed and then a desorption process starts. Information about the conformation of the adsorbed cationic polyacrylamide has also been obtained. The ellipsometry measurements indicate that the structure of the adsorbed layer changes as the adsorption process proceeds. The polymers adsorbed first to the surface give a more dense layer than the polymers adsorbed last. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Polyelectrolyte adsorption; Ellipsometry; Ionic strength; Silica; Cationic polyacrylamide; Flocculation

1. Introduction High molecular weight cationic polyacrylamides (C-PAM) are widely used as flocculants. * Corresponding author. Tel.: +46-8655-9175; fax: + 468655-9421. E-mail address: [email protected] (L. O8 dberg)

In the papermaking industry they are frequently used as retention aids. One factor which determines the efficiency of these polymers as retention aids is their conformation at the surface on which they adsorb. The conformation depends mainly on the charge density and the molecular weight of the polymer, the ionic strength of the solution and the surface charge density [1].

0927-7757/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 7 - 7 7 5 7 ( 9 9 ) 0 0 0 2 3 - 0

146

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

High molecular weight C-PAM flocculates mostly according to a bridging mechanism [2]. This means that a C-PAM molecule forms a bridge between two particles. To obtain such a bridge, the adsorbed molecule must have an extension that is at least of the same magnitude as the distance between the particles, which is often determined by the thickness of the electrical double layer. Even though a polymer may have an extended conformation in solution this does not mean that it is extended when adsorbed to a surface. When the polymer is adsorbed to a surface, the interaction with the surface is of great importance and will normally result in an altered conformation. A highly charged polymer will at low electrolyte concentration have an extended conformation in solution but a rather flat conformation at a surface. Two-component systems are often used as retention aids [3]. The most commonly used dual systems consist of a cationic polymer together with an anionic component. The anionic component can be either an anionic polymer, often a polyacrylamide, or a negatively charged particle. These particles are often called microparticles in the papermaking industry. In our study, we used a high molecular weight cationic polyacrylamide together with anionic silica particles. The size of the silica particles is of the order of nanometers and the silica is either non-aggregated or microaggregated, (Section 2). These silicas are known to have different flocculation efficiencies depending on the cationic polymer in the system. Microaggregated silica is preferentially used with C-PAM, whereas non-aggregated silica is reported to give the best effect with cationic starch [4]. These two-component systems of a cationic polymer and an anionic component have several advantages as retention aids, one example being a high shear resistance [5]. A high shear resistance is important since the hydrodynamic forces on a modern high speed paper machine are quite high. The microparticle systems also have a high reflocculation tendency even after long shearing times [6]. Other characteristic features of the microparticle systems are a high porosity in the sheets formed and an increased dewatering effect, both on the wire and in the press section [3].

The electrolyte concentration is an important factor influencing the polymer conformation. Several studies have recently been performed in this area [7–9]. An important trend in the papermaking industry is towards a more closed water system. In a closed system there will be an accumulation of electrolytes and substances that may interact with the added chemicals such as retention aids. The aim of this study has been to investigate the interactions between silica particles and adsorbed C-PAM layers and the effects of electrolyte on these systems. The interactions between an anionic polyacrylamide and C-PAM were also investigated. The ellipsometry technique has been chosen since it is a suitable method to obtain information about the in situ conformation of adsorbed polymer layers.

2. Experimental

2.1. Materials 2.1.1. Polymers The polymers used were a cationic and an anionic polyacrylamide from Allied Colloids, Bradford, UK. The polymer denoted C-PAM is a cationic copolymer of acrylamide and [(N,N,Ntrimethylammonio)propyl]-methylacrylamide (MAPTAC). The charge density was determined by polyelectrolyte titration with an anionic polymer, potassium polyvinylsulphate, KPVS (Wako Pure Chemicals, Japan), in 1 mM NaHCO3. Cationic orthotoluidine blue, OTB, was used as indicator [10]. From the experimentally determined charge density of 0.55 meq/g, the degree of substitution was calculated to be 0.043 for this polymer. The molecular weight (Mw) of the polymer was determined by size exclusion chromatography to be 5.9×106 [11]. Percol 173 is a high Mw anionic polyacrylamide (Mw \ 15× 106, intrinsic viscosity 13 dl/g) with a degree of substitution of 0.05, according to the manufacturer. The water used was purified with a modified Millipore Milli-Q purification system with a final 0.2-mm Zetapore filter. Inorganic

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

electrolytes of analytical grade were used in all experiments. The standard buffer was 1 mM NaHCO3 which gives pH 8. The buffer was present in all experiments. Stock solutions of the polymers were prepared in the following way. The polymers were first wetted for 2 min with a small amount of ethanol and then diluted with a small amount of 1 mM NaHCO3 and shaken vigorously for 2 min. After further dilution, the solutions were stirred with a magnetic stirrer for 2 h and then finally left to swell for at least 12 h before use.

2.1.2. Anionic colloidal silica products The microparticles used were a non-aggregated and a microaggregated anionic colloidal silica (ACS) denoted ACS I and ACS II, respectively. They were manufactured and characterized by Eka Chemicals AB (their trade names are BMA-9 and BMA-780, respectively). To obtain stable products the silica was modified, having approximately 10% of aluminol groups on the surface. ACS I is a non-aggregated anionic colloidal silica with a particle size of 5 nm and a charge density of 0.55 meq/g at pH 8. ACS II is a microaggregated ACS built up by 3 nm particles in an elongated structure. Assuming spherical aggregates the radius of ACS II is approximately 25 nm as determined by dynamic light scattering [12] and with a charge density of 0.8 meq/g. The charge densities were estimated by polyelectrolyte titration [10,13] with a cationic highly charged, low Mw polyelectrolyte (poly-3,6-ionene, Janssen Chimica). This type of microparticles and their properties are more thoroughly discussed in Ref. [4]. 2.1.3. Substrate surfaces The substrates used for adsorption were silica surfaces of 10× 15 mm. These substrates were cut from standard silicon wafers used in semiconductor processing. The surfaces are close to atomically flat and excellent substrates for optical studies of thin adsorbed films. To further enhance the sensitivity in the optical measurements, the substrates were thermally oxidized to give a silicon dioxide layer with a thickness of typically 30 nm. Before the measurements, the silicon wafers

147

were cleaned in H2O2:NH4OH:H2O (1:1:5) at 80°C for 5 min followed by a similar treatment in which the NH4OH was replaced by HCl [14]. The resulting surface is extremely hydrophilic with a contact angle for water close to zero. The surface is negatively charged (point of zero charge pH 3) due to the hydroxyl groups on the silicon dioxide.

2.2. Methods 2.2.1. Theory Ellipsometry is an optical technique which measures the polarization changes occurring at oblique reflection from a surface. These polarization changes are very sensitive to the presence of a thin film or a layer of adsorbed molecules. The physical quantity measured by ellipsometry is the complex reflectance ratio r=Rp/Rs where Rp and Rs are the complex reflection coefficients for light polarized parallel (p) and perpendicular (s) to the plane of incidence. r is often expressed in polar form as r= tan C exp(i D), where C and D are the ellipsometric angles determined by the ellipsometer. From C and D, it is possible to calculate properties of a thin surface film providing that the optical properties of the substrate and of the ambient medium are known, as well as the angle of incidence and the wavelength of the light. With a multi ambient media measurement approach [15,16] both the silicon and the silica layer were taken into account. Four-zone null ellipsometry was used to cancel most imperfections in the components of the apparatus [17]. After these determinations, the adsorption process was investigated. The results of C and D 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 by solving numerically the Drude equations as described in Ref. [14]. Surface concentrations were calculated according to de Feijter [18], using a refractive index increment of 0.162 cm3/g for C-PAM. The refractive index increment was measured by using an Abbe refractometer.

148

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

2.2.2. Instrument The ellipsometer used was an automated null ellipsometer (Rudolph Research, NJ, type 436) operating at 401.5 nm. A detailed description of the apparatus and the experimental setup is given by Landgren and Jo¨nsson [15]. In each experiment, the silica surface was put into a quartz cuvette with a magnetic stirrer and a volume of 5 ml. The polymer concentration of 10 mg/l was reached by adding a concentrated polymer solution (1 g/l) to the buffer solution in the cuvette. All measurements were made at room temperature (22°C).

3. Results and discussion

3.1. Adsorption of a C-PAM to a silica surface Fig. 1 shows the adsorption of C-PAM to a silica surface from a 10 mg/l solution in 1 mM NaHCO3. In the figure, the refractive index and layer thickness are plotted as a function of the adsorbed amount during the build-up of a complete polymer layer. The polymer concentration used was high enough to reach the plateau level. When a similar polymer was used there was no difference in adsorbed amount on a silica surface when the concentration was changed from 10 to 200 mg/l [8]. At the end of adsorption 2.19 0.2 mg/m2 was adsorbed to the surface, giving a layer thickness of approximately 30 nm. At the very lowest adsorbed amounts the sensitivity of the

Fig. 1. Thickness () and refractive index ( ) of an adsorbed C-PAM layer as a function of adsorbed amount; 1 mM NaHCO3 buffer, 10 mg/l C-PAM.

ellipsometric system is insufficient resulting in a large scatter in the experimental points. The adsorption of a polyelectrolyte at a low ionic strength to an oppositely charged surface can as a first approximation be regarded as an ion-exchange process. The polymer adsorbs to the surface until charge compensation is closely reached. The adsorbed amount of the polyelectrolyte is thus inversely proportional to the charge density of the polymer. Bolt [19] reports a charge density of 1.0 meq/m2 for silica at pH 8 and high electrolyte concentration determined with a titration method. The value at high electrolyte concentration should be chosen for comparison, since in our case the polyelectrolyte has neutralized the surface charge and the pH should be approximately the same at the surface as in the solution. This gives an adsorbed amount of 1.8 mg/m2 at charge neutralization which should be compared with our experimental value of 2.1 mg/m2. This slight charge overcompensation has been observed in other studies [9,20]. After the first polymer molecules have adsorbed a polymer layer with a thickness of approximately 15 nm is rapidly established, see Fig. 1. Thereafter more polymer is adsorbed while the thickness of the layer remains almost constant. This suggests that the molecules have the same conformation. This is also reflected in the constant increase in the refractive index during the adsorption. For the last part of the adsorption process another layer structure is obtained. It is evident in Fig. 1 that the layer thickness starts to increase much more rapidly while the last 0.3–0.4 mg/m2, corresponding to 20% of the adsorbed amount, of the polymer is adsorbed. Obviously this adsorption process gives a less dense polymer layer. The less dense layer is also reflected in the decrease in the refractive index. This structure of the adsorbed polyelectrolyte layer is also supported by [21] where exchange reactions for polyelectrolytes on polystyrene latex were investigated. The last adsorbed polymers ( 20%) were more easily exchanged with other polymers than the polymers first adsorbed. It should also be mentioned that these systems are not expected to be in complete thermodynamic equilibrium. There are very few studies on

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

the conformation changes for polyelectrolytes after the adsorption plateau is reached. However, in Ref. [9] the adsorption of a highly charged cationic polyacrylamide onto mica in 0.1 mM Na2SO4 was investigated by a surface force apparatus. Even after 5 days, the structure of the polymer layer was still changing. A complete layer of a C-PAM on a silica surface for a very similar polyelectrolyte was also studied by ellipsometry by O8 dberg et al. [8]. In their study the adsorbed amount was found to be 1.7 mg/m2 and the layer thickness 40 nm. The difference in the adsorbed amounts is explained by different values used for the refractive index increment. If the same increment is used, the same adsorbed amount is obtained. The discrepancy in the layer thickness is probably explained by the different measurement procedures used. In our investigation, we have used a four-zone measurement to cancel imperfections in the components of the apparatus. The silica and silicon layer have been determined separately instead of using a combined value (Section 2.2). These differences in the methods used could be an explanation for the different values obtained for the thickness. In [7], the adsorption of a less charged C-PAM (degree of substitution 0.034) and with a lower Mw, Mw =1× 106 was studied, and an adsorbed amount of 2.3 mg/m2 and a layer thickness of 18 nm at pH 5.6 and 1 mM 1:1 electrolyte were obtained from ellipsometry. This adsorbed amount also corresponds to an ion-exchange process with a slight overcompensation (5%) of C-PAM.

3.2. Interactions between adsorbed C-PAM and non-aggregated silica C-PAM was adsorbed to a silica surface from a 10 mg/l solution. The polymer solution was then removed from the cuvette by rinsing with buffer. A concentrated solution of non-aggregated silica sol (ACS I) was added, giving a concentration of 10 mg/l in the cuvette, in order to study the interactions between the C-PAM layer and the ACS I. Fig. 2 shows an adsorption of 0.5 mg/m2 of ACS I during the first couple of seconds and thereafter a slow desorption. The thickness in-

149

Fig. 2. Adsorbed amount ( ) and thickness () of a C-PAM/ ACS I layer shown as a function of time. The arrows mark the addition of an ACS I solution to a preadsorbed C-PAM layer.

creased rapidly from 25 to 70 nm becoming approximately 95 nm after 1 h. As in all other measurements, the experiment was performed in a 1 mM NaHCO3 buffer solution. Due to the complex system studied when ACS is added to the polyelectrolyte layer in the presence of electrolyte and also because of the large layer thickness observed (being comparable to the wavelength used in the ellipsometer), we aim primarily at a qualitative interpretation of the observed effects. The surface charge is often overcompensated when a polyelectrolyte adsorbs to a surface. The electrostatic forces between the stretched out and overcompensated polymer layer and the oppositely charged ACS are probably one driving force for the initial adsorption of approximately 0.5 mg/m2 of ACS I. The value of 0.5 mg/m2 has been calculated using the same refractive index increment as for the polymer. The adsorption will partly neutralize and hence give a less charged polymer layer. A less charged polymer generally results in a more expanded layer [22]. After the initial expansion of the layer and adsorption of the ACS, the expansion continues at a much slower rate and some desorption occurs. One explanation of the slow reconformation could be a slow diffusion of the ACS particles into the polymer layer. The effects of ACS should also be compared with the normal electrolyte effect which also gives an expanded layer and some desorption [7–9]. In our study, the desorption was however much slower. When the ACS I concentration was increased to 50 mg/l, the slow desorption contin-

150

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

ued and no significant difference was seen in the thickness (Fig. 2). The desorption had not ended after 90 min when the ACS I solution was changed to a buffer (not shown in the figure). Nor was any change observed in the layer thickness or in the adsorbed amount. This system with C-PAM and ACS I adsorbed to a silica surface in a 1 mM NaHCO3 buffer solution was thereafter subjected to higher electrolyte concentrations. The excess of ACS I was removed from the cuvette by rinsing with buffer, and a NaCl solution was added giving a concentration of 10 – 100 mM NaCl in the cuvette. The layer thickness decreased from 100 to 20 nm with increasing NaCl concentration, as shown in Fig. 3. With increasing electrolyte concentration, the layer thickness of an adsorbed polyelectrolyte layer will in general increase [7 – 9,22]. In our case, when ACS was added before the electrolyte concentration was raised, the opposite effect was observed (Fig. 3). When NaCl is added to this polyelectrolyte layer, the NaCl screens the interactions between the cationic polymer and the surface. It also screens the interactions within the layer, which in the present case are primarily the interactions between the negatively charged ACS and the polyelectrolyte. Obviously the latter effect is the more important. The layer thickness decreased gradually to 20 nm as the concentration was raised to 100 mM NaCl. At 100 mM NaCl

Fig. 3. Adsorbed amount ( ) and thickness () of a C-PAM/ ACS I layer shown as a function of time. The arrows indicate the time when NaCl is added to the cuvette. There was no ACS I in the solution when NaCl was added. The final arrow indicates the addition of ACS I to the solution.

the layer was even thinner than the original thickness before any addition of ACS I. The layer at 100 mM NaCl was however denser and had a higher refractive index. This observed influence of electrolyte concentration on the conformation could be an explanation of the observed difficulties in using non-aggregated ACS together with C-PAM as a retention aid system [4]. Microaggregated ACS adsorbed to C-PAM is not as sensitive to electrolyte, as will be shown below. The results in Fig. 3 also show an increased adsorbed amount during the addition of electrolyte. This could be due to incorporation of electrolyte in the adsorbed layer. It could also be a consequence of the simplified optical model used. After the C-PAM/ACS layer had contracted due to the addition of NaCl a test was carried out to see whether the layer could be made to expand by adding more ACS. A concentrated ACS I solution was added giving a concentration of 10 mg/l in the 100 mM NaCl solution in the cuvette. The layer thickness increased to 100 nm and the adsorption of ACS I increased to almost 14 mg/ m2, Fig. 3. One explanation for this large adsorption could be coagulation of ACS at the silica surface. But this does not seem to be probable, according to the literature. Iler [23] gives an empirical formula according to which ACS would gel at a sodium salt concentration of 0.28 M for a 10 mg/l ACS solution over a period of a few hours. Even with a concentration as high as 300 g/l, the ACS would gel first at a salt concentration of 0.13 M. This ACS concentration is as estimated from the ellipsometry measurements even higher than the surface concentration of silica in our system. In our case, we also used an alumina-modified sol which has an even higher stability against sodium chloride [23]. Since a large increase in adsorbed ACS I to the polyelectrolyte layer was observed when the NaCl concentration was increased, the adsorption of silica to the polyelectrolyte layer as a function of the electrolyte concentration was investigated systematically. The results are given in Fig. 4. The measurement was made by first adsorbing the C-PAM to the silica surface. The cuvette was thereafter rinsed with buffer. ACS I was added

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

Fig. 4. Adsorbed amount ( ) and thickness () of a C-PAM layer in the presence of a 10 mg/l ACS solution as a function of NaCl concentration. The measurements are made at least 20 min after the last NaCl addition.

giving a concentration of 10 mg/l. Finally the electrolyte concentration was raised gradually to 200 mM. The adsorption of ACS I increased with increasing electrolyte concentration up to approximately 50 mM where a plateau level was reached.

3.3. Interactions between adsorbed C-PAM and microaggregated silica The same procedures as above were used to study the interactions between microaggregated ACS (ACS II) and the polyelectrolyte layer. CPAM was adsorbed to a silica surface and, after rinsing with buffer ACS II was added to the cuvette giving a concentration of 10 mg/l. Fig. 5 shows that 1.5 mg/m2 of ACS II adsorbed to the C-PAM layer and that the layer expanded from 30 to 85 nm.

Fig. 5. Adsorbed amount ( ) and thickness () of a C-PAM/ ACS II (microaggregated ACS) layer as a function of time. A 10 mg/l ACS II solution was added to a preadsorbed C-PAM layer.

151

As in the case of ACS I, the anionic ACS II adsorbed to the cationically overcompensated CPAM layer, probably due to electrostatic interactions. Even though the microaggregated silica (ACS II) had a higher charge density than the non-aggregated (ACS I), the polyelectrolyte layer adsorbed more of the microaggregated silica. A reasonable explanation is that all negative sites on the ACS II are not available to interact with C-PAM because of the geometrical structure of the microaggregated ACS. This would give a less effective charge density for the ACS II. Another explanation could be a smaller entropy loss when the microaggregated silica is used than in the case of non-aggregated silica. The adsorption of ACS II is a slower process than the adsorption of ACS I, Fig. 2. This is in agreement with the larger size of ACS II, which gives a slower diffusion rate. During the experimental time of 2 h, no desorption was observed in the ACS II system, in contrast to the behavior with ACS I. Nor was any change observed in the system when the ACS II solution was exchanged for a buffer. At the end of the adsorption process, the layer thickness was of the same order as for the non-aggregated ACS, which means approximately 90 nm (not shown in the figure). The sensitivity to electrolyte was investigated as already described for ACS I. The excess of ACS II was removed from the cuvette by rinsing with buffer. A NaCl solution was thereafter added giving 10–100 mM NaCl in the cuvette. Fig. 6

Fig. 6. Adsorbed amount ( ) and thickness () for a C-PAM layer with preadsorbed ACS II as a function of time. There was no ACS II in the solution when NaCl was added to the cuvette. The additions of NaCl are indicated by arrows. The last arrow indicates an addition of a 10 mg/l ACS II solution.

152

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

shows that the layer thickness decreased from 90 to 60 nm. When the electrolyte concentration was raised a significantly smaller change in layer thickness and adsorbed amount was thus observed for ACS II than for ACS I. Since the layer thickness is of importance for flocculation, this could be one of the reasons why ACS II is more efficient than ACS I when it is used together with C-PAM as a retention aid. After the contraction of the C-PAM/ACS layer due to the addition of NaCl, we investigated whether the layer could be made to expand by adding more ACS, as was the case for the non-aggregated ACS. A 10 mg/l ACS II solution was added to the 100 mM NaCl solution in the cuvette. The layer thickness increased from 60 to 70 nm and the adsorbed amount increased to 6.0 mg/m2 at the end of the adsorption. It should perhaps once more be pointed out that the systems with ACS I and ACS II are both very complex and the optical model is simplified. Further we have used the same refractive index increment for ACS as for the polymer. Approximate measurements with an Abbe refractometer showed a lower increment of 0.07, which if used would give an even higher adsorbed amount of ACS. These remarks show that the results for the systems with ACS are only qualitative.

3.4. Interactions between adsorbed C-PAM and anionic polyacrylamide Fig. 7 shows the interactions between an anionic polyacryamide (A-PAM) and a C-PAM layer already adsorbed to the silica surface. Before the addition of A-PAM, the system was rinsed with buffer to remove the excess of C-PAM solution. Initially, there was an adsorption (0.5 mg/ m2) of A-PAM that after some time changed to a desorption. There was a small increase in the layer thickness simultaneously with the adsorption of A-PAM. When the desorption started, there was also a decrease of the polymer layer thickness. As for the experiments in Fig. 1, there is a large scatter in the experimental points during the initial adsorption of C-PAM due to the insufficient sensitivity of the ellipsometric system.

Fig. 7. Adsorbed amount ( ) and thickness () when adding A-PAM to a C-PAM layer. The arrow indicates the addition of a 10 mg/l A-PAM solution. Before the addition, the system was rinsed with a 1-mM NaHCO3 buffer.

The initial adsorption of A-PAM is due to electrostatic interactions between the positively charged C-PAM layer and the negatively charged A-PAM. These types of multilayers of polyelectrolytes have recently been studied [24,25]. In Ref. [24], different situations are distinguished depending on the charge density of the polymers used. In the case with more highly charged polymers (5– 20%) complexation occurs. When the polymer charge is in the lower part of this interval, only a few bonds remain to keep the complex attached to the substrate and desorption of the complex can take place. This is the situation in our study. First an adsorption of the A-PAM to the C-PAM occurs and this gives an increase in the adsorbed amount. Thereafter a desorption process is observed, which is explained by the formation of C-PAM/A-PAM complexes that desorb from the surface. Not only microparticle systems are used as dual retention aids. Cationic polymers together with high Mw A-PAM are also used. When these dual systems are used, a cationic polymer is first adsorbed onto the fibers and filler particles. A high Mw A-PAM is then added to form bridges between the cationic particles. For a properly working system, an adsorbed A-PAM with an extended conformation is needed. From our measurements it is difficult to draw any conclusions regarding the A-PAM conformation. A small expansion of

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

the layer is initially seen but, since the mean value is observed and the adsorption change is small, the conformation cannot be evaluated. The experiment indicates however that the present system is not a well working retention aid system due to the small layer expansion and also due to the desorption. It is therefore not surprising that in practice more highly charged medium Mw cationic polymers (e.g. poly-DADMAC) are used. This is also in accordance with [24] where it is shown that a more highly charged polymer will give a more stable system. We did however in this investigation stick to the C-PAM as the cationic polymer to be able to make comparisons with the microparticle systems. There could also be a kinetic explanation for the small adsorption of A-PAM. In Ref. [26], it was shown that more A-PAM was adsorbed to the preadsorbed C-PAM layer if the A-PAM was added shortly after the C-PAM. This does not give C-PAM time to change to a more flat conformation.

4. Conclusion The interactions between an adsorbed high Mw cationic polyacrylamide (C-PAM) and anionic components have been investigated by ellipsometry. A three-fold expansion of the C-PAM layer was observed when non-aggregated and microaggregated ACS was added at low electrolyte concentrations. When high additions of NaCl are made, there is a great decrease in layer thickness for a system with non-aggregated ACS. A system with microaggregated ACS particles is less affected by an increase in electrolyte concentration. This can be one reason why in practice microaggregated ACS gives better flocculation together with CPAM. More ACS is adsorbed to a C-PAM layer in the presence of NaCl than in absence of NaCl. When A-PAM is added to the C-PAM layer, the thickness and adsorbed amount are influenced much less than when ACS is added. A small increase in adsorption and layer thickness is first observed and then a desorption process starts. Information about the conformation of the adsorbed cationic polyacrylamide has also been ob-

153

tained. The ellipsometry measurements indicate that the structure of the adsorbed layer changes as the adsorption process proceeds. The polymers adsorbed first to the surface give a more dense layer than the polymers adsorbed last.

Acknowledgements We should like to thank Agne Swerin for stimulating discussions and Anthony Bristow for a linguistic revision of the manuscript

References [1] G.J. Fleer, M.A. Cohen Stuart, J.M.H.M. Scheutjens, T. Cosgrove, B. Vincent, Polymers at Interfaces, 1st ed., Chapman & Hall, London, 1993. [2] E. Dickinson, L. Eriksson, Adv. Colloid Interface Sci. 34 (1991) 1. [3] T. Lindstro¨m, Some fundamental chemical aspects on paper forming. In: C.F. Baker, V.W. Punton (Eds.), Fundamentals of Papermaking, Trans. Ninth Fundam. Res. Symp., Cambridge, Mech. Publ. Eng. Publ. Ltd, London, 1989, p. 311. [4] K. Andersson, E. Lindgren, Nord. Pulp Pap. Res. J. 11 (1) (1996) 15. [5] Britt, K.W. Tappi, 1973;56(83 – 86). [6] A. Swerin, G. Risinger, L. O8 dberg, J. Pulp Pap. Sci. 23 (1997) 374. [7] V. Shubin, P. Linse, J. Phys. Chem. 99 (4) (1995) 1285. [8] L. O8 dberg, S. Sandberg, S. Welin-Klintstro¨m, H. Arwin, Langmuir 11 (7) (1995) 2621. [9] M.A.G. Dahlgren, P.M. Claesson, R. Audebert, Nord. Pulp Pap. Res. J. 8 (1) (1993) 62. [10] H. Terayama, J. Polym. Sci.: Part B: Polym. Chem. 8 (2) (1952) 243. [11] A. Swerin, L. Wa˚gberg, Nord. Pulp Pap. Res. J. 9 (1) (1994) 18. [12] D. Biddle, C. Walldal, S. Wall, Colloids Surf. A: Physicochem. Eng. Aspects 118 (1996) 89. [13] D. Horn, Prog. Colloid Polym. Sci. 65 (1978) 251. [14] W. Kern, D.A. Puotinen, RCA Rev. 31 (2) (1970) 187. [15] M. Landgren, B. Jo¨nsson, J. Phys. Chem. 97 (8) (1993) 1656. [16] F. Tiberg, M. Landgren, Langmuir 9 (4) (1993) 927. [17] R.M.A. Azzam, N.M. Bashara, Ellipsometry and Polarized Light, North-Holland, Amsterdam, 1989. [18] J.A. de Feijter, J. Benjamins, F.A. Veer, Biopolymers 17 (1978) 1759. [19] G.H. Bolt, J. Phys. Chem. 61 (1957) 1166. [20] H. Tanaka, L. O8 dberg, L. Wa˚gberg, T. Lindsto¨m, J. Colloid Interface Sci. 134 (1) (1990) 219.

154

S. Stemme et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 155 (1999) 145–154

[21] H. Tanaka, A. Swerin, L. O8 dberg, Langmuir 10 (10) (1994) 3466. [22] H.G.M. van de Steeg, M.A. Cohen Stuart, A. de Keizer, B.H. Bijsterbosch, Langmuir 8 (1992) 2538. [23] R.K. Iler, The Chemistry of Silica, Wiley, New York, 1979.

.

[24] N.G. Hoogeveen, M.A. Cohen Stuart, G.J. Fleer, M.R. Bo¨hmer, Langmuir 12 (15) (1996) 3675. [25] G. Decher, Y. Lvov, J. Schmitt, Thin Solid Films 244 (1994) 772. [26] R. Aksberg, L. O8 dberg, Nord. Pulp Pap. Res. J. 5 (4) (1990) 168.