Adsorption of Poly (ethyleneimine) on Silica Surfaces: Effect of pH on

Unilever Research Port Sunlight, Quarry Road East, Bebington CH63 3JW, United Kingdom, and Department of ... Budapest 112, Hungary H-1518. Received ...
2 downloads 0 Views 70KB Size
5026

Langmuir 2004, 20, 5026-5029

Adsorption of Poly(ethyleneimine) on Silica Surfaces: Effect of pH on the Reversibility of Adsorption Ro´bert Me´sza´ros,*,†,‡ Imre Varga,‡ and Tibor Gila´nyi‡ Unilever Research Port Sunlight, Quarry Road East, Bebington CH63 3JW, United Kingdom, and Department of Colloid Chemistry, Lora´ nd Eo¨ tvo¨ s University, P.O. Box 32, Budapest 112, Hungary H-1518 Received February 13, 2004. In Final Form: April 2, 2004 The role of polymer charge density in the kinetics of the adsorption and desorption, on silica, of the polyelectrolyte poly(ethyleneimine) (PEI) was investigated by stagnation-point flow reflectometry. In the first series of experiments, PEI solutions were introduced at the same ionic strength and pH as the background solvent. It was found that the adsorbed amount of PEI increased by increasing pH. In the second series of investigations, several PEI solutions with ascending pH were introduced consecutively into the cell. In these cases, a stepwise buildup of the adsorbed amount was observed and the “final” adsorbed amounts were observed to be roughly equal with the adsorbed amounts of the first series of measurements at the same pH. Finally, adsorption/desorption experiments were performed where the preadsorption of PEI was followed by the introduction of PEI solutions of descending pH. No desorption was detected when the pH changed from pH ) 9.7 to pH ) 5.8. However, when there was a 9.7 f 3.3 or 5.8 f 3.3 decrease in the pH, the kinetic barriers of desorption seemed to completely disappear and roughly the same adsorbed amount as in the first series of experiments at pH ) 3.3 was quickly attained by desorption of the PEI. This study reveals the high impact of pH, affecting parameters such as charge density of the surface and polyelectrolyte as well as the structure of the adsorbed macromolecules, on the desorption properties of weak polyelectrolytes. The observed interfacial behavior of PEI may have some important consequences for the stability of alternating polyelectrolyte multilayers containing weak polyelectrolytes.

Introduction Poly(ethyleneimine)s (PEIs) are branched weak polyelectrolytes which are widely used as thickeners, flocculating agents, adhesives, and so forth, especially in the paper industry as well as in the home and personal care industries.1-2 In these applications, the dynamic and static interfacial properties of PEI are of paramount importance. The adsorption kinetics of PEI on silica was recently investigated by a reflectometric study which revealed the importance of an electrostatic barrier in the adsorption mechanism especially at low ionic strength and at low and moderate pH.3 In the same work, the dependence of the adsorbed amount on the pH and ionic strength suggested significant non-Coulombic affinity of the adsorbed PEI segments toward the silica surface. The adsorbed layer of PEI on silica was found to be very compact at low and moderate pH and at low ionic strength.4 At higher pH, the adsorbed layer is more extended but can be considered still quite compact. These observations also correlate with the adsorbed layer structure of PEI on mica reported by Claesson et al.5 These results indicate that the adsorbed branched PEI molecules significantly contract, perpendicularly to the surface, to increase the number of surface/segment contacts. * To whom correspondence should be addressed. E-mail: [email protected]. † Unilever Research Port Sunlight. ‡ Lora ´ nd Eo¨tvo¨s University. (1) Lindquist, G. M.; Stratton, R. A. J. Colloid Interface Sci. 1976, 55, 45. (2) Alince, B.; Vanerek, A.; van de Ven, T. G. M. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 954. (3) Me´sza´ros, R.; Thompson, L.; Bos, M.; de Groot, P. Langmuir 2002, 18, 6164. (4) Me´sza´ros, R.; Thompson, L.; Bos, M.; Varga, I.; Gila´nyi, T. Langmuir 2003, 19, 9977. (5) Claesson, P. M.; Paulson, O. E. H.; Blomberg, E.; Burns, N. L. Colloids Surf., A 1997, 123-124, 341.

If a polyelectrolyte adsorbs onto an oppositely charged surface, the surface charge may not only be neutralized but also reversed, depending on the adsorbed amount and charge density of the adsorbed macromolecules. In the case of PEI adsorption on silica and glasslike surfaces, the extent of the surface charge reversal is very large compared to that of other cationic polyelectrolytes, as was proved by electrokinetic and surface force measurements.3,6 This observation was interpreted according to the hyperbranched structure and the nonelectrostatic surface affinity for the PEI.3,6 This finding is the main reason behind the intensive use of PEI type polyelectrolytes in the recent studies of alternate polyelectrolyte multilayers.7-10 In this technology, the reversibility of polyelectrolyte adsorption is also a crucial issue, since during the removal of the supernatant polymer solution, the (multi)layer of polyelectrolytes must be attached irreversibly to the surface in order to stick to the next layer of a polyelectrolyte with opposite charges. Since electrostatic interactions act to keep the alternate layers of polyelectrolytes together, the charge density of these macromolecules and that of the surface as well as the ionic strength are of paramount importance in the stability of these multilayers.11-12 In the present paper, the adsorption/desorption kinetics of PEI on silica will be investigated at low ionic strength (6) Poptoshev, E.; Rutland, M. W.; Claesson, P. M. Langmuir 2002, 18, 2590. (7) Decher, G.; Hong, J.-D. Ber. Bunsen-Ges. Phys. Chem. 1991, 95, 1430. (8) Decher, G. Science 1997, 277, 1232. (9) Schwarz, S.; Eichhorn, K.-J.; Wischerhoff, E.; Laschewsky, A. Colloids Surf., A 1999, 159, 491. (10) Baba, A.; Kaneko, F.; Advincula, R. C. Colloids Surf., A 2000, 173, 39. (11) Hoogeveen, N. G.; Cohen Stuart, M. A.; Fleer, G. J. Langmuir 1996, 12, 3675. (12) Kovacevic, D.; van der Burgh, S.; de Keizer, A.; Cohen Stuart, M. A. J. Phys. Chem. B 2003, 107, 7998.

10.1021/la049611l CCC: $27.50 © 2004 American Chemical Society Published on Web 05/11/2004

Adsorption of Poly(ethyleneimine) on Silica

Langmuir, Vol. 20, No. 12, 2004 5027 When the medium (background solvent) flows toward the surface, there is an initial signal S0. When the source is switched to a solution of adsorbing polymer, the signal changes by ∆S. Under appropriate conditions,14-15 that is, when the adsorbed amount is not too large, the relative change of the signal is proportional to the adsorbed amount:

Γ)

Figure 1. Schematic representation of the chemical structure of PEI.

by reflectometry. It will be shown that the pH has a dramatic effect on the reversibility of PEI adsorption, and we discuss important consequences that this may have for the stability of alternating polyelectrolyte layers containing weak polybases or polyacids. Experimental Section Materials. The PEI with a mean molecular weight of 750 000 g/mol was purchased from BASF in the form of a 33 wt % aqueous solution. As shown schematically in Figure 1, this PEI sample is a hyperbranched polymer containing the primary, secondary, and tertiary amine groups in a 1:2:1 ratio.13 The pH of the solutions was adjusted with ACS reagents of HCl and (carbonate free) NaOH. The supporting electrolyte was NaCl. All of these chemicals were purchased from Sigma-Aldrich. During the experiments, ultraclean Millipore water was used for making solutions. Throughout the measurements, a constant ionic strength of 0.01 M NaCl was applied. As a surface substrate, strips of silicon wafers were used bearing a 100 nm layer of SiO2 on the top. The silica wafers were cleaned by putting them into concentrated persulfuric acid for 60 min. The purified samples were rinsed and then stored under Millipore water prior to the measurements. Reflectometry. Polyelectrolyte adsorption was measured by reflectometry using the standard method of Dijt et al.14-15 All the measurements were carried out at 295 K. In this method, the flux of the solution toward the solid substrate is controlled by a stagnation-point flow cell. In a steady-state situation, the flux can be described by the following equation:

J0 ) 0.776ν1/3R-1D2/3(R j Re)2/3c

(1)

where ν is the kinematic viscosity, R is the radius of the tube, D is the diffusion coefficient, R j is the dimensionless flow intensity parameter, and Re is the Reynolds number. During the experiments, a laser source emits a polarized beam which is reflected off the wafer in the reflectometric cell. The reflected light is split into its parallel and perpendicular components, which are detected separately with photodiodes. The signal S is defined as

S)f

Rp Rs

(2)

where f is an equipment constant, and Rp and Rs are the reflectivities of the parallel and perpendicular components. (13) BASF literature on LUPASOL. (14) Dijt, J. C.; Cohen Stuart, M. A.; Fleer, G. J. Adv. Colloid Interface Sci. 1994, 50, 79. (15) Dijt, J. C.; Cohen Stuart, M. A.; Hofman, J. E.; Fleer, G. J. Colloids Surf. 1990, 51, 141.

1 ∆S As S0

(3)

where As is the sensitivity factor. This factor depends on the angle of incidence θi, the wavelength of the laser beam λ, the refractive indices of the silicon, silica, adsorbed layer, and solution phase, the refractive index increment of the polymer solution (dn/dc), and the thickness of the silica layer and the adsorbed layer. The values used here were nSi ) 3.8, nSiO2 ) 1.46, dSiO2 ) 100 nm, nads ) 1.36, nsol ) 1.334, dn/dc ) 0.21 cm3/g,17 θi ) 70°, and λ ) 632.8 nm. The refractive index of the silicon and silica and that of the thickness of the silica layers (100 nm) were determined by independent ellipsometry measurements. The refractive index of the adsorbed layer was arbitrarily fixed at nads ) 1.36. (In our recent ellipsometric investigations, nads was varied between 1.345 and 1.37 in the investigated pH and ionic strength range.) By means of these sets of parameters and leaving the adsorbed layer thickness (dads) as a variable parameter, the proportionality constant of Γ/∆S ) 1/(AsS0) in eq 3 was calculated using the method of Hansen.16 This also means that although the adsorbed amount was determined, the actual values of the refractive index and the thickness of the adsorbed layer are not known but correlated according to the expression of Γ ) (nads -nsol)dads/(dn/dc).14,15 In the first series of experiments, the polymer solution was directly introduced after the medium at the same ionic strength and pH. During the next series of measurements, this procedure was followed by the incremental introduction of further PEI solutions with higher or lower pH. The effect of the pH of the background medium (0.01 M NaCl) on the reflectometric baseline was investigated but found to be negligible.

Results and Discussion Adsorption Kinetics. One of the most important factors governing the interfacial behavior of PEI on silica is the electrostatic interaction. The charge densities of PEI and of the silica surface vary conversely with the pH.3 The point of zero charge (pzc) of silica is around pH = 23,18 so that at high pH the silica surface is negatively charged. In contrast, the PEI segments are practically neutral at pH > 10.5 but roughly 70% of the amine groups are charged at pH = 3.3 Therefore, the variation of pH is expected to have pronounced consequences for the mechanism of adsorption/desorption. In Figure 2, the variation of PEI adsorption with time is shown at three different pHs, in 0.01 M NaCl. The adsorption rate seems to be constant at the beginning of the adsorption process, when the surface is barely occupied and the rate-determining step is the transport of the polymers from the bulk to the surface. At this stage, the rate of adsorption can be estimated according to the flux of the solutions in the stagnation-point flow cell from eq 1.15 As the surface coverage increases, the adsorption rate decreases and in a relatively short time scale the adsorbed amount seems to be independent of the time (“final” or “saturation” adsorption). The larger the charge density of PEI, the steeper the decrease of dΓ/dt compared to the initial adsorption rate. This behavior was interpreted by the buildup of an electrostatic barrier due to surface charge (16) Hansen, W. N. J. Opt. Soc. Am. 1968, 58, 380. (17) Park, I. H.; Choi, E. J. Polym. 1996, 37, 313. (18) Behrens, S. H.; Grier, D. G. J. Chem. Phys. 2001, 115, 6716.

5028

Langmuir, Vol. 20, No. 12, 2004

Figure 2. The adsorbed amount of PEI as a function of time in 0.01 M NaCl and at three pHs. In each case, the baseline was created with a 0.01 M NaCl solution at the same pH as that of the subsequently introduced PEI solution. The solid squares, open circles, and solid triangles refer to pH ) 9.7, 5.8, and 3.3, respectively. Having attained the “saturation” adsorbed amount, the medium was introduced again in order to test the potential desorption of PEI. cPEI ) 50 ppm in 0.01 M NaCl.

Figure 3. Stepwise addition of PEI solutions with ascending pH. The baseline was created with a 0.01 M NaCl solution at the same pH as that of the subsequently introduced PEI solution. Curve 2 with solid squares refers to the experiment where first a PEI solution at pH ) 3.3 was introduced (after the medium), which was followed by the addition of a PEI solution at pH ) 9.7. Curves 1 (open squares) and 3 (open triangles) are repeated from Figure 2 at pH ) 9.7 and pH ) 3.3, where no pH change occurs during the adsorption experiments. cPEI ) 50 ppm in 0.01 M NaCl.

overcompensation via the adsorption of the oppositely charged polyelectrolyte.3,19 This also means that despite the apparent time independence of the adsorbed amount at longer times, this “final” adsorbed amount might slightly underestimate the equilibrium adsorbed amount,19 as further, very slow adsorption may occur. Nevertheless, these adsorbed amounts can be roughly considered as the saturation or maximum adsorption at the given pHs.3 Thus, Figure 2 also shows that the “final” adsorbed amount increases with increasing pH. This pH dependence of PEI adsorption is in accordance with the predictions of the self-consistence field theories of equilibrium polyelectrolyte adsorption.20 In Figure 3 and Figure 4, the adsorption kinetics is shown for the subsequent introduction of two PEI solutions with ascending pH at the same ionic strength. In Figure 3, after the medium a PEI solution at pH ) 3.3 was (19) Cohen Stuart, M. A.; Hoogendam, C. W.; de Keizer, A. J. Phys.: Condens. Matter 1997, 9, 7767. (20) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces, 1st ed.; Chapman and Hall: London, 1993; Chapter 7.

Me´ sza´ ros et al.

Figure 4. Stepwise addition of PEI solutions with ascending and descending pH. The baseline was created with a 0.01 M NaCl solution at pH ) 3.3. At first, a PEI solution at pH ) 3.3 was introduced, which was followed by the addition of a PEI solution at pH ) 5.8. The “final” adsorbed amount of PEI (ΓPEI) in 0.01 M NaCl and at pH ) 5.8, taken from Figure 2, is indicated by a dotted line. Having attained the “final” adsorbed amount at pH ) 5.8, the PEI solution at pH ) 3.3 was introduced again into the cell. cPEI ) 50 ppm in 0.01 M NaCl.

introduced, which was followed by the addition of a PEI solution at pH ) 9.7. In Figure 4, after the PEI solution of pH ) 3.3, a PEI solution at pH ) 5.8 was introduced into the cell. (This was subsequently followed by the introduction of a pH ) 3.3 PEI solution. This process will be analyzed later.) In both cases, the ascending pH of the subsequent PEI solutions results in a stepwise buildup of the adsorbed amount and approximately the same “final” adsorbed amounts as in Figure 2 at pH ) 9.7 and pH ) 5.8, respectively, were attained. However, after the pH change the adsorption kinetics seems to be slightly different in that the rate of attainment of the “final” adsorbed amount slows down compared to those experiments with no pH changes. (The data of Figure 2 are also plotted in Figure 3 for comparison.). We believe this slowdown to be attributable to the time required for the relaxation of the more compact PEI layer adsorbed at the initial, lower pH, which must occur before further adsorption can take place. However, the differences in the charge generation dynamics of the silica surface and that of the adsorbed PEI molecules during the experiments of Figure 2 and Figure 3 may also be in connection with this observation. Desorption Kinetics. The desorption of polymers is known to be a very slow process due to the huge number of surface/segment contacts which can result in a significant kinetic barrier of desorption.21 This is particularly true for polyelectrolyte adsorption onto oppositely charged surfaces. In the adsorption experiments of Figure 2 after the attainment of the “final” adsorbed amount, the background solvent (with the same ionic strength and pH as that of the PEI solutions) was introduced again into the reflectometric cell. No desorption was observed (practically it was checked for 30 min but only the initial time range is shown in Figure 2). Similar results were observed in our recent ellipsometric study where the adsorbed amount of PEI on silica was found to be the same after 2 h of rinsing with the medium.4 In the case of PEI adsorption on silica, there is a significant nonelectrostatic attraction between the PEI segments and the silica surface3 which can further increase the kinetic barrier of desorption. (21) Hoogeveen, N. G.; Cohen Stuart, M. A.; Fleer, G. J. J. Colloid Interface Sci. 1996, 182, 146.

Adsorption of Poly(ethyleneimine) on Silica

Langmuir, Vol. 20, No. 12, 2004 5029

al. investigated the adsorption of different strong and weak polyelectrolytes on oxide surfaces21 as well as hydrophobically modified poly(acrylic acid) adsorption on a hydrophobic, polystyrene-coated silica.22 At small ionic strength, even a small increase in the charge density of the polymer resulted in a significant desorption and a pronounced adsorption/desorption hysteresis was found in a very large pH range including the highly charged states of the polymers as well. The differences between our work and the mentioned studies can be mainly attributed to the significant nonelectrostatic affinity of PEI to the silica as well as the opposite effect of the pH on the charge density of the silica surface and that of the adsorbed PEI molecules. Figure 5. Stepwise addition of PEI solutions with descending pH. The baseline was created with a 0.01 M NaCl solution at pH ) 9.7. The open squares refer to the experiment where at first a PEI solution at pH ) 9.7 was introduced into the cell after the medium, which was followed by the addition of a PEI solution at pH ) 5.8. The solid squares refer to the experiment where after the PEI solution at pH ) 9.7, a PEI solution at pH ) 3.3 was introduced into the reflectometric cell. For the sake of comparison, the “saturation” adsorbed amounts of PEI (ΓPEI) at pH ) 5.8 and at pH ) 3.3 (in 0.01 M NaCl), which were estimated according to Figure 2, were also indicated by dotted lines. cPEI ) 50 ppm in 0.01 M NaCl.

In the following, we will investigate the effect of pH (polymer and surface charge density variation) on the desorption process. In Figure 4, the stepwise buildup of the adsorbed layer and the attainment of the “final” adsorbed amount at pH ) 5.8 were followed by the incremental introduction of the PEI solution at pH ) 3.3. As a consequence, desorption of PEI took place and in a short time scale the “saturation” adsorbed amount, characteristic at pH ) 3.3, was nearly attained. In Figure 5, after the medium, a PEI solution at pH ) 9.7 was introduced into the reflectometric cell. Having attained the “final” adsorbed amount, PEI solutions at pH ) 5.8 and pH ) 3.3 were introduced into the cell in two separate sets of experiments. During the 9.7 f 5.8 pH change, no desorption was detected in the time scale of the experiments (practically it was checked for 30 min but only the initial time range is shown in Figure 5). Despite the fact that this pH change means an approximate 30% increase in the bulk protonation degree of PEI,3 which results in an increased repulsion between the adsorbed PEI segments, the desorption of PEI is still hindered. This indicates surprisingly high activation energy of desorption which might be in connection with the nonelectrostatic affinity of the PEI segments to the silica surface. On the other hand, when the final pH was 3.3 (5.8 f 3.3 pH change in Figure 4 and 9.7 f 3.3 pH change in Figure 5), a quick desorption takes place and the “final” adsorbed amount is in good agreement with that determined at the same pH in the adsorption experiments of Figure 2. This means that in these cases the activation energy of desorption is significantly reduced. Since pH ) 3.3 is close to the pzc of the silica, the surface is only slightly charged; therefore the number of oppositely charged PEI/silica surface contacts is significantly reduced. On the other hand, the PEI segments become highly charged at pH ) 3.3; therefore the repulsion between the adsorbed PEI segments can overcome the activation energy of desorption. These observations are different from other studies on polyelectrolyte adsorption reversibility. Cohen Stuart et

Summary The adsorption kinetics of PEI is governed by the transport from the bulk at the beginning of the adsorption process, whereas at longer times an electrostatic barrier to adsorption builds up. If the final state of the system is attained via ascending pH of the PEI solution, a stepwise buildup of the adsorbed layer takes place and the “final” adsorbed amount is found to be approximately the same as that found without pH change. On the other hand, during a descending pH change no desorption was observed between pH ) 9.7 and pH ) 5.8. Although this pH change means a significant increase in the bulk charge density of PEI (30%), the desorption of the PEI is still significantly hindered. This high activation energy for PEI desorption from the silica surface is thought to be due to the non-Coulomb affinity of PEI to the surface. However, in the pH cycles of 3.3 f 9.7 f 3.3 and 3.3 f 5.8 f 3.3 the adsorption of PEI was proved to be reversible. At pH ) 3.3, the silica surface is only slightly charged; therefore when the pH is lowered to this value, the barrier of desorption is overcome due to the decreased number of surface/PEI segment contacts and the increased repulsion between the recharged PEI segments in the adsorbed layer, which results in a quick and complete desorption of the polymer. In the fabrication of alternate polyelectrolyte multilayers containing weak polybases or acids, care must be taken during the adsorption/rinsing cycles. The potential pH difference between the rinsing medium (water) and the supernatant weak polyelectrolyte solutions might result in the disruption of the multilayer because of the desorption of the bound weak polyelectrolyte segments in the first layer. To get stable polyelectrolyte multilayers, the use of strong polyelectrolytes is recommended. Alternatively, a thorough control of the pH of the polymer solutions and that of the rinsing medium is necessary to prevent the potential desorption of the adsorbed weak polyelectrolytes. Acknowledgment. R. Me´sza´ros and Imre Varga are Be´ke´sy Gyo¨rgy Fellows of the Hungarian Ministry of Education which is gratefully acknowledged. The work is supported by the Hungarian OTKA Foundation under Project No. F034838 and by the Ministry of Education, Hungary, under Project FKFP 0051/2001. The authors are thankful to Laurie Thompson for his expert help and comments. LA049611L (22) Go¨bel, J. G.; Besseling, N. A. M.; Cohen Stuart, M. A. J. Colloid Interface Sci. 1999, 209, 129.