Added Surfactant Can Change the Phase Behavior of Aqueous

The phase behavior of aqueous mixtures of the “clouding” polymer ethyl(hydroxyethyl)cellulose (EHEC) mixed with colloidal particles and surfactant...
0 downloads 0 Views 208KB Size
Download PDF
Langmuir 2005, 21, 2743-2749

2743

Added Surfactant Can Change the Phase Behavior of Aqueous Polymer-Particle Mixtures Martin Olsson,† Go¨ran Bostro¨m,† Leif Karlson,‡ and Lennart Piculell*,† Division of Physical Chemistry 1, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden, and Akzo Nobel Surface Chemistry AB, SE-444 85 Stenungsund, Sweden Received November 1, 2004. In Final Form: January 10, 2005 The phase behavior of aqueous mixtures of the “clouding” polymer ethyl(hydroxyethyl)cellulose (EHEC) mixed with colloidal particles and surfactants has been studied. These types of mixtures are important in many technical formulations. Two types of particles, polystyrene latex and silica, and two types of EHEC, nonmodified EHEC (N-EHEC) and hydrophobically modified EHEC (HM-EHEC), were studied. The EHECs adsorb to both kinds of particles. Both the amount and the type of added surfactant were seen to dramatically influence the partitioning of the particles between the EHEC-rich and EHEC-poor phases of phase-separated mixtures (above the cloud point temperature). Surfactants that are known not to associate with the EHEC backbone, that is, nonionic surfactants and short-chain cationic surfactants, changed the interaction between EHEC and the colloidal particles from attraction to repulsion above a specific surfactant concentration, resulting in a change in the partitioning of the particles from the EHECrich to the EHEC-poor phase. No such particle inversion was observed for ionic surfactants that bind to the EHEC backbone. An analysis considering both the binding of surfactant to EHEC and the competitive adsorption of surfactant to the particle surfaces could rationalize all observations, including the large variations observed, among the studied mixtures, in the surfactant concentration required for particle inversion.

Introduction Aqueous mixtures of polymers, surfactants, and colloidal particles are widely used in applications such as paints, pharmaceutical products, and papermaking. Of central importance for these applications are the interactions of the polymer and surfactant molecules with the particle surface. Many previous studies have addressed the adsorption from polymer/surfactant mixtures onto solid surfaces.1-11 Depending on the interaction between the polymer and the surfactant molecules, two different situations can arise. A competition between the two components takes place if the polymer and the surfactant molecules do not associate. The substance with the largest affinity to the surface will be the one that is preferentially adsorbed. Alternatively, if the polymer and the surfactant molecules associate to each other, polymer-surfactant complexes are formed. At high surfactant concentrations, * Corresponding author. E-mail: [email protected]. Fax: +46 46 222 44 13. † Lund University. ‡ Akzo Nobel Surface Chemistry AB. (1) Shubin, V. Langmuir 1994, 10, 1093-1100. (2) Ma, Z.; Chen, M.; Glass, J. E. Colloids Surf., A 1996, 112, 163184. (3) Ghodbane, J.; Denoyel, R. Colloids Surf., A 1997, 127, 97-104. (4) Mears, S. J.; Cosgrove, T.; Obey, T.; Thompson, L.; Howell, I. Langmuir 1998, 14, 4997-5003. (5) Glass, J. E. Adv. Colloid Interface Sci. 1999, 79, 123-148. (6) Cosgrove, T.; Mears, S. J.; Obey, T.; Thompson, L.; Wesley, R. D. Colloids Surf., A 1999, 149, 329-338. (7) Lauten, R. A.; Kjøniksen, A.-L.; Nystro¨m, B. Langmuir 2000, 16, 4478-4484. (8) Rachas, I.; Tadros, T. F.; Taylor, P. Colloids Surf., A 2000, 161, 307-319. (9) Joabsson, F.; Thuresson, K.; Lindman, B. Langmuir 2001, 17, 1499-1505. (10) Rojas, O. J.; Ernstsson, M.; Neuman, R. D.; Claesson, P. M. Langmuir 2002, 18, 1604-1612. (11) Casford, M. T. L.; Davies, P. B.; Neivandt, D. J. Langmuir 2003, 19, 7386-7391.

that is, high degrees of surfactant binding, these complexes typically desorb from the surface. In this study, we have investigated systems containing ethyl(hydroxyethyl)cellulose (EHEC), surfactant, and colloidal particles. Two types of colloidal particles were used, that is, polystyrene latex and silica, as well as two types of EHEC, nonmodified EHEC (N-EHEC) and hydrophobically modified EHEC (HM-EHEC). The purpose of the study has been to examine how the interactions between EHEC and colloidal particles are influenced by the presence of surfactant. Industrially, EHEC is used, for example, as a thickener in water-borne paints, where it helps to control the rheology of the formulation through polymer-polymer interactions. Besides the thickening effect, the polymer is also expected to contribute to stabilizing the particles in the system. However, in some cases the opposite effect is seen. In the latter cases, flocculation of particles leads to unwanted effects such as changes in the rheology of the paint and poor color acceptance. A poor color acceptance means that a tinted paint shows variations in shade depending on the magnitude of the shear used during the application of the paint. Thus, to be able to control the interactions between the components in dispersions is of great importance for the preparation of formulations with desired properties. Owing to the importance of EHEC in industrial products, a large number of studies have been performed to investigate the interactions between EHEC and various surfactants.12-38 All common surfactants form mixed (12) Carlsson, A.; Karlstro¨m, G.; Lindman, B. Langmuir 1986, 2, 536-537. (13) Carlsson, A.; Karlstro¨m, G.; Lindman, B.; Stenberg, O. Colloid Polym. Sci. 1988, 266, 1031-1036. (14) Carlsson, A.; Karlstro¨m, G.; Lindman, B. J. Phys. Chem. 1989, 93, 3673-3677. (15) Karlstro¨m, G.; Carlsson, A.; Lindman, B. J. Phys. Chem. 1990, 94, 5005-5015. (16) Zana, R.; Binana-Limbele, W.; Kamenka, N.; Lindman, B. J. Phys. Chem. 1992, 96, 5461-5465.

10.1021/la0473254 CCC: $30.25 © 2005 American Chemical Society Published on Web 02/17/2005

2744

Langmuir, Vol. 21, No. 7, 2005

micelles with the hydrophobic groups on HM-EHEC. In addition, both anionic and certain cationic surfactants bind to the main chain of EHEC in N-EHEC and HMEHEC, whereas no such binding occurs for nonionic surfactants carrying oligo(ethylene oxide) headgroups. The interactions between EHEC and colloidal particles are far less studied. However, insights into these interactions can be obtained from studies on adsorption behavior of EHEC to macroscopic surfaces.39-43 Further, the adsorption behavior of EHEC onto colloidal particles in the presence of surfactant was investigated by light scattering by Lauten and co-workers.7 For the colloidal particles studied in the present paper, Olsson et al. showed that N-EHEC and HM-EHEC adsorb onto silica and polystyrene latex particles.44 Note that both EHEC and the surfactants show an affinity to the surfaces of the two particle types studied here, except that anionic surfactants do not adsorb to the negatively charged silica surface. An important property of EHEC is that it is a “clouding polymer”; that is, it phase separates upon heating. The (concentration-dependent) temperature at which an EHEC solution phase separates is referred to as the cloud point temperature, Tcp. Adding a third component to the solution may influence Tcp. A surfactant showing association with EHEC gives an initial decrease followed by an increase in Tcp at increasing surfactant concentrations.12,15,22,27 The effect on Tcp of a surfactant that does not associate is considerably smaller than for an associating one.27 Recently, we have shown that also addition of colloidal (17) Holmberg, C.; Nilsson, S.; Singh, S. K.; Sundelo¨f, L. O. J. Phys. Chem. 1992, 96, 871-876. (18) Zhang, K.; Karlstro¨m, G.; Lindman, B. Progress Colloid Polym. Sci. 1992, 88, 1-7. (19) Zhang, K.; Jonstro¨mer, M.; Lindman, B. J. Phys. Chem. 1994, 98, 2459-2463. (20) Olofsson, G.; Wang, G. Pure Appl. Chem. 1994, 66, 527-532. (21) Walderhaug, H.; Nystro¨m, B.; Hansen, F. K.; Lindman, B. Prog. Colloid Polym. Sci. 1995, 98, 51-56. (22) Thuresson, K.; Nystro¨m, B.; Wang, G.; Lindman, B. Langmuir 1995, 11, 3730-3736. (23) Holmberg, C. Colloid Polym. Sci. 1996, 274, 836-847. (24) Goldszal, A.; Costeux, S.; Djabourov, M. Colloids Surf., A 1996, 112, 141-154. (25) Thuresson, K.; So¨derman, O.; Hansson, P.; Wang, G. J. Phys. Chem. 1996, 100, 4909-4918. (26) Thuresson, K.; Lindman, B.; Nystro¨m, B. J. Phys. Chem. B 1997, 101, 6450-6459. (27) Thuresson, K.; Lindman, B. J. Phys. Chem. B 1997, 101, 64606468. (28) Evertsson, H.; Holmberg, C. Colloid Polym. Sci. 1997, 275, 830840. (29) Nahringbauer, I. Langmuir 1997, 13, 2242-2249. (30) Singh, S. K.; Nilsson, S. J. Colloid Interface Sci. 1999, 213, 133151. (31) Singh, S. K.; Nilsson, S. J. Colloid Interface Sci. 1999, 213, 152159. (32) Evertsson, H.; Nilsson, S. Carbohydr. Polym. 1999, 40, 293298. (33) Kjøniksen, A.-L.; Nystro¨m, B.; Lindman, B. Colloids Surf., A 1999, 149, 347-354. (34) Kjøniksen, A.-L.; Nilsson, S.; Thuresson, K.; Lindman, B.; Nystro¨m, B. Macromolecules 2000, 33, 877-886. (35) Nilsson, S.; Thuresson, K.; Lindman, B.; Nystro¨m, B. Macromolecules 2000, 33, 9641-9649. (36) Hoff, E.; Nystro¨m, B.; Lindman, B. Langmuir 2001, 17, 28-34. (37) Lund, R.; Lauten, R. A.; Nystro¨m, B.; Lindman, B. Langmuir 2001, 17, 8001-8009. (38) Ridell, A.; Evertsson, H.; Nilsson, S. J. Colloid Interface Sci. 2002, 247, 381-388. (39) Malmsten, M.; Lindman, B. Langmuir 1990, 6, 357-364. (40) Malmsten, M.; Claesson, P. M.; Pezron, E.; Pezron, I. Langmuir 1990, 6, 1572-1578. (41) Malmsten, M.; Claesson, P. M. Langmuir 1991, 7, 988-994. (42) Pezron, I.; Pezron, E.; Claesson, P. M.; Malmsten, M. Langmuir 1991, 7, 2248-2252. (43) Malmsten, M.; Tiberg, F. Langmuir 1993, 9, 1098-1103. (44) Olsson, M.; Joabsson, F.; Piculell, L. Langmuir 2004, 20, 16051610.

Olsson et al. Table 1. Polymer Dataa polymer N-EHEC HM-EHEC a

DSethyl MSEO MShydrophob 0.8 0.8

2.1 2.1

0.01

Mw (Da) 5.4 × 105 5.1 × 105

Tcp Mw/Mn (°C) 6 7

68 47

Tcp refers to an aqueous 1 wt % solution.

particles to which EHEC adsorbs, such as polystyrene latex or silica, decreases the Tcp.44 In this work, to monitor the interactions between EHEC and the colloidal particles in dispersions with added surfactant, the mixtures were heated above Tcp. Then the dispersions phase separated into two phases, one rich and one poor in EHEC, and the partitioning of the particles between the two phases was studied. As both the polystyrene latex and silica particles scatter light, their presence in a phase can easily be detected visually. This experimental procedure, thus, gives a simple and direct way to determine the net interaction between EHEC and the colloidal particles in the presence of surfactant. To obtain further insights into the interactions between EHEC and the colloidal particles, we investigated how the colloidal particles influenced Tcp of the EHEC solution in the presence of anionic and nonionic surfactants at various concentrations. Experimental Section Materials. The N-EHEC and HM-EHEC samples were produced by Akzo Nobel Surface Chemistry AB, Stenungsund, Sweden. The degrees of substitution of ethyl and hydroxyethyl groups onto the cellulose backbone have previously been determined by gas chromatography45,46 and are given as DSethyl and MSEO, respectively, referring to the average number of substituents per glucose unit. In addition, HM-EHEC contains hydrophobic nonylphenol groups with a substitution degree specified as MShydrophob. The two EHECs were purified before use as described previously.47 Their weight- and number-average molecular weights (Mw and Mn, respectively) were determined by size exclusion chromatography.44 The data on N-EHEC and HMEHEC are summarized in Table 1. Data on the various surfactants, including abbreviations used in this paper, are summarized in Table 2. Polystyrene latex particles with a mean diameter of 350 nm were purchased from Polysciences, Inc., Warrington, U.S.A., and obtained as a stock dispersion of 2.6 wt %. According to the manufacturer, the particle dispersion was stabilized by a minimum of surfactant. Silica particles with a mean diameter of 100 nm were obtained from Nissan Chemicals, Japan, as a stock dispersion of 40.5 wt %. No surfactant was included in the silica dispersion. The water in the experiments was of MilliPore quality. Methods. The studied dispersions were prepared by adding aqueous stock solutions of EHEC, surfactant, and colloidal particles into test tubes with Teflon screw caps. The EHEC concentration was 1 wt % unless otherwise indicated, while the surfactant concentration was varied over a wide range. The colloidal particles were polystyrene latex at a final concentration of 0.01-1 wt % or silica at a final concentration of 1 wt %. All dispersions were equilibrated on a tilting board at room temperature for at least 12 h. After the mixing procedure, the dispersions were heated above Tcp until a macroscopic phase separation occurred. The temperature for macroscopic phase separation was chosen individually for each system to be more than 10 °C above Tcp (except when Tcp > 85 °C) to obtain a volume of the EHEC-rich phase amounting (45) Stead, J. B.; Hindley, A. H. J. Chromatogr. 1969, 42, 470-475. (46) Karlson, L.; Joabsson, F.; Thuresson, K. Carbohydr. Polym. 2000, 41, 25-35. (47) Thuresson, K.; Karlstro¨m, G.; Lindman, B. J. Phys. Chem. 1995, 99, 3823-3831.

Polymer-Particle Mixtures Phase Behavior Changes

Langmuir, Vol. 21, No. 7, 2005 2745

Table 2. Surfactants Used systematic name sodium dodecyl sulfate sodium octyl sulfate sodium dodecyl-di(ethylene oxide)-sulfate hexadecyl trimethylammonium bromide dodecyl trimethylammonium bromide octyl trimethylammonium bromide hexyl trimethylammonium bromide octaethylene glycol mono n-dodecyl ether t-octylphenoxypolyethoxyethanol octyl-β-D-glucopyranoside a

abbreviation

cmca (mM) in water

manufacturer

NaC12S NaC8S NaC12E2S

8.1 133 3b

BDH Merck Kao Chemicals Gmbh

C16TABr C12TABr C8TABr C6TABr C12E8 Triton X-100 C8G1

0.9 16 220 580b 0.071 0.26 25c

Merck Tokyo Kasei Kogyo Tokyo Kasei Kogyo Tokyo Kasei Kogyo Nikkol Acros Chemicals Anatrace

Values from ref 48 except where otherwise specified. b Value from the supplier. c Value from ref 49.

to at least 10% of the total sample volume. The partitioning of the particles between the separated phases was determined visually. In addition, measurements of Tcp for 1 wt % EHEC solutions including different amounts of the various surfactants were performed. The solutions were prepared as described above except that no colloidal particles were added. The measurements of Tcp were done in a temperature-controlled water bath, in which the temperature was raised at a rate of approximately 0.5 °C/min. The Tcp was determined visually. The reproducibility of such determinations is within (0.5 °C.

Results EHEC-Surfactant Interactions. The starting point in these investigations was to study the interactions between the EHECs and surfactants without particles. A convenient way to do this is to study the variation in Tcp of EHEC solutions at increasing concentrations of added surfactant.12,15,22,27 Generally, a change in Tcp indicates that EHEC and the surfactant form mixed micelles. One of two effects may dominate, depending on the surfactant concentration, when mixed micelles are formed. At low degrees of surfactant binding, the polymer chains share bound micelles, and this surfactant-mediated attraction between the polymer chains leads to a drop in Tcp. At high degrees of surfactant binding, each micelle binds to only one chain, leading to an increase in water solubility and, hence, an increase in Tcp. The solubilizing power of ionic surfactants is especially strong, owing to the entropy of mixing of the surfactant counterions. Previous measurements of Tcp in mixtures of N-EHEC with NaC12S have shown that Tcp first decreased followed by an increase at increasing surfactant concentrations.22 In the present study, mixtures of N-EHEC with two additional anionic surfactants, NaC8S and NaC12E2S, are found to conform to the same behavior (see Figure 1). Also cationic surfactants such as CnTACl with different lengths of the hydrophobic chain have been shown to form mixed micelles with N-EHEC.27 Here, we have studied N-EHEC/CnTABr mixtures for different values of n, as illustrated in Figure 2. Only the two CnTABr’s with the longest alkyl chains show a distinct decrease in Tcp before the increase. This indicates that C8TABr and C6TABr do not form micelles at the N-EHEC molecules, and the increase in Tcp at quite high surfactant concentrations may possibly be viewed as a “salting-in” effect, similar to what has previously been found for salts such as NaI or NaSCN.15 The decreased tendency for association between EHEC and CnTABr’s with short hydrophobic chains is verified also in earlier work.15,50 (48) Mukerjee, P.; Mysels, K. J. Critical Micelle Concentrations of Aqeuous Surfactant Systems; National Bureau of Standards: Washington, DC, 1971. (49) Nilsson, F.; So¨derman, O.; Johansson, I. Langmuir 1996, 12, 902-908.

Figure 1. Cloud point curves for aqueous 1 wt % N-EHEC and increasing concentrations of NaC12S (filled circles; data from Thuresson et al.),22 NaC8S (open circles), or NaC12E2S (open squares). Full lines are drawn to guide to the eye. Vertical arrows indicate the cmc values.

Figure 2. Cloud point curves for aqueous 1 wt % N-EHEC and increasing concentrations of C16TABr (filled circles), C12TABr (open circles), C8TABr (open squares), or C6TABr (open diamonds). Full lines are drawn to guide to the eye. Vertical arrows indicate the cmc values.

The influences on Tcp by the different nonionic surfactants are presented in Figure 3. For the mixtures with C12E8, there is a large difference between N-EHEC and (50) Rose´n, O.; Sjo¨stro¨m, J.; Piculell, L. Langmuir 1998, 14, 57955801.

2746

Langmuir, Vol. 21, No. 7, 2005

Olsson et al.

Figure 4. Phase-separated dispersions of 0.1 wt % polystyrene latex particles in 1 wt % N-EHEC and 0.03 mM (left) or 0.3 mM (right) C12E8.

Figure 3. Cloud point curves for 1 wt % aqueous N-EHEC (a) or HM-EHEC (b) and increasing concentrations of C12E8 (filled circles; data from Thuresson and Lindman),27 C8G1 (open circles), or Triton X-100 (open squares). Full lines are drawn to guide to the eye. Vertical arrows indicate the cmc values.

HM-EHEC. An increase in Tcp at high surfactant concentrations occurs for HM-EHEC but not for N-EHEC.27 This indicates that C12E8 associates only to the grafted hydrophobic groups and not to the EHEC backbone. Mixtures of HM-EHEC with all studied nonionic surfactants showed a decrease in Tcp at low surfactant concentrations followed by a rather sharp increase at higher surfactant concentrations (see Figure 3b), an expected consequence of a mixed micellization between the polymer hydrophobes and the surfactant molecules. In mixtures with N-EHEC (Figure 3a), the result for C8G1 differed from that of the nonassociating surfactant C12E8. The cloud point of C8G1 decreased with increasing surfactant concentration roughly in the interval 10-100 mM surfactant, after which it increased again. This indicates that an association between N-EHEC and C8G1 occurred. This was confirmed by analysis of the surfactant concentration in phase-separated solutions of N-EHEC and C8G1, which showed that C8G1 was enriched in the viscous phase rich in N-EHEC. Surfactant Effects on EHEC-Particle Interactions. The next step in this study was to investigate the interactions between EHEC and colloidal particles in the presence of anionic, cationic, and nonionic surfactants, respectively. All studied dispersions were phase separated macroscopically. In the absence of added surfactant the

particles partitioned essentially quantitatively to the polymer-rich phase after macroscopic phase separation, because EHEC adsorbs to both types of particle. This behavior has been reported recently by two of the present authors.44 However, at increasing concentrations of added surfactants, an inversion of the partitioning of particles was seen in some of the systems so that the particles switched over to the polymer-poor phase. This particle inversion phenomenon was observed in the systems with nonionic or short-chained cationic surfactants but not in the other systems. When the inversion occurred, it did so sharply at a specific surfactant concentration, where virtually all particles changed the preferred phase. A further increase of the surfactant concentration had no effect on the partitioning of the particles; that is, the particles still preferred the EHEC-poor phase. Figure 4 shows representative photographs of samples from a system that showed a particle inversion. The surfactant concentration, cs,i, required for a particle inversion in the different systems was shown to depend on the types of EHEC, surfactant, and particle as well as on the concentrations of particles and EHEC. Figure 5a,b summarizes cs,i for all systems displaying a particle inversion. Each particle inversion concentration is given as a concentration interval in the figure, reflecting the difference in surfactant concentration between consecutive samples in the series. In the sample at the low concentration end of the interval, the particles still partitioned to the EHEC-rich phase. At the high concentration end, the particle inversion had occurred. For N-EHEC/nonionic surfactant/polystyrene latex mixtures, the critical micelle concentration (cmc) falls within the cs,i interval at a particle concentration of 0.1 wt %. A variation in the N-EHEC concentration between 0.2 and 2 wt % did not influence cs,i in the system N-EHEC/ C12E8/polystyrene latex at a particle concentration of 0.1 wt %. An increase in the polystyrene latex particle concentration in the system N-EHEC/C12E8/polystyrene latex to 1 wt % gave a larger cs,i, while a decrease to 0.01 wt % polystyrene latex particles gave the same cs,i as for a particle concentration of 0.1 wt %. In N-EHEC/C12E8/ silica mixtures, cs,i was found to be somewhat larger at a silica concentration of 1 wt % than in the corresponding system with polystyrene latex particles. A lower silica concentration than 1 wt % was not used owing to the low turbidity; it then became difficult to determine the partitioning of the particles visually. In HM-EHEC/nonionic surfactant/polystyrene latex mixtures, the cmc fell in the cs,i interval only for C8G1 at

Polymer-Particle Mixtures Phase Behavior Changes

Langmuir, Vol. 21, No. 7, 2005 2747

polystyrene latex, the cs,i interval was found to include the surfactant cmc. Effects of Added Particles on the Cloud Point Temperature. In a previous study, two of the present authors showed that adding polystyrene latex or silica particles to N-EHEC or HM-EHEC dispersions in the concentration interval of 0.1-1 wt % EHEC gave a decrease in Tcp of the EHEC solutions.44 This was a consequence of the interfacial free energy between the added particles and the surrounding solution, which favored phase separation. When repeating the same experiments here for the system N-EHEC/polystyrene latex particles with an addition of 1 mM C12E8, that is, above cs,i for the system, we found that Tcp of N-EHEC was no longer affected by the added particles. By contrast, the effect on Tcp by the polystyrene latex particles was still present in experiments performed in the presence of NaC12S at 2 or 7 mM. Discussion

Figure 5. Particle inversion concentrations (see text) for all systems showing particle inversion. (a) Mixtures with C12E8. (b) Mixtures with Triton X-100, C8G1, C8TABr, or C6TABr. PSL ) polystyrene latex. Bars indicate experimental cs,i intervals, and the open circles represent calculated values (see Discussion).

particle concentrations of 0.01 and 0.1 wt %. For C12E8 and for Triton X-100, the values of cs,i were considerably higher than the cmc. In all HM-EHEC/nonionic surfactant/ polystyrene latex mixtures, the particle inversion occurred close to the minima in the respective cloud point curves presented in Figure 3b. The particle inversion concentration depended on the HM-EHEC concentration, as shown in the system HM-EHEC/C12E8/polystyrene latex. This was different from what was found for the system N-EHEC/C12E8/polystyrene latex. For the system HM-EHEC/C12E8/polystyrene latex, dispersions with a polystyrene latex concentration of 1 wt % were also studied. Despite the increase in particle concentration, cs,i was identical to cs,i for 0.01 and 0.1 wt % particles. In the system HM-EHEC/Triton X-100/silica, no particle inversion was observed even at 15 mM Triton X-100. For the N-EHEC/charged surfactant/particle mixtures, the concentration of added surfactant was increased only up to a few times the cmc of the surfactant, because a further increase raised the cloud point temperature above 100 °C, making a macroscopic phase separation experiment unpractical. Most of these systems showed no particle inversion at a polystyrene latex concentration of 0.1 wt %, the exceptions being systems containing the cationic surfactants with the shortest hydrocarbon chain, C8TABr and C6TABr. In mixtures of the latter surfactants with

The results presented above demonstrate that added surfactants, in addition to influencing Tcp of an aqueous EHEC solution, may also influence the partitioning of added particles in the separating phases, giving rise to a sharp inversion in the partitioning at a certain surfactant concentration. This particle inversion concentration may vary with the concentrations of polymer and particles. Moreover, additive effects on Tcp can occur in the sense that added particles may additionally lower Tcp of a solution where Tcp is already lowered by added surfactant. However, the latter effect, at least at the low particle concentrations considered here, only appears when the polymer-particle interactions are attractive, that is, when the particles partition to the polymer-rich phase. This observation is in agreement with our previous observation and interpretation of a particle-induced phase separation in cases where adsorbing particles are added to polymer solutions near phase separation.44 We are convinced that the somewhat complex combined effects of particles and surfactants observed here may be quite important in applications where the stability of the dispersion is a major concern. Clearly, a predictive molecular interpretation should be valuable, and this will be our objective in the following discussion. Added surfactants are known to remove adsorbed polymers, such as EHEC, from surfaces. In terms of the polymer-particle interactions, the polymer removal implies a change from the situation of polymer adsorption to that of polymer depletion at the surface. It seems clear that this effect must be the origin of the inversion in the partitioning of particles from the polymer-rich to the polymer-poor phase. The change from attractive to repulsive polymer-particle interactions is also the reason added particles no longer give rise to a decrease in Tcp. What remains to be understood are the (quantitative) differences observed for the different surfactants and the correlations (sometimes) observed between the particle inversion and the cmc and/or the cloud point minimum for the various surfactants. To rationalize these trends, we must consider the (competitive) adsorption of the surfactant to the particles and the binding of surfactant to the polymer molecules. Both these processes consume surfactant molecules. Thus, at any surfactant content, we may express the total surfactant concentration as a sum of the free surfactant, the surfactant bound to the polymer, and the surfactant bound to the particle surface. Specifically, at the particle

2748

Langmuir, Vol. 21, No. 7, 2005

Olsson et al.

Figure 6. Schematic illustration of the situation at particle inversion. The EHEC molecules have just been removed from the particle surfaces, which are saturated with a layer of adsorbed surfactant molecules. For associating polymersurfactant pairs, surfactant molecules are also bound in micellelike aggregates at the EHEC molecules. The adsorbed and the polymer-bound surfactant molecules are in equilibrium with a bulk surfactant concentration, which is close to the cmc of the surfactant.

inversion concentration cs,i, we obtain

cs,i ) csf,i + csa,i + csb,i

(1)

where the subscripts f, a, and b denote “free”, “adsorbed” (to the particle), and “bound” (to the polymer), respectively. Our analysis above implies that the polymer has been removed from the particle surface at the particle inversion; hence, there is no cross term (no surfactant bound to polymers adsorbed to the particles). The situation is schematically illustrated in Figure 6. Both the adsorption and the binding are governed by the chemical potential of the surfactant (essentially, the monomer concentration of surfactant). The two processes may be described in terms of an adsorption isotherm and a binding isotherm, respectively, relating the surface coverage and the degree of surfactant binding to the free surfactant concentration. Experimental isotherms indicate that the removal of a surfactant-binding polymer from an attractive surface occurs when the free surfactant concentration is close to the cmc,1,4,7,10,11 that is,

csf,i ≈ cmc

(2)

At this point, we expect that the particle surface is saturated with surfactant so that

csa,i ) (Asp/As)cc

(3)

where Asp is the specific surface area of the particles, As is the area occupied by 1 mol of surfactant molecules at the particle surface at full coverage, and cc is the concentration of colloidal particles. Finally, we may express the concentration of polymer-bound surfactant molecules as

csb,i ) βicp

(4)

where cp is the concentration of polymer and βi is the degree of binding of surfactant to the polymer at particle inversion. Before pursuing this analysis any further, we note that both the terms csf,i and csb,i may differ greatly among the investigated surfactants, owing to the large variations in both the cmc and the affinity of the surfactant to the polymer. This is the reason cs,i varies so much between different surfactants. Anticipating the quantitative analysis below, we may further conclude that the term csa,i will be negligible for most of the systems investigated here,

because the particle concentrations were low. The exceptions occur for surfactants with both low cmc values and low affinities to the polymer. Thus, except for surfactants giving low inversion concentrations, we expect the particle inversion concentration to be independent of the particle concentration in our systems. Indeed, this is found for all systems except for the N-EHEC/C12E8/polystyrene latex mixtures. For surfactants with both high cmc values and a low affinity to EHEC, we obtain cs,i ≈ csf,i ≈ cmc, hence, the similarity between the inversion concentration and the cmc observed for the cases with polystyrene latex particles in mixtures of N-EHEC/C8G1 and HM-EHEC/ C8G1. The parameter βi requires a more detailed consideration. For N-EHEC and nonionic surfactants, βi ) 0, and csb,i vanishes altogether. By contrast, for ionic surfactants that bind to N-EHEC, βi may be large. Because we expect csf,i to be close to the cmc, we should expect βi to be close to its maximum value. (From the point where free micelles form, the chemical potential of the surfactant stays roughly constant at increasing surfactant concentration.) Estimates of the maximum binding of NaC12S to N-EHEC are available,25,51 and using these estimates, we obtain βi ≈ 4-4.5 mmol/g EHEC. Inserting this number into relations 1-4 above and neglecting csa,i, we predict cs,i ≈ 50 mM for NaC12S in a 1% N-EHEC solution. However, at this concentration, the cloud point of the system has already exceeded 100 °C. Hence, the particle inversion does not occur in the experimentally accessible concentration range. We propose that the same explanation holds also for the other ionic surfactants where no particle inversion was observed. If the above analysis is correct, we may nevertheless expect that there will be a repulsion between the NaC12Sdressed N-EHEC molecules and particles at sufficiently high NaC12S concentrations. If, furthermore, the particle and polymer concentrations are also sufficiently high, this repulsion should result in a depletion flocculation, even at temperatures where the NaC12S-dressed N-EHEC molecules are fully miscible with water. To check for this possibility, we prepared a sample containing 1 wt % EHEC, 1 wt % polystyrene latex particles, and 500 mM NaC12S and indeed observed at room temperature a phase separation that was segregative; that is, the low-viscous particle-rich phase was depleted in polymer. Next, we will turn to the mixtures of HM-EHEC and nonionic surfactants. Here, the surfactant binding occurs only to the polymer hydrophobes, resulting in mixed micelles of hydrophobes and surfactant molecules. Thus, we may express the concentration of polymer-bound surfactant at particle inversion as

csb,i ) βh,ich

(5)

where ch is the concentration of polymer hydrophobes (0.37 mM at 1% of HM-EHEC) and βh,i is the ratio of surfactant molecules to hydrophobes in the mixed micelles. An independent estimate of βh,i is problematic, because this value increases rapidly as the free surfactant concentration approaches the cmc.52,53 Here, we will instead use the experimental data and the assumption csf,i ≈ cmc to extract information on βh,i to see whether the numbers thus obtained are consistent and reasonable. We will only use (51) Rose´n, O.; Bostro¨m, M.; Nyde´n, M.; Piculell, L. J. Phys. Chem. B 2003, 107, 4074-4079. (52) Piculell, L.; Thuresson, K.; Ericsson, O. Faraday Discuss. 1995, 101, 307-318. (53) Piculell, L.; Guillemet, F.; Thuresson, K.; Shubin, V.; Ericsson, O. Adv. Colloid Interface Sci. 1996, 63, 1-21.

Polymer-Particle Mixtures Phase Behavior Changes

data for the surfactants with low cmc values (Triton X-100 and C12E8), where the term csb,i dominates the particle inversion concentration (in total seven different mixtures). We thus obtain βh,i ) 4.7 ( 1.9 for the investigated mixtures. These values are internally consistent, and the fact that the mixed micelles are dominated by the surfactant is consistent with a free surfactant concentration close to the cmc and a detachment of the polymer molecules from the particle surfaces. In this context, we also note that, for the nonionic surfactants and HM-EHEC, the particle inversion concentrations almost coincide with the respective minima in the cloud point curves (see Figure 3b). This seems entirely reasonable, because both phenomena reflect an increase in solubility of the polymersurfactant complex due to a breakup of the hydrophobic (mixed micellar) cross-links between the polymer molecules. Furthermore, independent viscosity experiments on similar mixtures of hydrophobically modified polymers and surfactants indicate that a significant breakup of mixed micellar cross-links occurs at similar values of βh,i.26,52,54 In Figure 5a,b, we summarize the above analysis by showing the calculated values of cs,i for all the cases where a particle inversion has been observed. In the calculations, we have used csf,i ) cmc, Asp ) 26 m2/g for silica, Asp ) 16 m2/g for polystyrene latex, and As ) 4 × 105 m2/mol for all surfactants. The values of Asp were determined by assuming smooth spherical particles with densities of 2300 kg/m3 for silica and 1050 kg/m3 for the polystyrene latex particles, while As is an average value based on the reported plateau adsorbed amounts in the interval 1.83.5 µmol/m2 for the surfactants C12E8, NaC12S, and Triton X-100.55-60 For βh,i we have used the fitted value 4.7; see above. The agreement between the calculated and the experimental values of cs,i is excellent in almost all cases. There remains one observation to explain. The nonionic surfactant Triton X-100 did not give rise to particle inversion in mixtures with silica particles and HM-EHEC, even at very high surfactant concentrations (≈50 × cmc). (54) Piculell, L.; Egermayer, M.; Sjo¨stro¨m, J. Langmuir 2003, 19, 3643-3649. (55) Levitz, P.; El Miri, A.; Keravis, D.; Van Damme, H. J. Colloid Interface Sci. 1984, 99, 484-492. (56) Tiberg, F.; Jo¨nsson, B.; Tang, J.; Lindman, B. Langmuir 1994, 10, 2294-2300. (57) Turner, S. F.; Clarke, S. M.; Rennie, A. R.; Thirtle, P. N.; Cooke, D. J.; Li, Z. X.; Thomas, R. K. Langmuir 1999, 15, 1017-1023. (58) Romero-Cano, M. S.; Martin-Rodriguez, A.; de las Nieves, F. J. J. Colloid Interface Sci. 2000, 227, 322-328. (59) Geffroy, C.; Stuart, M. A. C.; Wong, K.; Cabane, B.; Bergeron, V. Langmuir 2000, 16, 6422-6430. (60) Lin, S.-Y.; Dong, C.; Hsu, T.-J.; Hsu, C.-T. Colloids Surf., A 2002, 196, 189-198.

Langmuir, Vol. 21, No. 7, 2005 2749

We attribute this difference for silica particles to the fact that the nonionic surfactants adsorb as micelles to the hydrophilic silica surface.56,61,62 Thus, there will be a competition between free micelles and polymer-bound micelles to bind to the silica surface. Because a polymer molecule contains more than one micelle, the adsorption of polymer-bound micelles will be favored for entropic reasons, just as the adsorption of a homopolymer to a surface is favored over the adsorption of the corresponding monomer. As a consequence, a large concentration of free micelles is required to replace the polymer-surfactant aggregates from the surface. Conclusions (i) The interactions between EHEC and dispersed silica or polystyrene latex particles may change from attraction to repulsion by added surfactant. (ii) The change in interaction results in an inversion of the partitioning of the particles, from the EHEC-rich to the EHEC-poor phase, in phase-separated dispersions (above the cloud point). (iii) Even small amounts of added particles lower the cloud point of EHEC-surfactant solutions below the particle inversion concentration but not above. (iv) The particle inversion concentration can be predicted by considering three states of surfactant molecules in equilibrium at the particle inversion concentration: free (at a concentration close to cmc), adsorbed to the particle surface, or bound to the EHEC molecules. (v) The observed large variations in the particle inversion concentration among different surfactants are due to their large differences in the cmc and/or degree of binding to EHEC. (vi) Ionic surfactants that bind extensively to EHEC may not display a particle inversion, because the binding of surfactant to the polymer increases the cloud point above 100 °C before the particle inversion is expected to occur. Acknowledgment. The authors thank Krister Thuresson for valuable discussions. Anne Anderse´n and Siv Karlsson at Akzo Nobel Surface Chemistry, Stenungsund, Sweden, are acknowledged for the analysis of the surfactant content in selected solutions. This work was financed by the Center for Amphiphilic Polymers from Renewable Resources (CAP) and the Swedish National Research Council (NFR). LA0473254 (61) Levitz, P.; Van Damme, H.; Keravis, D. J. Phys. Chem. 1984, 88, 2228-2235. (62) Wijmans, C. M.; Linse, P. J. Phys. Chem. 1996, 100, 1258312591.