Mixed Solutions of Surfactant and Hydrophobically Modified Polymer

Susanne Nilsson*, Krister Thuresson, Per Hansson, and Björn Lindman. Physical Chemistry 1, Center for Chemistry and Chemical Engineering, Lund Univer...
0 downloads 0 Views 103KB Size
J. Phys. Chem. B 1998, 102, 7099-7105

7099

Mixed Solutions of Surfactant and Hydrophobically Modified Polymer. Controlling Viscosity with Micellar Size Susanne Nilsson,*,† Krister Thuresson,† Per Hansson,‡ and Bjo1 rn Lindman† Physical Chemistry 1, Center for Chemistry and Chemical Engineering, Lund UniVersity, P.O. Box 124, S-221 00 Lund, Sweden, and Department of Physical Chemistry, Uppsala UniVersity, P.O. Box 532, S-751 21 Uppsala, Sweden ReceiVed: February 22, 1998; In Final Form: June 3, 1998

The viscosity of mixtures of a hydrophobically modified cellulose polymer (HMHEC) and surfactants in aqueous solution was determined as a function of the concentration of mixed micelles. It was shown that, in solutions of low viscosity where the ratio between polymer hydrophobic tails and mixed micelles is low, a high viscosity can be recovered by increasing the surfactant aggregation number in the mixed micelles. The surfactant aggregation number of the mixed micelles was increased by adding either a screening electrolyte or an oppositely charged surfactant to solutions containing the hydrophobically modified polymer and an ionic surfactant. Time-resolved fluorescence quenching (TRFQ) was used to investigate the size of the mixed micelles.

Introduction Solutions of hydrophobically modified water-soluble polymers and surfactants have been widely investigated during the past decade. A reason for this is their frequent use as rheology modifiers.1-3 One class of such polymers consists of a (nonionic or ionic) water-soluble backbone onto which a low amount of hydrophobic side chains has been grafted.4-6 In dilute solutions the hydrophobic tails associate intramolecularly to minimize their water contact. In semidilute solutions intermolecular association also takes place which macroscopically results in an enhanced viscosity. It is well-known that the viscosity of semidilute solutions of associative polymers is sensitive to the presence of surfactants.1 Surfactant molecules are regarded to influence the cross-linking between different polymer chains due to the formation of mixed micelles composed of polymer hydrophobic tails and surfactant molecules. Several investigations have shown that the molar ratio between the hydrophobic polymer tails and the mixed micelles in which they participate is an important parameter for determining the rheological response of the solution.7-10 Typically, a low viscosity follows from a decreased connectivity between different polymer chains when the ratio becomes small. This may cause difficulties in applications such as water-based paint formulations where different kinds of surfactants are used for other purposes such as to stabilize pigments, latex particles, and fillers. This serves as the base for the present investigation where the surfactant aggregation number of the mixed micelles has been changed in a controlled way while monitoring the viscosity. As a model system we have chosen aqueous solutions of the nonionic polymer hydrophobically modified (hydroxyethyl)cellulose (HMHEC) and the anionic surfactant sodium dodecyl sulfate (SDS), which is a system well-known from literature.11-16 Pure SDS micelles grow when a screening electrolyte is added.17 A similar effect is expected for the mixed †

Lund University. Uppsala University. * To whom correspondence should be addressed.



micelles in the present system. Addition of an oppositely charged surfactant is known to increase the aggregation number of ionic micelles,18,19 and this was another method employed to control the size of the mixed micelles. Time-resolved fluorescence quenching (TRFQ) was used to follow the changes in the concentration of the mixed micelles as a function of screening electrolyte concentration or as a function of the fraction of cationic surfactant, while keeping the total surfactant concentration constant. Furthermore, the viscosity was obtained from oscillatory shear experiments and correlated to the concentration of mixed micelles. Experimental Section Materials. Hydrophobically modified (hydroxyethyl)cellulose (HMHEC) with the commercial name Natrosol Plus grade 330 was obtained from Aqualon. According to the manufacturer, HMHEC has a molecular mass of 250 000 and carries grafted C16 alkyl chains corresponding to a hydrophobic modification degree of 0.5-1 w/w %. The degree of substitution of hydroxyethyl groups per repeating anhydroglucose unit of the polymer equals 3.3. In the present investigation the polymer material was analyzed according to a description by Landoll4 and found to contain 1.2 w/w % of C16 alkyl chains of the (total) dry sample weight. This corresponds to 0.54 mm alkyl chains in a 1 w/w % aqueous polymer solution. Using the hydroxyethyl substitution degree as given by the manufacturer, 1.7 mol % of the anhydroglucose units contain alkyl chains. Prior to use HMHEC was extensively washed with acetone to remove unreacted alkyl chains, dried, dissolved in water to a concentration of 1 w/w %, and centrifuged at 10 000g to get rid of high molecular weight impurities (such as unreacted cellulose). Low molecular weight impurities (such as salt) were removed by dialysis against Millipore water in a Filtron Ultrasette device. The dialysis was performed until the expelled water showed a conductivity of less than 2 µS/cm. After freezedrying, the polymer was stored in a desiccator. Sodium dodecyl sulfate (SDS) (specially pure) from BDH, dodecyl trimethylammonium chloride (DoTAC) from TCI

S1089-5647(98)01237-1 CCC: $15.00 © 1998 American Chemical Society Published on Web 08/21/1998

7100 J. Phys. Chem. B, Vol. 102, No. 37, 1998

Nilsson et al.

Tokyo Kasei, and sodium chloride (NaCl), ACS reagent, from Sigma were used as supplied. Pyrene from Aldrich was recrystallized twice from ethanol. N-Dodecylpyridinium chloride (DoPC) from Aldrich was recrystallized several times from acetone. N-Cetylpyridinium chloride (CPC) (analytic grade) from Merck was used as supplied. Methods. Sample Preparation. The samples were prepared from stock solutions by weight. All of them had a polymer concentration of 1 w/w % (1 g of polymer per 100 g of solvent), which is in the semidilute region and well above the critical polymer concentration (ca. 0.2%) where intermolecular hydrophobic associations take place.11 The test tubes were of glass and sealed with Teflon-tightened screw caps. The samples were mixed by tilting the test tubes end over end several days at 25 °C. To prepare the samples for the TRFQ measurements which all contained the water-insoluble pyrene, an accurate amount of pyrene dissolved in ethanol was transferred to a beaker. Finally, after evaporating off the ethanol, proper amounts of HMHEC and Millipore quality water were added. All TRFQ samples contained 10 µm pyrene. The concentration of pyrene was chosen such that it was high enough to decrease the influence of fluorescent impurities originating from the polymer material but still small enough not to disturb the micelles. CPC was used as a quencher in systems containing DoTAC or NaCl since the long aliphatic chain of the cationic CPC implies a strong binding also to cationic micelles. On the other hand, DoPC was used in the reference system (HMHEC + SDS) in order not to affect the aggregation number at low surfactant concentrations. Viscosity. The viscosity measurements were performed in the oscillatory mode with a Carri-Med controlled stressed CSL 100 rheometer, which has an automatic gap setting. Two different cone and plate geometries, 4 and 6 cm in diameter, were used for high viscous and low viscous samples, respectively. All measurements were carried out at 25 °C, and the complex viscosity, η*, was calculated through

η* ) [(G′2 + G′′2)/(2πf)2]1/2

(1)

where G′ is the storage modulus, G′′ the loss modulus, and f the frequency. In most samples the reported value corresponds to the Newtonian plateau. However, in samples with high viscosity sometimes this limit was not reached, and therefore the viscosity is given at the lowest frequency measured (typically 0.005-0.01 Hz). This procedure results in slightly too low absolute values in samples with high viscosity, while the general trends remain unchanged. Fluorescence Quenching (TRFQ). Time-resolved fluorescence decay data were collected with the single-photon counting technique. A detailed description of the experimental technique and equipment is given by Almgren et al.20 The fluorescence from the probe (pyrene) was monitored at 395 nm following excitation at 320 nm. All measurements were carried out at 25 °C. From experiments using quenchers (DoPC or CPC), the concentration of hydrophobic domains, [M], was obtained from the estimated average number of quenchers per micelle (hydrophobic domain), 〈n〉, using

M ) Qm/〈n〉

(2)

Qm is the concentration of quencher (DoPC or CPC) in the micelles. All quenchers are assumed to be inside the micelles as long as there is surfactant present. CPC distributes to the micelles due to strong hydrophobic attractions, while DoPC was used in mixtures with SDS where the free concentration of both

cationic and anionic surfactants is expected to be very low. For the pure HMHEC sample the free concentration of quencher may become important (see below). Infelta21 proposed the following function describing the time evolution of the fluorescence signal, F(t), from a probe situated in small uniform micelles in the presence of quenchers.

F(t) ) A1 exp[-A2t + A3{exp(-A4t) - 1}]

(3)

A1 is the amplitude of the decay curve at the excitation event (t ) 0). In the case of stationary probe and quencher, i.e., when both compounds reside in the micelles considerably longer than the fluorescence lifetime of the probe (τ0), the other parameters can be interpreted as22 A2 ) 1/τ0, A3 ) 〈n〉, and A4 ) kq, where 〈n〉 is the average number of quencher per micelle and xkq is the deactivation rate in micelles containing x quenchers. For every sample, two different fluorescence decay curves were recorded, one without quencher which by fitting a singleexponential function to the decay curve gave the fluorescence lifetime of pyrene, τ0. The other decay curve, with quencher present, is fitted with eq 3. In one sample (1 w/w % HMHEC, with DoTAC to SDS molar fraction of r ) 0.25) with large nonuniform aggregates (as compared to the small micelles in the other samples), eq 3 resulted in poor fits. This is expected since it describes (in an approximate way) the quenching in small micelles. Instead 〈n〉 was evaluated graphically by plotting the logarithm of the relative fluorescence intensity against time for the quenched sample.23 An extrapolation of the single-exponential part of the quenched curve to time zero gives 〈n〉 as the difference in the amplitude between the extrapolated (linear) curve and the amplitude of the quenched curve. In principle, this method can be used once it has been confirmed that the micelles are discrete; i.e., the decay rate at long times is the same for the quenched and for the unquenched samples. Note that 〈n〉 depends on the distribution of the quencher between water and micelles, not on the quenching process. Results and Discussion General Discussion. We start this section with a short review of recent results from the literature in order to facilitate the interpretation of our data. It is generally found that the viscosity of a polymer solution increases strongly on hydrophobic modification provided the concentration is high enough to allow interpolymeric associations. However, this is only true within certain limits. Of course, the modification degree must exceed a certain threshold value; i.e., each polymer chain must contain at least two hydrophobic tails on average to have the possibility to form networks. Furthermore, Annable et al. showed that the length of the hydrophobic tail must exceed ca. 6 carbons to influence the viscosity in a solution of end-capped PEG polymers.24 It is known that there is also an upper limit in hydrophobic modification degree. A too high modification degree and/or too long hydrophobic tails will result in dissolution problems of the polymer material.16 Hydrophobic molecules tend to dissolve in the polymer hydrophobic domains. Especially surfactants are expected to bind to these regions.9 Typically the viscosity of a semidilute polymer solution of constant concentration is expected to first increase and then decrease as a function of added surfactant. The increase has been attributed to two effects. An addition of surfactant at low concentrations can be expected to increase both the strength of the connections and the concentration of mixed micelles. We refer the increased strength of the connections to

Solutions of Surfactant and Modified Polymer an improved packing of surfactant molecules in the mixed micellar corona. Such an improvement of the interface between the micellar core and the aqueous environment is expected to increase the residence time of the polymer hydrophobic tails in the micelles (see below). An increased concentration of mixed micelles is expected to cause a reorganization of polymer chains from intra- to interaggregates.25 Thus, the number of polymer strands that actually connects two mixed micelles and therefore participate in the network increases. The decrease in viscosity is generally regarded to be due to an untying process in which the connections between different polymer chains are broken when the concentration of mixed micelles become larger, and the average number of polymer hydrophobic tails in each micelle decreases.26 A general finding in an investigation by Piculell et al.26 was that the breakdown of the network was strongly correlated to a cooperative aggregation process of the surfactant molecules. In this context it may also be valuable to discuss the binding of SDS to hydrophobically modified ethyl(hydroxyethyl)cellulose (HMEHEC), which is a cellulose derivative related to HMHEC. We start to discuss the unmodified parent polymer (EHEC) which also interacts with SDS and promotes aggregation of the surfactant. At the critical aggregation concentration (cac) rather well-defined aggregates are regarded to form, and typically the aggregation number is smaller than in normal micelles (20-40 as compared to 60).9,27-30 The hydrophobic modification of HMEHEC provides nucleation sites at which individual surfactant molecules can bind. Thuresson et al. showed that the binding of SDS to HMEHEC could be divided into two regions.9,31 At low concentrations the binding was reminiscent of the adsorption to a hydrophobic macroscopic surface, and the binding could be described with a Langmuir adsorption isotherm. Therefore, at low concentrations a constant fraction (in this particular case about 0.5) of the added surfactant material binds to the hydrophobic moieties. On increasing the surfactant concentration the sites available for noncooperative binding become saturated, and the fraction of bound SDS decreases. When the free SDS concentration in the HMEHEC solution corresponds to the cac in the EHEC solution (3.5 mm), the isotherm is determined by a cooperative process, and aggregates resembling those in the EHEC solution form. HMHEC and SDS. The main purpose of the present study is to investigate the effect of salt and an oppositely charged cosurfactant on HMHEC-SDS mixtures. However, we start with a characterization of the system in the absence of additives. Figure 1 shows how the viscosity changes with the SDS concentration in 1 w/w % HMHEC solutions. The pronounced viscosity maximum at 7 mm SDS is in good agreement with the result by Tanaka et al.12 The estimated concentration of micelles (from TRFQ) is also presented in Figure 1. The nearly constant concentration of micelles to the left of the viscosity maximum confirms the existence of aggregated polymer side chains already in the absence of surfactant, as previously observed with HMEHEC. In further agreement with the latter system, the reduction of the viscosity at higher surfactant concentrations can be explained by an increased micelle concentration as the cac in the corresponding unmodified system is exceeded. In the present system the cac is about 6 mM as indicated by a significant drop in the viscosity observed in dilute HEC solutions.32 This value is higher than for the more hydrophobic EHEC, where a stronger interaction with the polymer is at hand. However, for both polymers, the maximum in viscosity in the modified system is positioned close to the cac in the unmodified system.

J. Phys. Chem. B, Vol. 102, No. 37, 1998 7101

Figure 1. Complex viscosity (filled circles) and concentration of micelles (open circles) as a function of SDS concentration in a 1 w/w % HMHEC solution. The lines only serve as a guide to the eye.

Before proceeding, we note that the concentration of micelles estimated from TRFQ in Figure 1 relies on the assumption that all quenchers (DoPC) are distributed in the micelles. As long as there are SDS bound to the micelles, the error will be negligible due to the favorable electrostatic interactions. In fact, a comparison of the estimates in the presence and in the absence of SDS (see Table 1) suggests that the concentration of free quencher is small also in the latter case. It is important to point out that the concentration of micelles as calculated from 〈n〉 is independent of the SDS binding. In contrast, to obtain the number of SDS per aggregate, Nagg, we need to know the binding isotherm. However, even with the isotherm at hand the binding at low SDS concentrations would be different in the presence of the quencher. Nevertheless, by keeping in mind the qualitative nature of the result, we find it is instructive to calculate the aggregation number assuming that the binding isotherm is the same as reported for HMEHEC/ SDS (see above),9 except that the cac is 6 mm (instead of 3.5 mm). The results for low and high SDS concentrations are given in Table 2. At intermediate concentrations no calculation was made due to uncertainties in the binding isotherm. The small aggregation numbers found at low SDS concentrations are in agreement with a noncooperative binding to aggregates formed by the polymer side chains. Note that, on a qualitative level, the same conclusion can be made even if all SDS is bound to the aggregates (multiply the aggregation numbers by two). The values obtained above 10 mm SDS agree very well with those reported in the HMEHEC/SDS system. In this range the influence of the quencher on both the binding and the aggregation number is expected to be small.33 The TRFQ measurements also provide information about the dynamics of the quenching process. Here, kq, the quenching frequency within the micelles, is sensitive to the size and the microviscosity of the micelles. In polymer/surfactant complexes, kq, which is independent of 〈n〉 , is generally found to be smaller than in the corresponding pure surfactant micelles even when the former aggregates are smaller.34 This effect, which is present also in HMHEC-SDS (Table 1), has been attributed to a decreased mobility within the polymer-surfactant complexes.35 The most striking observation in the present system is, however, the slow quenching in the polymerdominated aggregates. In fact, there seems to be a turn over to faster quenching rates at the onset of the cooperative binding

7102 J. Phys. Chem. B, Vol. 102, No. 37, 1998

Nilsson et al.

TABLE 1: Parameters Estimated from TRFQ Measurements on Mixtures of HMHEC-SDS, HMHEC-SDS-NaCl, and HMHEC-SDS-DoTAC sample

a

kq × 10-7/s-1

τ0/ns

quencher/mm

0 mm SDS 0.2 mm SDS 1.0 mm SDS 5.0 mm SDS 7.0 mm SDS 8.0 mm SDS 9.0 mm SDS 11.9 mm SDS 17.9 mm SDS

0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.2

1 w/w % HMHEC; DoPC 246 0.53 243 0.61 255 0.79 233 1.4 222 1.5 221 1.7 225 1.7 218 1.8 215 1.9

0 M NaCl 0.03 M NaCl 0.05 M NaCl 0.1 M NaCl 0.2 M NaCl 0.4 M NaCl 0.5 M NaCl

0.2 0.2 0.2 0.2 0.2 0.2 0.2

1 w/w % HMHEC + 15 mm SDS; CPC 214 2.1 205 1.9 204 1.9 200 1.7 201 1.7 202 1.6 203 1.5

0 DoTAC molar ratio 0.10 DoTAC molar ratio 0.20 DoTAC molar ratio 0.25 DoTAC molar ratio 0.90 DoTAC molar ratio 1.0 DoTAC molar ratio

1 w/w % HMHEC + 30 mm (SDS + DoTAC); CPC 0.5 217 2.3 0.5 207 1.4 0.5 189 0.67 0.5 178 0.45 0.5 177 1.6 0.5 191 4.4

χ2 a

〈n〉

1.14 0.989 1.08 1.31 1.02 0.950 1.03 1.07 1.01

1.17 0.989 0.728 1.05 1.24 1.39 1.28 1.01 0.658

0.085 0.101 0.137 0.095 0.162 0.144 0.156 0.198 0.304

1.00 0.852 1.14 1.08 1.01 1.09 1.09

0.789 0.920 0.982 1.08 1.20 1.36 1.41

0.254 0.217 0.204 0.186 0.166 0.148 0.142

1.08 1.17 1.68 3.25 1.07 1.19

0.889 1.61 3.18 4.35b 3.54 1.69

0.562 0.311 0.157 0.115 0.141 0.296

[M]/mm

Quality of fit. b Graphical evaluation gives 〈n〉 ) 4.55 and [M] ) 0.110.

TABLE 2: Calculated Surfactant Aggregation Numbers and the Number of Polymer Hydrophobic Tails per Mixed Micelle for the HMHEC-SDS, HMHEC-SDS-NaCl, and HMHEC-SDS-DoTAC Systemsa Nagg,s

nb

0 1 4 6 14 19 30 39

6.4 5.3 3.9 5.7 3.3 3.8 3.5 2.7 1.8

1 w/w % HMHEC + 15 mm SDS 6 35 2.3 58 1.6 65 0.97 75 0.67 86 0.44 98 0.40 103

2.1 2.5 2.6 2.9 3.3 3.6 3.8

sample

cac/mM 1 w/w % HMHEC

0 mm SDS 0.2 mm SDS 1.0 mm SDS 5.0 mmSDS 7.0 mm SDS 8.0 mm SDS 9.0 mm SDS 11.9 mm SDS 17.9 mm SDS 0 M NaCl 0.03 M NaCl 0.05 M NaCl 0.1 M NaCl 0.2 M NaCl 0.4 M NaCl 0.5 M NaCl

6 6 6 6 6

1 w/w % HMHEC + 30 mm (SDS + DoTAC) 0 DoTAC molar ratio 6 43 0.10 DoTAC molar ratio 0 96 0.20 DoTAC molar ratio 0 191 0.25 DoTAC molar ratio 0 261 0.90 DoTAC molar ratio 0 213 1.0 DoTAC molar ratio 19 37

1.0 1.7 3.4 4.7 3.8 1.8

a Note that the cac values used in the calculations are estimated (see text). b Number of polymer hydrophobic tails per hydrophobic domain.

(i.e., surfactant-dominated aggregates). This behavior, which was also found in HMEHEC/SDS,9 points to the difference in the nature of the two types of aggregates. Effect of Salt. In an investigation regarding the HMHECSDS system, Tanaka et al.12 found that NaCl only had a minor effect on the viscosity profile as a function of SDS in a solution containing 1 w/w % HMHEC. At first this is somewhat

unexpected because on addition of simple salt the relative reduction of the cac (for ionic surfactants) has been found to resemble that for the reduction of cmc.36,37 Thus, the concentration of free surfactant required to achieve cooperative surfactant binding can be expected to decrease substantially, and the viscosity maximum could be expected to be located at a lower total SDS concentration than in the salt-free system. However, except from a decreased cac,38 micellar growth is also promoted by the reduced electrostatic repulsion between the charged surfactant headgroups which follows from salt addition.39 From the viscosity profile’s point of view, this is expected to oppose the effect of a decreased cac and can therefore be the explanation for the observation by Tanaka et al.12 Indeed, in the present investigation the surfactant aggregation number Nagg of the mixed micelles was found to increase with increasing NaCl concentration in a system composed of 15 mm SDS and 1 w/w % HMHEC; see Table 2. As suggested by data in the literature, it was assumed that salt induces the same relative reduction of the cac as that of the cmc.37,40 This is expected to be a good assumption at the present SDS concentration where the effect of the hydrophobic modification of the polymer on the micelles is negligible. Throughout this section the polymer and surfactant concentrations were kept constant at 1 w/w % and 15 mm, respectively, and we report on variations in the viscosity and changes in the concentration of mixed micelles as a function of NaCl. In Figure 2 it can be seen that the viscosity increases with added NaCl up to a concentration of 0.1 M. At higher NaCl concentrations the viscosity decreases. In contrast, it was found that the concentration of mixed micelles continuously decreased with salt. The fastest decrease was observed at low NaCl concentrations. One explanation for the decrease in viscosity at high NaCl concentrations could be a too low concentration of mixed micelles which opposes a fully developed network due to intraaggregation of the polymer chains. However, this explanation seems less likely because a high viscosity could not be recovered by increasing the SDS concentration above

Solutions of Surfactant and Modified Polymer

Figure 2. Complex viscosity (filled circles) and concentration of micelles (open circles) versus the concentration of NaCl in samples with 1 w/w % HMHEC and 15 mm SDS. The lines are drawn to guide the eye.

Figure 3. Complex viscosity in mixtures of 1 w/w % HMHEC, SDS and NaCl.

Figure 4. Phase boundaries for the mixed cationic-anionic surfactant system at different molar fractions of DoTAC at a total SDS + DoTAC concentration of 30 mm.

15 mm while keeping the NaCl concentration constant (corresponding to an increase in the concentration of mixed micelles) (Figure 3). Furthermore, as mentioned above, Figure 1 indicates that the increase in viscosity on addition of SDS is not due to an increased concentration of mixed micelles. The same data plotted in a different way (see Figure 6) show that the viscosity increases abruptly with the concentration of mixed micelles in the HMHEC-SDS system. This behavior is not reproduced by the corresponding curve (viscosity as a function of concentration of mixed micelles) for the system where the NaCl

J. Phys. Chem. B, Vol. 102, No. 37, 1998 7103

Figure 5. Complex viscosity (filled circles) and concentration of micelles (open circles) versus the molar fraction of DoTAC at the total DoTAC + SDS concentration of 30 mm in a 1 w/w % HMHEC solution. Note the phase separation area at intermediate ratios. The lines only serve as a guidance for the eye.

Figure 6. Complex viscosity as a function of the concentration of micelles for three different systems. (A) 1 w/w % HMHEC + SDS. The changes with the SDS concentration (from left to right 0, 3, 5, 6, 6.5, 7, 7.6, 8, 9, 11, 15, 30, 50 mm) are depicted with triangles. (B) 1 w/w % HMHEC + 30 mm surfactant (SDS + DoTAC). Circles mark the variation with DoTAC to total surfactant molar fraction (r ) 0, 0.05, 0.10, 0.15, 0.20, 0.25, 0.28, 0.30, 0.90, 1.0). The hatched lines depict the phase separation area; the line is only drawn to make it easier to follow the trend curve. (C) 1 w/w % HMHEC + 15 mm SDS + NaCl. The variation with the NaCl concentration (0, 0.03, 0.05, 0.10, 0.20, 0.40, 0.50 M) is marked with filled diamonds. In samples where the viscosity was measured at a concentration intermediate to two concentrations at which TRFQ measurements were performed; the concentration of micelles, [M], was extrapolated.

concentration was changed (polymer and SDS concentrations kept constant). Here the variations were found to be more gradual. This points to a different mechanism (than stoichiometry) at high NaCl concentrations. We suggest that the decrease in viscosity at high concentrations of a screening electrolyte is due to attractive forces within the polymer-surfactant complexes which leads to a contraction. In line with this we found that on further addition of NaCl an associative phase separation occurred (at ca. 0.7-0.8 M NaCl).

7104 J. Phys. Chem. B, Vol. 102, No. 37, 1998 Thus, the solubility of the complex (and therefore the viscosity of the solution) is determined by an interplay between electrostatic repulsion and an attraction due to a hydrophobic effect. To summarize, as the system becomes microheterogeneous close to the phase separation limit, the viscosity decreases. Mixtures of Oppositely Charged Surfactants. Investigations concerning the phase behavior of mixtures of the oppositely charged surfactants dodecyl trimethylammonium bromide (DoTAB) and SDS have been reported by several authors.18,41,42 Techniques such as TRFQ, cryo-transmission electron microscopy (cryo-TEM), light scattering, and conductivity were used. It was found that at a total surfactant concentration of 50 mM a demixing with precipitation occurred in the interval r ) [DoTAB]/([SDS] + [DoTAB]) ) 0.38-0.88.41 Close to phase separation (from r ) 0.30-0.38) the solution was found to be slightly bluish and to contain larger aggregates, such as vesicles and threadlike micelles, probably in equilibrium with smaller micelles. Further away from the region of demixing the solution was isotropic and contained only spherical micelles. In the present investigation we have used a slightly different system where the cationic surfactant has Cl- as the counterion. The phase behavior at a total surfactant concentration of 30 mm (DoTAC + SDS) is presented in Figure 4. In qualitative agreement with the DoTAB-SDS system a turbid, bluish, viscoelastic solution was observed at mixing ratios between r ) 0.32 and r ) 0.38, and the precipitation region extends from r ) 0.38 to r ) 0.93. To the mixtures of DoTAC-SDS was added HMHEC in a concentration of 1 w/w %. It was found that the phase boundaries were slightly changed. The “turbidity boundary” moved from r ) 0.32 to r ) 0.30. On the DoTAC-rich side the precipitation region was decreased from r ) 0.93 to below r ) 0.90. At this stage we cannot decide whether this is a real effect or the shift mainly is due to slow kinetics. However, no phase separation was observed in any sample that was experimentally investigated by TRFQ or rheology. Another speculation is that anionic surfactants bind stronger to HMHEC than cationic ones which also may have influence on the phase behavior. In qualitative agreement with the results obtained in the HMHEC-SDS system on addition of NaCl, it was found that the viscosity increased on addition of the cationic surfactant to the HMHEC-SDS solution. The viscosity increase was correlated with a decreased concentration of mixed micelles (Figure 5). At a molar fraction of r ) 0.18, the viscosity exhibits a sharp increase, and at further addition of DoTAC a highly viscoelastic solution was formed. At a molar fraction corresponding to r ) 0.30 the turbid solution appears (see above). On the DoTAC-rich side the viscosity decreases at the same time as the concentration of micelles increases. As also mentioned in previous sections, the surfactant aggregation numbers can be estimated from the TRFQ data only if the free surfactant concentration is known. Three regions can be identified: the HMHEC system containing only SDS, the HMHEC system containing only DoTAC, and the HMHEC system containing a mixture of the two surfactants. In the first system the free SDS concentration was estimated to 6 mM (see above); in the second system the free DoTAC concentration can be estimated to be close to the cmc (20 mM). The latter is based on the assumption that DoTAC has a cac of ca. 20 mM in the EHEC system and HEC is expected to interact to a lower degree.31 In the third case the free surfactant concentration may be neglected because addition of an oppositely charged surfactant to a solution containing an ionic surfactant usually gives a

Nilsson et al. dramatic decrease in the cmc. Lucassen-Reynders et al.43 measured surface tension on mixtures of SDS and DoTAB. They found that the cmc40 decreased from 8.1 mM for SDS and 15 mM for DoTAB to below 0.1 mM already at minor addition of DoTAB to SDS. The estimated aggregation numbers at the different conditions are presented in Table 2. It was found that the aggregation number increases substantially when SDS is exchanged with DoTAC. It was also found that at the pure DoTAC solution (r ) 1.0) the concentration of micelles is less than that observed at the pure SDS solution (r ) 0). However, this can be explained by the difference in cmc for the two surfactants, 8.1 mM for SDS and 20.0 mM for DoTAC,40 and by the difference in aggregation number; see Table 2. It is important to note that at the highest SDS concentration that was used the concentration of (mixed) micelles is approximately the same as the concentration of polymer hydrophobic tails. Therefore, already at such rather low concentrations of SDS, free surfactant micelles may form. When the aggregates are too large (r ) 0.25), eq 3 is no longer suitable to fit the fluorescence decay data as suggested by the large χ2 value in Table 1. However, a graphical analysis as described in the Methods section gave quantitatively the same results as the evaluation with eq 3. From the high viscosity and the large aggregation number at surfactant ratios above r ) 0.25, we expect the samples to contain larger aggregates. This is supported by pictures taken with cryo-transmission electron microscopy (cryo-TEM44,45). The data, which will be published in a separate paper, reveals the presence of vesicles in samples at r ) 0.3. The existence of large aggregates in the solution is in agreement with findings by Kamenka et al. and Herrington et al. in surfactant mixtures of DoTAB and SDS (without polymer).18,41 Comparison between the Three Different Systems: HMHEC-SDS, HMHEC-SDS-NaCl, and HMHEC-SDSDoTAC. The results from the former sections are replotted and summarized in Figure 6, where the viscosity in the three different systems is given as a function of the concentration of mixed micelles. (These numbers can easily be transferred to an average number of polymer hydrophobic tails per mixed micelle (Table 2).) The most important result from Figure 6 is that the qualitative behavior is the same regardless of the molecular details of the mixed micelles. This fortifies conclusions drawn in previous investigations,8,10 i.e., that the stoichiometry in the mixed micelles is the most important factor in determining the rheology of the solution. However, there exist differences which suggest that other factors have an influence. In the reference system, the rheology can be expected to have contributions from an electrostatic swelling due to the charged surfactants which are bound in the mixed micelles (compare with the macroscopic swelling of covalently bonded gels upon addition of ionic surfactants46). The swelling of the polymer matrices can be expected to increase the viscosity. This is probably a reason to why the viscosity always is lower in solutions with screening electrolyte as compared to the solutions only containing SDS. In line with this and as was mentioned above, the decrease in viscosity at high concentration of electrolyte probably reflects the vicinity to the phase separation due to screening. Another reason to the lower viscosity than that observed in the reference system may be the larger surfactant aggregation numbers (see Table 2). Intraaggregation of polymer tails may be expected to be more frequent with a big mixed micelle (lowering the probability of interaggregation). Both the discussed mechanisms (lower electrostatic swelling

Solutions of Surfactant and Modified Polymer and larger aggregation numbers) may be responsible for the lower viscosity in the systems containing mixtures of oppositely charged surfactants as compared to the HMHEC-SDS system. Conclusions In this work we have controlled the viscosity by varying the aggregate size of the mixed micelles with the intention of providing principles of optimizing rheological properties of mixtures of hydrophobically modified polymers and surfactants. In the reference system (HMHEC + SDS) the maximum in viscosity appears in a small concentration interval. Already small amounts (20 mm) of surfactant decreases the viscosity dramatically from the maximum value (located at 7 mm). We have shown that it is possible to regain a higher viscosity simply by changing the stoichiometry between polymer hydrophobic tails and mixed micelles. From the present investigation it follows that this is an important parameter for the understanding of the rheological behavior, but it was also shown that other mechanisms had to be invoked. Most likely electrostatic effects as well as effects from large aggregation numbers of the mixed micelles have an influence. The behavior in viscosity and aggregation number of the HMHEC-SDS system was shown to be similar to that of the HMEHEC-SDS system earlier reported.9 In particular, the increase in viscosity mainly originates from an improved packing of the amphiphilic molecules in the surface of the mixed micelles. The decrease in viscosity at higher surfactant concentrations originates from a rearrangement such that the intermolecular connectivity decreases when the concentration of mixed micelles increases. Acknowledgment. Johan Borgstro¨m is gratefully acknowledged for his cryo-TEM expertise. We thank Eduardo Marques for many useful discussions on vesicles and oppositely charged surfactants. Thanks also to Ingegerd Lind for her kind technical assistance. Financial support from the NUTEK and Industry Sponsored Centre for Amphiphilic Polymers (CAP) is gratefully acknowledged. References and Notes (1) Glass, J. E. In Polymers in Aqueous Media; Glass, J. E., Ed.; American Chemical Society: Washington, DC, 1989; Vol. 223. (2) Schulz, D. N.; Glass, J. E. In Polymers as Rheology Modifiers; Schulz, D. N., Glass, J. E., Eds.; American Chemical Society: Washington, DC, 1991. (3) McCormick, C. L.; Bock, J.; Schulz, D. N. In Encyclopedia of Polymer Science and Engineering; Mark, H. F., Bikales, N. M., Overberger, C. G., Menges, G., Eds.; Wiley-Interscience: New York, 1989; Vol. 17, p 730. (4) Landoll, L. M. J. Polym. Sci., Part A: Polym. Chem. 1982, 20, 443. (5) Magny, B.; Lafuma, F.; Iliopoulos, I. Polymer 1992, 33, 3151. (6) Wang, T. K.; Iliopoulos, I.; Audebert, R. In Water Soluble Polymers. Synthesis, Solution Properties and Applications; Shalaby, S. W., McCormick, C. L., Butler, G. B., Eds.; ACS Symposium Series 467; American Chemical Society: Washington, DC, 1991; p 218. (7) Loyen, K.; Iliopoulos, R.; Audebert, R.; Olsson, U. Langmuir 1995, 11, 1053.

J. Phys. Chem. B, Vol. 102, No. 37, 1998 7105 (8) Sarrazin-Cartalas, A.; Iliopoulos, I.; Audebert, R.; Olsson, U. Langmuir 1994, 10, 1421. (9) Thuresson, K.; So¨derman, O.; Hansson, P.; Wang, G. J. Phys. Chem. 1996, 100, 4909. (10) Piculell, L.; Guillemet, F.; Thuresson, K.; Shubin, V.; Ericsson, O. AdV. Colloid Interface Sci. 1996, 63, 1. (11) Tanaka, R.; Meadows, J.; Phillips, G. O.; Williams, P. A. Carbohydr. Polym. 1990, 12, 443. (12) Tanaka, R.; Meadow, J.; Williams, P. A.; Phillips, G. O. Macromolecules 1992, 25, 1304. (13) Ka¨stner, U.; Hoffmann, H.; Do¨nges, R.; Ehrler, R. Colloids Surf. A: Physicochem. Eng. Aspects 1994, 82, 279. (14) Dualeh, A. J.; Steiner, C. A. Macromolecules 1991, 24, 112. (15) Dualeh, A. J.; Steiner, C. A. Macromolecules 1990, 23, 251. (16) Gelman, R. A. In TAPPI Proceedings of International Pulp Conference, Geneva, 1987, p 159. (17) Turro, N. J.; Yekta, A. J. Am. Chem. Soc. 1978, 100, 5951. (18) Herrington, K. L.; Kaler, E. W.; Miller, D. D.; Zasadzinski, J. A.; Shivkumar, C. J. Phys. Chem. 1993, 97, 13792. (19) Khan, A.; Marques, E. In Specialist Surfactants; Robb, I. D., Ed.; Blackie Academic and Professional, An Imprint of Chapman and Hall: London, 1996. (20) Almgren, M.; Hansson, P.; Mukhtar, E.; van Stam, J. Langmuir 1992, 8, 2405. (21) Infelta, P. P.; Gra¨tzel, M.; Thomas, J. K. J. Phys. Chem. 1974, 78, 190. (22) Tachiya, M. Chem. Phys. Lett. 1975, 33, 289. (23) Evans, D. F.; Wennerstro¨m, H. The Colloidal Domain; VCH Publichers: New York, 1994. (24) Annable, T.; Buscall, R.; Ettelaie, R.; Whittelstone, D. J. Rheol. 1993, 4, 695. (25) Annable, T.; Buscall, R.; Ettelaie, R.; Shepherd, P.; Whittlestone, D. Langmuir 1994, 10, 1060. (26) Piculell, L.; Thuresson, K.; Ericsson, O. Faraday Discuss. 1995, 101, 307. (27) Carlsson, A.; Karlstro¨m, G.; Lindman, B.; Stenberg, O. Colloid Polym. Sci. 1988, 266, 1031. (28) Holmberg, C.; Nilsson, S.; Singh, S. K.; Sundelo¨f, L.-O. J. Phys. Chem. 1992, 96, 871. (29) Kamenka, N.; Burgaud, I.; Zana, R.; Lindman, B. J. Phys. Chem. 1994, 98, 6785. (30) Olofsson, G.; Wang, G. Pure Appl. Chem. 1994, 66, 527. (31) Thuresson, K.; Lindman, B. J. Phys. Chem. B 1997, 101, 6460. (32) Goddard, E. D.; Hannan, R. B. In Micellization, Solubilization, and Microemulsions; Mittal, K. L., Ed.; Plenum Press: New York, 1977; Vol. 2, p 835. (33) Shiloach, A.; Blankschtein, D. Langmuir 1998, 14, 1618. (34) Zana, R.; Binana-Limbele´, W.; Lindman, B. J. Phys. Chem. 1992, 96, 5461. (35) Anthony, O.; Zana, R. Langmuir 1996, 12, 1967. (36) Rose´n, O.; Piculell, L.; Hourdet, D. Langmuir 1998, 14, 777. (37) Murata, M.; Arai, H. J. Colloid Interface Sci. 1974, 46, 475. (38) Lindman, B.; Wennerstro¨m, H. In Topics in Current Chemistry; Boschke, F. L., Ed.; Sprinder-Verlag: Heidelberg, 1980; Vol. 87, p 1. (39) Missel, P. J.; Mazer, N. A.; Benedek, G. B.; Carey, M. C. J. Phys. Chem. 1983, 87, 1264. (40) Mukerjee, P.; Mysels, K. J. CMC of Aqueous Surfactant Systems; NSRDS-NBS 36; U.S. Government Printing Office: Washington, DC, 1971. (41) Kamenka, N.; Chorro, M.; Talmon, Y.; Zana, R. Colloids Surf. 1992, 67, 213. (42) Chorro, M.; Kamenka, N. J. Chim. Phys. 1991, 88, 515. (43) Lucassen-Reynders, E. H.; Lucassen, J.; Giles, D. J. Colloid Interface Sci. 1981, 81, 150. (44) Dubochet, J.; Adrian, M.; Chang, J. J.; Homo, J. C.; Lepault, J.; McDowall, A. W.; Schultz, P. Q. ReV. Biophys. 1988, 21, 129. (45) Bellare, J. R.; Davis, H. T.; Scriven, L. E.; Talmon, Y. J. Electron Microsc. Tech. 1988, 10, 87. (46) Rose´n, O.; Piculell, L. Polym. Gels Networks 1997, 5, 185.