Adsorption of water-soluble polymers with surfactant character.

Journal of Colloid and Interface Science 307 (2007) 398–404 www.elsevier.com/locate/jcis

Adsorption of water-soluble polymers with surfactant character. Adsorption kinetics and equilibrium properties Ana M. Díez-Pascual a , Aurora Compostizo a , Amalia Crespo-Colín a , Ramón G. Rubio a,∗ , Reinhard Miller b a Departamento de Química Física I, Facultad de Química, Universidad Complutense, 28040 Madrid, Spain b Max-Planck Institut für Kolloid- und Grenzflächenforschung, Forschungscampus Golm, D14476 Golm, Germany

Received 4 October 2006; accepted 30 November 2006 Available online 8 December 2006

Abstract A comparative study between Langmuir and Gibbs monolayers of a hyperbranched polyol, poly(propylene glycol) homopolymers, and poly(propylene glycol)–poly(ethylene glycol) copolymers with different structure and molecular weight, is reported. Dynamic surface tension (DST) and surface pressure measurements have been carried out to characterize these amphiphilic water-soluble polymers. The adsorption kinetics results are consistent with a rapid diffusion stage followed by a slow reorganization at the air–water interface. The characteristic times of these steps, calculated by the Joos model, point out differences among the polymers in the diffusion rate and rearrangement mechanisms for diluted solutions. Short time analysis of DST data leads to diffusion coefficients in qualitative agreement with the diffusion times calculated with Joos’ model. Spread monolayers remain stable for long periods of time. The desorption process seems quite inoperative. As a consequence, the surface pressure of the spread monolayers can be studied over a broad surface concentration range. 2D first-order phase transitions have been evidenced from plateaux observed in Langmuir and Gibbs isotherms. It has been found that Gibbs monolayers lead to lower surface tension states than the Langmuir ones. © 2006 Elsevier Inc. All rights reserved. Keywords: Adsorption; Surfactant polymer; Air–water interface; Monolayer; Diffusion; Phase transition; Kinetics; Isotherms

1. Introduction In the last decades the need to build up nanomaterials with controlled shape and size [1–3], has increased the interest in understanding and predicting the structure and properties of the monolayers, specially on the polymeric materials [4–6]. It is well known that soluble polymers with hydrophilic groups are able to form stable monolayers at the air–water interface [7,8], either by spreading the polymer on the interface (Langmuir monolayers) [9], or by adsorption of the polymer from the bulk (Gibbs monolayers) [10]. In fact, in Langmuir monolayers it is possible to increase the surface concentration at a rate higher than the one required by the polymer molecules to diffuse into the bulk and dissolve. In this case the corresponding * Corresponding author.

E-mail address: [email protected] (R.G. Rubio). 0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.11.056

spread monolayer will behave as if the polymer were insoluble [11–15]. The fact that spread monolayers remain insoluble for long periods of time suggests a high asymmetry in the adsorption– desorption kinetics. The formation of entangled “gel-like” monolayers has been proposed as the possible source of the almost irreversible adsorption of polymers and proteins at the liquid–air surface. However, the detailed mechanism remains still unknown, and some experimental results indicate that the real picture might not be that simple. On one hand, the washingout experiments of Svitova and Radke point out that Gibbs monolayers of Pluronic F-68 [a symmetric triblock copolymer of PEG-b-PPG-b-PEG, PEG being poly(ethylene glycol), and PPG being poly(propylene glycol)] show partial solubility at high concentrations [12]. Also the oscillatory barrier experiments of Muñoz et al. [16] suggest that, when subject to low-frequency mechanical perturbations, the diluted Gibbs monolayers of Pluronic F-68 show an adsorption–desorption

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

kinetics that is compatible with the Lucassen–van den Tempel model [17]. This solubilization process is more evident for concentrated Langmuir monolayers of the same copolymer. It must be stressed that most of the theoretical models aiming at describing the adsorption/desorption kinetics of polymers at fluid surfaces assume that it is an equilibrium reversible process. From a technological point of view, this phenomenon could help to stabilize highly diluted emulsions and colloidal dispersions, which would be of a great interest in food technology, pharmacy and paint industry. The origin of such an extra stability would be related to the mechanical properties (elasticity and viscosity) of the polymer monolayers that would help to prevent the coalescence process. From the mechanical point of view, both Langmuir and Gibbs monolayers were found to be equivalent for diluted states (low surface pressures, Π , or high surface tensions, γ ) for monolayers of Pluronic F68 [11]. However, the situation seems to be more complex for concentrated states of the monolayers [14]. In effect, the surface elasticity of the spread films of Pluronic F-68 tends to a maximum at high surface concentrations, Γ , and probably dissolve in water at values of Π  19 mN/m. Nevertheless, the surface elasticity of adsorbed films decreases again after the maximum, thus suggesting the formation of a loose surface structure [14]. A majority of the papers studying the apparent irreversible adsorption of synthetic polymers focus on the behaviour of Pluronic F-68. This introduces a further complication because of the existence of plateaux [regions in which Π(Γ ) is almost constant]. These plateaux have been frequently related to surface phase transitions [16–21]. Any structural change of the surface film may modify the kinetics of the adsorption/desorption mechanism. Furthermore, it would change the mechanical properties of the interface [14,22,23], thus affecting its influence in emulsion stabilization. In the case of Pluronic F-68 a simple model of structural changes was suggested in order to explain the surface phase transitions [11,12]. Such a model strongly relies on the different hydrophobicity of the central and the lateral blocks of the polymer chain. Thus it seems reasonable to carry out similar studies on soluble polymers with different molecular structures but similar chemical nature. In this paper we will study the adsorption/desorption kinetics of PPG homopolymers and PPG–PEG copolymers with different molecular weights and geometries. The results will be compared to those obtained for a hyperbranched polyol. In a subsequent paper, the mechanical properties of the monolayers of these polymers at the air–liquid surface will be reported.

399

Table 1 Chemical characteristics of the two random copolymers Copolymer

%PEG

Ni a

Mw (g/mol)

d (g/cm3 )

PPG-r-PEG (I) PPG-r-PEG (II)

75 75

189 39

12000 2500

1.092 1.095

a N : average polymerization degree. i

acteristics of these copolymers are summarized in Table 1. A hyperbranched polyol, named generation 2, with the structure [O[CH2 C(C2 H5 )(CH2 O–)2 ]2 A4 B8 ], where A = [COC(CH3 )– (CH2 O–)2 ] and B = [COC(CH3 )(CH2 OH)2 ]; Mw = 1750 g/ mol, purchased from Sigma-Aldrich (USA). Double distilled and deionised water by Milli-Q system (resistivity greater than 18.0 M cm), was used to prepare the aqueous solutions, and the subphase in the case of the spread monolayers. Surface tension of the aqueous polymer solutions were measured by the Pt-Wilhelmy plate method [24–28], with a commercial Krüss K10 digital tensiometer. Successive readings of the instantaneous surface tension γ (t) were taken at 1 s intervals, after a fresh surface was formed by aspiration of the air–solution interface. The glass measuring cell was designed to minimize the solvent evaporation during the experiments. The experimental accuracy of the measurements was ±0.01 mN/m, and the reproducibility of the results was better than ±0.03 mN/m. Surface tension experiments were also carried out using a home-made bubble tensiometer. The cell was tightly sealed to avoid any evaporation, thus allowing us very long (six days) stabilization periods. Blank experiments with pure water were performed to check that no leaks existed in the set up. Surface pressure/area (Π/A) isotherms were obtained using a KSV Minitrough surface film balance. The entire system was covered with a metacrylate box in order to prevent environmental pollution. Monolayers were obtained by spreading small amounts of polymer solutions in pure chloroform, on a fresh aqueous subphase contained in a teflon trough. Temperature was kept constant at 298.15 ± 0.01 K in the three techniques by circulating thermostated water from an external bath. The samples were prepared by weight in an analytical balance of precision ±0.01 mg; very low concentration solutions were prepared by dilution. 3. Results and discussion

2. Experimental

3.1. Dynamic surface tension γ (t)

Different polymers of the highest purity available have been characterized in this study: Three linear homopolymers poly(propylene glycol), PPG, H[OCH(CH3 )CH2 ]n OH, with molecular weights Mw = 400, 2000 and 4000 g/mol, respectively, supplied by Polysciences (Germany). Two random copolymers, poly(ethylene glycol– propylene glycol), with the structure H(OCH2 CH2 )x [OCH2 – CH(CH3 )]y OH, from Sigma-Aldrich (USA). The main char-

To analyze the mechanisms for polymer adsorption at the air–water interface, we have measured time-resolved surface tension (DST) for the indicated polymers. Fig. 1 shows the γ (t) curves for PPG (Mw = 4000 g/mol) at different polymer bulk concentrations. Similar results were obtained for other polymers (see supporting material) The DST experiments show the bimodal character of adsorption process. As it will be discussed below, there is a fast diffusive step towards the interface,

400

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

Fig. 1. DST experiments at T = 298.15 K for adsorbed monolayers of PPG (Mw = 4000 g/mol). Bulk concentrations c [mM]: (Q) 3.67 × 10−8 ; (!) 4.01 × 10−7 ; (a) 4.01 × 10−6 ; (+) 1.00 × 10−4 ; (P) 2.56 × 10−4 ; (2) 5.04 × 10−4 ; (E) 2.00 × 10−3 ; (") 4.01 × 10−3 ; (F) 4.02 × 10−2 ; (e) 8.02 × 10−1 ; (×) 4.01 × 10−0 ; () 3.52 × 101 ; (1) 1.04 × 102 .

followed by a slow molecular reorganization process at the surface. Except for the most diluted concentrations, the initial values are lower than the water surface tension, due to a first diffusion step in the adsorption kinetics which is too fast to be detected by the plate technique. It has to be noted that the time needed to reach the equilibrium surface tension γeq ranges from a few minutes for the more concentrated samples, to more than two days for the more diluted ones. It must be stressed that even for the long-time runs shown in Fig. 1 (up to 55 h) the final equilibrium state was not reached yet. Similar conclusions have been previously reported in the literature [11,29,30]. For a given family of polymers, the time required to achieve γeq increased with the polymer molecular weight. For a given adsorption time, the hyperbranched polyol shows stronger surface tension decrease than the linear polymers with similar molecular weight, which reveals its higher surfactant character. 3.2. Surface pressure/area isotherms Fig. 2 shows Π –A isotherms for spreading monolayers at 298.15 K for the polymers studied in this work. Three different regions can be observed in the curves of the copolymers. Similar shapes were reported in the symmetric triblock copolymer Pluronic F-68 [11,16], for which the pseudo-plateaux were interpreted as surface phase transitions. It was considered that in the high area region (Π < 10 mN/m) the polymer chains remain adsorbed at the interface, forming a film of thickness similar to the size of the monomer. As the area per monomer is decreased, part of the chains desorb into the subphase. This process takes place at almost constant Π (adsorbed chain to mushroom phase transition). The second pseudo-plateau was interpreted as the mushroom-to-brush phase transition, where long tails protrude into the subphase while they remain anchored to the surface by the more hydrophobic PPG block. The

Fig. 2. Langmuir isotherms Π–A at T = 298.15 K for the indicated polymers.

interpretation of the shape of the isotherms seems to be more difficult for the present systems because no hydrophobic block exists. The ellipsometric results for PPG-r-PEG Mw = 12000 (not shown) point out that, as in the case of F68 [16], there is a sharp change of thickness at Π ≈ 18 mN/m from ca. 0.3 to ca. 1.0 nm. The random character of the copolymer makes it hard to assign the plateau and the thickness increase to the formation of a brush phase. In the case of the PPG’s only a pseudo-plateau is observed. This again is consistent with the results reported in Ref. [16] where the increase of the PPG/PEG ratio in two Pluronics also lead to the supression of the Π ≈ 10 mN/m pseudo-plateau. In the present polymers the ellipsometric thickness shows a smooth increase with surface concentration, and the pseudo-plateaux take place at lower values of Π , and at higher surface concentrations, as the molecular weight of the polymer is decreased. The non-existence of pseudo-plateau in the hyperbranched polymer seems to support the idea that those transitions are linked to changes of the shape of the polymer molecules. Finally, at high concentrations (area per monomer smaller than 10 Å2 ), the isotherms reach a maximum value of surface pressure, and the polymer starts to solubilize. This saturation value increases with the molecular weight, and for a given value of Mw it is higher for PPG than for the copolymers. This is consistent with the higher hydrophobicity of PPG with respect to PEG. As it will be discussed below, the adsorption kinetics results suggest the existence of a rather slow reorganization process of the polymer chains at the interface. A comparison of the isotherms obtained by spreading known amounts of polymer, and by continuous compression of the barriers of the Langmuir trough may help in pointing out the existence of such a process. The continuous compression experiments were carried out at the slowest rate of the barriers of our balance (1 mm/min ⇔ 0.75 cm2 /min). The two types of experiments coincide only for very diluted monolayers, the compression experiments leading to higher values of Π for states above 2 mN/m (see the supporting material). If the motion of the barriers is stopped at a

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

Fig. 3. Oscillatory experiments performed on a Langmuir monolayer of PPG (Mw = 4000 g/mol) for Π = 16.5 mN/m. The amplitude of the area oscillation was 5%, and the frequency was 0.05 s−1 . The symbols are experimental data, and the line is the fit to a sinusoidal function.

state with Π > 10 mN/m during the compression, the system relaxes toward the equilibrium state obtained by the addition method. Similar results were reported by Cammarata et al. [31] for monolayers of insoluble polymers. This suggests that, once the polymer coils contact each other in the monolayer, there is a relaxation process too slow for the polymer chains to complete it at the rate of area change imposed by the barriers. This clearly points out that the continuous compression method leads to non-equilibrium states for most of the concentration range. Similar results were obtained for the other polymers studied. Two types of tests were performed in order to demonstrate the absence of any detectable desorption in the Langmuir monolayers. Stable hysteresis cycles were found at two different barrier rates (1 and 5 mm/min). On the other hand, oscillatory barrier experiments were carried out for Π = 5 and 16.5 mN/m, at two different amplitudes, and for frequencies 10  ω (mHz)  100. Fig. 3 shows a typical result for PPG (Mw = 4500 g/mol). The stability of the system’s response demonstrates that no solubilization takes place. A similar study was reported previously for monolayers of Pluronic F-68 [22]. 3.3. Adsorption kinetics A quantitative description of adsorption kinetics is frequently based on the Ward and Tordai model [17,32,33]. This diffusion-controller adsorption model assumes that the step of transfer from the subsurface to the interface is fast compared to the transport from the bulk to the subsurface. The governing equation is ∂c(z; t)/∂t = D∇ 2 c(z; t)

at z > 0, t > 0,

(1)

assuming z as the diffusion direction perpendicular to the surface. To solve the transport problem, boundary conditions have to be defined: c = cs at t → ∞ and z = 0, being c polymer concentration on the bulk, cs bulk concentration in the subsurface

401

Fig. 4. Diffusion coefficients at 298.15 K for the indicated polymers, obtained from the short-time analysis, Eq. (3) (open symbols), and from surface-rheology experiments (closed symbols). Lines show a slightly increase on D with the surface concentration.

and D the polymer diffusion coefficient in the bulk. The solution of Eq. (1) is known as Ward–Tordai equation: Γ (t) = 2c(Dt/π)

1/2

− 2(D/π)

1/2

t 1/2 cs (t − τ ) dτ 1/2 .

(2)

0

At short times and very low concentrations [34,35], the surface tension can be calculated by: γ (t) ∼ = γ (0) − 2nRT c(Ds t/π)1/2

at t → 0.

(3)

At long times another asymptotic solution can be obtained [34,35]:  1/2 nRT Γ 2 π + at t → ∞, γ (t) ∼ γ (4) = eq c 4Dl t where n = 1 for none ionic surfactants; n = 2 for ionics surfactants. γ0 and γeq , are the solvent and equilibrium surface tensions respectively, and Γ the surface adsorption. The diffusion coefficient Dl could be obtained from the slopes at t −1/2 → 0 if surfaces adsorption data were available; Γ might be estimated assuming ideal behaviour, which is incorrect in aqueous polymer solutions. Nevertheless the diffusion coefficient Ds has been calculated from the slopes of the γ vs t 1/2 plots, using Eq. (3). Some calculated diffusion coefficients are shown in Fig. 4, together with those obtained from the frequency dependence of the elasticity in combination with the Lucassen–van den Tempel theory [17]. These results will be fully discussed in a future publication. The fact that both methods, based on completely independent data, lead to similar diffusion coefficients within the experimental error suggests that the use of Eq. (3) is adequate for the present systems. A slight increase on D is observed when the surface concentration rises. As it could be expected [36], within the same family, polymers with higher molecular weight have lower values of D. The hyperbranched polyol diffuses faster due to its more globular geometry.

402

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

Once the polymer reaches the subsurface, the transfer to the surface obeys a first-order kinetics [37,38]: dΓ (c; t)/dt = Jads (c) − Jdes (c),

(5)

where the adsorption and desorption fluxes can be written as: Jads (c) − Jdes (c) = ka cs (Γ∞ − Γ ) − kd Γ.

(6)

In this mass balance equation, ka and kd are the adsorption and desorption kinetic constants respectively, which accounts for the activation energy of both processes. The diffusion controlled adsorption followed by molecular reorientation at the interface has been theoretically described by Joos [39–41]:   γ (t) ∼ = γeq + (γ0 − γeq − B) exp(−4t/πτD )1/2 + B × exp(−t/τR ),

(7)

where τD and τR are, respectively, characteristic times of the diffusion and reorganization process; τD is related to the diffusion coefficient D by [32]:   1 dΓ 2 τD = (8) . D dc B is an amplitude which determines the last step limiting character over the diffusion step. When B = 0 and τR τD , adsorption can be considered as purely diffusion-controlled. B > 0 indicates the existence of two consecutives stages, in a way that if τR > 10τD , two well differentiated steps are detected in the adsorption kinetics. One must not forget that the Joos equation is only approximate for surfactant solutions, and it might be questionable for polymer solutions. However, Ref. [16] showed that it gave consistent results for F68 solutions. Since the molecular weights and the chemical nature of the present systems is similar to that of F68, one may expect that the Joos equation will perform equally well. Equation (7) was found to fit the experimental γ (t) data within the experimental uncertainty. In all cases studied in this work, B takes positive values pointing out the bimodal adsorption mechanism (see supporting material). B seems to be independent of the polymer’s molecular weight and geometry. B increases as the polymer concentration decreases, which, according to the model, suggests a prevailing diffusion control for the most concentrated solutions. For all polymers studied, τD vanishes at the highest concentrations (see supporting material). This corresponds to a saturated monolayer, where the surface concentration remains constant even though the bulk concentration increases. At these high concentrations, the calculated relaxation times are only a few seconds, and close to the time resolution of the experiments, thus being difficult to distinguish among the different polymers. Besides, under these conditions the overlapping concentration c∗ is exceeded [42,43], so the systems are probably three dimensional intercrossed networks. In the low concentration range, τD increases with Mw for a given family of polymers. When polymers of similar Mw are compared [e.g., PPG (Mw = 2000) and the copolymer of Mw = 2500], similar values of τD are found. However, smaller τD ’s are obtained for the hyperbranched polymers, as expected

Fig. 5. Equilibrium surface tension values calculated from the fits of DST data to Eq. (7). The empty circles correspond to the t = 144 h runs carried out with the bubble tensiometer for two of the polymer solutions.

for a more compact globular structure. These conclusions are qualitatively similar to the ones derived from the γ (t) curves. Once polymer molecules have reached the subsurface by diffusion, the adsorption at the interface may imply a reorganization process (conformational change of the polymer coil, rearrangements of the molecules already adsorbed, etc.). At high concentrations τR shows a constant value, which is independent of Mw and of the macromolecular geometry (see supporting material). For diluted solutions, a strong hyperbolic decay is shown as concentration goes up, which can be explained according to Eq. (6). The global process includes adsorption from the subsurface, were concentration is cs , and possible desorption in equilibrium. It is considered as a first-order kinetic process, with a relaxation time given by τR = 1/(kA cS + kD ); cS should change linearly with c, which justifies the τR behaviour found. As a consequence of the macromolecular structure, three groups can be distinguished in the reorganization dynamics. The hyperbranched polyol shows the fastest reorganization, due its more compact structure with less degrees of freedom. Copolymers are much slower than homopolymers, although all have linear geometry. The differences can be originated by the random distribution of the comonomers, which generates areas with higher PEG concentration, able to enter into the subphase. Therefore, the number of monomers in the subsurface may be larger, difficulting the reorganization of the neighbour comonomers. Besides the polymer reptation (motion along the confining tube) [44–46], molecular rearrangements and chain rotation are needed, so the reorganization process takes longer than for homopolymers. No correlation between τR and the polymer molecular weight has been found. The equilibrium surface tension values, γeq , calculated by fitting the experimental γ (t) curves to Eq. (7), are plotted as a function of concentration (Gibbs isotherms) in Fig. 5. In all the cases the calculated errors are lower than 2%. Because the Joos equation is approximated, and the γeq values are signifi-

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

403

Fig. 6. Comparison between Langmuir and Gibbs isotherms, left and right respectively, at T = 298.15 K, for the indicated systems.

cantly different than the long-time values of γ (t) curves (see Fig. 1), it is important to ensure that the fitted values of γeq correspond to real equilibrium surface tensions. The surface tensions measured with the bubble tensiometer for six-days aged surfaces are included also in Fig. 5 for two of the solutions. The agreement with the values obtained from the fits to Eq. (7) is remarkable, thus supporting the validity of the use of Joos’ equation for these systems. All curves show a strong decrease at very low concentration, which reveal the great surfactant character of these systems. As in the case of Langmuir monolayers (Fig. 2), the curves show different pseudo-plateaux. The influence of Mw on the value of γ of the plateaux, and the relative positions of the PPG polymers and the copolymers is qualitatively the same than the one observed in the Langmuir monolayers. However, some differences are observed in the hyperbranched polymer, where pseudo-plateaux are clearly observed in the Gibbs monolayers but not in the Langmuir ones. Any detailed comparison between the two types of isotherms would only be possible if the values of Γ (c) were available for the adsorption monolayers. The application of Gibbs adsorption equation (with the ideal solution assumption) is questionable for relatively concentrated polymer solutions. It is easy to see that its application to the data shown in Fig. 5 for the PPGs lead to Γ vs log c curves that present two well defined maxima, as it was also the case for Pluronic F68 [11]. In any case, the values of γ corresponding to both types of monolayers allow one to see that they do not correspond to the same thermodynamic states (see Fig. 6). In fact, it seems that the minimum value of γ (maximum value of Π ) that can be achieved in the Langmuir monolayers, where they start to solubilize, correspond to the saturation point of the Gibbs monolayers. These results may be interesting in explaining the washing-out experiments of Svitova and Radke [12], and the role of polymer surfactants as emulsion stabilizers. In effect, for concentrated polymer solutions the surface tension may be as low as 35–40 mN/m, thus stabilizing the emulsion droplets. After dilution γeq will rise. However, we have shown that for thermodynamic states corresponding to γ ≈ 50 mN/m or higher, the surface layer behaves as insoluble (in absence of external mechanical disturbances).

As a consequence the emulsion droplets may remain covered by a polymer surfactant layer that prevents coalescence, even for highly diluted systems. The different regimes seem to have a big influence on the thermodynamic interface state, as well as the adsorption kinetics [17]. The presence of such structures at the interface generates a barrier to the adsorption of new polymer molecules; diffusion from the bulk to the surface is especially difficult when the brush is made. Such adsorption barriers are taken into account by Joos model, which assumes that, at a freshly formed surface, the diffusion stage is very fast, while after a certain surface coverage is reached, an adsorption barrier is generated, so the transfer of molecules between the subsurface and the interface is delayed. 4. Conclusions Dynamic surface tension experiments have revealed the bimodal character of the adsorption process, with two consecutive steps: a rapid diffusion stage followed by a slow reorganization at the interface. The diffusion coefficients, obtained from the short time analysis of DST data, are in good quantitative agreement with the values calculated from the frequency dependence of the elasticity and the Lucassen–van den Tempel theory. The diffusion coefficient decreases as the molecular weight increases. The diffusion and reorganization relaxation times, calculated according to Joos equation, show a remarkable decay for diluted solutions, and reach a constant value at high concentrations that is independ of the geometry or the molecular weight. The concentration dependence of the equilibrium surface tension, obtained from the DST data, shows the existence of two surface phase transitions for all the polymers studied. Langmuir monolayers of the polymers have also been studied. The monolayers are stable over long time periods in spite that the polymers are water soluble. The minimum surface tension that can be obtained in the Langmuir monolayer of a given polymer, before solubilization, is very close to the saturation surface tension that can be obtained for the adsorption monolayers. This

404

A.M. Díez-Pascual et al. / Journal of Colloid and Interface Science 307 (2007) 398–404

may help in understanding the ability of surfactant polymers for stabilizing dilute emulsions. Acknowledgments This work was supported in part by MEC through Grants VEM2003-20574-C03-03/Inter and FIS2006-12281-C02-01. The work of A.M.D-P was supported by a FPU fellowship from MEC. We are grateful to Dr. F. Monroy for helpful discussions. Supporting material The online version of this article contains additional supporting material. Please visit DOI:10.1016/j.jcis.2006.11.056. References [1] D.F. Evans, H. Wennerström, The Colloidal Domain, second ed., Wiley, New York, 1999. [2] R.A.L. Jones, Soft Machines. Nanotechnology and Life, Oxford Univ. Press, Oxford, 2004. [3] G.A. Ozin, A.C. Arsenanlt, Nanochemistry. A Chemical Approach to Nanomaterials, RSC Publishing, Cambridge, 2005. [4] B. Jönsson, B. Lindman, K. Holmberg, B. Kronberg, Surfactants and Polymers in Aqueous Solution, Wiley, Chichester, 1998. [5] K. Esumi, Polymer Interfaces and Emulsions, Marcel Dekker, New York, 1999. [6] R.A.L. Jones, R.W. Richards, Polymers at Surfaces and Interfaces, Cambridge Univ. Press, Cambridge, 1999. [7] D.J. Kuzmenka, S. Granick, Macromolecules 21 (1998) 779; D.J. Kuzmenka, S. Granick, Polym. Commun. 29 (1998) 64. [8] G.J. Fleer, M.A. Cohen-Stuart, J.M.H.M. Schentiens, T. Cosgrove, B. Vincent, Polymers at Interfaces, Chapman & Hall, London, 1993. [9] G.L. Gaines, Insoluble Monolayers at the Liquid–Gas Interface, Wiley, New York, 1966. [10] K. Tajima, K. Muramatsu, T. Sasaki, Bull. Chem. Soc. Jpn. 43 (1975) 1991. [11] M.G. Muñoz, F. Monroy, F. Ortega, R.G. Rubio, D. Langevin, Langmuir 16 (2000) 1083. [12] T.F. Svitova, C.J. Radke, Ind. Eng. Chem. Res. 44 (2005) 1129. [13] T.F. Svitova, M.J. Wetherbee, C.J. Radke, J. Colloid Interface Sci. 261 (2003) 170. [14] B.A. Noskov, S.-Y. Liu, G. Loglio, R.G. Rubio, R. Miller, Langmuir 22 (2006) 2647.

[15] E.M. Freer, K.S. Yim, G.G. Fuller, C.J. Radke, J. Phys. Chem. B 108 (2004) 3835; E.M. Freer, K.S. Yim, G.G. Fuller, C.J. Radke, Langmuir 20 (2004) 10159. [16] M.G. Muñoz, F. Monroy, F. Ortega, R.G. Rubio, D. Langevin, Langmuir 16 (2000) 1094. [17] S.S. Dukhin, G. Kretzschmar, R. Miller, Dynamics of Adsorption of Liquid Interfaces, Elsevier, Amsterdam, 1995. [18] T. Radeva, Adsorption of Polymers at Colloid Surfaces, in: Encyclopedia of Surface and Colloid Science, Marcel Dekker, Santa Barbara, CA, 2002. [19] S.T. Milner, Science 251 (1991) 905. [20] S. Alexander, J. Phys. 38 (1977) 977, 983. [21] B.H. Cao, M.W. Kim, Faraday Discuss. 98 (1994) 245. [22] M.G. Muñoz, F. Monroy, P. Hernández, F. Ortega, R.G. Rubio, D. Langevin, Langmuir 19 (2003) 2147. [23] Ch. Kim, H. Yu, Langmuir 19 (2003) 4460. [24] L. Wilhelmy, Ann. Phys. 119 (1863) 177. [25] D.O. Jordan, J.E. Lane, Austral. J. Chem. 17 (1964) 7. [26] J.T. Davies, E.K. Rideal, Interfacial Phenomena, Academic Press, New York, 1961. [27] I. Langmuir, J. Am. Chem. Soc. 39 (1917) 1848. [28] B.A. Noskov, G. Loglio, R. Miller, J. Phys. Chem. B 108 (2004) 18615. [29] B.A. Noskov, Adv. Colloid Interface Sci. 95 (2002) 237. [30] B.A. Noskov, S.N. Nuzhnov, G. Loglio, R. Miller, Macromolecules 37 (2004) 2519. [31] V. Cammarata, L. Atanasoska, L.L. Millar, C.J. Kolaskie, B.J. Stallman, Langmuir 8 (1992) 876. [32] V.B. Fainerman, D. Mobius, R. Miller, Surfactants: Chemistry, Interfacial Properties Applications, Elsevier, Berlin, 2001. [33] A.F. Ward, L. Tordai, J. Chem. Phys. 14 (1946) 453. [34] V.B. Fainerman, A.V. Makievski, R. Miller, Colloids Surf. A 87 (1994) 61. [35] J. Eastoe, J.S. Dalton, Adv. Colloid Interface Sci. 85 (2000) 103. [36] I.N. Levine, Physical Chemistry, fifth ed., McGraw–Hill, New York, 2004. [37] P. Joos, Dynamic Surface Phenomena, VSP, Utrecht, 1999. [38] M. Volmer, Z. Phys. Chem. 115 (1925) 253. [39] G. Serrien, P. Joos, J. Colloid Interface Sci. 139 (1990) 149. [40] G. Serrien, G. Geeraerts, L. Ghosh, P. Joos, Colloids Surf. A 68 (1992) 219. [41] C. Barentin, J.F. Joanny, Langmuir 15 (1999) 1802. [42] P.G. de Gennes, Scaling Concepts in Polymer Physics, Cornell Univ. Press, Ithaca, NY, 1967. [43] R. Koningsveld, H. Berghmans, R. Simha, W.H. Stockmayer, J. Phys. Chem. B 108 (2004) 16168. [44] G.B. McKenna, G. Hadziioannou, P. Lutz, G. Hild, C. Strazielle, C. Straupe, P. Rempp, A.J. Kovacs, Macromolecules 20 (1987) 498. [45] P.G. de Gennes, J. Chem. Phys. 52 (1971) 572. [46] M. Doi, S.F. Edwards, The Theory of Polymer Dynamics, Clarendon Press, Oxford, 1986.