Self-Assembly of Charged Amphiphilic Diblock Copolymers with

Nov 24, 2010 - †Complex Fluids Laboratory, CNRS UMR 166, Rhodia Research Center, 350 George Patterson Boulevard,. Bristol, Pennsylvania 19007 ...
0 downloads 0 Views 3MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Self-Assembly of Charged Amphiphilic Diblock Copolymers with Insoluble Blocks of Decreasing Hydrophobicity: From Kinetically Frozen Colloids to Macrosurfactants ,^

)

M. Jacquin,† P. Muller,†,‡ H. Cottet,§ and O. Theodoly*,†,

)

† Complex Fluids Laboratory, CNRS UMR 166, Rhodia Research Center, 350 George Patterson Boulevard, Bristol, Pennsylvania 19007, United States, ‡Universit e de Strasbourg, Institut Charles Sadron, CNRS UPR 22, 23 rue du Loess, BP 84047, F-67034 Strasbourg Cedex 2, France, §Institut des Biomol ecules Max Mousseron (UMR 5247 CNRS - Universit e de Montpellier 1 - Universit e de Montpellier 2), place Eug ene Bataillon CC 1706, 34095 Montpellier Cedex 5, France, Laboratoire Adh esion et Inflammation, INSERM U600, CNRS UMR 6212, Case 937, 163 Avenue de Luminy, Marseille F-13009, France, and ^Aix/Marseille Universit e, Facult e des Sciences/de M edecine ou de Pharmacie, Marseille F-13000, France

Received August 25, 2010. Revised Manuscript Received October 25, 2010 We have investigated the self-assembly properties in aqueous solution of amphiphilic diblock copolymers with insoluble blocks of different hydrophobicity and demonstrated that the condition to obtain dynamic micelles is to design samples with insoluble blocks of low enough hydrophobicity. We focus here on results with new water-soluble amphiphilic diblock copolymers poly(diethyleneglycol ethylether acrylate)-b-poly(acrylic acid), or PDEGA-b-PAA. The physical characteristics of PDEGA-b-PAA micelles at high ionization have been determined by small angle neutron scattering (SANS). We show that PDEGA-b-PAA samples form micelles at thermodynamic equilibrium. The critical micelle concentrations (CMCs) decrease strongly with ionic strength and temperature due to a solvent quality decrease for, respectively, the corona and the core. This behavior of reversible aggregation is remarkable as compared to the behavior of kinetically frozen aggregation that has been widely observed with samples of similar architecture and different hydrophobic blocks, for example, poly(styrene)-b-poly(acrylic acid), PS-b-PAA, and poly(butyl acrylate)-bpoly(acrylic acid), PBA-b-PAA. We have measured the interfacial tension between water and the homopolymers PDEGA and PBA at, respectively, 3 and 20 mN/m at room temperature, which permits one to estimate the energy cost to extract a unimer from a micelle. The results are consistent with a micelle association that is fast for PDEGA-b-PAA and kinetically frozen PBA-b-PAA. Hence, PDEGA-b-PAA samples form a new system of synthetic charged macrosurfactant with unique properties of fast dynamic association, tunable charge, and water solubility even at temperatures and NaCl concentrations as high as 65 C and 1 M.

1. Introduction Water-soluble amphiphilic diblock copolymers are made of one hydrophobic block and one hydrophilic block connected by a covalent bond. Like low molecular weight surfactants, these macrosurfactants self-assemble in aqueous solution and form micelle-like aggregates. They are interesting for applications, as they may combine properties of low molecular weight surfactant and macromolecules. Typical targeted applications of macrosurfactants are stabilization boosting of suspensions, emulsions, or foams, encapsulation and delivery, surface charge modification, surface wettability enhancement, or antifouling and antisoiling treatments.1,2 As compared to small molecular weight surfactants, macrosurfactants permit in principle to achieve higher stability of selfassembled aggregates, larger loading capacity of active into aggregates, longer range of repulsion between colloidal particles in suspension, and more efficient anchoring on hydrophobic surfaces. Although charged water-soluble amphiphilic block copolymers have proven useful for various applications, it has also been recognized that most macrosurfactant systems are generally characterized by very slow kinetics of spontaneous adsorption *To whom correspondence should be addressed. Tel: þ33 (0)4 91 82 88 69. Fax/Tel: þ33 (0)4 91 82 88 51. E-mail:[email protected]. (1) Amphiphilic Block Copolymers. Self assembly and applications; Alexandridis, P., Lindman, B., Eds.; Elsevier: Amsterdam, 2000. (2) Muller, P.; Sudre, G.; Theodoly, O. Langmuir 2008, 24, 9541–9550.

Langmuir 2010, 26(24), 18681–18693

toward hydrophobic interfaces. In other words, they have a very low surface activity or have even no surface activity at all.3-9 A remarkable exception to this rule is the Pluronics family. The reason is clarified in this paper. Note also that Pluronics are neutral and we focus here on charged polymers. In all applications where a fast adsorption kinetics is required, small-molecule surfactants are today the only solution. In this sense, there is a great need to improve the surface active properties of macrosurfactants in order to open many areas of applications. This challenge has recently received an important new insight.9 A new family of charged amphiphilic diblock copolymers has been synthesized, poly(diethyleneglycol ethylether acrylate)-blockpoly(acrylic acid), or PDEGA-b-PAA, which has a fast and reversible adsorption kinetics. It has been shown that the adsorption dynamics of this macrosurfactant was significantly improved (3) Johner, A.; Joanny, J. F. Macromolecules 1990, 23, 5299–5311. (4) Matsuoka, H.; Maeda, S.; Kaewsaiha, P.; Matsumoto, K. Langmuir 2004, 20, 7412–7421. (5) Amiel, C; Sikka, M; Schneider, J. W.; Tsao, Y. H.; Tirrell, M; Mays, J. W. Macromolecules 1995, 28, 3125–3134. (6) Matsuoka, H.; Matsutani, M.; Mouri, E.; Matsumoto, K Macromolecules 2003, 36, 5321. (7) Kaewsaiha, P.; Matsumoto, K.; Matsuoka, H. Langmuir 2005, 21, 9938– 9945. (8) Bijsterbosch, H. D.; Cohen Stuart, M. A.; Fleer, G. J. Macromolecules 1998, 31, 9281–9294. (9) Theodoly, O.; Jacquin, M.; Muller, P.; Chhun, S. Langmuir 2009, 25, 781– 793.

Published on Web 11/24/2010

DOI: 10.1021/la103391p

18681

Article

due to the fast association dynamics of molecules in solution. These results indicate that the problem of efficient surface activity is narrowly linked to the problem of the dynamics of self-assembly in bulk solution. The present paper focuses on the self-assembly properties of this new PDEGA-b-PAA macrosurfactant. The aim is to demonstrate how the self-assembly dynamics of macrosurfactants in bulk solution can be controlled by the chemical nature of the insoluble block. Many water-soluble amphiphilic block copolymers, for example, poly(styrene)-b-poly(acrylic acid) or PS-b-PAA and poly(tert-butylstyrene)-b-poly(stryrene sulfonate) or PtBS-b-PSS,3,7,5 form frozen aggregates in solution. The high glass transition temperature Tg of the insoluble block of PS or PtBS has often been inferred to explain the irreversibility of the self-assembly in solution. However, recent experiments with samples of poly(n-butyl acrylate)-b-poly(acrylic acid) or PBA-b-PAA, whose insoluble blocks (PBA) have a Tg as low as -55 C, have shown that such samples form also irreversible aggregates in solution.10-12 These results proved that the hydrophobicity of the insoluble block was a major factor favoring a slow association dynamics in solution, as predicted theoretically.3,13,14 Measurements of the interfacial tension between the insoluble block of PBA and the solvent have yielded quantitative evidence that the interfacial tension can induce an energetic barrier to unimer-micelle exchange that is large enough to explain the presence of kinetically frozen aggregates in PBA-b-PAA samples.10 The same argument can also explain the frozen aggregation of other systems such as PS-b-PAA.23 Following this idea, decreasing the interfacial tension between the insoluble core of aggregates and the solvent seems an interesting strategy to achieve a thermodynamically stable exchange between aggregates and unimers in diblock copolymer solutions. Several ways permit to reduce the coresolvent interfacial tension between the insoluble core of polymer aggregates and the solvent. Addition of a nonselective cosolvent is efficient,15-18 but we are interested here in the case of aqueous solutions. Addition of small-molecule surfactants has also been shown to promote reversible aggregation of macrosurfactant aggregates.15,19 However, the properties of mixtures of low molecular weight surfactants and macrosurfactants (for example, the competitive adsorption toward interfaces) are complex and the expected advantages of using a macrosurfactant may be jeopardized by the presence of low molecular weight surfactants. The natural route to control core-solvent interfacial tension consists of synthesizing new diblock with insoluble blocks of more moderate hydrophobicity.20,9,21 This strategy is dependent on synthesis constraints. Lejeune et al.21 used atom transfer (10) Jacquin, M.; Muller, P.; Futterer, T.; Talingting-Pabalan, R.; Berret, J. F.; Cottet, H.; Theodoly, O. J. Colloid Interface Sci. 2007, 316, 897–911. (11) Colombani, O.; Bukhardt, M.; Drechsler, M.; Ruppel, M.; Schumacher, M.; Gradzielski, M.; Schweins, R.; M€uller, A. H. E. J. Phys. Chem. B 2009, 113, 4218–4225. (12) Colombani, O.; Ruppel, M.; Schumacher, M.; Pergusov, D.; Schubert, F.; M€uller, A. H. E. Macromolecules 2007, 316, 897–911. (13) Nose, T.; Iyama, K. Comput. Theor. Polym. Sci. 2000, 10, 249–257. (14) Haliloglu, T.; Bahar, I; Erman, B.; Mattice, W. L. Macromolecules 1996, 29, 4764–4771. (15) Van Stam, J.; Creutz, S.; De Schryver, F. C.; Jerome, R. Macromolecules 2000, 33, 6388–6395. (16) Lund, R.; Willner, L.; Stellbrink, J.; Radulescu, A.; Richter, D. Phys. B 2004, 350, e909–e912. (17) Lund, R.; Willner, L.; Monkenbusch, M.; Panine, P.; Narayanan, T.; Colmenero, J.; Richter, D. Phys. Rew. Lett. 2009, 102, 188301. (18) Cristobal, G.; Berret, J.-F.; Chevallier, C.; Talingting-Pabalan, R.; Joanicot, M.; Grillo, I. Macromolecules 2008, 41, 1872–1880. (19) Jacquin, M.; Muller, P.; G.; Cottet, H.; Crooks, R.; Theodoly, O. Langmuir 2007, 23, 9939–9948. (20) Jacquin, M. These de doctorat, Universite de Paris 6, 2006. (21) Lejeune, E.; Drechsler, M.; Jestin, J.; M€uller, A. H. E.; Chassenieux, C.; Colombani, O. Macromolecules 2010, 43, 2667–2671.

18682 DOI: 10.1021/la103391p

Jacquin et al.

radical polymerization (ATRP) to synthesize diblocks of P(BAs-AA)-b-PAA, where the insoluble block is a statistical copolymer of BA and AA. We have used the macromolecular design via interexchange of xanthate MADIX (which allows a greater choice of monomer than ATRP) to synthesize new diblock copolymers of PDEGA-b-PAA. As shown in this paper, the insoluble block of PDEGA has an interfacial tension with water of only a few mN/m at room temperature and a lower critical solubility temperature (LCST) of 4 C. The moderate and tunable hydrophobicity of PDEGA makes it a good candidate to orient the association properties of PDEGA-b-PAA toward fast and thermodynamically stable dynamics. Herein, we compare the self-assembly properties in aqueous solution of a series of amphiphilic diblock copolymers PS-b-PAA, PBA-b-PAA, P(BA-s-DEGA)-b-PAA, and PDEGA-b-PAA. These diblocks differ only by the chemical nature of the insoluble block, which is less and less hydrophobic from PS, PBA, PBA-sPDEGA, to PDEGA. The self-assembly properties of PS-b-PAA and PBA-b-PAA, which form kinetically frozen micelles in solution, have been detailed elsewhere,22,23 and the results presented here focus mainly on PDEGA-b-PAA samples. We first discuss the synthesis conditions and the purification of the products. We then investigate the specific properties of h-PDEGA homopolymer versus water and h-PAA homopolymer via the building of the ternary phase diagram and the determination of the interfacial tensions and Flory parameters. We then present the study of the aggregation properties in melt and aqueous solutions by X-ray and neutron scattering and discuss the results in comparison with the systems of PS-b-PAA and PBA-b-PAA.

2. Experimental Section Materials. Acrylic acid (AA, Aldrich, 99%), n-butyl acrylate (BA, Aldrich, 99%), di(ethyleneglycol) ethylether acrylate (DEGA, Aldrich, 90%), and deuterated acrylic acid d4 (Polymer Source, D4AA) were used as received. The controlling agent xanthate Rhodixan A1 (2-mercaptopropionic acid, methyl ester, o-ethyl dithiocarbonate) was provided by Rhodia.24 The initiator was 2,20 azobis(2-methylbutanenitrile) (AMBN, DuPont Vazo 67). Reactions were performed in ethanol (pure grade, Aldrich 459844), which is a good solvent for both PDEGA and PAA, and is usable at industrial scale. Polymerization. All polymers were synthesized via controlled radical polymerization process MADIX. For the first block, a solution of monomers in ethanol at a concentration of 40 wt % was bubbled with ultrapure N2 and heated at 70 C in the presence of xanthate. A shot of initiator was then added to initiate the reaction. A molar ratio of xanthate over initiator of 10, nXa = 10ninitiator was used for all syntheses. The ratio of the monomer mass introduced versus the number of xanthate moles fixed the targeted molar mass of the polymers chains to M target ¼

mmonomer mmonomer  nXa þ ninit nXa

ð1Þ

The reaction times were optimized to stop reaction as soon as the residual amount of monomers was lower than 1% of the initial amount. In the case of statistical copolymer synthesis of BA and DEGA, it was checked by GC-Head space that the consumption rate was comparable for both monomers. After completion of the first block growth, the second monomer was added as a 40 wt % (22) Bendejacq, D.; Ponsinet, V.; Joanicot, M. Langmuir 2005, 21, 1712–1718. Bendejacq, D.; Joanicot, M.; Ponsinet, V. Eur. Phys. J. E 2005, 17, 83–92. (23) Jacquin, M.; Muller, P.; Futterer, T.; Talingting-Pabalan, R.; Berret, J. F.; Cottet, H.; Theodoly, O. J. Colloid Interface Sci. 2007, 316, 897–911. (24) Jacquin, M.; Muller, P.; Lizaragua, G.; Bauer, C.; Cottet, H.; Theodoly, O. Macromolecules 2007, 40, 2672–2682.

Langmuir 2010, 26(24), 18681–18693

Jacquin et al.

Article

Table 1. Chemical Characteristics of P(BA-s-DEGA)-b-PAA Samples 3k-12ka sample

Mw1

ΦBA

Ip1

(3k-s-0k)-12k 5400 1.5 1 (2.25k-s-0.75k)-12k 3840 2.2 0.83 (1.5k-s-1.5k)-12k 4300 1.74 0.61 (0.75k-s-2.25k)-12k 4100 1.76 0.35 (0k-s-3k)-12k 4000 1.98 0 a Mw1 and Ip1 are the molecular weight and the index of polydispersity determined by GPC of the first block of BA-s-DEGA, and ΦBA is the molar fraction of BA in the first block determined by NMR.

Table 2. Chemical Characteristics of Polymer Samplesa sample

Mw1

Ip1

XNMR

XLC-PEAT h-PDEGA

XCE h-PAA

6k-24k 6830 3.5 5 65 C) and high ionic strength (NaCl 1 M). They are made of nontoxic and biocompatible chemical constituents and are efficient suspension stabilizing agents.52 This set of characteristics makes PDEGA-b-PAA samples potentially interesting for many surface modification applications.9

We have studied the reversible micellization of a novel series of diblock copolymer samples poly(diethyleneglycol ethylether

Acknowledgment. We thank Rhodia, Inc. for financial and technical support, and particularly M. Morvan, T. Fuetterer, S.

(50) Astafieva, I.; Fu Zhong, X.; Eisenberg, A. Macromolecules 1993, 26, 7339. (51) Cottet, H.; Gareil, P.; Guenoun, P.; Muller, M.; Delsanti, M.; Lixon, P.; Mays, J. W.; Yang, J. J. Chromatogr., A 2001, 939, 109.

(52) Valignat, M.-P.; Theodoly, O.; Crocker, J. C.; Russel, W. B.; Chaikin, P. M. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 4225–4229.

18692 DOI: 10.1021/la103391p

Langmuir 2010, 26(24), 18681–18693

Jacquin et al.

Article

Chhun, H. Mauermann, D. Radtke, and D. Bendejacq. We are also grateful to L. Porcar and S. Kline from NIST, USA, to F. Cousin from LLB, France, for help with SANS experiments, and to M. Rawiso from ICS, Strasbourg, France, for helpful discussions on the SANS scattering data.

Appendix In this work, we have tested several fitting procedures to analyze the SANS scattering data of micelle solutions. For polymer micelles with large core size and diluted corona, the scattering intensity in the q range 0.003-0.1 A˚ -1 is often dominated by the core scattering25,23 and the data can be fitted by a form factor of spheres: PðqÞ ¼



9 ðqRÞ

6

2 sinðqRÞ - qR cosðqRÞ

ð7Þ

where R is the core radius. To fit precisely the core-corona structure, the adapted models depend on the state of the corona. The urchin-like model35,36 is adapted when the corona chains are stretched, whereas the model proposed by Pedersen and Gerstenberg26,27 (PG model) permits to fit data of monodisperse micelles surrounded by a corona of Gaussian chains: PðqÞ ¼ N 2 βs 2 F s ðq, RÞ þ Nβc 2 F c ðq, Rg Þ þ NðN - 1Þ βc 2 S cc ðq, Rg , RÞ þ 2N 2 βc βs S sc ðq, Rg , RÞ ð8Þ where the subscripts s and c refer to uniform spherical micelle cores and attached Gaussian chains, R is the core radius, Rg = (LlK/6)0.5 is the radius of gyration of the Gaussian chains of the hydrophilic blocks, L is the contour length, lK is the Kuhn length, N is the micelle association number, βx= ΔFxVx is the total excess scattering length of a chain in the core (x = s) or in the corona (x = c), ΔFx is the excess scattering length density, Vx is the volume of the block, Fs is the normalized self-correlation term for a uniform sphere, Fc is the self-correlation term for a Gaussian chain, Ssc corresponds to the interference cross term between the sphere and the Gaussian chain starting at the surface of the sphere, and Scc is the interference term between different Gaussian chains attached to the surface of the sphere. This formula corresponds to eq 10 from ref 27, for the special case where the distance from the core surface to the center of the Gaussian chains is equal to Rg, that is, parameter R of ref 27 is equal to 1. Since polymeric micelles are known to be polydisperse, we have extended the PG model to polydisperse systems. For both the PG model and the polydisperse spheres model, we consider a Gaussian distribution in number of core radii. We call Rc the average core radius and ΔRc the standard deviation. The contribution P(q) of a population of micelles of core size R is calculated using eq 7 and eq 8. The total scattering is calculated by addition of Pi(q) for all sizes with a weight proportional to the number density n(R) of an object of radius R. R I poly ðqÞ ¼

R PðR, qÞ

nðRÞ dR with nðRÞ ¼ ntotal V total

exp

- ðR - Rc Þ2 2ΔRc 2 pffiffiffiffiffiffi 2πΔRc

!

ð9Þ (53) Willner, L.; Poppe, A.; Allgaier, J.; Monkenbusch, M.; Richter, D. Europhys. Lett. 2001, 667–773. (54) Dong, Y.; Sundberg, D. C. J. Colloid Interface Sci. 2003, 258, 97–101. (55) Rager, T.; Meyer, W. H.; Wegner, G. Macromolecules 1997, 30, 4911–4919. (56) Rager, T.; Meyer, W. H.; Wegner, G. Macromol. Chem. Phys. 1999, 200, 1672–1680.

Langmuir 2010, 26(24), 18681–18693

The aggregation number of a micelle of radius R can be deduced from the ratio of the volume of the core to the volume of the PDEGA block Vc when the core is made of pure PDEGA. The most noticeable difference between the three main models described so far is the behavior at high q: while the polydisperse sphere model has a q-4 dependence in the high q range, the urchin-like model has a q-1 dependence and the Pedersen-Gerstenberg a q-2 dependence, originating from the scattering of individual hydrophilic blocks. Since, in our case, the micelle cores of PDEGA can incorporate a significant amount of water, we added the fraction XcDEGA of DEGA in the core (the rest being water) as an additional parameter of the fitting functions. The effective volume of a PDEGA block is multiplied by 1/XcDEGA, the scattering length density is multiplied by XcDEGA, and the aggregation number N is given by 4πR3 with 3V c ðX cDEGA Þ V DEGA V c ðX cDEGA Þ ¼ and X cDEGA ΔFc ðX cDEGA Þ ¼ ΔFDEGA X cDEGA

NðR, X cDEGA Þ ¼

ð10Þ

Note that, for a given aggregation number N, the total scattering of the core ΔFcVc is independent of XcDEGA. We impose in all fits that the distribution (Rc, ΔRc) is consistent with the experimental volume fraction ΦDEGA of hydrophobic diblock, that is R 4π 3 R X cDEGA nðRÞdR V DEGA ¼ R 3 φDEGA ¼ V total V total   ntotal 4πX cDEGA 3 Rc þ 3Rc ΔRc 2 ¼ V total 3  FDEGA

φmass diblock M DEGAblock 100M diblock

ð11Þ

where φmass diblock is the mass fraction in wt % of the polymer. The last equality in eq 12 is based on the approximation that the mean densities of the water and PAA blocks remain close to the density of pure water (1 g/cm3). As we consider the possibility of the presence of water inside the core of the micelles, a natural consequence is to consider that such swollen cores present local fluctuations of composition. A standard Ornstein-Zernike approach is used to describe such fluctuations: PðqÞ ¼

Aoz Afit ¼ ðχ - χC Þ þ Boz q2 1 þ Bfit q2

ð12Þ

where χ and χc are the Flory and critical Flory parameters, and Aoz and Boz are parameters that depend on solution characteristics. Equation 12 shows that the scattering term due to fluctuations can become large as we are close to the LCST of PDEGA polymer. As we have no access to χ and χc, we use the two parameters Afit and Bfit of eq 12 to fit the contribution of concentration fluctuations. Supporting Information Available: Plots of SANS data of a PDEGA-b-d3PAA 8k-8k solution and of the best fits obtained by the Pedersen-Gerstenberg model with polydisperse micelles and without adding the core fluctuation term. There is clearly a deficit of scattering at high q values. This material is available free of charge via the Internet at http://pubs.acs. org. DOI: 10.1021/la103391p

18693