Wormlike Aggregates from a Supramolecular Coordination Polymer

Sep 19, 2007 - The formation of wormlike micelles in mixed systems of a supramolecular coordination polymer Zn-L2EO4 and a diblock copolymer ...
0 downloads 0 Views 514KB Size
11662

J. Phys. Chem. B 2007, 111, 11662-11669

Wormlike Aggregates from a Supramolecular Coordination Polymer and a Diblock Copolymer Yun Yan,† Nicolaas A.M. Besseling,† Arie de Keizer,*,† Markus Drechsler,‡ Remco Fokkink,† and Martien A. Cohen Stuart† Laboratory of Physical Chemistry and Colloid Science, Wageningen UniVersity, Dreijenplein 6, 6703 HB Wageningen, The Netherlands, and Makromolekulare Chemie II, UniVersity of Bayreuth, 95440 Bayreuth, Germany ReceiVed: March 6, 2007; In Final Form: July 11, 2007

The formation of wormlike micelles in mixed systems of a supramolecular coordination polymer Zn-L2EO4 and a diblock copolymer P2MVP41-b-PEO205 is investigated by light scattering and Cryo-TEM. By direct mixing at a stoichiometric charge ratio, the above mixtures proved to be capable of formation of spherical micelles with a radius of about 25 nm (Yan et al. Angew. Chem., Int. Ed.; 2007, 46, 1807-1809). Lately, we find wormlike micelles with a hydrodynamic radius >150 nm in a mixture with excess positive charge, that is, a negative charge fraction f- < 0.5. The transformation between wormlike and spherical micelles can be realized by variation of the mixing ratio through different protocols. Upon addition of negatively charged Zn-L2EO4 to a mixture with excess positively charged P2MVP41-b-PEO205, most of the wormlike micelles are transformed into spherical ones; upon addition of positively charged P2MVP41-b-PEO205 to a mixture of pure spherical micelles, wormlike micelles can be produced again. The effect of sample preparation protocol, sample history, and concentration on this transformation process is systematically reported in this article. A possible mechanism for the formation of wormlike micelles is proposed.

Introduction Nanostructures based on electrostatic interaction, namely, complex coacervate core micelles (C3Ms),1-3 or “block ionomer complexes” (BICs),4 or “polyion complex micelles” (PIC micelles)5 have attracted increasing attention in the field of nanotechnology over the recent decade. This class of micelles is formed from mixed solutions of a polyion-neutral diblock copolymer and an oppositely charged polyelectrolyte (or an oppositely charged polyion-neutral diblock copolymer), as illustrated in Scheme 1. One of the many possible areas of application of C3Ms is their potential for mimicking biological phenomena,6 because most of the non-viral vectors used for gene delivery are made of complexes between negatively charged DNA and oppositely charged objects that can be either amphiphiles7 or homo- or copolymers.8,9 The common feature of all these systems, including C3Ms, is that they result from electrostatic selfassembly. Obviously, if charged compounds, such as proteins and DNA, are used as homo-polyelectrolytes, they can be transported to target places in the form of C3Ms, when mixed with a neutral-oppositely charged diblock copolymer. These charged compounds will stay in the micellar core, which is stabilized by the electroneutral blocks of the diblock copolymer. Recently,10 we demonstrated the formation of a new kind of spherical C3Ms, consisting of a diblock copolymer [poly(2vinyl-N-methyl pyridinium iodide)-b-(ethylene oxide) (P2MVP41b-PEO205), see Scheme 2A] and a coordination polymer consisting of zinc ions connected by ditopic ligands based on terdendate ligand groups.11-13 The bisligand consists of pyridine* Corresponding author. E-mail: [email protected]. † Wageningen University. ‡ University of Bayreuth.

SCHEME 1: Schematic Representation of Formation of a Complex Coacervate Core Micellea

a The core is formed by complexation of the anionic homopolymer (green) and the cationic block (red) of the diblock copolymer. The corona consists of the electroneutral hydrophilic blocks (blue) of the diblock copolymer.

SCHEME 2: Structure of (A) P2MVP41-b-PEO205 and (B) L2EO4

2,6-dicarboxylic acid groups connected at the 4-position of the pyridine ring to a four-ethylene oxide (4-EO) spacer; they are denoted as “L2EO4” (rather than “C4” as in refs 11-13), as illustrated in Scheme 2B. At concentrations lower than 10 mM, the 1:1 coordination polymer between zinc and L2EO4 exists mainly in the form of small rings with a low polymerization

10.1021/jp0718146 CCC: $37.00 © 2007 American Chemical Society Published on Web 09/19/2007

Wormlike Aggregates from Zn-L2EO4 and P2MVP41-b-PEO205

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11663

SCHEME 3: Illustration of the Formation of Complex Coacervate Core Micelles in the Zn-L2EO4/ P2MVP41-bPEO205 Mixed Systema

a For simplicity reasons, only one Zn-L2EO4 chain is illustrated in the core.

degree (oligomers). At higher concentrations, linear chains with a higher polymerization degree (polymers) predominate.11-13 The complexes can be regarded as polyelectrolytes: each coordination center carries two elementary negative charges. Upon mixing with an oppositely charged diblock copolymer, the local concentration of the coordination oligomers is greatly enhanced due to complexation with the cationic polymer. This high local concentration of the coordination compounds leads to an increase of the degree of the polymerization, so that oligomer rings transform into polymers. The length of these coordination polymers thus meets the requirements to form C3Ms with the diblock copolymer, as illustrated in Scheme 3.10 This new class of micelles containing heavy metal ions opens up new avenues to the application of C3Ms. For instance, heavy metal ions can be enriched and transported by this kind of vehicles. Therefore, it is of both fundamental and practical interest to get a better understanding of these systems. The present study aims at demonstrating that not only spherical micelles but also large particles, probably wormlike micelles, are formed in the mixed systems of coordination polymer and diblock copolymer provided positive charges are in excess. Wormlike aggregates are reported for the first time in the family of complex coacervate core micelles, or block ionomer complexes, or polyion complex micelles systems. In addition, the formation of spherical micelles and wormlike ones can be controlled simply by varying the mixing ratio.

Figure 1. Variation of scattered light intensity (open symbols in (a)) and mean hydrodynamic radius (filled symbols in (b)) of the particles in the titration of various P2MVP41-b-PEO205 solutions with [-] ) 4.26 mM Zn-L2EO4. The light intensities in (a) do not represent the real scaling for the sake of clearance.

scattering cell, and the temperature was controlled within (0.5 °C using a Haake C35-F8 thermostat. Titrations were carried out using a Schott-Gera¨te computer-controlled titration setup to control sequential addition of titrant and cell stirring. The concentrations of both the titrant and titrated solutions are expressed in terms of charge concentration, micromoles per liter. The charge concentration of the titrant is always about 5 times higher than that of the titrated solution. After every dosage, the solution is automatically stirred with a stirring bar for 60 s, during which no LS data are taken. Then after a waiting time of 30 s, the LS measurement is started. After each titration step, the laser-light-scattering intensity (I) at 90° and the intensity auto correlation function were recorded. The data were analyzed by CUMULANT fitting16 yielding a first moment Γ h (decaying time) and a second moment µ2. From the first moment Γ h , the average diffusion coefficient is obtained:

D h )Γ h /q2

(1)

where q is the scattering vector: Experimental Materials. The diblock copolymer used in this study is poly(2-vinyl-N-methyl pyridinium iodide)-b-poly(ethylene oxide) (P2MVP41-b-PEO205), which is quaternized from poly(2-vinylpyridine)-b-poly(ethylene oxide) (P2VP41-b-PEO205) (Polymer Source, Mw/Mn ) 1.03, Mw ) 13.3 K), following the procedure described elsewhere.14 The degree of quaternization is higher than 90%.15 The coordination complex between zinc(II) and 1,11-bis(2,6dicarboxypyridin-4-yloxy)-3,6,9-trioxaundecane bisligand molecules (Zn-L2EO4) was prepared according to literature.11 Piperazinebis(ethanesulfonic acid) (PIPES, 98%) was obtained from Aldrich. Sodium hydroxide (NaOH) and zinc nitrate (Zn(NO3)2‚6H2O) were analytical grade. Bidistilled water was used throughout the experiments. All solutions were prepared in 20 mM PIPES-NaOH buffer at pH 5.4. Methods. Light-Scattering Titration. Light-scattering (LS) measurements were performed with an ALV laser-light-scattering apparatus, equipped with a 400 mW argon ion laser operating at a wavelength of 514.5 nm. A refractive index matching bath of filtered cis-decalin surrounded the cylindrical

q ) 4πn sin(θ/2)/λ

(2)

where n is the refractive index of the solvent, λ the laser wavelength, and θ the angle of detection. Using the StokesEinstein equation, the average hydrodynamic radius is obtained: where k is the Boltzman constant, T the absolute temperature, and η the viscosity of the solvent.

Rh ) kT/6πηD h

(3)

The polydispersity index (PDI) is defined as

PDI ) µ2/Γ h2

(4)

Variations in Rh and I are studied as a function of the mole fraction of negative charge, f-. For a system with positive and negative chain molecules, f- is defined as

f- )

[-] [-] + [+]

(5)

11664 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Yan et al.

where [-] and [+] are the molar concentrations of charges for each of the chains. In order to correct for dilution, we report a normalized intensity I for all experiments obtained from the measured light intensity Im in the following way:

I ) Im(V0 + Vt)/V0

(6)

where V0 and Vt are the initial and titrant volumes, respectively. By means of the CONTIN method,17,18 the distribution of Rh can be obtained. This method yields a distribution G(Γ) from any field correlation function that can be described as n

g1(t) )

G(Γn)e-Γ t ∑ τ)ι n

(7)

Analogous to eqs 1 and 3, each relaxation rate Γn can be related to a diffusion coefficient:

Dn ) Γn/q2

(8)

each with an associated hydrodynamic radius. For angular-dependent DLS, ten correlation functions g2(t) were recorded at 7 angles θ, from 45 to 135° in increments of 15°, to evaluate the angular dependence of the diffusion coefficient. Depolarized Light Scattering. The depolarization ratio R ) I|/I⊥ was measured with the same light-scattering instrument as above. Linearly polarized argon laser light with the laser source being polarized perpendicular with respect to the scattering plane was scattered at 90°. The only modification was that we placed a polarizer before the photomultiplier detector. By changing the polarizer orientation from perpendicular (⊥) to parallel (|), we can measure the depolarization by the particles in the solution. Cryogenic Transmission Electronic Microscopy (Cryo-TEM). A few microliters of sample were placed on a bare copper TEM grid (Plano, 600 mesh), and the excess liquid was removed with filter paper. This sample was cryo-fixed by rapidly immersing it into liquid ethane cooled to -170 to -180 °C in a cryo-box (Carl Zeiss NTS GmbH). The specimen was inserted into a cryotransfer holder (CT3500, Gatan, Munich, Germany) and transferred to a Zeiss electron microscope EM922 EFTEM (Zeiss NTS GmbH, Oberkochen, Germany). Examinations were carried out at temperatures around -180 °C. The TEM was operated at an acceleration voltage of 200 kV. Zero-loss filtered images were taken under reduced dose conditions (500-2000 e/nm2). All images were recorded digitally by a bottom-mounted CCD camera system (UltraScan 1000, Gatan) and processed with a digital imaging processing system (Digital Micrograph 3.9 for GMS 1.4, Gatan). Results and Discussion Titration of P2MVP41-b-PEO205 with L2EO4. In Figure 1, we show three LS titration curves obtained by adding a Zn-L2EO4 solution to a P2MVP41-b-PEO205 solution. In all the systems, we observed an increase of scattered light intensity with the addition of Zn-L2EO4 (increasing f-), which indicates the formation of colloidal particles upon mixing the two agents. The maxima of the scattered light intensity show up at f- ≈ 0.5 as expected, because in a typical LS titration experiment for a C3M system2 a sharp peak of scattered light intensity I always shows up at f- ≈ 0.5, where [-] ≈ [+]. In addition, two remarkable features are also seen in Figure 1. First, the light intensity levels off beyond f- ) 0.5 instead of decreasing steeply, as is usually found for C3Ms formed with ordinary

Figure 2. Comparison of CONTIN distribution of particles with variation of the negative charge fraction f-. The numbers following each symbol are the different f- in titration of [+] ) 0.86 mM P2MVP41-b-PEO205 with [-] ) 4.26 mM Zn-L2EO4.

(covalent) polyelectrolytes. This means that the excess Zn-L2EO4 coordination polymer has almost no effect on the formed particles. We can easily understand that an excess of Zn-L2EO4 coordination polymer does not destroy the C3Ms as conventional covalent homopolymers do;1-3 this is because free Zn-L2EO4 in our experimental conditions exists in the form of oligomers, which can be regarded as indifferent small ions rather than polyions,11 and therefore are not capable of forming soluble complex particles as other covalent polyelectrolytes do in solution.2 Second, it is remarkable that before the light intensity reaches a maximum, a small shoulder is often seen at f- ≈ 0.3-0.4, the size of which depends on experimental conditions. It is not easy to get a reproducible shoulder, even in carefully repeated experiments; this can be seen in Figure 1a, where we present data for two parallel experiments for a 0.88 mM P2MVP41-b-PEO205 solution to which we added the same solution of Zn-L2EO4 coordination polymer. In one experiment the shoulder is distinctly weaker than in the other. Correspondingly, the CUMULANT mean hydrodynamic radius reveals that bigger particles appear before the system completely reaches a constant particle size beyond f- g 0.5. We can read from Figure 1b that the particles initially formed in all the three mixed systems are rather big, but the average size decreases sharply with increasing f- in the range of f- ) 0.25-0.5. For all systems, the mean hydrodynamic radii of the particles become constant after f- g 0.5. The presence of the large aggregates was further analyzed with the CONTIN method. In Figure 2, we show the comparison of the CONTIN results at different f- in the titration of [+] ) 0.86 mM P2MVP41-b-PEO205 with [-] ) 4.26 mM Zn-L2EO4 (Figure 1, squares). We can see clearly that, at f- ) 0.32, there are two families of particles, one representing smaller ones of Rh ≈ 25 nm, and the other comprising bigger ones of Rh ≈ 150 nm. At f- ) 0.32, the height of the peak from the larger particles is about 80% of that of the smaller ones. By increasing f- to 0.36, the relative intensity of the bigger ones decreases to 40% of that of the smaller ones. At f- ) 0.49, the relative amount of the bigger ones is further lowered to 20%. At the same time, the average size of these large particles decreases slightly. It is obvious that the relative amount of the large objects is decreasing with f- approaching 0.5. But even at f- ) 0.86, there are still some residuals of them with even larger radii. This means that the large particles disintegrate slowly, which indicates that there is probably a time effect when they are forming. Therefore we decided to pay attention to the timedependent light intensity during the titration process. In Figure 3, we show how the scattered light intensity varies with time during the step-by-step addition of a Zn-L2EO4 solution ([-] ) 4.12 mM) to a P2MVP41-b-PEO205 solution ([+] ) 0.85 mM). After each addition the mixture was stirred for

Wormlike Aggregates from Zn-L2EO4 and P2MVP41-b-PEO205

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11665

Figure 4. Variation of the size distribution of the particles with time during the titration of [+] ) 0.85 mM P2MVP41-b-PEO205 solution with [-] ) 4.12 mM Zn-L2EO4. Data correspond to f- ) 0.33.

Figure 3. Variation of the scattered light intensity with time after subsequent additions of Zn-L2EO4 ([-] ) 4.12 mM) to a solution of P2MVP41-b-PEO205 (initial [+] ) 0.85 mM). The f- values after additions are indicated next to the data points. The corresponding titration curve is not shown. The gray circles connected by the dotted line indicate the approximate relaxation time.

60 s, after which it was allowed to come to rest in another 10 s. Then the scattered intensity was followed in time, yielding the traces in Figure 3. Therefore, the first available measuring time after each addition is 70 s. The intensity at t ) 0 is the equilibrium intensity of the last addition (equilibrium intensity of the previous step). It is found that at f- e 0.33, I gradually increases with time until it levels off. The relaxation time seems larger for lower f-. At 0.33 < f- e 0.4, I initially increases, then decreases before it levels off. Upon further addition of ZnL2EO4 to reach 0.43 < f- < 0.47, the initial increase is so fast that we could not obtain data points during the increase (increase already happens during mixing, at which no stable I values can be obtained). What we see in Figure 3 is that I simply decreases with time before it levels off. After f- g 0.50, the scattered light intensity shows no time effect any more and remains stable, suggesting that equilibrium is reached. The poor reproducibility of the shoulder in the light-intensity curve is therefore probably caused by a subtle history effect. The titration steps and waiting times were different for the different titrations of Figure 1. Time effects during micelle formation have been observed both for surfactant19 and C3M systems composed of oppositely charged covalent homopolymer and diblock copolymer pairs.3 The variation of scattered light intensity with time in Figure 3 reflects changes of the size distribution of the particles. Shown in Figure 4 are the CONTIN results of the size distribution variation with measuring time at f- ) 0.33. It is clearly seen that the size of the big particles shifts from 75 to 108 nm within 4 min. At times longer than 4 min, the mean size of the big particles does not shift, but the peak significantly broadens, which implies more contributions of the big particles to the scattered light intensity in agreement with the observed increase of intensity at time longer than 4 min. Titration of Zn-L2EO4 with P2MVP41-b-PEO205. Results of “reverse” LS titration experiments, that is, addition of (relatively concentrated) P2MVP41-b-PEO205 to the dilute ZnL2EO4 solution are shown in Figure 5. We read from Figure 5a that with increasing P2MVP41-b-PEO205 (decreasing f-, read from right to left in the figure), the scattered light intensity increases for solutions of various initial Zn-L2EO4 concentration and reaches a maximum at f- ≈ 0.5 as expected. However, with this “reverse” titration experiment, we observe a systematic

Figure 5. Variation of scattered light intensity (open symbols in (a)) and mean hydrodynamic radius (filled symbols in (b)) of the particles in the titration of various Zn-L2EO4 solutions with P2MVP41-b-PEO205 solutions. For titration of [-] ) 0.87, 1.74, and 1.80 mM Zn-L2EO4 solutions, [+] ) 4.34 mM P2MVP41-b-PEO205 solution was used; for titration of [-] ) 4.26 mM Zn-L2EO4 solution, [+] ) 21.58 mM P2MVP41-b-PEO205 solution was used.

decrease with initial Zn-L2EO4 concentration of the f- value at which the peak maximum appears. We conclude that the peak at f- ≈ 0.5 in Figure 5a should correspond to the one at f- ≈ 0.5 in Figure 1a. Excitingly, upon further decrease of f-, a very pronounced increase of the scattered light intensity was observed again at f- < 0.3 for all the systems. The intensity of the second peak in this experiment is at least 3 times higher than the first one. Actually, the second peak at f- < 0.3 for the [-] ) 4.26 mM Zn-L2EO4 system is already too strong to be properly measured. Precipitation occurs in the sample before a real scattering peak was found. Like the position of the first peak, the absolute position of the second peak also shifts systematically with concentration. The large peaks at f- < 0.3 in Figure 5a should correspond to the shoulder at f- ) 0.32-0.35 in Figure 1a. That is to say, the weak shoulder in Figure 1a is now successfully enlarged in Figure 5a. The difference of their strength can be explained by the concentration difference at f< 0.3. In Figure 1a, the negative charge concentration [-] ) 0.40 mM at the shoulder position for the titration of 0.88 mM P2MVP41-b-PEO205 with 4.26 mM Zn-L2EO4 solution. While, it is 2.09 mM when the titration is done reversely with addition of 4.34 mM P2MVP41-b-PEO205 to the [-] ) 1.74 mM Zn-L2EO4 solution. That is to say, the concentrations are more than

11666 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Yan et al.

Figure 6. CONTIN size distribution of particles in the reverse titration of a Zn-L2EO4 solution ([-] ) 1.74 mM) with a P2MVP41-b-PEO205 solution ([+] ) 4.32 mM).

Figure 7. Comparison of the CONTIN distribution of particles: (a) f- ) 0.5, direct mixing; (b) f- ) 0.33, direct mixing; (c) f- ) 0.5, prepared by addition of Zn-L2EO4 to a directly mixed f- ) 0.33 solution; (d) f- ) 0.33, prepared by addition PMVP-b-PEO to a directly mixed f- ) 0.5 solution.

5 times larger in the reverse titration. That is one reason why the second scattering peak is so much more pronounced when the reverse titration was done. Another reason might be the difference in the sample history, because the number of the C3Ms before the large objects show up is obviously different in the two sets of titration experiments. The sample history does play a role in the aggregation of the system, as we may also infer from other indications described below. The magnitude of the scattering peak at f- < 0.33 depends strongly on concentration. For an initial negative charge concentration of 0.87 mM, this peak does not show up clearly, but it is already quite pronounced at an initial [-] ) 1.74 mM, and becomes huge at initial [-] ) 4.26 mM. When P2MVP41-b-PEO205 is added to a solution of Zn-L2EO4, the mean hydrodynamic radii of the particles are of a constant value of 25 nm in a broad range of f- (0.35-0.90), independent of concentration. This size is far smaller than those in Figure 1b, which is because of the absence of large aggregates under these conditions. The CONTIN size distribution indeed confirms this (Figure 6a). However, in the regime where the scattered light intensity shows a second peak (f- < 0.35), the CONTIN results clearly indicate a bimodal distribution. In Figure 6b, CONTIN size distributions for f- ) 0.30, 0.27, and 0.22 are shown. It is interesting to find that at f- ) 0.30, where the scattered light rises (circles in Figure 5a), large particles with a radius of about 154 nm show up. Upon further addition of P2MVP41-b-PEO205 to f- ) 0.27, where the second peak appears, the bigger particles grow to 164 nm. As f- decreases to 0.22, at which the light intensity has decreased to half the

intensity of that at maximum, CONTIN analysis shows that the size of the bigger particles shrinks from 192 to 164 nm, and the peak for the smaller particles is now centered at 20 nm. Therefore, the sharp decrease with decreasing f- of the second scattering peak in Figure 5a is caused by disintegration of bigger particles into smaller ones. DLS Measurement upon Direct Mixing. According to the titration experiments discussed so far, large aggregates are formed at f- < 0.4, no matter which protocol is used. In Figure 7, parts a and b, we show the CONTIN results of the size distribution of the C3Ms formed at f- ) 0.5 and 0.33, respectively, by one-step addition of [-] ) 4.26 mM Zn-L2EO4 solution to 0.86 mM P2MVP41-b-PEO205 solution, that is, direct mixing of the two components. Both solutions were left standing for 1 h before subjecting them to DLS measurement. We see only one population of particles in the f- ) 0.5 system with PDI lower than 5%, but two populations are observed in the f- ) 0.33 solution. The shape and position of the peaks in both systems do not change with time, even when investigated 6 months later. This result confirms once more that the large aggregates do exist at f- around 0.33. To prove that the large particles cannot completely disappear once formed, we added Zn-L2EO4 to the above f- ) 0.33 system to obtain a f- ) 0.5 solution. Then we always reach a system containing these large objects (Figure 7c). In contrast, if a f- ) 0.33 solution is prepared by addition of extra P2MVP41-b-PEO205 to the directly mixed f- ) 0.5 solution, two families of particles are observed (Figure 7d). Sometimes the two families are found to be

Wormlike Aggregates from Zn-L2EO4 and P2MVP41-b-PEO205

Figure 8. Angular dependence of the apparent particle radius (a) and the self-diffusion coefficient (b) in the P2MVP41-b-PEO205/Zn-L2EO4 mixed systems. Filled squares: f- ) 0.33; open squares: f- ) 0.5. In both systems, an initial [+] ) 0.86 mM P2MVP41-b-PEO205 and an initial [-] ) 0.87 mM Zn-L2EO4 solution were used as stock solutions.

incompletely separated into two single peaks, which may be due to the rather short measuring time. Angular-Dependent DLS. To get a better understanding of the large objects in the system, we carried out angular-dependent dynamic light-scattering experiments. Shown in Figure 8a is the comparison of the angular (q2)-dependent radii of the particles at f- ) 0.33 and 0.50. It is clear that the apparent average size of particles in the f- ) 0.33 solution decreases

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11667 from 136.6 to 48 nm as the detection angle varied from 45 to 135°, but those in the f- ) 0.5 systems remain almost unchanged. Because bigger particles scatter strongly at small angles, we can conclude that bigger ones indeed exist in the f) 0.33 system, whereas they are absent at f- ) 0.5. The dependence of the self-diffusion coefficient D on the square of the scattering vector, q2, is shown in Figure 8b. It is known20 that asymmetric particles always give rise to a dependence of D () Γ/q2) on q2, but for spherical particles, the D () Γ/q2) values should be independent of the scattering vector, because of the undetectable rotational motion. In Figure 8b (open squares), an obvious dependency of D on the scattering angle is seen for a directly mixed f- ) 0.33 mixture, while that for the f- ) 0.5 mixture (Figure 8b, filled squares) shows almost no dependence on q2. This result indicates that the big aggregates in the f- ) 0.33 system are probably not spherical. However, one should keep in mind that high polydispersity of the particles can also lead to an angular-dependent D and Rh; therefore, we decided to carry out another experiment, namely, depolarized light scattering, to get further information about the large objects. Depolarized Light Scattering. Depolarized light scattering was measured to further confirm the existence of nonspherical objects in the f- ) 0.33 solution. If spherosymmetric particles exist in the solution, the depolarized scattered intensity is almost undetectable, since the rotation of the spheres is isotropic. In contrast, if asymmetric particles, such as rods or worms, exist in the solution, there is significant depolarized intensity due to the anisotropic rotation of the rods or worms. The scattered light intensities were recorded at both a perpendicular (I⊥) and parallel (I|) polarizer position. The ratio R ) I|/I⊥ characterizes the relative amount of large asymmetric particles in the system.21,22 We find that R for the f- ) 0.33 system is 2.6 × 10-3, whereas

Figure 9. Cryo-TEM image of the aggregates in the mixture of P2MVP41-b-PEO205/Zn-L2EO4. (a) 0.02% P2MVP41-b-PEO205 with [+] ) 0.44 mM, f- ) 0.33; (b) 0.04% P2MVP41-b-PEO205 with [+] ) 0.87 mM, f- ) 0.33; (c) 0.02% P2MVP41-b-PEO205 with [+] ) 0.44 mM, f- ) 0.50; (d) 0.02% P2MVP41-b-PEO205 with [+] ) 0.44 mM, f- ) 0.80. The scaling bar represents 200 nm.

11668 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Figure 10. Titration of [+] ) 0.43 mM P2MVP41-b-PEO205 with 4.35 mM L2EO4 free ligands. Squares: f- is calculated considering 4charges for every L2EO4; triangles: considering 2- charges for every L2EO4.

that for the f- ) 0.5 system is 0.63 × 10-3. The former is more than 4 times larger than the latter. This is further evidence that the large aggregates in the f- ) 0.33 system are indeed asymmetric. Unfortunately, we cannot study a system containing only the big particles because they always show up together with the small ones. This makes the characterization of the big particles by static light scattering (SLS), small-angle neutron scattering (SANS), or small-angle X-ray scattering (SAXS) difficult. Therefore, we tried to examine the morphology of the larger ones by Cryo-TEM. Cryo-TEM. In Figure 9 we show the Cryo-TEM results. The contrast is enhanced compared to normal C3Ms owing to the presence of heavy metal ions in the complex coacervate cores.3 Hence, the dark objects in sight are interpreted as the complex coacervate cores, and the bright region separating them is interpreted as the hydrated coronas. Interestingly, we observe wormlike objects, which we shall tentatively denote as “micelles”, in the f- ) 0.33 system. As shown in Figure 9a,b, some wormlike micelles are found coexisting with spherical micelles. In a solution with 0.02% P2MVP41-b-PEO205 ([+] ) 0.43 mM), the wormlike micelles are about 100-150 nm in length, but their cross-sectional radius is almost the same as that of spherical micelles. It seems that the wormlike micelles are aggregated spherical micelles. We can even clearly see that they are actually a series of “spherical beads” connected tightly, like a necklace. As the concentration of the P2MVP41-b-PEO205 is increased to

Yan et al. 0.04% in Figure 9b, the number of wormlike micelle beads increases as compared to Figure 9a. As the wormlike particles become more pronounced in Figure 9b, their length is also increased to as much as 200 nm (Figure 9b). In contrast, only spherical micelles are seen in the f- ) 0.5 solution (Figure 9c). As f- increases to 0.80, the spherical micelles are still preserved and their size is almost unchanged (Figure 9d). The spherical micelles observed in Figure 9c,d are of the same size as those in Figure 9a,b, indicating they are similar species. A significant feature of Figure 9 is the dense regular packing of the spheres and wormlike micelles. We believe that this dense regular packing is induced by the sample preparation process, as is also found in other systems.23 The solution itself, of course, is a dilute dispersion of micelles. However, the observation of only spherical micelles at f- g 0.5 gives us confidence that the objects we see in the cryo-TEM do occur in the original sample; the mean separation between particles of about 50 nm in Figure 9c is an indication that the average radius of these micelles is about 25 nm, which agrees well with the hydrodynamic radius obtained from DLS. Taking into account the observation of big aggregates by DLS and the presence of asymmetric particles confirmed by depolarized laser light scattering, we can draw as a conclusion that the wormlike micelles observed by cryo-TEM must be real objects in the investigated solution. Since the average radii of the spherical and wormlike micelles from the cryo-TEM are all about 7-10 nm, and recalling the hydrodynamic radius of 25 nm for these micelles, the thickness of the corona is thus found to be 15-17 nm. Possible Mechanism for the Formation of Wormlike Micelles. The formation of wormlike micelles around f- ) 0.33 does not occur for ordinary C3Ms with comparable chain length;2,3 it must somehow be due to the special properties of the coordination complexes. In an attempt to explain their appearance, we assume that, possibly, fully charged free ligand molecules L2EO44- (instead of a coordination complex of [ZnL2EO4]2- with two elementary negative charges at each coordination center) are playing some role in the mixed system, although the weakly acidic pH condition in our experiments does not favor such an assumption. Therefore, we performed a light-scattering titration of a P2MVP41-b-PEO205 with free L2EO4 ligands without metal ions. Shown in Figure 10 is the variation of the scattering intensity of a P2MVP41-b-PEO205 solution upon addition of L2EO4. The

SCHEME 4: Illustration of the Bridging Effect of Zn-L2EO4 Coordination Oligomers (small rings) between Positively Charged Pre-micellesa

a The shape for the pre-micelles is not certainly of such regular geometry, and the distance between two pre-micelles bridged by Zn-L2EO4 does not represent the actual distance. The conformation of the Zn-L2EO4 bridges in the wormlike micelles, as well as the positive charge density of the pre-micelles, is also an illustration; it does not represent the real case.

Wormlike Aggregates from Zn-L2EO4 and P2MVP41-b-PEO205 two sets of symbols represent f- values calculated assuming that an L2EO4 carries either 4 or 2 negative charges. A modest increase of the scattered light intensity is observed, but the absolute scattering intensity remains very low. We have shown in a previous report that this slight increase of scattering is negligible when compared with the increase in the case of formation of true C3Ms.10 The scattering plateau shows up at f- ) 0.6 if we consider fully dissociated L2EO4, but almost exactly at f- ) 0.5 if half-dissociation of L2EO4 is assumed. Therefore, we can exclude the possibility that full dissociation of L2EO4 occurs in the presence of P2MVP41-b-PEO205. Hence, to explain the wormlike micelle formation, we must explore other directions. As a second possible reason for the formation of asymmetric wormlike micelles in the P2MVP41-b-PEO205/Zn-L2EO4 mixed system, we consider a “multivalent ion” effect of Zn-L2EO4 complexes. In this context, it is important to mention the fact that wormlike micelles were always observed only when bivalent metal ions, such as Zn2+, Fe2+, Ni2+, and Co2+, were used to construct the coordination polymer. For such ions the coordination “polymers” themselves are small ring-shaped species under the condition of the present experiments.11 In contrast, these wormlike micelles were always absent if trivalent ions, such as Nd3+ and La3+, were used to form the coordination polymer, where the coordination polymer exists in the form of networks.12 Therefore, it must be the bivalent coordination polymer that is responsible for the formation of the wormlike micelles. It has been reported24,25 that in the presence of multivalent ions, such as La3+ and Th4+, oppositely charged polyelectrolytes undergo chain precipitation at medium multivalent ion concentration due to a kind of “bridging” effect. Too low or too high concentrations of ions lead to a homogeneous solution due to the repulsive interaction caused by an excess of charges. It is possible that the small ring-shaped [Zn2(L2EO4)2]4ions act like a tetravalent ion in the mixed P2MVP41-b-PEO205/ Zn-L2EO4 system. The region where we find large particles is probably the medium multivalent ion concentration regime that causes phase separation in polyelectrolyte-multivalent ion mixtures. These complex phases are less dense than those formed with polymeric coordination polymers, so that their volume/interface area ratio is also larger. In condition with the stabilizing PEO block, this may well lead to non-spherical wormlike micelles, such as those we observed in Figure 9a, b. A different way to see this is as a “bridging” effect of ZnL2EO4 complexes between the positively charged pre-C3Ms. From the cryo-TEM results, it seems that the wormlike large aggregates are strings of micelles. At f- ) 0.2-0.4, there are already some positively charged pre-micelles in the solution. Due to the lack of enough negatively charged coordination polymer, some positively charged pre-micelles have to share few [Zn2(L2EO4)2]4-, and the small ringlike structures of ZnL2EO4 enable themselves this possibility, as illustrated in Scheme 4. Therefore, it is plausible that the wormlike micelles, or micellar strings, are either under way in splitting up into small spherical micelles during increasing f-, or in disintegrating to electrostatic complexes during decreasing f-. A change in composition easily leads to the breakup of the micellar strings. Although it is still not clear yet why large aggregates are formed in the P2MVP41-b-PEO205/Zn-L2EO4 mixed system, the present study will no doubt build valuable experience on further understanding this novel metal ions containing C3M systems. The transitions between small and large aggregates are based on subtle interactions and is an interesting topic for further study. We expect significant progress with reversible and controllable

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11669 molecular metal containing assemblies from these investigations. The morphology-controllable aggregate formation in this kind of systems is an enrichment to the family of self-assembly. Conclusions Large aggregates are found in mixed systems of the oppositely charged bivalent coordination polymer Zn-L2EO4 and the diblock copolymer P2MVP41-b-PEO205. These large aggregates revealed by cryo-TEM are wormlike micelles or strings of spherical micelles. At negative charge fraction f- < 0.4, they always show up together with spherical micelles. Once formed, they do not completely disintegrate. There is a narrow f- range where the size varied with f-. Acknowledgment. The authors thank Dr. A. T. M. Marcelis (Laboratory of Organic Chemistry, Wageningen University, Dreijenplein 8, 6703 HB Wageningen, The Netherlands) for help with the synthesis of L2EO4 bisligands. Financial support is from the EU POLYAMPHI/Marie Curie program (RT6-2002, proposal 505027) and SONS Eurocores program (Project JA016SONS-AMPHI). M.D. gratefully acknowledges financial support by the Deutsche Forschungsgemeinschaft (SFB 481). References and Notes (1) Cohen Stuart, M. A.; Hofs, B.; Voets, I. K.; de Keizer, A. Curr. Opin. Colloid Interface Sci. 2005, 10, 30-36. (2) Van der Burgh, S.; de Keizer, A.; Cohen Stuart, M. A. Langmuir 2004, 20, 1073-1084. (3) Cohen Stuart, M. A.; Besseling, N. A. M.; Fokkink, R. G. Langmuir 1998, 14, 6846-6849. (4) Kabanov, A. V.; Bronich, T. K.; Kabanov, V. A.; Yu, K.; Eisenberg, A. J. Am. Chem. Soc. 1998, 120, 9941-9942. (5) Harada, A.; Kataoka, K. Science 1999, 283, 65-67. (6) Thu¨nemann, A. F.; Mu¨ller, M.; Dautzenberg, H.; Joanny, J.-F.; Lo¨wen, H. AdV. Polym. Sci. 2004, 166, 113-171. (7) Koltover, I.; Salditt, T.; Ra¨dler, J. O.; Safinya, C. R. Science 1998, 281, 78-81. (8) Kakizawa, Y.; Kataoka, K. AdV. Drug DeliVery ReV. 2002, 54, 203222. (9) Boussif, O.; Lezoualch, F.; Zanta, M. A.; Mergny, M. D.; Demeneix, B.; Behr, J. P. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 7297-7301. (10) Yan, Y.; Besseling, N. A. M.; de Keizer, A.; Marcelis, A. T. M.; Drechsler, M.; Cohen Stuart, M. A. Angew. Chem. Int. Ed. 2007, 46, 18071809. (11) Vermonden, T.; Van der Gucht, J.; De Waard, P.; Marcelis, A. T. M.; Besseling, N. A. M.; Sudho¨lter, E. J. R.; Fleer, G. J.; Cohen Stuart, M. A. Macromolecules 2003, 36, 7035-7044. (12) Vermonden, T.; Van Steenbergen, M. J.; Besseling, N. A. M.; Marcelis, A. T. M.; Hennink, W. E.; Sudho¨lter, E. J. R.; Cohen Stuart, M. A. J. Am. Chem. Soc. 2004, 126, 15802-15808. (13) Vermonden, T.; de Vos, W. M.; Marcelis A. T. M.; Sudho¨lter, E. R. Eur. J. Inorg. Chem. 2004, 2847-2852. (14) Biesalski, M.; Johannsmann, D.; Ru¨he, J. J. Chem. Phys. 2004, 120, 8807-8814. (15) Biesalski, M.; Ru¨he, J. Macromolecules 1999, 32, 2309-2316. (16) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814-4820. (17) Provencher, S. W. Comput. Phys. Commun. 1982, 27, 229-242. (18) Provencher, S. W. Comput. Phys. Commun. 1982, 27, 213-227. (19) Besseling, N. A. M.; Cohen Stuart, M. A. J. Chem. Phys. 1999, 110, 5432-5436. (20) Xu, R.; Winnik, M. A.; Hallett, F. R.; Riess, G.; Croucher, M. D. Macromolecules 1991, 24, 87-93. (21) Van de Hulst, H. C. Light Scattering by Small Particles; Wiley: New York, 1957. (22) Kerker, M. The Scattering of Light and Other Electromagnetic Radiation; Academic Press: New York, 1969. (23) He, Y.; Li, Z.; Simone, P.; Lodge, T. P. J. Am. Chem. Soc. 2006, 128, 2745-2750. (24) Delsanti, M.; Dalbiez, J. P.; Spalla, O.; Belloni, L.; Drifford, M. In Macro-Ion Characterization from Diluted Solutions to Complex Fluids; Schmitz, K. S., Ed.; ACS Symp. Ser. 1994, 548, 381-392. (25) Delacruz, M. O.; Belloni, L.; Delsanti, M.; Dalbiez, J. P.; Spalla, O.; Drifford, M. J. Chem. Phys. 1995, 103, 5781-5791.