Synthesis of Anisotropic Nanoparticles by Seeded ... - ACS Publications

Key Centre for Polymer Colloids, School of Chemistry F11, UniVersity of Sydney, NSW 2006, Australia. ReceiVed January 1, 2006. In Final Form: February...
0 downloads 0 Views 196KB Size
Langmuir 2006, 22, 4037-4043

4037

Synthesis of Anisotropic Nanoparticles by Seeded Emulsion Polymerization Eric B. Mock,† Hank De Bruyn,‡ Brian S. Hawkett,‡ Robert G. Gilbert,*,‡ and Charles F. Zukoski† Department of Chemical and Biomolecular Engineering, UniVersity of Illinois, Urbana, Illinois 61801, and Key Centre for Polymer Colloids, School of Chemistry F11, UniVersity of Sydney, NSW 2006, Australia ReceiVed January 1, 2006. In Final Form: February 23, 2006 Anisotropic polystyrene nanoparticles of diameters below 0.5 µm were prepared by coating the surface of crosslinked polystyrene latex particles with a thin hydrophilic polymer layer prior to swelling the particles with styrene and then initiating second-stage free-radical polymerization. Conditions were found so that all particles had uniform asymmetry. The effect of surface chemistry on the development of particle anisotropy during seeded emulsion polymerization of sub-0.5 µm diameter particles was studied. The extent and uniformity of the anisotropy of the final particles depended strongly on the presence of the hydrophilic surface coating. Systematic variation of the degree of hydrophilicity of the surface coating provided qualitative insight into the mechanism responsible for anisotropy. Conditions were chosen so that the surface free energy favored the extrusion of a hydrophobic bulge of monomer on the hydrophilic surface of the particle during the swelling phase: the presence of a hydrophilic layer on the particle surface causes this asymmetry to be favored above uniform wetting of the particle surface by the monomer. Kinetic effects, arising from the finite time required for the seed to swell with the monomer, also play a role.

Introduction Anisotropic particles in a colloidal suspension are expected to pack into configurations considerably different from those of spherical particles. In the simplest instance, the anisotropy is purely volume excluded, while, in more complex cases, the open packing configurations arise from particles that experience anisotropic physical interactions (for example, charge variations or as in protein crystals whose microstructure is controlled by heterogeneous surface interactions). Methods of synthesizing anisotropic particles are of interest in developing novel colloidcrystalline structures and in studying the effects of shear. Here, we explore the use of seeded emulsion polymerization to create particles that are anisotropic in shape but are sufficiently uniform that they order when the particle volume fraction is raised. A seeded emulsion polymerization technique has been described by Sheu et al. that results in anisotropic particles with characteristic sizes of 5-10 µm.1 The methodology introduced in the present paper extends the techniques of Sheu et al. by the addition of a grafted hydrophilic layer on the surface of the particles, which will be seen to enable uniformly asymmetric particles of submicron size to be obtained reproducibly. The method of Sheu et al. consists of synthesizing crosslinked seed particles that are swollen with monomer and then raising the reaction temperature and initiating free radical polymerization. The resulting asymmetric particles can be viewed as partially penetrating spheres of different sizes. The anisotropy was shown to arise from the immiscibility of the monomer in cross-linked polymer networks. Many applications of anisotropic particles are enhanced by characteristic diffusion times, td, less than 500 ms, where the diffusion time is defined as a2/6D, with a being the particle’s hydrodynamic radius and D being the particle * Corresponding author. E-mail: [email protected]. Telephone: +61 2 9351 3366. Fax: + 61 2 9351 8651. † University of Illinois. ‡ University of Sydney. (1) Sheu, H. R.; El-Aasser, M. S.; Vanderhoff, J. W. J. Polym. Sci., Part A: Polym. Chem. 1990, 28, 629-651.

diffusivity. In low-viscosity continuous phases such as water, this occurs for particles in the size range below 500 nm in diameter. The method of Sheu et al. resulted in particles with td > 5-10 s that settle easily under gravity, unless specific care is taken to density-match the particles. Here we describe an extension of the method of Sheu et al. to prepare uniform anisotropic particles with sizes in the 100-300 nm range, these being sufficiently small to remain colloidally stable indefinitely. We report that anisotropy is enhanced through a surface modification where the cross-linked spherical seed particle surface is coated with a thin hydrophilic layer prior to swelling particles with styrene. After initiating polymerization by the addition of an oil-soluble initiator, bulges are formed on the parent seed spheres. Systematically changing the type and degree of hydrophilic coating applied to the spheres results in anisotropic nanoparticles with varying sharpness of delineation between the seed particle and the bulge.

Thermodynamic and Kinetic Factors Governing Particle Asymmetry Thermodynamics of Swelling. Sheu et al. demonstrated that raising the temperature of monomer-swollen cross-linked polystyrene spheres causes the monomer to phase-separate from the seed particle as a nonuniform bulge out of the particle. When the polymerization reaction is initiated, the phase-separated bulge polymerizes, resulting in anisotropic particles. Sheu et al. developed a thermodynamic model that accounts for swelling as a result of balancing the free energy of mixing the monomer with the polymer, the elastic energy resulting from stretching the chains within the cross-linked seed particles and the surface tension of the particle/continuous phase interface. As the temperature is increased, the elastic force increases, resulting in a lower equilibrium swelling ratio (monomer mass in swollen particle/ unswollen seed particle mass). The equilibrium swelling ratio results in the swollen particle expelling monomer, the result being the development of a monomer bulge out of each particle. This process can be qualitatively understood by considering seed particles in a solution saturated with a monomer. Under

10.1021/la060003a CCC: $33.50 © 2006 American Chemical Society Published on Web 03/31/2006

4038 Langmuir, Vol. 22, No. 9, 2006

Mock et al.

these conditions, the particles swell until the monomer in the particles is in equilibrium with the monomer in the saturated aqueous phase. Following the thermodynamic model developed by Sheu et al., for hydrophobic monomers such as styrene, where the solubility is low, the free energy of the monomer within a polymer network relative to the free energy of the monomer in a monomer droplet in solution, ∆G h m,p, can be approximated as the sum of three terms: the mixing of monomer and polymer, ∆G h m, the elastic energy of the cross-linked polymer network, ∆G h el, and the interfacial tension between the particle and water, ∆G h t:

∆G h m,p ) ∆G h m + ∆G h el + ∆G ht

(1)

Substituting the Flory-Huggins expression for ∆G h m,2 the Flory-Rehner equation for ∆G h el,2 and the Morton equation for ∆G h t,3 the following formula is derived:

∆G h m,p ) RT[ln(1 - νp) + νp + χmpνp2] + RTNVm(νp1/3 /2νp) + 2Vmγ/a (2)

1

Here, R is the ideal gas constant, T is the absolute temperature, νp is the volume fraction of polymer in the swollen seed particle, χmp is the monomer-polymer interaction parameter, N is the effective number of chains in the network per unit volume, Vm is the monomer molar volume, γ is the interfacial tension between the particle and water, and a is the radius of the swollen seed particle. From eq 2, it is apparent that the ∆G h el and ∆G h t terms make positive contributions to ∆G h m,p, thus encouraging seed particle contraction, while the ∆G h m term will make a negative contribution to ∆G h m,p, thereby encouraging seed particle expansion. The conceptual framework for generating anisotropic particles is that, when seed particles are exposed to a monomer at low temperatures, the particles begin to swell toward an equilibrium size determined when ∆G h m,p ) 0. Furthermore, the degree to which the monomer and seed particle separate will depend strongly upon the interfacial tensions of the particle and the aqueous phase, γP,A, the particle and the monomer, γP,M, and the monomer and the aqueous phase, γM,A. Assuming that the interface between the seed particle and the expelled monomer is flat, and performing a force balance, an equation analogous to Young’s equation for a liquid drop on a solid surface in the presence of a vapor is obtained:

γP,A ) γP,M + γM,A cos θ

(3)

where θ is the contact angle between the monomer and the seed particle. As the contact angle becomes larger, the bulge extending from the seed particle will be more pronounced, and the region of phase separation, where the spherical seed particle ends and the bulge begins, will be more readily distinguished. Rearranging eq 3 in terms of contact angle yields

θ ) cos

- γP,M γM,A

γ -1 P,A

(4)

From eq 4 it can be seen that the contact angle will increase, and thus the bulge will become more distinct, as the difference between γP,A and γP,M decreases. (2) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (3) Morton, M.; Kaizerman, S.; Altier, M. W. J. Colloid Sci. 1954, 9, 300312.

Figure 1. Schematic of different states of the monomer-seed particle system as it is (a) very favorable, (b) mildly favorable, (c) not very favorable, and (d) not at all favorable for the expelled monomer to wet the cross-linked seed particle surface.

A significant innovation in the present work is to coat the particles with a grafted layer of hydrophilic polymer (poly(vinyl acetate) and poly(acrylic acid)). The rationale for this is to make it thermodynamically unfavorable for the monomer to wet the particle surface: increasing γP,M. This will increase the asphericity of the monomer bulge, as well as the asphericity of the final particle after this bulge is polymerized. This is because it will be thermodynamically unfavorable (through the appropriate value of χ) for the polymer formed by polymerizing the hydrophobic monomer in the bulge to be compatible with the hydrophilic surface of the seed particle. A number of possible swelling geometries are illustrated in Figure 1. If surface wetting is completely unfavorable, then the expelled monomer should form a droplet of its own in the aqueous phase, as suggested in the sketch of Figure 1d. The hydrophobicity of the expelled monomer is, however, modified by the presence of surfactant that will reduce the interfacial energy between the continuous phase and the monomer. The energy difference between the two extreme cases of complete wetting and no wetting, illustrated in Figure 1a,d, may be expressed as

E ) 4π [γM,A(aF2 - aM2) + (γP,M - γP,A)aP2]

(5)

Here, aF, aM, and aP are the radii of the final swollen particle, of the monomer droplet, and of the original cross-linked seed particle, respectively. When the energy difference between complete wetting and no wetting, E, is positive, the system will move toward no wetting of the seed particle with monomer, with larger values of E indicating a larger driving force for the separation of the expelled monomer and the cross-linked seed particle. Substituting eq 3 for γP,A yields

E ) 4πγM,A(aF2 - aM2 - aP2 cos θ)

(6)

Ignoring, for the moment, the difference in densities between styrene and polystyrene (∼14%) and assuming ideal mixing, then

aF3 ) aP3 + aM3

(7)

Substituting eq 7 into eq 6 and dividing through by 4πγM,AaP2 leads to an expression for a dimensionless energy ED ) E/4πγM,AaP2,

ED )

(aP3 + aM3)2/3 aP2

-

aM2 aP2

- cos θ

(8)

Assuming that the conversion of styrene to polystyrene seed particles is complete, and that the seed particles are dense, the swelling ratio may be expressed as

Synthesis of Anisotropic Polystyrene Nanoparticles

S)

aM3FM

Langmuir, Vol. 22, No. 9, 2006 4039

(9)

aP3FP

where FM and FP are the densities of the monomer and the polymer, respectively (0.909 and 1.05 g cm-3 in the present case). Rearranging eq 9 in terms of aM yields

(

) ( )

FP ED ) 1 + S FM

2/3

FP - S FM

2/3

- cos θ

(10)

This expression for ED illustrates that, for a given swelling ratio, as surface wetting becomes less favorable (i.e., as ED increases from negative to positive), the region of phase separation between the bulge and the seed particle will become more distinct (i.e., θ will increase). This supports earlier arguments from eq 4 that the wetting of the seed particle surface by the monomer, and thus the distinctness of the bulge, is highly dependent on the difference of the interfacial tensions of the particle and the aqueous phase, and that of the particle and the monomer. Kinetic Aspects of Swelling. The rate of monomer swelling of the seed particles to form a bulge depends on the rate of transport through the water phase (which is fast,4 provided that the shear in the reactor provides a sufficient rate of mass transfer) and diffusion into the initially dry polymer seed particle, a process which can be quite slow in systems without cross-linking (e.g., ref 5), and which can be even slower in cross-linked systems. As the seed particles swell with monomer, the cross-linked polymer chains become solvated and relax into an equilibrium configuration. Interestingly, at room temperature, it has been shown that the particles overswell, that is, in the presence of monomer, particles first swell and then shrink, expelling monomer.1 This process is explained by polymer chains in the cross-linked seed particles being far from equilibrium. Thus, phase separation occurs as a result of particles containing chains far from equilibrium. Because of phase separation at room temperature or because of increases in elastic energy upon raising the temperature, the cross-linked polymer chains relax, causing the seed particle to contract, thus driving the monomer back into solution. If the contraction takes place rapidly enough, the seed particle will shrink, leaving a monomer droplet attached to the seed particle. The shape of this droplet will be determined by the monomer/polymer wetting properties, the kinetics of the polymer network contraction, and the kinetics of monomer polymerization. For asymmetric particles, this shape should be a bulge, rather than a uniform coating (wetting) of the seed particle. Kinetic Aspects of the Polymerization of Asymmetrically Swollen Particles. To synthesize a latex of uniformly asymmetric polymer particles from a spherical particle that is asymmetrically swollen with monomer, there are two essential criteria: (a) that no new particles be formed (i.e., avoidance of secondary nucleation; these new particles would not be cross-linked and would thus be spherically symmetric) and (b) that the asymmetry of the bulge of the monomer is not lost upon polymerization. When an oil-soluble initiator (azobisisobutyronitrile (AIBN) in this study) is added to the suspension of styrene swollen seed particles suspended in water, it will enter the seed particles and generate radicals. As in any emulsion polymerization, these radicals may propagate in the particle or undergo transfer to a monomer, forming a styryl radical that may exit the particle. There will also be a small but significant fraction of the AIBN (4) Gilbert, R. G. Emulsion Polymerization: A Mechanistic Approach; Academic Press: London, 1995. (5) van der Zeeuw, E.; Sagis, L. M. C.; Koper, G. J. M. Macromolecules 1996, 29, 801-803.

Figure 2. Scanning electron micrographs of (a) 18% DVB crosslinked seed and (b) 18% DVB cross-linked seed after 5:1 nominal swelling and polymerization.

in the water phase. These desorbed styryl radicals, and/or the radicals resulting from aqueous-phase AIBN dissociation, may then cause new (secondary) spherical particle formation by propagation with styrene in the aqueous phase followed by homogeneous or micellar nucleation; alternatively, these aqueousphase radicals can reenter another seed particle. To prevent the formation of secondary particles during the seeded growth step, we added sodium nitrite as an aqueousphase inhibitor, a variation of the technique developed by the Lehigh group.6 Experiment Seed Particle Preparation. Monodisperse polystyrene seed particles of ∼250 nm diameter, cross-linked with divinylbenzene (DVB), were prepared using emulsion polymerization (Figure 2a). A 1 L round-bottom flask equipped with a poly(tetrafluoroethylene) (PTFE)-coated stir bar was immersed in a constant-temperature bath at 80 °C. Immediately afterward, deionized water (400 mL) was added to the flask and allowed to reach bath temperature. The reactor setup was then charged with styrene (42.3 g, Sigma-Aldrich, 99% purity grade), sodium dodecyl sulfate (0.500 g, Bio-Rad, electrophoresis purity reagent grade) dissolved in 50.0 mL of deionized water, followed by 50.0 mL of rinse water, and the desired amount of DVB (Aldrich, 55% mixture of isomers tech. grade), with enough stirring to form a medium-sized vortex. The quantity of DVB added to the flask in this study was 18% of the total styrene monomer mass added in this emulsion polymerization step, unless stated otherwise. After waiting 1 h for the contents of the flask to mix and reach bath temperature, we added potassium persulfate (1.55 g, Fisher Scientific, (6) Sudol, E. D.; El-Aasser, M. S.; Vanderhoff, J. W. J. Polym. Sci., Part A: Polym. Chem. 1986, 24, 3515-3527.

4040 Langmuir, Vol. 22, No. 9, 2006 99.5% purity grade) dissolved in 75.0 mL of deionized water, and the reaction was allowed to proceed for 24 h. The diameter of the seed particles obtained using Brookhaven Instruments fiber optics quasi-elastic light scattering (FOQELS) was 249 nm. The size was also estimated to be 250 ( 8 nm by scanning electron microscopy (SEM; Hitachi S4700), measuring particle diameters at different points around the SEM sample grid (SPI Supplies, Formvar coated). SEM sample grids were prepared by placing a drop of the latex taken directly from the reaction vessel after the reaction was complete (sampled using a pipet placed below the surface) onto a grid laying on filter paper (Whatman, Grade 1) to assist in removing water and drying the sample. While SEM is not an optimal way to measure particle size, it is the simplest method for obtaining the external morphology of aspherical particles, and so the size from this method is reported for consistency with later results. Average values of the diameters from SEM for the seed particles are in acceptable accord with those measured by FOQELS. Seed particle suspension volume fractions were determined by weight loss upon drying ∼5 mL of the suspension at 60-70 °C, using a polymer density of 1.05 g cm-3. Addition of the Hydrophilic Layer. Two types of hydrophilic layers were used to coat the surfaces of the spherical cross-linked seed particles. The first hydrophilic layer, and the one on which this study mainly focused, consisted of vinyl acetate (Aldrich, 99+% purity grade) initiated with potassium persulfate and grown in situ, thereby grafting sulfate ions to the particle surface, further enhancing its hydrophilicity. The second type of hydrophilic layer resulted from the copolymerization of acrylic acid (Sigma-Aldrich, 99% purity grade) and styrene to form an electrosteric “hairy” layer as described below. The standard amount of either chemical species used to coat the seed particle surfaces was 3.56 × 10-21 g of vinyl acetate or acrylic acid per square nanometer of particle surface. For vinyl acetate, a surface coverage of 1.78 × 10-21 g/nm2 of the particle surface was also used as discussed below. Using the particle diameter and the volume fractions of the crosslinked seed particle suspensions obtained via dry weight measurements, the necessary amount of vinyl acetate or acrylic acid to achieve the desired particle surface coating was determined for each batch of cross-linked seed particles as σApVtφ/Vp. Here, σ is the desired surface coating (typically 3.56 × 10-21 g/nm2 of particle surface), Ap is the surface area of a single particle, Vt and φ are the total volume and volume fraction of the suspension to which the vinyl acetate or acrylic acid was to be added, respectively, and Vp is the volume of a single particle. For vinyl acetate addition to the seed particle surfaces, as with the seed particle synthesis, a round-bottom flask equipped with a PTFE-coated stir bar was immersed in a constant-temperature bath at 80 °C. Next, the desired amount of seed particle suspension was introduced to the flask and allowed to reach bath temperature. Potassium persulfate (1 g/150 g deionized water) was charged to the reaction flask, the mass being 25% of the mass of vinyl acetate necessary for the desired surface coverage. Upon adding the potassium persulfate, the vinyl acetate was added over four intervals spanning 45 min, so that, initially, around 25% of the vinyl acetate was added, and then, about every 15 min, approximately 25% more was added. After the last vinyl acetate addition, the reaction was allowed to continue for 24 h. To add acrylic acid to the seed particle surfaces, the same procedure as that used for vinyl acetate addition was employed, except that all of the acrylic acid was added at once. Additionally, when the acrylic acid was charged into the reactor, an equal mass of styrene was added; this is so the acrylic acid/styrene copolymerization will result in a blocky copolymer of these two monomers, the first part of which will largely be acrylic acid units formed by rapid aqueousphase polymerization, and the second part will largely comprise styrene units, formed when the aqueous-phase propagating radical picks up enough styrene units so that its radical end becomes hydrophobic, enters the particle, and continues polymerizing in this styrene-rich phase. The result is a poly(acrylic acid) moiety anchored

Mock et al. to the particle (e.g., refs 7 and 8). It is noted that the characterization of such a layer (e.g., the size of this corona) is notoriously difficult (e.g., ref 9) and is really only possible when such a layer is deliberately synthesized so as to have a monodisperse hydrophilic component (e.g., using controlled radical polymerization10,11). Suspensions with either type of hydrophilic layer attached to the spherical seed particles were placed in SpectraPor 4 dialysis tubing (molecular weight cutoff 12 000-14 000) and dialyzed for 24 h against deionized water, with the deionized water dialysate being changed four times during the dialysis. The seed particle suspension initially had a pH of ∼2, and, after the addition of acrylic acid and dialysis, it had a pH of 3.2. Once the dialysis was complete, for every milliliter of suspension, 0.021 mL of a 1 M sodium phosphate (Fisher Scientific, 99.6% purity grade) solution was added so that the pH of the final suspension was 6.2. Suspension volume fractions for particles coated with either type of hydrophilic coating were determined by weight loss upon drying as described above. Seed Particle Swelling. Once the cross-linked seed particles had been coated, anisotropic nanoparticles were prepared on the basis of our extension of the seeded emulsion polymerization technique of Sheu et al.1 A 500 mL flat-bottom jar equipped with a PTFEcoated stir bar was immersed in a constant-temperature bath at room temperature. Next, 25.0 mL of the seed particle latex with a hydrophilic layer attached was discharged into the jar. Using the volume fractions of cross-linked seed particles coated with hydrophilic chains, the amount of styrene necessary to swell the particles with a chosen ratio of swelling monomer mass to polystyrene particle mass was determined by SφFVt, where S is the desired swelling ratio, and φ, F, and Vt are the volume fraction, density, and total volume of the suspension to be swollen with monomer, respectively. This amount of styrene was then added to the suspension under vigorous stirring, and the swelling was allowed to continue for 22 h. Swelling ratios of styrene mass/polystyrene particle mass of 1:1, 3:1, 5:1, 7:1, and 9:1 were explored in this study. Second-Stage Polymerization. After the seed particles had been swelling with styrene for 22 h, the constant-temperature bath was heated to 80 °C. After waiting 2 h for the temperature bath and contents of the reaction vessel to reach 80 °C, we added 0.250 g of sodium dodecyl sulfate and 0.200 g of sodium nitrite (Fisher Scientific, 99.5% purity grade), both dissolved in 15.0 mL of deionized water, followed by 30 mL of rinse water. When 1 h had passed to allow the reactants to mix together and reach bath temperature, 0.0400 g of AIBN dissolved in 2.00 mL of styrene was charged to the vessel, and the reaction was allowed to proceed for 24 h. Anisotropic particles were made under low shear (magnetic stirrer bar) using 2, 6, 9, 12, 15, 18, and 21% DVB cross-linked seeds; vinyl acetate coverages of 1.78 × 10-21 and 3.56 × 10-21 g/nm2 or an acrylic acid coverage of 3.56 × 10-21 g/nm2; and nominal swelling ratios of 5:1, 7:1, and 9:1. A second series was made under higher shear (250 rpm impeller) as discussed below.

Results and Discussion Initially, a series of experiments were carried out with crosslinked seed particles that did not have a surface coating (Figure 2a). While some anisotropy was achieved under these conditions (18% DVB; 5:1 swelling ratio), the phase separation between the spheres and the bulges was unclear, and the particles ended up looking egg-shaped or ellipsoidal at best, typified by Figure 2b. As outlined above, we hypothesized that, upon being expelled from a hydrophilically coated seed particle, it is more favorable (7) Bassett, D. R.; Hoy, K. L. In Polymer Colloids II; Fitch, R. M., Ed.; Plenum: New York, 1980; pp 1-25. (8) Coen, E. M.; Lyons, R. A.; Gilbert, R. G. Macromolecules 1996, 29, 51285135. (9) De Bruyn, H.; Gilbert, R. G.; White, J. W.; Schulz, J. C. Polymer 2003, 44, 4411-4420. (10) Ferguson, C. J.; Hughes, R. J.; Nguyen, D.; Pham, B. T. T.; Hawkett, B. S.; Gilbert, R. G.; Serelis, A. K.; Such, C. H. Macromolecules 2005, 38, 21912204. (11) Thickett, S. C.; Gilbert, R. G. Macromolecules 2006, 39, 2081-2091.

Synthesis of Anisotropic Polystyrene Nanoparticles

Langmuir, Vol. 22, No. 9, 2006 4041

Figure 3. Scanning electron micrographs of (a) an 18% DVB seed coated with 3.56 × 10-21 g/nm2 of vinyl acetate; (b) the same seed after 5:1 nominal swelling and polymerization; (c) the same seed after 9:1 nominal swelling and polymerization; (d) an 18% DVB seed coated with 3.56 × 10-21 g/nm2 acrylic acid followed by 5:1 nominal swelling and polymerization; and (e) an 18% DVB seed coated with 1.78 × 10-21 g/nm2 vinyl acetate followed by 5:1 nominal swelling and polymerization.

for the hydrophobic styrene monomer to form a bulge than to wet the surface of the existing seed particle. To test this hypothesis, the seed particles were coated with a hydrophilic layer in accordance with eq 4, so as to decrease the interfacial tension between the seed particle and aqueous phase, γP,A, and to increase the interfacial tension between the seed particle and the expelled monomer, γP,M. This was initially done by adding vinyl acetate in the presence of a potassium persulfate initiator. The hydrophilic sulfate radicals generated from the initiator will propagate with the vinyl acetate in the water phase (the solubility of vinyl acetate in water is 0.3 M at 50 °C12), until the resulting chain reaches a sufficient degree of polymerization (∼8 for vinyl acetate13) that it enters a particle13 and continues to propagate with the vinyl acetate in the interior of the particles. The vinyl acetate/potassium persulfate system was chosen partly because of the relative hydrophilicity of vinyl acetate (compared to styrene) but mainly because the high entry efficiency13 ensures that most of the sulfate radicals produced by decomposition of the persulfate are grafted to the particle surface. To avoid depletion of aqueous-phase vinyl acetate, which is necessary to maintain high entry efficiency, the vinyl acetate was added in four intervals over 45 min. The same procedures were followed using acrylic acid instead of vinyl acetate. To anchor the acrylic acid, an equal amount of styrene was added to the reactor at the same time. Without the (12) Yeliseeva, V. I. In Emulsion Polymerization; Piirma, I., Ed.; Academic Press: New York, 1982; pp 247-288. (13) Maxwell, I. A.; Morrison, B. R.; Napper, D. H.; Gilbert, R. G. Macromolecules 1991, 24, 1629-1640.

additional styrene, a propagating radical will not enter the seed particle, but will rather polymerize in the aqueous phase. As mentioned earlier, the pH of the seed particle latexes was around 2, indicating that the acrylic acid would be protonated, which was desirable because it reduces the hydrophilicity of the AAco-styrene oligomers, thereby increasing the probability of entry and therefore that of grafting of the oligomers to the particles. The pH was raised to 6.2 before the anisotropic growth stage to maximize the hydrophilicity of the particle surface. Typical SEM micrographs (Figure 3b) show that seed particles coated with acrylic acid were more anisotropic at the same nominal swelling ratio than seed particles coated with vinyl acetate (Figure 3d). Furthermore, as the amount of vinyl acetate coating the particles is decreased, the particles become less anisotropic, as illustrated in Figure 3a,e. In all the latexes grown from hydrophilically modified crosslinked seeds under low shear conditions, one sees a spherical particle that is nearly the same size as that of the original seed, either with or without a bulge of smaller or commensurate size (typified by Figure 3b-e). It should be emphasized that the sample placed on the SEM sample grids, from which these micrographs were obtained, was taken directly from the reaction vessel, and no steps were taken to filter or clean the sample. If all of the added monomer were polymerized within the bulge, one would expect a bulge significantly larger than the seed. It is thus immediately apparent from these micrographs that not all of the added monomer is polymerized in the seed particles. This is postulated to arise because of an incomplete transfer of monomer to the seed (a point explored experimentally later in

4042 Langmuir, Vol. 22, No. 9, 2006

Mock et al.

Figure 4. Scanning electron micrographs of 2% DVB seed particles coated with 3.56 × 10-21 g/nm2 vinyl acetate and swollen with swelling ratios of (a) 1:1 in a flask and impeller setup, (b) 3:1 in a flask and impeller setup, (c) 3:1 in a flask and impeller setup and mixed with toluene, (d) 7:1 in a flask and impeller setup, and (e) 7:1 in a jar and stir bar setup.

this section), and/or because of incomplete conversion, and/or because of polymerization in the monomer droplets. Droplet polymerization can lead to coagulum, and, indeed, in all cases, significant coagulum was formed during the polymerization process. Thus, the nominal swelling ratios only indicate the total amount of monomer in the system, and though there was a general trend toward larger bulges with increasing nominal swelling ratios, the effect is qualitative at best. Figure 3b-e may be considered representative of all the low shear anisotropic latexes. Upon performing numerous repeats, the length that the bulge protruded from the seed particle remained constant, demonstrating that the degree of anisotropy, as assessed visually from the micrographs, was reproducible. The aim of the present paper is the preparation of small anisotropic and uniform particles, and it will be seen that the addition of a hydrophilic layer enables this goal to be achieved, with the accompanying formation of coagulum that can be easily removed. The origin of the lack of complete monomer conversion solely within the seed particles was not pertinent to this aim, and was therefore not investigated further. As discussed above, the sizes of the seed and the formed particle suggest that much of the added monomer was not polymerized within the swollen seed. One possible reason for this is that the shear system did not provide a sufficient rate of mass transfer for equilibrium swelling of the seed particles in the allocated time. When the reaction was scaled up by a factor of 30 in a 3 L flask fitted with a glass impeller with a PTFE paddle, attached to a Glas-Col 099D GT31 stirrer system running at ∼250 rpm, which is expected to provide much more efficient

shear, anisotropy was dramatically increased. Figure 4a shows the results of swelling seed particles cross-linked with 2% DVB with a swelling ratio of 1:1 using this experimental configuration. The resulting particles were dumbbell-shaped, suggesting that the mixing of the monomer and the seed particles was essentially complete for this experimental configuration (i.e., all of the monomer diffused into the seed particles). Figure 4b,d shows the results of swelling the same seed particles with swelling ratios of 3:1 and 7:1, where the diameter of the larger sphere on the anisotropic particles was observed to be larger than that of the original seed particle, while the smaller sphere was similar in size to the original seed particle, indicating that the seed particles could grow bulges more than twice the original seed volume, provided that mixing was adequate. The simple stirrer configuration with poor shear also gave rise to secondary nucleation of spherical spheres mixed with the anisotropic particles, as seen in Figure 4e for seed particles crosslinked with 2% DVB and swollen with a swelling ratio of 7:1. This secondary nucleation was not seen in the system with good shear. The latter latex is indeed sufficiently uniform that it undergoes an order/disorder phase transition, as evidenced by the appearance of the type of iridescence also often seen in uniform spherically symmetric polymer colloids. To test the hypothesis that anisotropy is the result of a phase separation instead of a permanent distortion of the seed particle during swelling, experiments were carried out in which the seed particles cross-linked with 2% DVB were swollen with a nominal swelling ratio of 3:1 and polymerized. The polymer produced during the polymerization was then extracted with toluene to

Synthesis of Anisotropic Polystyrene Nanoparticles

remove at least some of the non-cross-linked material formed in the second-stage polymerization, as follows. About 0.4 g of this anisotropic-particle latex was placed in a 7 mL vial with 0.9 g toluene, and the vial was then capped and vigorously shaken for 2 min. The resulting particles were examined with SEM. Figure 4c shows that, when the anisotropic particles were exposed to toluene, spherical particles similar in size to the original seed particles resulted, suggesting that anisotropy results due to the expulsion of monomer and subsequent polymerization, not from permanent deformation of the cross-linked seed particles.

Conclusions On the basis of thermodynamic considerations, conditions have been designed to create anisotropic latex particles synthesized via seeded emulsion polymerization; the conditions are that the seed particle be cross-linked (as in the earlier work of Sheu et al.1) and coated with a thin layer of hydrophilic polymer to encourage bulge formation rather than uniform wetting by the monomer. The results were investigated with SEM. Interfacial tensions between the seed particle and water, between the seed particle and the monomer, and between the monomer and water, all played an important role in determining the morphology of the final particles. This was illustrated by varying the type and quantity of the hydrophilic layer attached to the seed particles. With no surface coating, anisotropic particles were ellipsoidal, showing limited demarcation between the phaseseparated bulge and the seed particle. When the amount of this hydrophilic layer was increased, particle anisotropy increased. This behavior is consistent with surface wetting arguments accounting for the energy difference between complete seed particle surface wetting by the monomer and absolutely no surface wetting by the monomer. The result was a latex of small and uniformly asymmetric particles, wherein one could see a spherical

Langmuir, Vol. 22, No. 9, 2006 4043

region corresponding to the original cross-linked seed, forming a homogeneous part of an asymmetric larger particle of a shape that varied with preparation conditions. Not all the added monomer was polymerized within this asymmetric particle, which can be ascribed to incomplete conversion and/or polymerization within the monomer droplets. It was found that swelling conditions (e.g., shear) could have a significant effect, due to the time required for equilibrium swelling of the cross-linked particles. The results of this study may be applied to the synthesis of uniform anisotropic nanoparticles exhibiting varying degrees of separation between the seed and the bulge. The resulting anisotropic nanoparticles are well suited for use in the construction of novel microstructures, where properties may be tuned by using different degrees of anisotropic particles (e.g., photonic band gap crystals, sensors, etc.). Acknowledgment. This material is based upon work supported by the U.S. Department of Energy, Division of Materials Sciences under Award No. DEFG02-91ER45439, through the Frederick Seitz Materials Research Laboratory at the University of Illinois at Urbana-Champaign. We also acknowledge support from the Nanoscale Science and Engineering Initiative of the National Science Foundation under NSF Award No. DMR0117792. Research for this publication was carried out in the Center for Microanalysis of Materials, University of Illinois, which is partially supported by the U.S. Department of Energy under grant DEFG02-91-ER45439. H.D.B and R.G.G. gratefully acknowledge the support of a Discovery grant from the Australian Research Council. The Key Centre for Polymer Colloids was established and supported under the ARC’s Research Centres program. LA060003A