Kinetic Studies on the Sulfate Radical-Initiated Polymerization of Vinyl

Jul 19, 2005 - The rate constants measured for the reactions of the surface transients with 4-VP are 1 order of magnitude higher than those of VA, des...
0 downloads 0 Views 213KB Size
Langmuir 2005, 21, 8001-8009

8001

Kinetic Studies on the Sulfate Radical-Initiated Polymerization of Vinyl Acetate and 4-Vinyl Pyridine in the Presence of Silica Nanoparticles Paula Caregnato,† Galo Carrillo Le Roux,‡ Daniel O. Ma´rtire,*,† and Mo´nica C. Gonzalez*,† Instituto de Investigaciones Fisicoquı´micas Teo´ ricas y Aplicadas (INIFTA), Facultad de Ciencias Exactas, Universidad Nacional de La Plata, C. C. 16, Suc. 4, (1900) La Plata, Argentina, and Escola Politecnica da USP, Departamento de Engenharia Quimica, Sa˜ o Paulo, Brazil Received March 7, 2005. In Final Form: June 6, 2005 The photolysis of silica suspensions of pH ∼8 containing peroxodisulfate ions leads to the generation of two “surface transients” with a distinct spectrum and reactivity. Time-resolved and continuous irradiation experiments of similar dispersions also containing variable concentrations of vinyl acetate (VA) or 4-vinyl pyridine (4-VP) allowed the evaluation of the contribution of silica/water interfacial reactions to the kinetics and structural pattern of polymers synthesized using sulfate radicals as initiators. The rate constants measured for the reactions of the surface transients with 4-VP are 1 order of magnitude higher than those of VA, despite the fact that both species show similar reactivity in homogeneous solution toward sulfate radicals. It is suggested that both the sorption capacity and the different specific interactions with the silica surface of 4-VP and VA contribute to the observed reaction rates. Micrometer-sized latex particles of 4-VP and VP showed higher stability and more homogeneous size distributions when obtained in the presence of silica nanoparticles. Under the experimental conditions required for obtaining polymer particles, both the contribution of the described interfacial reactions and the effect of silica adsorbed monomer on the initiation steps of the polymerization may be neglected. The importance of in situ adsorption of the oligomer/polymer chains to silica NP during the polymerization propagation steps in determining the particle morphology is discussed.

There is growing interest in structured organicinorganic hybrid materials with tunable properties and well-defined multidimensional architectures.1,2 In particular, composites based on nanosized inorganic particles are an emerging field because of their importance in diverse applications in optics, electronics, engineering, and chemical sensing in biosciences. Composite materials conformed by specific interactions between polyelectrolytes and colloids are also frequently found in the environment. The most prominent inorganic material used for composite synthesis is unmodified or modified silica. Silanol groups and siloxane bridges determine the surface properties of silica, which offers a versatile reactive surface easily forming hydrolyzable Si-O-C and stable C-Si bonds with organic molecules.1,3 Therefore, a variety of modifications can directly and durably be incorporated into the SiO2 network. The most important aspects of silanol chemistry are the relatively high acidity of the group and the strong tendency to hydrogen bond to each other and to other species containing suitable hydrogen bonding sites. For example, silanols act as centers of adsorption of molecules capable of forming H bonds or of undergoing donor-acceptor interactions. Two main strategies may be followed to obtain organized colloidal polymer/silica materials: (i) Adsorption of a * Corresponding authors. E-mail: [email protected], [email protected]. † Universidad Nacional de La Plata. ‡ Escola Politecnica da USP. (1) Kickelbick, G. Prog. Polym. Sci. 2003, 28, 83. (2) Mori, H.; Mu¨ller, A. H. E.; Klee, J. E. Polym. Mater. Sci. Eng. 2004, 90, 329-330. (3) Currie, E. P. K.; Norde, W.; Cohen Stuart, M. A. Adv. Colloid Interface Sci. 2003, 100-102, 205-265.

monomer on the unmodified particle surface, followed by polymerization in the adsorbed layer. A variation of this strategy considers the covalent or ionic modification of the surface reactive groups that then serve as initiators of polymerization (surface-initiation approach). (ii) Manipulation of experimental conditions such as temperature, pH, and salt content to favor the occurrence of specific intermolecular interactions (hydrogen bonding, acid-base and ionic interactions) between the colloid and the polymer leading to a self-organized system.2 An interesting example of the latter approach is the adsorption of long poly(ethylene oxide) chains on small silica particles by means of hydrogen bonding between surface silanols and ether oxygen in the chain.3 The number of poly(ethylene oxide) chains adsorbed on silica decreases with increasing pH because of the dissociation of the silanol groups at the higher pH (pKa ≈ 4.5-6.5 and 8.5-9.0 for geminal and single-surface silanols, respectively4). The polymerization of VM via free radical chemistry has been reported to offer several advantages in the context of colloidal nanocomposite synthesis. Monomer and/or oligomer adsorption on the silica surface seems to be an important prerequisite for successful composite formation. Solutions of pH 10 and the presence of minor amounts of comonomer 4-vinylpyridine (4-VP) were decisive variables for obtaining stable and monodisperse polymer emulsions derived from hydrophobic monomers such as styrene. Samples under identical experimental conditions but at lower pH (7 and 3) showed large polydisperse emulsions and larger numbers of aggregates. In the absence of 4-VP, very large but stable particles in coexistence with a majority of unattached silica particles were reported. It (4) Caregnato, P.; Mora, V. C.; Le Roux, G. C.; Ma´rtire, D. O.; Gonzalez, M. C. J. Phys. Chem. B 2003, 107, 6131-6138.

10.1021/la0506170 CCC: $30.25 © 2005 American Chemical Society Published on Web 07/19/2005

8002

Langmuir, Vol. 21, No. 17, 2005

Caregnato et al.

Table 1. Manifold of the Most Important Reactions Taking Place after the UV Irradiation of Peroxodisulfate Dispersions of Colloidal Silica (Ludox) in the Presence of Small Amounts of Either VA or 4-VPa

a

R:

was suggested that the excellent performance of 4-VP at pH 10 is related to a strong acid-base interaction between the basic vinyl monomer (pKa of the conjugated acid is 5.62) and the acidic silica sol.5,6 Therefore, 4-VP would be preferentially adsorbed onto the surface of the silica sol and could act as a particulate emulsifier with polymerization conceivably occurring in the styrene-rich monomer droplets stabilized by the silica particles. However, the final particle size and silica content of 4-VP/SiO2 composites were reported to be limited by the aqueous solubility of the 4-VP monomer but insensitive to the analytical concentrations of 4-VP and silica sol. The latter observation and the use of the water-soluble peroxodisulfate initiator suggested an aqueous-phase homogeneous nucleation with the incorporation of the silica particles within the growing polymer latex as polymerization proceeds. Despite increasing interest in nanocomposite synthesis and research, much effort is still needed to fully understand the influence of physical and/or chemical interactions between the organic monomers and the silica surface on the properties and structure of hybrid organicinorganic composites. Because colloidal silica has a very high specific surface area and is optically transparent in the near-UV and visible regions, it is a suitable model surface for investigating interfacial reaction kinetics involving photochemically generated species. In our previous work, we investigated the reaction of sulfate radicals, SO4•- generated by reaction (R1) in Table 1, with the surface of suspended silica nanoparticles (NP). A kinetic analysis of the experimental results indicates that the interaction of SO4•- radicals with the NP surface leads to the formation of an adduct (reaction (R2)), with λmax ≈ 330 nm. The (5) Percy, M. J.; Barthet, C.; Lobb, J. C.; Khan, M. A.; Lascelles, S. F.; Vamvakaki, M.; Armes, S. P. Langmuir 2000, 16, 6913-6920. (6) Michailidou, V.; Armes, S. P.; Perruchot, C.; Watts, J. F.; Greaves, S. J. Langmuir 2003, 19, 2072-2079.

NP-sulfate radical adducts (NPS•) react with geminal and single SiO- sites, yielding SiO• surface defects (reactions (R3a) and (R3b), respectively) showing an absorption maximum around 600 nm. To explore the surface modification of silica NP by vinyl compounds and to gain information on the chemistry and kinetics involved during nanocomposite formation using sulfate radicals as initiators of the polymerization, in the present article we investigate the reactions of the surface transients on silica NP with vinyl monomers, VM. Two different VM were selected: 4-VP and vinyl acetate, VA. The former was proposed in the literature to undergo acidbase interactions with the silica surface in alkaline media. However, VA is a weak Lewis base; therefore, its interaction with the silica NP surface is expected to be of lesser significance. For that purpose, time-resolved and continuous irradiation experiments with silica dispersions of pH ∼8 containing variable concentrations of the VM and sodium peroxodisulfate were performed. To further understand the effect of the specific interactions toward the generation of hybrid organic-inorganic composites, adsorption experiments were also carried out. Experimental Methods Sodium peroxodisulfate (p.a. Merck), sodium hydroxide (99% GC Aldrich), 4-VP (95% GC Aldrich), and VA (99% Aldrich) were used as received. Distilled water (>18 MΩ cm-1, 10-5 M showed the fast depletion of sulfate radicals in less than 100 µs and the formation of a stable reaction product (vide infra). (19) Percy, M. J.; Michailidou, V.; Armes, S. P.; Perruchot, C.; Watts, J. F.; Greaves, S. J. Langmuir 2003, 19, 2072-2079. (20) Kato, K.; Uchida, E.; Kang, E.-T.; Uyamaa, Y.; Ikada, Y. Prog. Polym. Sci. 2003, 28, 209-259. (21) Zemel, H. Ph.D. Thesis, Carnegie-Mellon University, Pittsburgh, PA, 1976. (22) Tiarks, F.; Landfester, K.; Antonietti, M. Langmuir 2001, 17, 5775-5780.

Langmuir, Vol. 21, No. 17, 2005 8005

Figure 4. Dependence of kapp on [VA] for 330 nm transients obtained from experiments at 293.2 K with 0.025 g/L NP(Ludox) suspensions of pH 8.3 containing 5 × 10-3 M S2O82-. Error bars in kapp are on the order of the symbol size. The dashed lines stand for the 95% confidence interval. (Inset) Dependence of the initial amplitude, ∆A0, on [VA] for the experiments of the main plot. The solid line stands for calculated values (see text).

The calculated values for the reaction rate constants are, within experimental error, independent of pH and of the same order as that reported for VA and for many other vinyl derivatives.9 Therefore, the pyridine moiety is not directly involved in the reaction between sulfate radicals and 4-VP. Instead, the addition of sulfate radicals to the double bond, reaction 4, seems to be the main reaction step. Therefore, sulfate radicals are efficient initiators of 4-VP polymerization (k64-VP ) (1.6 ( 0.6) × 109 M-1 s-1 at 293 K) independent of pH in the range 4-7. Reaction Kinetics of the Transients Generated at the Silica/Water Interface with VA. Time- resolved absorption curves at detection wavelengths in the range from 290 to 360 nm were obtained from experiments performed with 0.025 g/L silica suspensions containing 1 × 10-3 M S2O82- and 1 × 10-5 M e [VA] < 5 × 10-5 M, as shown in Figure 2. The decay rate of the transient follows a first-order law (see solid lines in Figure 2) with an apparent rate constant, kapp, increasing with [VA] as shown in Figure 4. The slope of the plot of kapp versus [VA] yields a second-order rate constant of k4-VA ) (8.5 ( 2) × 107 M-1 s-1 for the reaction between the NPS• adduct and VA, reaction (R4). The initial amplitude of the signals, ∆A0, decreases with increasing [VA], as shown in the Figure 4 inset. Within experimental error, no signal is observed for [VA] g 1 × 10-4 M. As already described in our previous publication,4 transient formation at 600 nm is fast compared to its decay rate; therefore, its formation and decay kinetics may be independently analyzed. The decay of the traces in this wavelength range (signals not shown) may be well fitted to a biexponential function, eq 1. In this equation, subscripts s and g stand for the slow- and fast-decaying components, respectively.

∆A ) Cg e-kgt + Cs × e-ks t

(1)

A biexponential analysis of the traces obtained in the presence of increasing [VA] yields kg and ks, which are also increasing with [VA] as shown in Figure 5. However, the sum of Cg and Cs decreases with increasing [VA] (Figure 5 inset). The slopes of the lines kg and ks versus [VA] yield the absolute rate constants kgVA ) (2.5 ( 1) × 107 M-1 s-1 and ksVA ) (5 ( 7) × 105 M-1 s-1 for the reaction between SiO• surface groups and VA, reactions (R5a) and (R5b), respectively. Reaction Kinetics of the Transients Generated at the Silica/Water Interface with 4-VP. Time-resolved absorption curves at detection wavelengths in the range from 290 to 360 nm were obtained from experiments performed with 0.025 g/L silica suspensions containing 5

8006

Langmuir, Vol. 21, No. 17, 2005

Caregnato et al.

Figure 5. Dependence of kg (b) and ks (O) on [VA] for experiments obtained at 600 nm and 293.2 K with 0.025 g/L NP (Ludox) suspensions of pH 8.3 containing 5 × 10-3 M S2O82-. The error bars represent standard deviations. (Inset) Dependence of Cg + Cs on [VA] for the experiments of the main plot. The solid line stands for calculated values (see text).

Figure 6. Plot of kapp vs [4-VP] for experiments at 292.3 K with [NP] ) 0.025 g/L and [S2O82-] ) 1 × 10-3 M at pH 8.3 and 293.2 K. Error bars in kapp are on the order of the symbol size. The dashed lines stand for the 95% confidence interval. (Inset) Dependence of the initial amplitude, f(λ), and final absorbance, h(λ) for λ ) 320 nm, on [4-VP] for the experiments of the main plot. The solid line stands for calculated values (see text).

× 10-3 M S2O82- and 5 × 10-7 M < [4-VP] < 7.5 × 10-6 M. The experimental absorption traces at a given wavelength of analysis, A(λ), were fit according to eq 2.

A(λ) ) f(λ) × e-kappt + h(λ)

(2)

The calculated constant kapp is independent of λ but linearly dependent on [4-VP], as shown in Figure 6. The slope of the plot of kapp versus [4-VP] yields the secondorder rate constant k44-VP ) (7.3 ( 0.5) × 108 M-1 s-1 for the reaction between the NPS• adduct and 4-VP. The absorption term h(λ) increases with increasing [4-VP], as shown in the Figure 6 inset, and remains constant even several minutes after irradiation with the flash of light. The initial amplitude of the signals, f(λ), decreases with increasing [4-VP], as also shown in the Figure 6 inset. Flash irradiation of 0.025 g/L silica suspensions containing [S2O82-] ) 1 × 10-3 M and [4-VP] > 1 × 10-5 M shows the formation of stable reaction products, as indicated by the observed differences in the absorption spectrum of the suspensions before and after irradiation; see Figure 7. The absorption profile of h(λ) obtained from the transient signals agree, within experimental error, with the absorption spectrum of the reaction products, as shown in the Figure 7 inset. The latter observations clearly indicate that the observed final absorption h(λ) is due to a stable reaction product. It is interesting that the irradiation of solutions without suspended silica NP showed a prompt and stable increase in absorbance due to the formation of reaction products with an absorption

Figure 7. Absorption spectra (1 cm optical path) of 0.025 g/L silica (Ludox) suspensions containing 5 × 10-3 M S2O82- and 2.5 × 10-5 M 4-VP at pH 8.2 and 293.2 K. (A) Before irradiation and (B) after 10 flashes of irradiation. (Inset) Absorption spectrum of h(λ) obtained from similar solutions to those of the main plot both in the presence (b) and absence (O) of 0.025 g/L suspended silica NP. The solid line corresponds to the normalized absorption spectrum of the main plot.

Figure 8. Dependence of kg (b) and ks (O) on [4-VP] for experiments obtained at 600 nm and 293.2 K with 0.025 g/L NP (Ludox) suspensions of pH 8.3 containing 5 × 10-3 M S2O82-. (Inset) Dependence of Cg + Cs on [4-VP] for the experiments of the main plot. The solid line stands for calculated values (see text).

spectrum coincident with that of h(λ), as also shown in the inset of Figure 7. The traces obtained at 600 nm after irradiation of silica suspensions containing 5 × 10-3 M S2O82- and increasing concentrations of 4-VP, such as those shown in the Figure 2 inset, may also be well fit to eq 1. The values of kg and ks increase with monomer concentration, as shown in Figure 8. However, the sum Cg + Cs decreases with increasing [4-VP] as shown in the Figure 8 inset. The slopes of the lines kg and ks versus [4-VP] yield the absolute rate constants kg4-VP ) (5.0 ( 2.5) × 107 M-1 s-1 and ks4-VP ) (4 ( 2) × 106 M-1 s-1 for the reaction between 4-VP and SiO• surface groups originating from geminal and single silanols, reactions (R5a) and (R5b), respectively. Insight into the Reaction Mechanisms. Reactions (R1)-(R6) in Table 1 conform to the simplest manifold of reactions necessary for interpreting the observed kinetic trends. An analysis of the reaction scheme indicates that the VM compete with the silica NP for sulfate radicals (reactions (R6) and (R2)). In fact, the reaction between sulfate radicals and the VM is very efficient (Table 1), and for the monomer concentrations used in the flash irradiation experiments, the product k6 × [VM] is in the range of (1.5-15) × 104 s-1. Because under our experimental conditions the depletion of sulfate radicals by the surface of the silica NP proceeds with a pseudo-first-order rate constant of 104 s-1,4 the reaction of sulfate radicals with the VM in the liquid phase is favored at high monomer concentrations. The decrease in the initial absorbance observed for the formation of the 320 nm adduct (NPS•)

Radical Polymerization in the Presence of Silica

Langmuir, Vol. 21, No. 17, 2005 8007

with the monomer concentration for both VA and 4-VP (see solid lines in the insets of Figures 4 and 6) may be due to these competing reactions. If we assume that the flash emission is a δ function producing sulfate radicals and that only the competing reactions (R6) and (R2) are of significance immediately after irradiation, the resolution of the kinetic equations yields eq 3 for the initial absorption of the NPS• adduct, ANPS. The latter assumption is based on the decay (reactions (R3) and (R4)) that is one order slower than for the generation of NPS•.

ANPS ) NPSl

[

k2[SO4•-]0[NP]

]

k2[NP] + k6[VM]

(3)

where NPS and l stand for the absorption coefficient of the NPS adduct and the optical path length of the cell, respectively. Initial sulfate radical anion concentrations, [SO4•-]o ) (5-6) × 10-6 M, were estimated from experiments under identical conditions but in the absence of NP and VM (450(SO4•-) ) 1600 M-1 cm-1).4 From this value and those of k2 and k6 taken from Table 1, eq 3 was used to calculate the dependence of the initial transient absorbance at 320 nm. The predicted values for VA and 4-VP (solid lines in the insets of Figures 4 and 6, respectively) show excellent agreement with the observed experimental trends. The initial absorbance of the signals at 600 nm also decreases with [VM], as shown in the insets of Figures 5 and 8 for VA and 4-VP, respectively. The latter curves may be reasonably fitted (see solid lines in the corresponding Figures) with an equation similar to eq 3 where the product NPS[SO4•-]0 is replaced by a constant term C. The observed trend is in agreement with SiO• surface defects being formed from the reaction of the NPS adduct with the deprotonated silanols;4 therefore, a decrease in the concentration of the NPS adducts leads to a diminished concentration of SiO•. The constant factor C therefore stands for the absorption coefficient of the SiO• times its yield of formation. Surface adsorption of the VM may also inhibit the reaction of sulfate radicals with the NP surface, leading to a diminution of the initial NPS absorbance with increasing [VM]. However, the good agreement between predicted and experimental initial absorbances at 320 nm (vide supra) indicate that the 4-VP and VA adsorptions on silica do not exert any significant effect on the formation efficiency of the NPS• adduct. The previous discussion indicates that for [VM] g 1 × 10-5 M and [NP] ) 0.025 g/L the most important reaction of sulfate radicals taking place is that with the VM in solution. Moreover, the product absorbance, h(λ), linearly increases with increasing [4-VP] (Figure 6 inset), as expected from its formation from reaction (R6) in the aqueous phase and in agreement with the coincidence of the h(λ) spectrum with that observed in the absence of NP. The NPS adduct is much less reactive with VM than sulfate radicals. The condition of the surface radical of the NPS adduct is expected to limit its reaction rates toward substrates in solution because the immobilization of a radical at a surface reduces access from solution and lowers the rate from a collisional encounter in the absence of adsorption.23,24 The rate of reaction at an interface, k4theor, (23) Astumian, R. D.; Schelly, Z. A. J. Am. Chem. Soc. 1984, 106, 304-308. (24) Shield, S. R.; Harris, J. M. J. Phys. Chem. B 2000, 104, 85278535.

with one reactant immobilized at a spherical NP surface can be derived, eq 4, from the Smoluchowski equation for uniformly reacting molecules applied to the VM in solution and the NP sphere. The effect of anisotropic reactivity of the NP due to the adsorbed radical is introduced by a steric factor f 25

k4theor )

4πNARNP+VMDNP+VM -Ea/kT fe 1000

(4)

where NA is Avogadro’s number, RNP+VM is the sum of the radii of the silica NP and that of the VM, and DNP+VM is the mutual diffusion coefficient given by the sum of the individual diffusion coefficients for NP and VM. The exponential term corrects for the fraction of collisions that do not result in reaction and therefore considers the activation energy for the reaction. Taking RNP+VM ) 37.5 Å, DNP+VM ≈ DVM in the range from 1 × 10-5 to 0.5 × 10-5 cm2 s-1 as observed for most molecules in water at 298 K,26 and f ) 9 × 10-3 as calculated from the model of reactive patches for a sulfate radical adsorbed on the surface of a silica NP,25,27,28 we estimate k4theor ) (2.51.25) × 108 × exp(-Ea/RT) M-1 s-1. Therefore, assuming similar activation energies for the homogeneous and heterogeneous processes, the reactions between VM and the NPS adduct are expected to be around 6 to 12 times smaller than the corresponding reactions of the VM with sulfate radicals in solution. However, intense, mobile surface adsorption of the monomers is expected to favor reaction (R4) and therefore increase the observed decay rates. The reaction rate of 4-VP with the NPS adduct is 1 order of magnitude higher than that of VA. A similar appreciation is true for the reaction of VM with the oxygen surface defects of silica NP because higher rate constants are observed for the reactions of 4-VP with single and geminal SiO• radicals than for VA. The reported rate constants for the addition reactions of sulfate radicals to 4-VP and VA in homogeneous solutions are very similar (this wοrk, ref 9); accordingly, reaction rates for the addition of surface transients to both monomers are also expected to be of similar magnitude. The observed trend is in agreement with the higher adsorption capacity of silica for 4-VP (13 µmol/m2) than for VA (maximum expected 0.4 µmol/m2, see above) favoring the encounter between reactant molecules. The different specific interactions of VM with the silica surface may also contribute to the reaction rate affecting the stability of the activated complex. The higher electronic density on the aromatic ring due to the specific interaction between 4-VP and the silica surface14 should favor electrophilic reactions involving the vinyl moiety or with the ring itself. However, silica adsorption of VA lowers the electronic density on the acetate group, thus acting as an electron-withdrawing substituent for the vinyl moiety. The reactivity of the vinyl group toward electrophilic reactants will therefore be diminished. The observed trend in the magnitude of the rate constants, also in agreement with the expected effect of the specific interactions of 4-VP and VA on the rate constants of reactions 4 and 5, further supports an electrophilic nature for these reactions. Latex Particle Characterization. VA Latex. The irradiation of aqueous suspensions containing 10-3 M (25) Barzykin, A. V.; Shushin, A. I. Biophys. J. 2001, 80, 2062-2073. (26) Levine, Ira N. In Fisicoquı´mica, 4th ed.; McGraw Hill: Madrid, 1996; Vol. 2. (27) Churio, M. S.; Brusa, M. A.; Grela M. A. Photochem. Photobiol. Sci. 2003, 2, 754-758. (28) In agreement with the NPS adduct being thought of as an adsorbed sulfate radical on the silica NP, see ref 4.

8008

Langmuir, Vol. 21, No. 17, 2005

Caregnato et al.

Figure 10. TEM images of 4-VP polymers synthesized from dispersions containing 48.75 g/L 4-VP, 0.50 g/L sodium peroxodisulfate, and 40 g/L Ludox colloidal silica.

Figure 9. Amplified TEM images of VA polymers as obtained after 30 min of irradiation of solutions containing 18.6 g/L VA and 0.186 g/L sodium peroxodisulfate in the presence (left) and absence (right) of 15 g/L dispersed Ludox colloidal silica. Dynamic light scattering weight-sized distributions of the particle diameters as well as those for pure silica (dark gray bars) are also shown.

S2O82-, [VA] < 0.05 M, and silica NP in the range of 7.815 g/L at 298 K does not show the formation of latex particles of higher size than those of the bare silica NP. For [VA] ) 0.1 M, a milky-white latex is formed with lower particle sizes, rather homogeneous size distributions, and higher colloidal stability than particles formed in experiments under identical conditions but in the absence of silica (Figure 9). Under the latter initial polymerization conditions, a maximum of 0.18 g VA is expected to be adsorbed on the silica NP surface, which stands for 1% of the analytical VA concentration. The TEM images in Figure 9 obtained after sample centrifugation and further dissolution of the precipitated solid do not provide evidence for the presence of silica dispersed within the latex particles. However, TEM images of latex particles prepared in the presence of silica but without the centrifugation-dissolution treatment of the samples showed the presence of free silica NP and latex particles covered by adsorbed silica NP (images not shown). 4-VP Latex. Colloidal, stable milky-white particles were obtained from solutions containing 48.75 g/L VP, 0.5 g/L S2O8Na2, and 40 g/L silica NP. The aqueous solubility of 4-VP is 2.9% w/w at neutral pH and 293 K,29 and a maximum capacity of 48 mg of 4-VP is expected to be adsorbed per gram of silica (vide supra). Thus, in the initial stage of our latex syntheses 4-VP is partially dissolved in the aqueous phase (60%) and around 6% adsorbed at the surface of the silica sol, and the remaining monomer (34%) appears as small drops. TEM micrographs of the 4-VP latex are shown in Figure 10. The particle size distribution of the latex is reasonably uniform with particle diameters in the range of 70 to 80 nm with diffuse surface boundaries. The obtained images do not provide evidence for the presence of silica dispersed within the latex particles as reported for 4-VP polymer(29) Ma, Q.; Gu, L.; Ma, S.; Ma, G. J. Appl. Polym. Sci. 2002, 83, 1190-1203.

silica nanocomposites prepared under emulsion conditions.6 However, TEM images of latex particles synthesized in the absence of silica (not shown) showed a large distribution of particle sizes and poor colloidal stability because considerable agglomeration is observed. Our flash-photolysis results indicate that the reaction of sulfate radicals with VM in the liquid phase is favored at [VM] > 10-5 M in experiments containing [NP] on the order of 0.025 g/L. Therefore, to generate surface defects capable of reacting with VM, a ratio of 4 × 10-6 mol VM/ m2 silica should not be surpassed. However, our polymerization experiments show that [VM] > 0.10 M is required for synthesizing size-measurable latex particles. Under the latter conditions, the maintenance of small 4-VP to NP ratios would require the use of NP concentrations higher than those of the Ludox stock suspensions. Therefore, ratios of 4 × 10-5 mol VM/m2 silica were used. Under the latter conditions, our results do not support the synthesis of a polymer initiated by surface radicals but rather by homogeneous reactions in the aqueous phase. However, the presence of silica NP clearly favors the stability and homogeneous size distribution of the polymeric particles. The importance of the specific interactions of oligomer/polymer chains to silica NP during the polymerization propagation steps in determining the morphology of the latex particles needs to be further investigated. In fact, a simple estimation may show the increasing significance of the specific adsorption of propagation chains as polymerization proceeds. As already discussed, approximately 1% of the VA monomers are H-bonded to surface silanols through the carbonyl groups of the acetate moiety during the initial steps of the polymerization experiments. The carbonyl groups of the acetate substituents of VA polymer molecules should also be able to H bond to silanol groups, and as a first approximation, it seems reasonable to assume an equal adsorption capacity for all carbonyl groups in the reaction mixture. Therefore, the number of acetate groups H bonded to silanols remains unchanged throughout the polymer synthesis though the number of molecules adsorbed on the silica NP steadily increases (i.e., all polymer chains containing more than 100 VA units are adsorbed on silica because at least one of its acetate groups is H bonded to a surface silanol). A similar argument is applicable to 4-VP. The latter discussion gives a rough view of the adsorption events in the reaction mixture but clearly indicates that polymer growth

Radical Polymerization in the Presence of Silica

is conditioned by the specific interactions with the silica NP either because silica maintains the polymer chains dispersed in the aqueous phase or because of a different kinetics consequence of the generation of surface-trapped organic radicals during the polymerization propagation steps. Acknowledgment. This research was supported by Agencia Nacional de Promocio´n Cientı´fica y Tecnolo´gica, Argentina (ANPCyT), PICT 2003 no. 06-14508. We thank the Argentinean Secretarı´a de Ciencia y Te´cnica (SECYT)

Langmuir, Vol. 21, No. 17, 2005 8009

and the Brazilean Coordenac¸ a˜o de Aperfeic¸ oamento de Pessoal de Nı´vel Superior (CAPES) for granting the collaboration project BR/PA03-EXIV/020. M.C.G. is a research member of the Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas (CONICET). D.O.M. is a research member of the Comisio´n de Investigaciones Cientı´ficas y Te´cnicas de la Pcia. De Buenos Aires (CICPBA). P.C. is a doctoral fellow from CICPBA. LA0506170