Particle morphology development in hybrid miniemulsion polymerization.

Incompatibility between polymer phases resulting from hybrid miniemulsion polymerization of acrylic monomers in the presence of alkyd resin leads to interesting particle morphologies. In this paper, morphology was deduced through cross-comparison of results from several forms of microscopy. For the combination of methyl methacrylate and alkyd, a derivative of core/shell morphology was observed through the combination of transmission electron microscopy, scanning electron microscopy, and spin diffusion NMR. A raspberry-like shell was found to form on the hybrid particle surface consisting of a full coverage of small (roughly 25 nm) polymethyl methacrylate spheres anchored to the particle surface through grafting with the alkyd core. Migration of the spheres to that surface is thought to be induced by phase separation, and the size of the spheres precludes their origin from homoparticles from homogeneous nucleation. Homopolymethyl methacrylate particles were also detected in the particle distribution, resulting from the aqueous-phase initiator and hydrophilicity of methyl methacrylate monomer. For copolymer/alkyd systems (either methyl methacrylate/butyl acrylate/acrylic acid/alkyd or methyl methacrylate/butyl acrylate/alkyd), more traditional core/shell morphologies were observed with a lesser degree of homonucleated particles. A significantly different result was found in the combination of butyl acrylate and alkyd, resulting in a continuous particle-phase of polylbutyl acrylate and small internally dispersed island domains of alkyd. This is likely due to the lesser incompatibility between polybutyl acrylate and alkyd along with their similar hydrophobicity and glass transition temperatures. A higher degree of grafting between the alkyd and polybutyl acrylate also contributed to the compatibility between the two components, when compared to hybrid methyl methacrylate/alkyd systems.

Keywords: Hybrid miniemulsion, particle morphology, polymerization, alkyds

**********

Recent efforts have been made to combine the properties of water-based coating systems and traditionally solvent-based systems into a unified, one-component, hybrid water-based system. (1-8) These efforts have evolved in light of growing environmental awareness in addition to the synergistic benefits possible from hybrid products. The concept and implementation of hybrid polymerization is not a new field; various forms of graft copolymers consisting of a vinyl polymer component and an unsaturated polymer have existed for decades. Carrying out these graft polymerizations in an aqueous environment negates the traditional use of organic solvent. Miniemulsion polymerization seems the most viable means of creating these hybrid polymers, as droplet nucleation affords an emulsion-like environment lacking in monomeric or polymeric mass transport across the continuous aqueous phase. (1) Conventional emulsion polymerization of some systems has resulted in phase separation of the resinous component, which, by definition, implies that grafting did not occur. (1) Conversely, hybrid polymerizations performed via a miniemulsion have resulted in in-situ grafting of free radical-grown polymers with various polymeric resins, water-based in both process and product. (1-10) The resultant polymer particle morphology can yield information regarding the reaction mechanism, interaction of the grafting components, and the performance of the hybrid polymer in application. Morphology that results from polymerization gives an indication of the compatibility of comprising components as well as insight into the mechanism by which they interacted. Various tools exist for morphology determination; rarely is one technique conclusive on its own, so a number of methods must often be used to gain a reasonable understanding of particle morphology.

The objective of this article is to explore the morphological features resulting from a series of acrylic/alkyd hybrid graft-copolymers prepared via a miniemulsion.

EXPERIMENTAL

Chemicals and Equipment

McWhorter Chemicals (now part of Eastman Chemicals) supplied the alkyd resins used in these studies (Duramac 204-2768, medium alkyd, soya and linseed oil based). Monomers used were obtained from Aldrich Chemical Co. and included methyl methacrylate (MMA), butyl acrylate (BA), and acrylic acid (AA). The monomers were first distilled to remove any inhibitors present. The surfactant was sodium dodecylsulfate (SDS) and the initiators used included potassium persulfate (KPS), 2,2'-azobisisobutyronitrile (AIBN), and benzoyl peroxide (BPO), all supplied from Aldrich and used as received. Hydroquinone, for quenching, was supplied by Fisher Chemical Co., and the water used throughout the work was deionized. Osmium tetroxide (4 wt% solution in water) was obtained from Electron Microscopy Sciences Co. and used as received. Hexadecane (HD), used as the costabilizers for the miniemulsions, was obtained from Aldrich.

Dynamic light scattering (DLS) was performed on a Protein Solutions DynaPro-LSR instrument with high-resolution analysis over the 1 nm to 1 [micro]m radius size range. Samples were diluted at least 200:1 (vol:vol, water:sample) in deionized water. Dynamics version 5.26.53 was used for data analysis.

[FIGURE 1 OMITTED]

Transmission electron microscopy (TEM) was performed on a JEOL 100C TEM (100kV). Samples were prepared by diluting latex 100:1 (vol:vol) in deionized water and staining with three drops of a 4 wt% solution (in water) of osmium tetroxide. Samples were allowed to stand and react overnight with the stain before depositing a single drop on a 200-mesh nickel grid coated with formvar. Ambient evaporation of the drop deposited the polymer particles onto the TEM grid.

Scanning electron microscopy (SEM) was performed on a Hitachi S800 FEG scanning electron microscope at a resolution better than 3 nm. Vacuum dried latex samples were adhered to a sample stub with conductive tape and then plasma sputtered with a thin layer of gold for enhanced conductivity.

Atomic force microscopy measurements were made using a Nanoscope IIIa AFM obtained from Digital Instruments. Silicon cantilever tips (100 [micro]m, labeled TESP 7) were used in tapping mode to obtain height and phase measurements. The tip was a silicon cantilever with a spring constant of 20-100 nm and resonant frequency of around 300 Hz. Relevant settings were as follows: integral gain 0.3116, proportional gain: 0.400, amplitude setpoint: 1.290V, height range: 0-500 nm, and phase range: 0-180[degrees].

[.sup.1]H and [.sup.13]C NMR spectra were recorded on a Bruker AM400 spectrometer with a standard Bruker [.sup.1]H-[.sup.13]C double resonance MAS probe head. All samples were spun at a frequency of 3 kHz. The 90[degrees] pulse lengths were in the range of 3.5 to 4.0 [micro]s. The spin-diffusion experiments were carried out with selection of the component using the dipolar filter which consists of a pulse sequence with twelve 90[degrees] pulses separated by a delay time, [t.sub.d], varying between 10 and 50 [micro]s. Typically, 100-200 scans for the [.sup.1]H spectra were accumulated with a repetition time of two seconds. All chemical shifts are given relative to tetramethyl silane (TMS) as the external reference.

Miniemulsion Preparation

Monomer was mixed with alkyd HD (and AIBN, if applicable). This organic phase was mixed until completely homogeneous. While stirring vigorously with a magnetic stirrer, the organic phase was added to a solution of SDS in deionized water. The flask continued stirring to emulsify for 60 min after which a sonicating probe (Fisher 300W Sonic Dismembrator used at 70% output [maximum output of 300W] for durations of 25 min) was immersed into this pre-emulsion. During this process, the miniemulsion was cooled by a water bath to keep its temperature below 20[degrees]C. The miniemulsion was then transferred into a round bottom glass reactor equipped with a reflux cooler and stirred by an impeller. The whole system in a water bath was held at reaction temperature (70[degrees]C) by a thermocouple. A water-soluble initiator was injected into the reactor, upon reaching temperature, to start polymerization from a small solution dissolved in deionized water. Polymerization was carried out under a nitrogen atmosphere. At regular time intervals, samples were taken for various analyses. A general recipe is given in Table 1.

RESULTS AND DISCUSSION

Transmission Electron Microscopy Analysis

TEM can be used in the determination of particle surface properties and internal morphology. Contrast can be tailored to a heterogeneous system with osmium tetroxide staining techniques where the reactivity of a stain is preferential towards the phase or component containing residual double bonds. The acrylic portion of the hybrid particles will be lacking in residual double bonds as they have been consumed by polymerization. The alkyd component, however, contains residual double bonds, even after grafting (as verified by both NMR and IR, and so will complex with the osmium tetroxide, forming dark regions on the TEM. (11)

MMA/Alkyd Hybrid System (TEM)

The first system evaluated is a hybrid constituting methyl methacrylate and alkyd resin. Figures 1a and 1b are examples of the particle morphology for poly(alkyd-g-MMA) graft copolymers prepared via miniemulsion polymerization. The alkyd component is much more hydrophobic than methyl methacrylate and resides in the core, as evidenced by the darker hue of the particle center from the Os[O.sub.4] stain. Hydrophobicity is not the only motivation for alkyd to reside in the core; the acrylic polymer formed is also incompatible with alkyd. Higher esters, along with linseed oil, are nonsolvents for poly(methyl methacrylate) (PMMA). (12) Alkyds are essentially higher esters and linseed oil comprises a large concentration of the oil used in preparation of the particular alkyd used in the current example. Thus, as more MMA is converted to polymer, the tendency for that component to phase separate from the alkyd becomes greater. Since the MMA monomer preferentially resides closer to the outer shell of the particle, due to its hydrophilicity, that phase separation process creates the driving force for the typical core/shell particle morphology to develop.

Hybrid miniemulsion polymerization of MMA in the presence of alkyd generally leads to a final polymer distribution of poly(alkyd-g-MMA) copolymer, neat alkyd, and homo-PMMA. This has been confirmed by gel permeation chromatography (GPC) molecular weight profiles. (13) This is partially due to the low grafting efficiency (transfer followed by termination) of the methacrylate chain to alkyd but is also due to the propensity for MMA to be involved in homonucleation in the aqueous phase. (14) Particles derived from homonucleation will be free of alkyd, since the hydrophobicity of alkyd prevents it from diffusion across the aqueous phase to these particles. (1) This, combined with incomplete grafting within the original hybrid particles, leads to the production of some degree of PMMA homopolymer. (13) In Figure 1B, smaller, homoparticles of PMMA are shown. (The presence of both hybrid particles and homoparticles of PMMA is confirmed by DLS particle size data for this material, as shown in Table 2.) The darker and larger particles are predominantly alkyd with an acrylic shell and the faint smaller ones are PMMA particles from homonucleation.

[FIGURE 2 OMITTED]

To resolve the PMMA particles in Figure 1B, contrast had to be enhanced to such a level resulting in a barely discernible PMMA shell on the hybrid particles. This is why the darker particles incorrectly seem to be neat alkyd particles. With a less dramatic contrast, in the same image file, the hybrid particle shell is distinguishable. The points to note from the images in Figure 1 is that the hybrid particles exhibit typical core/shell morphology, with alkyd as the core and a dominantly acrylic phase as the shell, and that there is clear visual evidence of some degree of smaller homoPMMA particles derived from homonucleation.

MMA/BA/AA/Alkyd Hybrid System (TEM)

With the morphology for the simple MMA/alkyd system established, the study was expanded to the common acrylic phase of MMA, BA, and AA. This copolymerization, of course, took place in the presence of alkyd resin via a miniemulsion.

In the acrylic copolymer/alkyd system, shown in Figure 2, the core/shell morphology was again observed. The acrylic copolymer consisted of a common industrial ratio of 49:50:1 (MMA:BA:AA, wt:wt:wt). The total oil-phase was a 50:50 (wt:wt) ratio of acrylic to alkyd. A slightly larger acrylic shell is observable in Figure 2 relative to Figure 1A, likely due to less homonucleated acrylic polymer. Compared to MMA, BA is quite water-insoluble and, thus, is not prone to homonucleation. (12,14) Homogenous nucleation is suppressed by the fact that the concentration of MMA with respect to the total oil-phase is halved compared to the aforementioned MMA/alkyd system, due to the presence of BA.

[FIGURE 3 OMITTED]

MMA and BA have reactivity ratios predicting a random copolymer with some compositional drift. One would then expect the shell to contain a relatively homogeneous distribution of MMA and BA. In Figure 2B the close hexagonal packing of particles is due to the higher concentration of polymer in the deposited TEM sample, as compared to that used in Figure 2A. Core/shell morphology remains evident in the image and is emphasized in the close-up shown in Figure 2C. The acrylic shell is slightly distorted due to damage imparted by the electron beam, but the general spherical particle shape is relatively intact. The glass transition temperature ([T.sub.g]) of MMA-co-BA is much lower than that of MMA, due to the low [T.sub.g] of PBA, (-56[degrees]C). (12) The relatively soft polymer allows the electron beam of the TEM to distort the particle shape if the residence time under the beam is too long.

[FIGURE 4 OMITTED]

BA/Alkyd Hybrid System (TEM)

The next system was one in which only BA was used for the acrylic component. This system differs from the MMA/alkyd system in several respects.

(1) BA is essentially water insoluble (12) and will not be involved in homonucleation. The developing hybrid particle will always have concentrations identical to the formulated recipe, assuming no interparticle monomer transport. In reality, there will be transfer of monomer from unnucleated droplets, but nonetheless when compared to the MMA/alkyd system intraparticle component concentrations will be more similar to the recipe due to the lack of homonucleation. In addition, the water-insolubility of BA will lessen the tendency for it to migrate to the particle/water interface compared to MMA.

(2) BA has a much stronger tendency to form grafts with alkyd than MMA. BA reacts quite easily with alkyd, mostly through direct addition to an alkyd fatty acid double bond while direct addition in MMA is sterically hindered, and grafting must proceed by hydrogen abstraction. (13)

(3) The [T.sub.g] of PBA is -56[degrees]C compared 105[degrees]C for PMMA. (12) (A good estimate for the [T.sub.g] of the alkyd used is-30[degrees]C.) The low [T.sub.g] of PBA translates to a hybrid particle [T.sub.g] well below the reaction temperature of 70[degrees]C. This allows both components to be quite mobile. In the MMA/alkyd system, the high [T.sub.g] of PMMA immobilizes that component at high monomer conversion.

The BA/alkyd particle morphology is shown in Figure 3. Given the [T.sub.g] of this material, particle shape distortion due to the electron beam is not surprising. Nevertheless, the general spherical shape is clear. The gray coloring surrounding the particle is not a latex component; at this high magnification, it is an artifact of the support grid. What is significant is that in this system, the PBA component has become the continuous particle phase and the alkyd component small island domains dispersed throughout the particle. This results from the fact that PBA and alkyd are more compatible than PMMA and alkyd. Since PBA reacts quite easily with alkyd, mostly through direct addition to an alkyd fatty acid double bond, the resultant graft copolymer becomes bulkier and less mobile than free PBA chains. The thermodynamic equilibrium structure then becomes small alkyd island domains dispersed throughout a more mobile PBA continuous particle matrix.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Image contrast is adjusted in Figure 4 to emphasize only the alkyd component, washing the remaining image information into the background. Circular illustrations were added describing the particle dimension so that when the high contrast images on the right are compared with originals on the left, the total particle shape is understood. In performing the image contrast adjustment, some of the small alkyd domains in the top right image of the BA/alkyd system were also lost to the background, but the alkyd distribution throughout the particle is still clear. When compared to the alkyd location in the MMA/alkyd system (the lower half of Figure 4) the complete distribution of alkyd throughout the BA/alkyd particles becomes unmistakable. The MMA/alkyd system leads to a true core/shell distribution and BA/alkyd tends towards a complete distribution of small alkyd island domains within a continuous PBA particle matrix.

Scanning Electron Microscopy Analysis

Unlike TEM, scanning electron microscopy (SEM) records only film surface and topography. With SEM, electrons raster over the sample surface, whereas in TEM those electrons are transmitted through the entire sample thickness. (11) For most polymers imaged by SEM, a thin layer of gold must be sputtered over the dried polymer film and a small layer of conductive tape placed between the sample and support stub. Both of these preparative measures are to increase conductivity of the sample required to gain an image in the SEM. Polymers, in most cases, are insufficiently conductive on their own to be able to be imaged by SEM.

In the MMA/alkyd system, TEM revealed a core/shell morphology of the hybrid particles. SEM (Figure 5) provides information about particle morphology that was not observed by TEM. The image in Figure 5 is of the standard MMA/alkyd system initiated by potassium persulfate (KPS), the same system shown in Figure 1. Small light colored particles are observed dispersed over a matrix of semicoalesced darker particles. Compared to the reference bar at the bottom right of the image, those small particles can be estimated to be on the order of 25 nm in diameter. MMA is fully miscible with alkyd, but PMMA is not compatible with this particular alkyd. Therefore, as MMA is converted to polymer, it will gradually be less compatible with the alkyd and phase separation occurs. Since most of the hydrophilic MMA will tend to reside closer to the particle surface, the PMMA formed will create a shell. Many PMMA chains are grafted to the alkyd, though, so the formation of grafted bumps occurs on the particle surface instead of complete PMMA phase rejection. The result is an effectively continuous shell reminiscent of the surface of a raspberry.

At first, this seems in conflict with what is observed in Figure 1A by TEM. On the contrary, the results are consistent with each other when one considers that in TEM, electrons are transmitted through the whole thickness of the polymer and the image records data throughout. With a raspberry-like morphology, TEM electrons would combine information through the overlap of each bump, thus creating the mirage of a complete homogeneous shell. Referring again to Figure 1A, one notices that the particle surface is not smooth. Since SEM electrons only raster over the particle or film surface, the bumps are observed distinct from each other in Figure 5. Although there is some degree of coalescence, individual particles are easily observed with concentrated PMMA bumps over their entire surface.

This morphology is not restricted to MMA/alkyd systems initiated by KPS. When the oil-soluble initiator 2,2'-azobisisobutyronitrile (AIBN) is substituted for KPS, identical results ensue. Figure 6 shows the same morphology as in Figure 5 even with the oil-soluble initiator; individual particles are discernable with a high concentration of hard spherical bumps over their surface. Again, these raspberry bumps make up the outer shell of the MMA/alkyd hybrid particle and form on the surface due to the incompatibility between PMMA and the alkyd core. The estimation of PMMA bump diameter of roughly 25 nm can be determined from Figure 6B. The choice of initiator (water soluble KPS or oil soluble AIBN) does not appear to affect particle morphology.

[FIGURE 7 OMITTED]

Samples are not preferentially stained in SEM, as compared to TEM, and, thus, the dark regions in Figure 6 should not be considered evidence of alkyd. In fact, they are evidence of solvent escape from larger hybrid particles. The alkyd component is shipped at approximately 80% solids in solvent. During hybrid latex film formation, any solvent present in the alkyd phase will attempt to diffuse out of the particles. In certain cases, this can result in particle rupture leaving a crater behind and a partially empty particle. Further evidence of this is discussed in the next section.

SEM of the BA/alkyd system did not give useful results. Under the SEM electron beam, these soft particles are quickly distorted and lose their original shape. Coalesced and distorted particles were found offering no beneficial morphological insight for that system.

[FIGURE 8 OMITTED]

Atomic Force Microscopy Analysis

Atomic force microscopy (AFM) is a useful tool for the characterization of polymer surfaces. In this analysis, a probe is mounted on a thin cantilever and deflection of that cantilever offers a detectable signal. AFM provides a resolution of the same order as SEM (approximately 4 nm). In tapping mode, a high frequency alternating voltage vibrates the lever at a frequency at or near resonance. The resultant force gradient in the region where the tip is oscillating causes the resonant frequency to change, which in turn changes the amplitude and phase of the oscillations compared to the fixed drive frequency. In this manner, both surface height and phase data can be simultaneously collected. This is quite useful for the hybrid particles, providing height data to corroborate with data found by SEM along with phase data that can potentially resolve soft versus hard areas on the particle surface. In the MMA/alkyd system, a soft phase would correspond to the alkyd component and a hard phase to the PMMA phase.

The AFM height image in Figure 7 was taken on the same material as the SEM image in Figure 6C (MMA/alkyd with AIBN as the initiator). The raspberry surface observable in SEM is not resolved in the AFM data of Figure 7. Films for AFM analysis were prepared by spin coating onto a cleaned silicon wafer. The color gradient bar to the left of the AFM image indicates that the largest height measured between the film substrate and particle peak diameter was on the order of 250 nm. This agrees well with DLS and SEM data. In the Figure 7 AFM image, those same dark areas observed in the Figure 6C SEM image can now be attributed to holes from solvent escape as dark portions of the AFM images correspond to heights of less than 70 nm.

The phase image in Figure 8 (right) offers better morphological data than the height data. Hybrid particles, those that are roughly 200-250 nm in diameter, show evidence of solvent escape in more than several particles. This phase data more clearly shows the shape and dimension of the craters formed and one might even be able to observe the indications of a collapsed spherical shape to one resembling a doughnut. Smaller particles are also observed in Figure 8. As described earlier, many of these are considered to be homoPMMA particles derived from homonucleation.

Soft systems, such as BA/alkyd, are not well characterized by AFM, as tip interaction with a soft phase sample produces results strongly influenced by tackiness or other extraneous influences, and so no AFM results are available for such systems.

NMR Spin Diffusion Analysis

The TEM analysis allows one to determine the overall morphology of the particles, however, information about internal structures and interfaces or interphases are not accessible by microscopic measurements. Here, solid-state NMR measurements, namely spin-diffusion measurements, were shown to be suitable for the analysis of such internal structures. In combination with the information obtained by microscopic methods (TEM, SEM, and AFM), a detailed picture of the morphology of the particles' morphology can be given. For the performance of a spin-diffusion experiment, it is necessary to generate a nonequilibrium magnetization in the sample that contains at least two different components or phases with different NMR parameters. (17) The nonequilibrium can be achieved by using a dipolar filter or a chemical shift filter. (18) In the first case, the phases or components of the sample under study show different mobilities; in the latter case, the components or the phases can be distinguished by the chemical shifts in the proton spectrum. The dipolar filter can sensitively be used for the selection of regions with different mobilities. It is based on the application of multiple pulse homonuclear decoupling. By using different filter strengths, it is possible to distinguish between regions of different mobilities. The strength of the filter can be adjusted by changing the pulse length [t.sub.d] in the filter sequence. Short [t.sub.d] times represent a soft or weak filter; with increasing [t.sub.d] times the filter strength increases. With a soft filter, only the magnetization of rigid components with strong dipolar coupling is suppressed. The magnetization of all other components is detected after application of such a filter. With increasing filter strengths, the magnetization of components with higher mobility, and, therefore, less dipolar coupling, is also suppressed. In other words, the stronger the filter, the more mobility is required for detection after the application of the filter.

[FIGURE 9 OMITTED]

The abundance of magnetization in one phase and the lack in the other phase leads to a transfer of magnetization from one phase to the other. The magnetization transfer, called spin diffusion, takes place until the magnetization has reached equilibrium. The time for this process depends on the domain size and the spin-diffusion coefficient in both phases. It can be described with Fick's law: <[x.sup.2]>=a.D.t, where a is a factor which depends on the geometry of the system (a = 1 for a lamellar, a = 2 for a cylindrical, or a = 3 for a spherical system), D is the spin-diffusion coefficient and t is the diffusion time. (18) The spin-diffusion coefficients can be determined from the line width in the proton spectrum. (19) To get information about the magnetization of each phase versus time, the magnetization distribution in both phases can be read out in an NMR spectrum. Therefore, one can follow the decay of magnetization in one phase or the increase of magnetization in the other phase.

[FIGURE 10 OMITTED]

The intensity of the components has to be corrected by [T.sub.1] and is then plotted versus the square root of the mixing time ([t.sub.m.sup.1/2]). A typical curve is shown in Figure 9. For a quantitative evaluation, two values can be determined from the graph for interpretation of the data and for calculation of the volume to interface ratio, ([V.sub.tot]/[S.sub.tot])[[phi].sub.A]: the final value and [t.sub.m.sup.*1/2] (see Figure 9). For a certain filter strength, the final value corresponds to the amount detected as mobile with this filter. For the comparison of systems with varying core/shell content, it is more convenient to compare the different ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] values because then only the change of one component is monitored. (20) The ratio ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] can be calculated as:

[[V.sub.tot]/[S.sub.tot]][[phi].sub.A] = [1/[[phi].sub.B]][2/[square root of [pi]]][[[square root of [D.sub.A][D.sub.B]]]/[[square root of [D.sub.A]+[square root of [D.sub.B]]]]][square root of [t*.sub.m]] (1)

where

[V.sub.tot] is the total volume of the particle (in [nm.sup.3]);

[S.sub.tot] is the total interface area between the phases A and B (in [nm.sup.2]);

[[phi].sub.A], [[phi].sub.B] is the fraction of the phases A and B;

[D.sub.A], [D.sub.B] are the spin-diffusion coefficients of the phases A and B; and

[t*.sub.m] is the mixing time during which spin diffusion takes place.

Figure 10 shows the spin-diffusion curves for the sample MMA/alkyd (initiated by AIBN) measured at an elevated temperature of 50[degrees]C. The temperature of 50[degrees]C was chosen because at room temperature the mobility of the alkyd component was still too low to obtain a mobility difference that allowed the selection of this component, which is the prerequisite for the spin diffusion experiment. In these experiments four different filter strengths were chosen with [t.sub.d] = 10, 20, 40, and 50 [micro]s. The two important parameters of the curves are the final value and the slope, which is connected with [t*.sub.m]. The final value (listed in Table 3) corresponds directly to the mobile phase amount detected for each filter strength. The expected final value is, independent of the filter strength, 0.5 if the mobile alkyd and rigid PMMA were perfectly phase-separated. However, the spindiffusion curves sink (the slope increases and the final value decreases) with increasing filter strength, indicating that the mobile alkyd and the rigid PMMA influence each other, leading to mobilities within a wide range inside the sample.

At a weak filter strength ([t.sub.d] = 10 [micro]s), the final value is 0.70 which means that 30% of the magnetization of the entire particle is suppressed. This 30% can be attributed to rigid, pure PMMA, which is not affected by the mobile alkyd component. It therefore represents about 60% of the PMMA present in the particle. With increasing filter strength, not only the PMMA, but also the immobilized alkyd of an existing interphase, is suppressed, leading to the final value lower than the ones expected from the core-shell ratio of 0.5. The phase amount that is suppressed at the selected filter strength corresponds to a certain mobility. At [t.sub.d] = 10 [micro]s, phases with a low mobility are suppressed, at [t.sub.d] = 20 [micro]s, those with a slightly higher mobility are suppressed, and so on. The differences of the final values from one filter strength to the next one indicate the phase amount of different mobilities (see Table 4). It can be seen that in the sample MMA/alkyd (AIBN) high phase amounts with low mobility (corresponding to filter strength 0-10 [micro]s) are present indicating an efficient phase separation. At a slightly stronger filter ([t.sub.d] = 20 [micro]s), already 50% of the polymer is suppressed. At higher filters ([t.sub.d] [greater than or equal to] 40 [micro]s), the final value decreases down to 0.44 indicating that 44% of the particle is indeed highly mobile which can be assigned to the alkyd (The fraction of highly mobile alkyd can be calculated as the final value at [t.sub.d] = 50 [micro]s divided by the final value which is expected from the core/shell ratio (in this case 0.5). Only 12% of the alkyd is not detected at that high filter strength because it is immobilized in the interphase.

[FIGURE 11 OMITTED]

The comparison of the spin-diffusion curves for the MMA/alkyd samples initiated with different initiators (AIBN and KPS) shows that for weak filters (10 [micro]s) and for strong filters (50 [micro]s), the spin-diffusion curves differ slightly (see Figure 11). In the case of the KPS initiated sample, the final value obtained at a weak filter ([t.sub.d] = 10 [micro]s) is lower than that in the case of the AIBN initiated sample, indicating that more PMMA is detected as rigid (about 72% of the PMMA). At a filter strength of 50 [micro]s, 39% of the latex particles is detected as highly mobile phase. This means that compared to the AIBN initiated sample, the amount of immobilized alkyd increased from 12 to 22%.

It can be seen that in the sample MMA/alkyd (KPS), higher phase amounts with low mobility (corresponding to filter strength 0-10 [micro]s) are present than those in the AIBN initiated sample (0.36 instead of 0.30). At a slightly stronger filter ([t.sub.d] = 20 [micro]s), even 58% of the polymer is suppressed. At higher filters ([t.sub.d] [greater than or equal to] 40 [micro]s), the final value decreases down to 0.39 indicating that 39% of the particle is indeed highly mobile alkyd. From the differences for the final values at a filter strength of 10 and 50 [micro]s, it is now possible to determine the amount of polymer (acrylic and alkyd together) forming the interphase within a particle, as shown in Table 5. For the AIBN initiated particles, 26% of the particle is interphase with a PMMA/alkyd composition of 77 to 23 (which is calculated from the fact that 20% of the particle is interfacial acrylic and 6% interfacial alkyd). For the KPS initiated particles, 25% of the particles consist of interphase with a composition richer in alkyd (56 to 44).

From the values [t.sub.m.sup.*1/2] and the final values, the ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] were calculated and reported in Table 3. Since the final value corresponds to the phase amount of the detected phase, different zones can be detected. If the final value is higher than the expected one (e.g., for the latex MMA/alkyd (AIBN) up to [t.sub.d] = 20 [micro]s), there is information about the overall structure. If the final value is smaller ([t.sub.d] > 20 [micro]s), small structures in the interphase are characterized. The entire curves tell something about distribution of mobile and rigid areas in the particles. Therefore, the larger the ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] values in comparison to the corresponding phase amounts, the larger the structures built by this phase amount; whereas, the smaller the ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] values (at higher [t.sub.d]), the smaller are the highly mobile phase structures in the particles. In the case of the AIBN or KPS initiated particles, at weak filters, sizes of 25 and 17 nm respectively, (see Table 3) were detected, which fits perfectly to the values detected in the SEM pictures. (Note that in this case of the NMR experiment, the entire interphase is encountered to the domain sizes.)

To obtain the values for the entire structure of the particles, ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] for the appropriate filter strength, which represents the expected core/shell phase ratio, has to be determined by interpolation. (21) The values for the AIBN and KPS initiated samples are 10 and 11 nm, respectively, always far below the value of ideal core/shell particles (30 nm). With the knowledge of the TEM picture which shows clearly a core/shell structure, one can now get the following information about the NMR spin diffusion measurements. The particles do not consist of ideal core-shell structure, instead there are many small PMMA domains (in the case of KPS, smaller than in the case of AIBN) which are homogeneously distributed in the core. Pure MMA and pure alkyd phases exist, however, there are also interfacial regions which can be quantified in detail. Whereas in the microscopy differences between these two samples initiated with different initiators are difficult to detect, it is now the advantage of the NMR spin-diffusion technique to detect distinct differences in the morphologies and especially to quantify the interfacial regions.

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

The first two samples consisted of two component particles, the core and the shell were homopolymers (PMMA and alkyd), whereas the two other samples consisted of a homopolymer (BA)-alkyd phase and a copolymer phase of MMA and BA (MMA/BA/alkyd hybrid system) or MMA, BA, and AA (MMA/BA/AA/alkyd hybrid system). In the case of the copolymer samples, the structure inside the copolymer phase led to a much more complex morphology of the entire particle.

Comparison of spin-diffusion curves for the different samples in Figures 12 and 13 showed that for the weak filters (10 [micro]s), the first decay of the MMA/BA/AA/alkyd particles was quite similar to MMA/alkyd (AIBN) sample. Only the first slope was steeper for the copolymer particle, indicating smaller domains (see also Table 3, the ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] value decreased significantly to 10.2 nm) and might be explained with an increased compatibility between the core and shell components. The spin-diffusion behavior of the MMA/BA/alkyd hybrid system differed a lot, indicating a large change in the morphology in this sample. In this case, the final value was quite high (0.79) indicating a high mobilization of the rigid shell component. Only 21% of the material was detected as rigid. The ([V.sub.tot]/[S.sub.tot])[[phi].sub.A] value was calculated to be as large as 96 nm, which speaks indeed for large phases and a deviation from the core-shell structure. (See also Table 6.) At a stronger filter (50 [micro]s), the same behavior can be seen. Whereas the spin-diffusion curve of the MMA/BA/AA/alkyd system is very similar to the homopolymer/alkyd samples, the curve of the MMA/BA/alkyd hybrid system has a decreased slope and a higher final value of 0.49 (see Table 3). Whereas the interpolated filter strength for the expected value of 0.5 is between 16 and 20 [micro]s for the MMA-alkyd samples, for the copolymer particles, the value of 48.3 [micro]s is much higher for the MMA/BA/alkyd hybrid system.

The amount of polymer (acrylic and alkyd) forming the interphase is 27 and 30%, respectively, for latexes consisting of MMA/BA/AA/alkyd and MMA/BA/alkyd (see Table 5). Again, the interphase includes both the acrylic and the alkyd. However, for the copolymer particles, the influence of PBA makes the quantitative determination of each phase difficult and can only be a rough value. Thus, the homopolymer/alkyd particles exhibit a high fraction of highly mobile alkyd, consistent with a strong phase separation. The copolymer/alkyd particles exhibit a higher amount of interfacial polymer, suggesting less phase separation.

Whereas the morphology of the MMA/BA/AA/alkyd hybrid particles has many similarities to the homopolymer/ alkyd latex particles, the structures in the MMA/BA/alkyd hybrid system are significantly different. The BA/alkyd particles were not analyzed by NMR spin diffusion since mobility differences, which are essential for the method, were difficult to create.

CONCLUSIONS

The combination of data from a series of techniques was used to characterize the resultant particle morphology of hybrid miniemulsion polymerization. Hybrid systems comprised of the MMA and alkyd components led to raspberry-like core/shell morphology. The internal core was composed primarily of alkyd, as indicated in TEM through staining, where the acrylic PMMA polymer resided on or near the outer-particle shell. TEM confirmed the location of the acrylic component and SEM revealed that the shell was actually a concentrated surface comprised of small, roughly 25 nm diameter, PMMA spheres. This is not surprising when one considers that although MMA and alkyd are miscible, PMMA and alkyd are incompatible hence creating a phase separation induced PMMA migration away from the alkyd core towards the particle surface. Grafting between the two components anchors the raspberry bumps to the particle to keep from complete separation.

TEM, SEM, and AFM all confirmed the presence of homoPMMA particles in addition to the hybrid particles. This also is not surprising considering the use of a water-soluble initiator and the hydrophilicity of MMA monomer. The hydrophobicity of the alkyd component prevents it from diffusing across the aqueous phase to these newly formed particles leading to their homogeneous character. NMR analysis confirms significant phase separation in the PMMA/alkyd polymer.

Copolymer/alkyd systems including the soft BA component were also found to develop core/shell morphology. The compatibilizing effect of the BA component on MMA is likely what helped the particles form more typical smooth shell characteristics. NMR again confirms significant phase separation. However, when the hybrid system was comprised of a pure BA acrylic phase (BA/alkyd), a completely different morphology was observed. PBA and alkyd also are not fully compatible, but their differences are not as dramatic as those of MMA and alkyd. This is partially due to their similar hydrophobicity and glass transition temperatures. The resultant hybrid particle morphology for that system was observed to be a continuous particle-phase of PBA with small internally dispersed island domains of alkyd.

Table 1--Generalized Recipe for All Experiments

Component                               Concentration

Deionized water                         200 parts
Monomer (a)                              20 parts
Alkyd                                    20 parts (b)
Sodium dodecylsulfate (SDS)               0.01 mol/L water
Potassium persulfate (KPS)                0.01 mol/L water
2,2'-azobisisobutyronitrile (AIBN) (c)    0.5 wt% based on monomer
hexadecane                                4 wt% based on monomer

(a) Methyl methacrylate (MMA), butyl acrylate (BA), where appropriate.
(b) Accounting for solvent present in alkyd.
(c) Where appropriate.

Table 2--Particle Size from Dynamic Light Scattering

                  Avg. Hybrid Particle   Avg. Homoparticle
System                 Diameter          Diameter (if present)

MMA/alkyd (KPS)         201 nm                55 nm
MMA/alkyd (AIBN)        216 nm                42 nm
MMA/BA/AA/alkyd         234 nm                50 nm
MMA/BA/alkyd            220 nm                56 nm
BA/alkyd                220 nm                --

Table 3--Phase amounts ([[phi].sub.A]) of the mobile phase (= final
values), [t.sub.m.sup.*1/2], and the corresponding ([V.sub.tot]/
[S.sub.tot])

                                  Filter Strength
                                    10 [micro]s
                                                      [V.sub.tot]/
                                 [t.sub.m.sup.*1/2]  [S.sub.tot]
Sample            [[phi].sub.A]  [m[s.sup.1/2]]      [[phi].sub.A] [nm]

MMA/alkyd
 (AIBN)               0.70            27.6                24.3
MMA/alkyd
 (KPS)                0.64            23.1                16.9
MMA/BA/AA/alkyd
 hybrid system        0.59            15.9                10.2
MMA/BA/alkyd
 hybrid system        0.79            76.4                95.9

                                Filter Strength:
                                  20 [micro]s
                                                    [V.sub.tot]/
                                [t.sub.m.sup.*1/2]  [S.sub.tot]
Sample           [[phi].sub.A]  [m[s.sup.1/2]]      [phi]A [nm]

MMA/alkyd
 (AIBN)               0.50           18.6               9.8
MMA/alkyd
 (KPS)                0.42           17.7               8.0
MMA/BA/AA/alkyd
 hybrid system        0.37           10.2               4.3
MMA/BA/alkyd
 hybrid system        0.65           32.0              24.1

                                Filter Strength:
                                  40 [micro]s
                                                    [V.sub.tot]/
                                [t.sub.m.sup.*1/2]  [S.sub.tot]
Sample           [[phi].sub.A]  [m[s.sup.1/2]]      [[phi].sub.A] [nm]

MMA/alkyd
 (AIBN)              0.44           16.9                 8.0
MMA/alkyd
 (KPS)               0.40           15.7                 6.9
MMA/BA/AA/alkyd
 hybrid system       0.32            9.9                 3.8
MMA/BA/alkyd
 hybrid system       0.55           25.4                14.9

                                Filter Strength:
                                  50 [micro]s
                                                    [V.sub.tot]/
                                [t.sub.m.sup.*1/2]  [S.sub.tot]
Sample           [[phi].sub.A]  [m[s.sup.1/2]]      [[phi].sub.A] [nm]

MMA/alkyd
 (AIBN)              0.44            7.7                 3.6
MMA/alkyd
 (KPS)               0.39           12.1                 5.2
MMA/BA/AA/alkyd
 hybrid system       0.32            6.4                 2.5
MMA/BA/alkyd
 hybrid system       0.49           13.0                 6.7

                 Corresponding Filter   ([V.sub.tot]/[S.sub.tot])
                 Strength to Reach      [[phi].sub.A] Interpolated
                 Calculated  Expected   and from the Measurement
                 Final Value
Sample               [[micro]s]                 [nm]

MMA/alkyd
 (AIBN)                 20.0                     9.8
MMA/alkyd
 (KPS)                  16.4                    11.4
MMA/BA/AA/alkyd
 hybrid system          14.1                     7.8?
MMA/BA/alkyd
 hybrid System          48.3                     8.3?

Table 4--Phase Amount Corresponding to Each Filter Strength
Representing Certain Mobility

Sample             0-10 [micro]s  10-20 [micro]s  20-40 [micro]s

MMA/alkyd (AIBN)       0.30          0.20             0.06
MMA/alkyd (KPS)        0.36          0.22             0.02
MMA/BA/AA/alkyd
  hybrid system        0.41          0.22             0.05
MMA/BA/alkyd
  hybrid system        0.21          0.14             0.10

                                   Fraction of Highly
Sample             40-50 [micro]s    Mobile Alkyd

MMA/alkyd (AIBN)        0                 88%
MMA/alkyd (KPS)         0.01              78%
MMA/BA/AA/alkyd
  hybrid system         0                 --
MMA/BA/alkyd
  hybrid system         0.06              --

Table 5--Amounts of the Polymer Forming the Interphase. The Amount of
Polymer Forming the Interphase (Acrylics and Alkyd) Is Obtained from
the Differences of the Final Values of the Spin-Diffusion Curves

                                Interphase
Sample                          Acrylics+Alkyd  Acrylics  Alkyd

MMA/alkyd (AIBN)                  0.26           0.20     0.06
MMA/alkyd (KPS)                   0.25           0.14     0.11
MMA/BA/AA/alkyd hybrid system     0.27           0.09     0.18
MMA/BA/alkyd hybrid system        0.30           0.29     0.01

Table 6--Theoretically Calculated [V.sub.tot]/[S.sub.tot] [[PHI].sub.a]
Values for Three Different Structures for Isolated Latex Particles with
a Core Diameter Of 200 nm

                                  [V.sub.tot]/[S.sub.tot]
                                  [[phi].sub.A] (theoretical)
Structure                                   [nm]

Core/shell                 [??]              33
Hemispheres (10% covered)  [??]             340
Half-covered               [??]              66

References

(1) Wang, S.T., Schork, F.J., Poehlein, G.W., and Gooch, J., "Emulsion and Miniemulsion Copolymerization of Acrylic Monomers in the Presence of Alkyd Resin," J. Appl. Polym. Sci., 60, 2069 (1996).

(2) Nabuurs, T., Baijards, R.A., and German, A.L., "Alkyd-Acrylic Hybrid Systems for Use as Binders in Waterborne Paints," Prog. Org. Coat., 27, 163 (1996).

(3). Wu, X.Q., Schork, F.J., and Gooch, J.W., "Hybrid Miniemulsion Polymerization of Acrylic/Alkyd Systems and Characterization of the Resulting Polymers," J. Polym. Sci., (Polymer Chemistry), 37, 4159-4168 (1999).

(4) Tsavalas, J.G., Gooch, J.W., and Schork, F.J. "Water-Based Crosslinkable Coatings via Miniemulsion Polymerization of Acrylic Monomers in the Presence of Unsaturated Polyester Resin," J. App. Polym. Sci., 75, 916-927 (2000).

(5) van Hamersveld E., van Es, J., and Cuperus, F.P., "Oil-Acrylic Hybrid Latexes as Binders for Waterborne Coatings," Colloid Surf. A: Physicochem. Eng. Asp., 153, 285 (1999).

(6) Dong, H., Gooch, J.W., and Schork, F.J., "Water-Borne Oil-Modified Polyurethane Coatings via Hybrid Miniemulsion Polymerization," J. Appl. Polym. Sci., 76, 105-114 (2000).

(7) Kawahara, H., Goto, T., Ohnishi, K., Ogura, H., Kage, H., and Matsuno, Y., "Preparation Of Epoxy Resin/Acrylic Composite Latexes by Miniemulsion Polymerization Method," J. Appl. Polym. Sci., 81, 128 (2001).

(8) Jeong P., Dimonie, V.L., Daniels, E.S., and El-Aasser, M.S., Proc. American Chemical Society, Div. Polymeric Materials: Science and Engineering, 82 (2000) 305.

(9) Karlsson, O., Hassander, H., and Wesslen, B., "Influence of Seed Polymer Molecular Weight on Polymerization Kinetics and Particle Morphology of Structured Styrene-Butadiene Latexes," J. Appl. Polym. Sci., 77, 297 (2000).

(10) Landfester, K., "Polyreactions in Miniemulsions," Macromol. Rap. Comm., 22 (12) (2001), 896-936.

(11) Sawyer, L.C. and Grubb, D.T., Polymer Microscopy, Chapman and Hall, London, 1987.

(12) Brandrup J. and Immergut, E.H., Polymer Handbook, Second Edition, John Wiley & Sons, New York, 1975.

(13) Tsavalas, J.G., Luo, Y., and Schork, F.J., "Grafting Mechanisms in Hybrid Miniemulsion Polymerization," Polym. Mater. Sci. Eng., 87, 507 (2002).

(14) Gilbert, R.G., Emulsion Polymerization: A Mechanistic Approach, Academic Press, London, 1995.

(15) Slawinski, M., Strategic Aspects of the Incorporation of Acrylic Acid in Emulsion Polymers, Ph.D. Thesis, Eindhoven Institute of Technology, 1999.

(16) Kirsch, S., Landfester, K., Shaffer, O., and El-Aasser, M.S. "Particle Morphology of Carboxylated Poly(N-Butyl Acrylate)/Poly(Methyl Methacrylate) Composite Latex Particles Investigated by TEM And NMR," Acta Polym., 50, 357 (1999).

(17) Schmidt-Rohr, K. and Spiess, H.W., Multidimensional Solid-State NMR and Polymers, Academic Press, San Diego, 1994.

(18) Clauss, J., Schmidt-Rohr, K., and Spiess, H.W., "Determination of Domain Sizes in Heterogeneous Polymers By Solid-State NMR," Acta Polym., 44, 1 (1993).

(19) Spiegel, S., Schmidt-Rohr, K., Boeffel, C., and Spiess, H.W., "Proton Spin Diffusion Coefficients of Highly Mobile Polymers," Polymer, 34, 4566 (1993).

(20) Mellinger, F., Wilhelm, M., Spiess, H.W., Baumstark, R., and Haunschild, A., "Quantitative Determination of Dynamic Heterogeneities in Core-Shell Lattices By 1h Solid-State NMR," Macromol. Chemistry & Physics, 200, 719 (1999).

(21) Landfester, K., Dimonie, V.L., and El-Aasser, M.S., "The Evaluation of the Size and the Structure of the Interphase in Composite Particles Containing a Macromonomer Studied by Solid-State NMR," Macromol. Chem. Phys., 203, 1772 (2002).

John G. Tsavalas** and F. Joseph Schork ([double dagger])--Georgia Institute of Technology*

Katharina Landfester--Max Planck Institute of Colloids and Interfaces ([dagger])

*School of Chemical Engineering, Atlanta, GA 30332-0100, [e-mail]; [e-mail].

([dagger]) Golm 14424 Potsdam, Germany; [e-mail].

** Current address: The Dow Chemical Company, 1702 Building, Midland, MI 48674.

([double dagger]) Author to whom correspondence should be addressed.