Controlled coagulation of emulsion polymers.
Keywords: Stabilization, acrylics, latexes, colloids, emulsions, coagulation, powder properties
Although most emulsion polymers are used in the aqueous state, some applications require that the polymer be isolated and used as a dry powder. Examples include dry coatings formulations and additives for plastics formulations. Two common isolation processes are spray drying and coagulation, both of which tend to yield relatively broad particle size distributions. This paper discusses a unique type of controlled coagulation process which yields dense, spheroidal aggregates. These aggregates can then be dried to form a powdered product. Applications include those requiring very good powder flow properties in addition to the usual benefits of unique morphology and composition afforded by emulsion polymerization.
Diffusion and Reaction Limited Aggregation
Conventional coagulation involves adding a coagulant (e.g., C[a.sup.++] ions) to the latex in excess of the critical coagulation concentration (CCC). This causes coagulation of the latex particles in a matter of milliseconds and is termed diffusion limited aggregation (DLA). In DLA latex particles form aggregates as quickly as they can diffuse together and collide. (1) Each collision results in the particles "sticking" together. If slightly less coagulant than the CCC is used, then the coagulation rate is slowed to the timescale of seconds instead of milliseconds and mixing can be used to control aggregate size and shape. This is termed reaction limited aggregation (RLA) and the latex particles may collide several times before "sticking" together. (2,3) The result is a denser, more uniform aggregate than is formed in the DLA process. (4-7) RLA usually occurs in the region between 5-20% volume solids and is quite unexpected and unique. Below about 5% solids there is incomplete coalescence and above about 20% solids the viscosity during coagulation becomes prohibitively high.
Coagulation Rate Theory
The differences between DLA and RLA can best be explained using the DLVO theory of particle interaction energies (8,9) and the Smoluchowski theory of coagulation rates. (10) In DLVO theory, the total interparticle potential energy for two particles approaching one another, [V.sub.tot], is expressed as
[V.sub.tot] = [V.sub.vdw] + [V.sub.elec] (1)
where [V.sub.vdw] is the van der Waals attractive energy and [V.sub.elec] is the electrical repulsive energy. [V.sub.vdw] is a function of the polymer composition (i.e., its Hamaker constant) and the interparticle distance. [V.sub.elec] is a function of the interparticle distance and the particle size and its surface potential, usually approximated by the zeta potential.
By carefully adjusting the amount of coagulant in the process, the zeta potential is affected and consequently the potential energy barrier height, [V.sub.max], is thereby adjusted to yield a slow, controlled coagulation rate. Figure 1 shows the interparticle potential energy curves for a 100 nm diameter latex with no coagulant (-55 mv), enough coagulant to be in the RLA regime (-30 mv), and enough coagulant to exceed the CCC and reach the DLA regime (-18 mv). In this example, for RLA, [V.sub.max] is 11 kT, whereas it is 0 kT for the DLA regime and 57 kT for the system without any added coagulant. According to Smoluchowski theory, the coagulation half-life, [t.sub.1/2], for DLA having a particle number density of [n.sub.0] (expressed as number of particles per cc) is given by the following equation:
[FIGURE 1 OMITTED]
[t.sub.1/2] = 2 X [10.sup.11]/[n.sub.0] (2)
For a 100 nm diameter latex at 10% solids, the [t.sub.1/2] would be about 0.001 sec. Smoluchowski theory does not strictly apply to such concentrated systems, but it has proven to be a useful approximation. The stability ratio, W, is the ratio of the DLA (i.e., fast) to RLA (i.e., slow) coagulation rates and is expressed by the following equation:
W = exp ([V.sub.max]/kT) (3)
For RLA of the 100 nm latex, the [V.sub.max] is 11 kT, so equations (2) and (3) can be used to get an approximate [t.sub.1/2] of about 60 sec. The RLA regime occupies a narrow band of zeta potential-particle size space (Figure 2). Therefore, for a given latex particle size, the zeta potential must be adjusted carefully within this narrow band to yield a slow, controlled coagulation.
[FIGURE 2 OMITTED]
Aggregate Particle Size Distributions
Controlled coagulation in the RLA regime yields much narrower particle size distributions than are achievable with DLA. Figure 3 is a comparison of representative distributions both having an average size of about 200 [micro]m in diameter. Note the large amount of oversized and undersized particles in the DLA curve compared to the RLA curve. Aggregate morphology also differs significantly.
The difference in size distribution and morphology between RLA and DLA aggregates is striking. Figure 4 contains micrographs of the two types of aggregates. Note the narrow size distribution and spheroidal shape of the RLA aggregates. In contrast, the DLA aggregates are irregularly shaped, with a high number of oversized and undersized particles. The density of RLA aggregates is also much higher than for DLA aggregates. This is in accordance with theory relating to aggregate formation in RLA and DLA, (1-3) but the density is even higher than theory predicts, due to latent mobility of the primary particles in the aggregates.
The latexes used in this study were produced by redox initiated seeded emulsion polymerization run under monomer-starved conditions to ensure that the composition remained constant throughout preparation. The surfactant was 1% sodium dodecylsulfate based on monomer. Monomers included butyl acrylate (BA) and methyl methacrylate (MMA), but no acid (methacrylic acid (MAA) or acrylic acid (AA)). Table 1 lists the copolymer compositions along with the particle diameters determined by photon correlation spectroscopy and glass transition temperatures ([T.sub.g]) determined by differential scanning calorimetry.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
All RLA coagulations were run in beakers stirred with a 45[degrees] pitched-blade turbine agitator. The beakers were set in a thermostatted bath for temperature control. All latexes were coagulated at 10% polymer solids with CaC[l.sub.2] solution at pH=4. Coagulant concentration was adjusted to keep the coagulation rate within the RLA regime. Zeta potentials were measured using a Malvern Zetasizer[TM] III. Final aggregate particle size was determined by light scattering and confirmed by microscopy.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
RESULTS AND DISCUSSION
As one would expect, mixing has a significant effect on aggregate size during the RLA process. The data are linear on a plot of log aggregate size versus agitator speed (Figure 5). The higher the speed, the smaller the aggregate size. The shearing action of the mixing tends to break up the weaker aggregates and the collisions of aggregates tend to densify them. (4,6) This is depicted schematically in Figure 6, which is based on observations made with optical and scanning electron microscopes. Because of the weak repulsive energy barrier in RLA, the individual latex particles have some time to rearrange into very dense aggregates before the final structure is set. (11)
Latex Composition and Coagulation Temperature
One of the principal latex properties that affects the DLA aggregate is the copolymer composition and its resulting [T.sub.g]. Clearly the temperature at which the process is run will also be a significant factor. At a given process temperature, the aggregate size increases steeply as the %BA increases and the [T.sub.g] of the copolymer is exceeded (Figure 7). This makes sense because the polymer is "stickier," allowing easier interchain penetration from colliding particles, when the [T.sub.g] is exceeded. For example, the 100 MMA latex ([T.sub.g]=117[degrees]C) shows no temperature effect within the process temperature range of 10-50[degrees]C, whereas for 50 BA/50 MMA ([T.sub.g]=13[degrees]C) there is a difference in the aggregate size between 10 and 50[degrees]C process temperature.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Latex Particle Size and Coagulation Temperature
The other principal latex property which could affect aggregate size is the latex particle size. To test this idea, 50 BA/50 MMA latexes of particle diameters 76, 125, and 300 nm were coagulated at constant agitation speed at 10, 25, and 50[degrees]C. As expected, increasing temperature increases aggregate size (Figure 8). An unexpected result is that the latex particle size had no effect on the aggregate size. Apparently the micro-eddy size created in the turbulent mixing of the process is also an important factor in determining aggregate size, especially when latex particle size is the main property being varied.
[FIGURE 9 OMITTED]
RLA produces dense, spheroidal aggregates having a narrow size distribution. One of the major advantages this gives over conventional DLA is that the flow properties of the dried polymer powder are greatly improved. In addition, it is safer to handle due to the reduced dustiness caused by fines. Figure 9 shows the results of a powder flow test clearly demonstrating the superior flow and decreased dustiness of RLA-produced powder over DLA-produced powder.
When the coagulation rate of latexes is slowed so that the coagulation time is in the range of about 1-120 sec, reaction limited aggregation can occur. This can unexpectedly result in the formation of dense, uniform spheroidal aggregates. Both process and latex variables can affect the aggregate size. The following conclusions have been drawn based on controlled coagulation (i.e., RLA) process studies of these variables presented here:
(1) Unlike conventional coagulation (i.e., DLA), RLA allows dense, spheroidal aggregates of latex particles to be formed in the 5-20% polymer solids range.
(2) Increasing mixing rate decreases aggregate size and increases aggregate density.
(3) Increasing coagulation temperature (above the copolymer [T.sub.g]) increases aggregate size.
(4) Reducing the [T.sub.g] of the copolymer increases aggregate size.
(5) Varying the latex particle size has little effect on the aggregate size.
(6) Dried aggregates from RLA flow better and are less dusty than those formed from DLA.
Table 1 -- Characteristics of Latexes Used for DLA Studies Composition Particle Diameter (nm) [T.sub.g] ([degrees]C) 0 BA/100 MMA 104 117 25 BA/75 MMA 121 62 35 BA/65 MMA 120 42 50 BA/50 MMA 76,125,300 13 60 BA/40 MMA 111 -5
The author would like to thank the Rohm and Haas Co. for support and for permission to publish this work.
* P.O. Box 904, Spring House, PA 19477-0904; firstname.lastname@example.org.
(1) Botet, R. and Jullien, R., "Size Distribution of Clusters in Irreversible Kinetic Aggregation," J. Phys. A: Math. Gen., 17, 2517 (1984).
(2) Botet, R. and Jullien, R., "Hierarchical Model for Chemically Limited Cluster-Cluster Aggregation," J. Phys. A: Math. Gen., 17, L639 (1984).
(3) Family, F., Meakin, P., and Vicsek, T., "Cluster Size Distribution in Chemically Controlled Cluster-Cluster Aggregation," J. Chem. Phys., 83, No. 8, 4144 (1985).
(4) Yusa, M., "Mechanisms of Pelleting Flocculation," Int. J. Miner. Process, 4, 293 (1977).
(5) Yusa, H., Hoshino, M., and Isaka, H. (to Kureha Kagaku Kogyo Kabushiki Kaisha) UK Patent GB 2157297 (1988).
(6) Iwasaki, T., Momose, K., and Sakabe, H., "Particle Formation from Polymer Emulsion under Slow Coagulation Condition," Kagaku Kogaku Ronbunshu, 16, No. 4, 627 (1990).
(7) Hopper, M., Patel, R., and Kmiecik-Lawrynowicz, G. (to Xerox Corp.) U.S. Patent 5525452 (1996).
(8) Derjaguin, B. and Landau, L., Acta Physicochim., USSR, 14, 633 (1941).
(9) Vervey, E. and Overbeek, J., Theory of the Stability of Lyophobic Colloids, Elsevier, Amsterdam, the Netherlands, 1949.
(10) von Smoluchowski, M., Bull. Int. Acad. Polon. Sci., Classe Sci. Math. Nat., 184 (1903).
(11) Mills, P., Goodwin, J., and Grover, B., "Shear Field Modification of Strongly Flocculated Suspensions--Aggregate Morphology," Colloid Polym. Sci., 269, 949 (1991).
Edward Kostansek -- Rohm and Haas Company*
|Printer friendly Cite/link Email Feedback|
|Date:||Jan 1, 2004|
|Previous Article:||Optimizing nitroxide-mediated miniemulsion polymerization processes.|
|Next Article:||Knowledge-based control of emulsion polymerization: tailoring adhesive properties.|