Functional latex and thermoset latex films.

We review advances in the design and development of functional latex particles that can be used to form crosslinked coatings. Our emphasis is on understanding fundamental principles, of the formation and aging of latex films, of crosslinking of polymer films, of the reaction mechanisms that lead to bond formation, and of the competition between bond formation and polymer diffusion in latex films. These principles form the basis for the design of modern coatings that combine high performance with environmental compliance.

Keywords: Emulsion polymers, polymer diffusion, energy transfer, neutron scattering, latex film formation, crosslinking, gel content, melamine, acetoacetoxy esters, carbodiimide, aziridine, oxirane, alkyd resins, autooxidative cure

**********

In a previous series of reviews, Bufkin and Grawe (1) did a masterful job of covering a broad variety of crosslinking technologies available before 1978. Their technology series described the many crosslinking chemistries that produced homogeneous or interfacially crosslinked coatings. This review discusses more recent developments in crosslinking technology, and the balance between crosslinking and film formation rates that affect film properties. Commercial crosslinking chemistries are emphasized at the expense of more esoteric crosslinking chemistries. For this review, latex is defined as a waterborne dispersion of polymer particles prepared using emulsion polymerization. Monomers include styrene, methacrylates, acrylates, vinyl esters, or any monomers that polymerize via emulsion polymerization. A waterborne coating is defined as an organic film formed from waterborne particles over any substrate. Thus, an organic film cast from an acrylic latex over textile fibers would be considered a coating.

The coating markets that require crosslinking and film formation are large, to say the least. The total market for waterborne products for architectural, OEM products, and special purpose coatings such as industrial maintenance, automotive refinish, and traffic paints is estimated to be more than two billion pounds of dry resin per year. The current markets for waterborne resins used in the U.S. are shown in Figure 1. The large volume of waterborne polymers used in these market, provides a strong incentive for designing crosslinking chemistries that optimize film formation and coating performance.

This review is divided into six parts. Following the introduction, in order to put thermoset coatings into their proper perspective, we provide a historical review of latex technology, crosslinking, and film formation. The next section, Crosslinking Theories, presents an overview of theories that describe the properties of crosslinked networks. In the Film Formation Mechanism section, we review briefly the mechanism of film formation from latex dispersions, emphasizing the importance of polymer diffusion in latex films for the development of useful mechanical properties. This section also focuses on the relationship between film formation and crosslinking. The fifth section, Crosslinking Chemistries, provides a detailed overview of current chemistries employed in crosslinking latex systems. The final section is a Summary, in which we try to bring together the various topics presented in the article. We emphasize our view, that if synthesis and applications chemists are to be able to design and formulate waterborne coatings with optimum properties, the science of crosslinking and film formation must be understood in the context of their relationship to one another.

HISTORICAL PERSPECTIVE

In 1912, Kurt Gottlob (2) received the first patents for the invention of emulsion polymerization using diene monomers. Another 20 years passed before Luther and Hueck (3) received a patent for their work in emulsion polymerization (Figure 2). The efforts of Gottlob, Luther, and Hueck created the need to develop crosslinking chemistries and film formation theories for coatings based on latex technology. Melamines, initially developed as crosslinkers for solventborne polymers in the mid 1930s, became the mainstay crosslinking technology in the late 1950s for acrylic latex. Melamine crosslinking technology is still used today. (4)

Dow Chemical Company commercialized the first latex in 1946 using styrene and butadiene as the monomers. (5) This introduction was subsequently followed by AC-33, an all-acrylic latex manufactured by Rohm and Haas Company (6) and UCAR WC-130, a vinyl acetate-based latex manufactured by Union Carbide Corporation. (7) As shown in Figure 2, before the commercialization of acrylic and vinyl acetate-based latex, Flory (8) developed his theories on the swelling of network structures and gelation. In addition, Harkins, Smith, and Ewart (9) published the first theory on emulsion polymerization and Bradford and co-workers (10) published their dry-sintering theory of latex film formation.

[FIGURE 2 OMITTED]

The commercialization of synthetic polymer latex, however, created a need to improve the understanding of film formation of heterogeneous particles dispersed in water and to invent new crosslinking chemistries for these latex materials. Increasing concern about the environment, as well as technological advantages, has resulted in a continuous shift from solventborne to waterborne polymers. In 1967, the Federal Government enacted the Clean Air Act, and established the Environmental Protection Agency (EPA). In addition, the oil embargo in the early 1970s resulted in a worldwide energy shortage. The result was a search for crosslinking technologies that produced lower formaldehyde emissions from coatings, and provided cure at lower temperatures and faster than the "then" conventional crosslinking technologies. As shown in Figure 2, because of outside pressure to the coatings industry, research in developing new crosslinking technology for waterborne coatings began in earnest in the mid 1970s and continues today. The result was the development of crosslinking technologies that produced lower formaldehyde emissions from coatings, and provided cure at lower temperatures and faster than previous crosslinking technology.

The initial theories of film formation focused on particle compaction. In the dry-sintering theory proposed by Bradford et al., (11) the major contribution that drives particle compaction to a fully dense film is the surface tension of the polymer. According to Brown, (12) particle compaction resulted from the capillary forces that develop at the surface of the drying latex where capillary forces were proportional to the water-air interfacial tension and inversely proportional to latex particle size. Vanderhoff (13) proposed that particle deformation and densification resulted from water interfacial tension that provided the forces necessary for particle compression. Sheetz (14) suggested that the evaporation of water from latex films resulted in stresses that compressed particles. Routh and Russel (15) provided an up-to-date discussion of these various mechanisms, and developed a model which allows one to understand the circumstances under which each mechanism may predominate.

Voyutskii was the first to suggest that autohesion or polymer diffusion was necessary to develop mechanical properties. (16) He referred to autohesion as a mechanism that leads to the healing of weak particle boundaries. Studies by electron microscopy show that after latex particles compact and densify, the cells formed by the particles continue to change, leading to a continuous, smooth film. (11,17) The advent of surface-force microscopies such as atomic force microscopy, first applied to latex films in 1992, (18) allows one to monitor the time evolution of the film surface, which flattens as the film is allowed to age. These results suggest that molecular rearrangement or polymer diffusion occurs between particles. During the 1980s and 1990s, direct measurement of the diffusion of polymer chains in latex films became possible through the application of two techniques: small angle neutron scattering (SANS) and direct nonradiative energy transfer (DET). Hahn et al. (19) and Sperling et al. (20) monitored the diffusion of deuterated poly(butyl methacrylate) and deuterated polystyrene, respectively, using SANS. More recently, Winnik (21) developed the DET method for diffusion measurements of latex polymers. These techniques allow one to correlate the diffusion of polymer molecules across the intercellular boundaries in a latex film with the development of the mechanical strength of the film. They also open the possibility that one could monitor this diffusion in a thermoset latex film in which diffusion competes with the development of crosslinks.

[FIGURE 3 OMITTED]

CROSSLINKING THEORIES

Thermoset coatings have a number of important performance advantages over thermoplastic coatings. These include improved mechanical properties and increased resistance to solvent. For example, the shear and tensile moduli of the polymer increase in proportion to the crosslink density, and its degree of swelling decreases when exposed to solvent. It is well known that thermoplastic polymers that are soluble in a particular solvent lose their solubility when they are crosslinked. Two observations are made in the laboratory: crosslinked polymers swell to an equilibrium value that depends upon the interaction between the solvent and polymer, and often a measurable amount of polymer is not attached to the crosslinked matrix. As a result, this free polymer will dissolve in the solvent. The first observation can be quantified as the swell ratio Q the equilibrium volume of polymer swollen by solvent divided by the volume of the dry polymer. The second observation can be quantified as gel fraction, the fraction of crosslinked polymer that does not dissolve in a given solvent. The important molecular parameter connecting the shear modulus and the swell ratio is the crosslink density, characterized by the mean molecular weight between crosslinks, [M.sub.c].

Dynamic mechanical measurements on crosslinked elastomers at low deformation in extension provide the tensile (Young's) modulus [E.sub.o]. Oscillatory shear experiments yield the shear modulus [G.sub.0]. These terms are related, with [G.sub.0] = (1/3)[E.sub.o]. Simple theories of rubber elasticity provide a simple relationship between [G.sub.0] and the mean molecular weight [M.sub.c] between crosslinks. In the coatings field, one traditionally uses the predictions of the affine network model for which

[G.sub.0] = [v.sub.e]RT (1a)

[v.sub.e] = 1000 [rho]/[M.sub.c] (1b)

In these expressions, R is the gas constant, [rho] is the density of the bulk polymer (in g/mL), and T is the absolute temperature. From the magnitude of [v.sub.e], the molar concentration of mechanically effective chains, one can calculate [M.sub.c], or correspondingly, the mean number of monomers between crosslinks [N.sub.c] = [M.sub.c]/[M.sub.0], where [M.sub.0] is the monomer molar mass.

The main assumption of the affine network model is that the ends of the network strands (the crosslink junctions) are fixed in space and are displaced affinely with the whole network. The affine deformation model overestimates the stress (and the modulus) because the junctions are rigidly embedded in the network. In real networks, the junctions are not fixed in space but are free to fluctuate around their average positions. These fluctuations lower the free energy of the system by reducing the cumulative stretching of the network strands. In the phantom network model, the junctions inside the network are free to fluctuate. To keep the network from collapsing, the boundary conditions of the model require the junctions at the surface of the network to be fixed to the nonfluctuating boundary. (22) The phantom network model underestimates the stress (and the modulus) because the junctions are fixed only at the surface of the network and the chains are free to pass through one another. For this model, the magnitude of the modulus is reduced by a factor of (1 - 2/f). where f is the functionality of the junctions.

[FIGURE 4 OMITTED]

[G.sub.0] = [1000[rho]RT/[M.sub.e]](1 - 2/f) (2)

Crosslinking chemistries used in typical coatings commonly have functionalities of 3 or 4, although melamine formaldehyde crosslinking can have a theoretical functionality of 6. For high crosslink densities (small values of [M.sub.c]) the affine model represents the better description of network properties.

Measurements of [E.sub.o] can be carried out on freestanding elastomers films (i.e., films above their glass transition temperature [T.sub.g]). Oscillatory shear experiments using a parallel plate geometry can be used to follow the evolution of [G.sub.0] as a function of cure for resins like epoxy and melamine formaldehyde that cure to a high crosslink density. An alternative method to determine the crosslink density is to carry out solvent swelling experiments, often on free-standing films.

Flory-Rehner Theory

When a polymer of dry volume [V.sub.dry], crosslinked in the bulk state, is exposed to a solvent, the network absorbs solvent and swells to its equilibrium volume [V.sub.eq]. The swell ratio Q is defined as [V.sub.eq]/[V.sub.dry]. The volume fraction [[PHI].sub.2] of polymer in the swollen gel is equal to [Q.sup.-1]. In 1943, Flory and Rehner (23,24) published a theoretical description of the swell ratio, in which the chemical potential for solvent penetration into the polymer was balanced against the elastic deformation of the network. To describe the elastic response of the network, Flory and Rehner applied the affine network model leading to the expression.

[FIGURE 5 OMITTED]

[M.sub.c.sup.Aff] = [-[rho][V.sub.1]([1/[Q.sup.1/3]] - [2/fQ])]/[ln(1 - [[PHI].sub.2]) + [[PHI].sub.2] + [chi] [[PHI].sub.2.sup.2]] (3)

where [V.sub.1] is the molar volume of the solvent, and [chi], the Flory-Huggins parameter, characterizes the interaction energy between a solvent molecule and a monomer unit of the network polymer. The treatment of polymer swelling has been reviewed in several recent publications by Erman and Mark. (25,26) They show that in terms of the phantom network model,

[M.sub.c.sup.Ph] = [-[[rho][V.sub.1]/[Q.sup.1/3]](1 - [2/f])]/[ln(1 - [[PHI].sub.2]) + [[PHI].sub.2] + [chi] [[PHI].sub.2.sup.2]] (4)

Since methyl ethyl ketone (MEK) is used in the coatings industry to determine the solvent resistance of cured coatings, it is a useful solvent to measure the gel content and swell ratio of a crosslinked polymer. We use the properties of MEK as a solvent to demonstrate the swelling behavior of crosslinked polymers described by equation (3). The molar volume of MEK is 89.6 mL/mole. To use equations (3) or (4), one must independently determine the value of the [chi] parameter for the polymer in methyl ethyl ketone. For MEK and a styrene-acrylic polymer prepared by emulsion polymerization, an [chi]-parameter of 0.48 has been reported by Taylor and Bassett. (27)

[FIGURE 6 OMITTED]

The following assumptions are usually made when using equation (3): First, one assumes that the Flory [chi] parameter is a constant, and that it is not a function of [[PHI].sub.2], the volume fraction of the polymer in the swollen gel. Second, one assumes that if the [chi] parameter has been determined for a given polymer, then small changes in the composition of the polymer by chemically incorporating a small level of crosslinker into the matrix of the polymer do not affect the magnitude of [chi].

Others (28) have shown that the [chi]-parameter is not always constant as predicted by theory, but can vary nonlinearly with the volume fraction of polymer in the solvent. In particular, if the polymer or solvent possesses a significant dipole, [chi] may change with concentration. Further insights into crosslinking systems are possible, however, if one examines graphically the behavior predicted by equation (3). As shown in Figure 3a, as the [M.sub.c] of a crosslinked polymer decreases, the swell ratio difference between data plotted with different [chi]-parameter decreases. Figure 3b shows a comparison of data plotted using the affine Flory model and the phantom model. Again, as the [M.sub.c] decreases the swell ratio difference between the two models decreases for a given [chi]-parameter. As a result at low swell ratios, the [M.sub.c] can be estimated from the swell ratio with little knowledge of the absolute value of the [chi]-parameter. Swell ratios < 5 are values typical for waterborne coatings crosslinked to give solvent resistance. One measure of solvent resistance is the ability of a crosslinked film to with-stand the pressure and back and forth movement of a cloth saturated with a solvent. The saturated cloth is moved back and forth over the film until the substrate is exposed. One back and forth stroke is referred to as a double rub. For example, for a crosslinked film with a swell ratio < 5 and a gel fraction > 0.9 in methyl ethyl ketone, MEK double rubs > 100 for coatings with a dry film thickness of 1 mil (25 [micro]m) are typically obtained.

Gel Fraction

In a partially crosslinked polymer, the fraction of polymer that is insoluble in a good solvent for the corresponding linear polymer is called the "gel" fraction. The polymer that is not attached to the network dissolves in the solvent, and is referred to as the "sol" fraction. Flory has shown that the soluble fraction of polymer in equilibrium with the gel is related to the polydispersity of the polymer and the crosslinked density. (29) Equation (5) shows Flory's final derivation that relates the weight fraction of soluble material, [W.sub.s], to the weight fraction of species composed of y units, [w.sub.y], and the fraction [f.sub.x] of monomer units involved in crosslinks. Equation (5) cannot be solved explicitly for [W.sub.s]. However, the summation can be performed graphically using trial values of [W.sub.s] or by integration if [w.sub.y] can be expressed in a suitable analytical form.

[FIGURE 7 OMITTED]

[W.sub.s] = [[infinity].summation over (y=1)] [w.sub.y][1-[f.sub.x](1-[W.sub.s])][.sup.y] (5)

Equation (6) shows the relationship of the gel fraction of polymer, [W.sub.g], to the soluble fraction of polymer, [W.sub.s].

[W.sub.s] = 1 - [W.sub.g] (6)

Substitution of equation (6) into equation (5) gives equation (7), the relationship between gel fraction, polydispersity, and crosslink density.

[W.sub.g] = 1 - [[infinity].summation over (y-1)] [w.sub.y] (1 - [f.sub.x][W.sub.g])[.sup.y] (7)

Reiser and Pitts (30,31) at Eastman Kodak used equation (7) to develop a gel fraction model for negative photoresists. A suitable analytical expression was found for [w.sub.y], and equation (6) was cast into a form suitable for numerical integration. Computer plots showed the gel fraction of exposed polymers as a function of photospeed (and crosslink density) and a polydispersity index parameter. For the purposes of this review, their polydispersity index parameters were converted to express the numbers in terms of classical polydispersity terminology. Their results are plotted in Figure 4.

Figure 4 shows that you need more than one crosslinked monomer unit per weight-average polymer chain to obtain a measurable amount of gel fraction. Note that the number of crosslinks per chain is based on the weight-average molecular weight and not the number-average molecular weight of the polymer. The main point of this plot is to show that as the polydispersity increases more crosslinked monomer units are needed per polymer chain to obtain a given gel fraction.

For latex systems, the measurement of swell ratio and gel fraction allows one to determine the homogeneity of crosslinking within the cured film. For example, a crosslinked polymer with a gel fraction > 0.90 is fairly uniformly crosslinked whether the swell ratio is high or low. A gel fraction and a swell ratio of 0.8 and 8, respectively, show that 20% of the polymer is not chemically bound to the matrix of film. If a film with a swell ratio of 8 is re-cured to reduce the swell ratio to, for example, 4, and one finds little increase in the gel fraction, then the crosslinking in the film is not homogeneous, or the initial uncured polymer has a very high polydispersity. In summary, for the efficient crosslinking of monomers that participate in the crosslinking reaction, a polymer with low polydispersity is highly desirable, particularly if the monomer participating in the crosslinking reaction is expensive.

Other Crosslinking Models

Miller and Macosko (32) used a recursive approach and a probability generating function to compute network parameters such as the weight fraction of soluble material, the weight fraction of polymer that is part of the network, but not elastically active (chain ends), the concentration of junction points, the entanglement trapping factor, and the concentration of elastically effective network strands (crosslink density). They assumed the ideal network formation of Flory and Stockmayer, and published a FORTRAN program to calculate the network parameters.

Bauer and co-workers (33,34) used methods developed by Macosko and Miller to calculate the elastic effective crosslink density, [C.sub.el], for hydroxy-based low molecular weight polymers cured using hexamethoxymethyl melamine. They showed that for baked automotive coatings, [C.sub.el] correlated well with the physical properties of baked paint coatings. For [C.sub.el] values < 0.8 X [10.sup.-3] mol/g, the coatings tended to be under-cured, but for coatings with [C.sub.el] values > 1.25 X [10.sup.-3] mol/g, intercoat adhesion failure became a problem. Thus, the [C.sub.el] range defined an acceptable cure window. Bauer further developed Bauer's and Dickie's methodology and published a program that calculates the weight of the soluble fraction and the effective crosslink density past the gel point of a crosslinked coating. (35)

FILM FORMATION MECHANISM

In this section we consider the mechanism through which latex films form and their properties develop. This process can be divided into a number of stages, as indicated in Figure 5. As water evaporates from the dispersion, the particles come into proximity. When the forces accompanying drying exceed the modulus of the particles, particle deformation occurs to yield a void-free film comprised of space-filling polyhedral cells. At this stage, the film is still mechanically weak. In the final stage, film maturation involves polymer diffusion across the intercellular boundary to provide the entanglements that give strength to the film. These stages are not always well separated, and within each stage subtle features operate that often differ from system to system. Some factors (uniform particle size, low ionic strength) promote ordering of the dispersion in the liquid state. (36) It is likely that order in the fluid phase at high solids will strongly influence particle packing in the film. This is particularly important in films prepared from mixtures of latex ("latex blends") of different sizes or different compositions, where localization of the components or phase separation in the dispersion carries forward into the film. (37)

Transmission electron microscopy (TEM) images of freshly formed films allow one to see the cellular structure of the film, and freeze-fracture TEM images of these films indicate that fracture occurs predominately along the interface between adjacent cells. (17, 38) As the films are aged or annealed, there is a growth in the tensile strength of the films. At the same time, the cellular structure fades, and fracture becomes cohesive. In the language of polymer welding, the interfaces are said to "heal."

All latex particles have polar or ionic groups at their surface to provide colloidal stability. Many of the properties of latex coatings depend upon the final location of these groups in the film. In some systems, this polar layer surrounding the latex particles is sufficiently thick that it forms a continuous membrane in the freshly formed film. An example is shown in Figure 6 where the TEM images from the work of Distler and Kanig (39) show poly(butyl acrylate) (PBA) copolymer films prepared from a latex that contains 2% N-methylol-acrylamide (NMA), and 1% acrylic acid. As one can see in Figure 6A, the film has a regular but heterogeneous structure, in which polar polymer forms a web-like interconnected membrane separating the individual cells. In the dry state, the films are transparent and tough, with an extension-to-break of 400%. Upon exposure to water, they become turbid and weak. With light scattering, the wet films exhibit Bragg peaks, implying that the polar membranes are continuous, swollen with water, and regularly spaced. In the dry film, the boundary between the cells did not disappear even over two years of aging, even though the aging temperature (ambient temperature) was 50[degrees]C greater than the [T.sub.g] of the polymer. The authors comment that a similar web-like structure was found in films heated to induce crosslinking of the NMA groups. When a sample of the film shown in Figure 6A was stretched by 150% and sectioned and stained, the TEM image shown in Figure 6B was obtained.

[FIGURE 8 OMITTED]

More recently, Chevalier et al. (40) examined the membranes in films formed from linear styrene-butyl acrylate copolymer latex particles that were stabilized by a layer of essentially pure poly(acrylic acid) (PAA). These authors used a combination of TEM experiments, rehydration experiments with [D.sub.2]O, and SANS measurements to show that membranes in their films persist until heated above the [T.sub.g] of the membrane polymer. The break-up of the membrane becomes an important step in bringing the particle cores into intimate contact so that polymer inter-diffusion between the cores of adjacent cells can occur. (41)

Our emphasis in this review is on the maturation process through which the dry film evolves to achieve its desired mechanical properties.

Measuring Polymer Diffusion in Latex Films

Two techniques are now available to measure the rate of polymer diffusion, SANS and DET. Both techniques involve labeling the latex polymer to obtain contrast. In SANS experiments, one monitors the dimensions of deuterated microspheres introduced in small amounts into a dispersion of similar but unlabeled latex particles. In the drawing in Figure 7, the deuterated microsphere is labeled d, and the unlabeled surrounding matrix is denoted as h. The angular dependence of the scattering intensity provides the data from which one calculates the radius of gyration of the deuterated objects in matrix. As deuterated polymer diffuses, the particle radius grows in size until the component chains become randomized in the matrix. This technique was originally employed by Hahn et al. (19) at BASF to study diffusion rates in poly(butyl methacrylate) (PBMA) latex films, and by Klein, Sperling, and co-workers at Lehigh (20, 42) to study interdiffusion in melt-pressed films of polystyrene (PS) microspheres. Subsequently, Sperling et al. (43) examined films prepared from PS particles prepared by a miniemulsification technique from polymers of narrow molecular weight distribution. This methodology allowed them to examine interdiffusion rates for polymers of different chain lengths. Their most impressive result demonstrates, as predicted by theory, (44) that the film achieves full mechanical strength when the diffusion length is comparable to the radius of gyration of the latex polymer. (45-47)

[FIGURE 9 OMITTED]

In ET experiments, a fraction (typically half) of the latex particles are labeled with a small fraction of one fluorescent dye such as phenanthrene (Phe) or naphthalene (N) that can serve as a donor in an energy transfer experiment. The remaining particles are labeled with a second dye such as anthracene (An) that can act as the acceptor. In Figure 7B, the donor dye is denoted as D, and the acceptor dye as A. One excites the sample with a pulse of light at a wavelength absorbed selectively by the donor and measures the influence of the acceptor on the fluorescence decay rate of the donor. ET can occur when the dyes are in close proximity. For the donor-acceptor pair Phe/An, the maximum distance over which nonradiative energy transfer can occur is about 40 [Angstrom]. In the initial film, the dyes are confined to separate cells; ET can occur only across the interface between adjacent cells. (48) Polymer diffusion leads to mixing of Phe- and An-labeled polymers, leading to an increase in the quantum efficiency ([[PHI].sub.ET]) of energy transfer.

[FIGURE 10 OMITTED]

[GRAPHIC OMITTED]

[[PHI].sub.ET] = 1 - [[[infinity].[integral].0][I.sub.D](t')dt']/[[[infinity].[integral].0][I.sub.D.sup.0](t')dt'] (8)

[f.sub.m] = [[[PHI].sub.ET](t) - [[PHI].sub.ET](0)]/[[[PHI].sub.ET]([infinity]) - [[PHI].sub.ET](0)] (9)

To determine [[PHI].sub.ET], one measures donor fluorescence decay profiles [I.sub.D](t') following excitation of the film sample with a nanosecond or picosecond pulse of light. The decay profiles are fitted to an appropriate function and integrated to obtain the area under the decay curve. [[PHI].sub.ET] values are calculated as shown in equation (8), by comparison of the integrated normalized decay profile with that measured for a sample containing donor but no acceptor. While [[PHI].sub.ET] can in principle be determined with a simple fluorescence spectrometer, artifacts in calculating [[PHI].sub.ET] from fluorescence intensities obtained from film samples make the fluorescence decay method much more reliable. From data analysis based on equations (8) and (9), one obtains the parameter [f.sub.m], which represents the fractional growth in energy transfer due to polymer diffusion. This parameter tracks the extent of intercellular polymer diffusion, but the relationship between the parameter [f.sub.m] and the mass fraction or volume fraction of mixing is not simple. Nevertheless, even for broadly polydisperse systems, one can calculate from the experimental data mean apparent diffusion coefficients (49) that allow one to interpret quantitatively the changes in polymer diffusion rate with a change in temperature or with a change in additives such as coalescing solvents. Much more information about the distribution of donors and acceptors in the system is available by fitting the decay curve to models, which treat explicitly the distribution profiles created as a consequence of diffusion, (50) but this topic is beyond the scope of the current review.

This technique has been used to study temperature and molecular weight effects, (51) as well as the influence of polymer composition, (52) coalescing aids (53,54) and nonionic surfactants (55) on polymer interdiffusion rates. For example, high molecular weight PBMA, with [T.sub.g] = 30[degrees]C, has a negligible diffusion rate at ambient temperature. In contrast, a low molecular weight sample ([M.sub.w] = 34,000 ([M.sub.w]/[M.sub.n] = 2.5)) with [T.sub.g] = 21[degrees]C, has a measurable diffusion rate at 22[degrees]C. (56) ET measurements have also been employed to examine the effect of latex structure (particularly core-shell morphology) on the interdiffusion process. (57,58) SANS and ET measurements are to some extent complementary, in that they sample polymer diffusion on somewhat different length scales. In the ET experiment, the magnitude of [f.sub.m] approaches unity for polymer diffusion over a length comparable to the radius of the latex particles from which the film was formed. Thus, this technique is particularly sensitive to the early stages of polymer diffusion across the intercellular boundaries. The SANS experiment probes diffusion over longer length scales. Joanicot et al. (59) used SANS measurements to monitor polymer diffusion in films formed from latex particles with a layer of PAA at the surface. Annealing of the films at temperatures below the breakup temperature of the PAA membranes separating the individual cells led to diffusion of low molar mass polymers, but this had no effect on the mechanical strength of the film. Only temperatures that led to fragmentation of the membranes caused interdiffusion of the high molar mass constituents of the latex. At this stage the films became resistant to moisture. Kim and Winnik (60) used ET experiments to measure polymer diffusion in films formed from latex particles with a shell enriched in methacrylic acid. Here, the two phases (the -COOH-rich phase and the base polymer phase) were not so sharply delineated. Polymer diffusion was retarded but not suppressed. Neutralization of the acid groups with metallic hydroxides had a much more significant effect on retarding polymer diffusion in these films. This effect is likely connected to ionomer interactions in the rather diffuse membranes present in these latex films.

Film properties develop through the process of intercellular polymer diffusion, but little is known about the polymer diffusion rates in commercial latex coatings. When these coatings are formulated, these issues are subsumed into the more vague concept of "degree of coalescence." Formulators often judge the degree of coalescence of a latex film in terms of the difference between the cure temperature and the minimum film forming temperature (MFFT) of the dispersion. The MFFT measurement is carried out by drawing down a film of the dispersion onto a metal bar subjected to a thermal gradient and then passing dry air over the dispersion until the film is dry. The MFFT is taken to be the minimum temperature where one observes a clear and crack-free film. It is not uncommon in the coatings industry to assume that a substantial extent of polymer diffusion takes place at temperatures above but not far removed from the MFFT. A clear film is observed, and the visual signature of coalescence is assumed to indicate the healing of the interfaces in the film. This assumption is not correct for high performance coatings. For high molecular weight polymers designed for high performance applications, satisfactory diffusion rates are obtained only at temperatures 40[degrees]-50[degrees]C above the polymer [T.sub.g]. Thus, the cure temperature must be substantially larger than the [T.sub.g] of the polymer. To achieve this situation, one can choose a suitably high cure temperature, or one can add a plasticizing solvent to the latex polymer to lower its [T.sub.g]. For example, Taylor and Klots (61,62) modeled the plasticization and evaporation of plasticizing solvents from latex films prepared from a styrene/ethyl acrylate-based unstructured latex. To determine the initial [T.sub.g] of plasticized films, they determined the [T.sub.g]s (Table 1) and relative evaporation rates for commonly used plasticizing solvents in the coatings industry. The solvents [T.sub.g]'s varied from as high as -67[degrees]C to as low as -129[degrees]C. The average [T.sub.g] value for the solvents studied was -96[degrees]C.

Since volatile solvents evaporate from organic polymeric films, the [T.sub.g] depression of polymeric films by plasticizers decreases as the film ages. This phenomenon decreases the diffusion coefficient of polymer chains across interfacial boundaries. If the solvents leave too fast, adequate polymer chain mixing may not occur between particles for adequate properties to develop. Taylor and Klots also defined the [T.sub.g] evolution profile for a plasticized polymer with adequate corrosion resistance properties. The [T.sub.g] evolution of a film containing four solvents at 23[degrees]C was calculated and experimentally determined by differential scanning calorimetry (DSC). The initial [T.sub.g] of the film was -27[degrees]C. As the film aged, it initially exhibited first-order volatility-controlled solvent loss, followed by diffusion-controlled solvent loss. After seven hours, the rate of solvent loss in the film began a transition from volatility control to diffusion control at a film [T.sub.g] of 7[degrees]C (or 14[degrees]C below room temperature). Control experiments showed that when the difference between ambient temperature and the initial [T.sub.g] of the plasticized coating was less than 50[degrees]C, a significant decrease in salt-fog resistance (corrosion resistance) was observed. This work illustrates the difficulty in designing zero-VOC maintenance coatings prepared from waterborne latexes. In the context of the current concern about VOC emissions, an interesting strategy is to choose a crosslinking agent (such as hexamethoxymethylmelamine, HMMM) that also acts as plasticizer, promotes the rate of polymer diffusion, and ultimately reacts to become part of the coating. This strategy is discussed in the Crosslinking Chemistries section.

[FIGURE 11 OMITTED]

[GRAPHIC OMITTED]

For the purpose of this review article, we use the images of the film sections shown in Figure 6 to pose some interesting and as yet unanswered questions about the crosslinked network in this type of heterogeneous latex film. First, we would like to know whether crosslinking is confined to individual cells or whether it spans the intercellular boundary layer. Second, we would like to have information about the extent of intercellular polymer diffusion in these films. Is there interdiffusion in the boundary layer of the polar polymer at the particle surface? Is there interdiffusion of core polymer across the boundary layer? Finally, we would like to know if the wet-strength of the film and its solvent resistance could be improved by introducing a crosslinking agent that would specifically crosslink the web-like membrane in the boundary layer between the cells.

Crosslinking and Polymer Diffusion

In many thermoset latex systems, reactive functionality is present in all of the particles in the system. When these films undergo crosslinking, the final film properties are likely to be very sensitive to the relative rates of polymer diffusion and the crosslinking reaction. If the crosslinking rate is much faster than polymer diffusion, crosslinking will occur primarily within individual cells. The final film will resemble the film that one would obtain if one induced crosslinking prior to film formation. For example, precrosslinking latex particles models the effect of premature crosslinking on film formation and film properties.

Zosel and Ley (63) examined the mechanical properties of films formed from lightly crosslinked PBMA particles. They showed that these films remained brittle and were unable to develop significant tensile strength, even with extensive annealing at 60[degrees]C above the polymer [T.sub.g]. In contrast, films prepared from linear PBMA developed considerable tensile strength when subjected to a similar annealing treatment. These authors expressed the view that the difference between their films that developed toughness upon annealing and those that remained brittle is related to the crosslink density. They argued that films become brittle when the mean molecular weight between crosslinks [M.sub.c] becomes smaller than the mean molecular weight between entanglements [M.sub.e]. In their experiments, values of [M.sub.e] were determined from the plateau modulus of the linear uncrosslinked polymer.

A rather different result on films formed from crosslinked latex was reported by Tamai et al. (64) They examined a series of donor and acceptor dye-labeled P(BMA-BA) latex ([T.sub.g] [approximately equal to] 10[degrees]C) containing different amounts of ethylene glycol dimethacrylate (EGDMA), a crosslinking agent. They found that tough elastomeric films were formed, even for latex particles containing 4 mol% crosslinker. Here significant intercellular polymer diffusion occurred. Since each particle was fully crosslinked, this diffusion must be limited to the dangling ends of the polymer network. One imagines that the origin of the mechanical strength is related to the length of the dangling ends, which diffuse across the interface and penetrate into the matrix of the adjacent cell. The major difference between the Zosel and Lye experiments and those of Tamai et al. is the [T.sub.g] of the latex polymer. In the case of crosslinked PBMA, the tensile measurements were carried out below the [T.sub.g] (30[degrees]C) of the polymer, and the films were brittle. The tensile measurements on P(BMA-BA-EGDMA) latex films were carried out above the polymer [T.sub.g].

In spite of their good tensile properties, the P(BMA-BA-EGDMA) films disintegrated when exposed to organic solvents. The dangling ends which provide adhesion in the dry film lose their effectiveness when the polymer is swollen by solvent. To obtain films with good solvent resistance, polymer chains must not only span the interface, they must be locked in place through crosslinks on both sides of the interface. From this perspective, one concludes that for thermoset latex films to develop useful mechanical properties, solvent resistance, and low permeability to water, polymer diffusion must precede crosslinking. We summarize this idea in the drawing in Figure 8.

[FIGURE 12 OMITTED]

[GRAPHIC OMITTED]

Experimental Determination of Crosslinking and Polymer Diffusion Rates

Over the past several years, there has been a substantial effort, primarily from Winnik and co-workers in Toronto, to measure the rates of polymer diffusion in thermoset latex films, and to compare these with the rates of crosslink formation. A review of these experiments has recently appeared. (65) Here, we present a single example to show the type of data one can obtain and to illustrate how the formation of long chain branches and crosslinks affect the rate of polymer diffusion in the films. We present data from the work of Liu et al. (66) who examined poly(butyl acrylate-co-methyl methacrylate-co-isobutoxymethyl acrylamide) (PBA-MMA-IBMA) latex films containing a 4:1 weight ratio of BA:MMA and 2 wt% of the reactive IBMA functionality. The polymer had a [T.sub.g] of 10[degrees]C. The chemistry of IBMA crosslinking will be described in the Crosslinking Chemistries section. The films studied by Liu et al. were prepared from an equal mixture of 140 nm diameter donor- (Phe) and acceptor- (An) labeled particles. The Phe- and An-labeled particles were prepared by the incorporation of monomers 1 and 2 into the respective emulsion polymerization reactions (Figure 9).

In Figure 10 we show their results, which compare several features of the film maturation. This figure compares two films, with and without crosslinking chemistry. The PBA-MMA film has the same molecular weight and molecular weight distribution ([M.sub.w] = 200,000; [M.sub.w]/[M.sub.n] [approximately equal to] 2) as the PBA-MMA-IBMA film, which also contains 0.5 wt% of p-toluenesulfonic acid (PTSA) as an acid catalyst. The upper figure emphasizes the early time behavior for samples annealed at 80[degrees]C. One sees for the PBA-MMA-IBMA film that the gel content grows slowly over the first 10-15 min of annealing. The growth in [f.sub.m] for this film is significant, but slower than that for the nonreacting PBA-MMA film. In the lower figure, one sees on a longer time scale that the gel content of the PBA-MMA-IBMA film reaches its maximum value of 88% after 40 min, but the swell ratio continues to decrease, reaching a value of about 5 after 100 min. The latex particles were prepared from a common gel-free PBA-MMA seed latex representing 12% of its mass. Thus, 88% gel content implies complete gelation of the IBMA-containing polymer. The extent of mixing for the PBA-MMA film continues to grow and eventually reaches 1.0. In contrast, [f.sub.m] for the PBA-MMA-IBMA film reaches 0.43 and ceases to grow. Gelation has brought further diffusive mixing to a halt. Other experiments by Liu et al. compare PBA-MMA-IBMA films with and without acid catalyst. The acid-free films, at 80[degrees]C, remain gel free and undergo diffusion at the same rate as the PBA-MMA films formed from a polymer of similar molecular weight. The acid-containing films form gel and experience a reduction in the extent of polymer diffusion that depends on the nature of the acid catalyst, its concentration, and the annealing temperature.

Theory of Polymer Diffusion vs Crosslinking

A proper theory of polymer diffusion and crosslinking in latex films would describe not only the coupling of the diffusion rate and the chemical reaction rate, but it would also describe how the formation of long chain branches and crosslinks affect the adhesion at the intercellular interface. Two recent papers from the de Gennes group in France (67,68) make a major contribution toward reaching this goal. Their contribution was based in part on previous work (69) to describe the strength of adhesion between crosslinked elastomers, inspired in part by the classic experiments by Ahagon and Gent (70) on the adhesion between slabs of crosslinked siloxane elastomers as a function of crosslink density. In this model, dangling ends are seen to play the role of connector molecules. They promote adhesion without necessarily forming entanglements. The energy of adhesion, which resists crack propagation, is largely dominated by the friction associated with pulling out chains that span the interface. When individual chains can cross the interface multiple times, the strength of adhesion is strongly enhanced. Ji and de Gennes (71) suggest that this adhesion is so strong that if the system were in the glassy state, chain fracture might predominate over chain pullout.

[GRAPHIC OMITTED]

[GRAPHIC OMITTED]

Aradian et al. (67) considered a system in which the crosslinking agent and the reactive functional groups are uniformly distributed in the newly formed film. To simplify the complex problem associated with diffusion coupled with a chemical reaction, they restrict their model to a latex polymer with a unique length containing a relatively small fraction of reactive groups along the backbone. A key assumption is that the introduction of a single long chain branch, formed through a chemical reaction between two polymer chains, would prevent further diffusion of both polymers. They showed that diffused chains make a large contribution to the strength of the interface if they became anchored in crosslinks on both sides of the interface.

In a second paper, (68) they consider the evolution of properties that depend upon interfaces. They examine the interfacial adhesion energy G between two neighboring particles in a coating, and show how the interfacial adhesion evolves as a function of the key control parameter [alpha]. As shown in equation (10), [alpha] is defined as the ratio between [T.sub.dif], the time characteristic of the diffusion across the interparticle boundary of polymer molecules within a radius of gyration of the interface, and [T.sub.rxn], the reaction time required to observe one reaction per chain in the system.

[alpha] = [T.sub.dif]/[T.sub.rxn] (10)

After this reaction time, [T.sub.rxn], every chain in the system has undergone branching at least once and the final state of the film is reacted. For entangled chains, the time necessary for a chain to diffuse over a distance comparable to its size is the reptation time, [T.sub.rep] = [[tau].sub.0] [N.sup.3]/[N.sub.e] = [T.sub.dif], where N is the polymer chain length, [N.sub.e] is the number of monomers between entanglements, and [[tau].sub.0] is a microscopic time typical of monomer motion. [T.sub.rxn] is the time characteristic of the crosslinking reaction. The major assumption of their analysis is that a single reaction will introduce a long-chain branch that will stop further diffusion of the reacted polymer. When the reaction is fast (and thus, [alpha] [much greater than] 1) very little diffusion precedes crosslinking. When [alpha] is small, diffusion dominates. Many chains diffuse across the interface and form bridges connecting crosslinks on opposite sides of the original interface.

The interfacial adhesion energy is the energy one has to provide to open a fracture between two polymer pieces in contact. Because of viscoelastic dissipation, this energy depends upon the rate at which the fracture is propagated. Aradian et al. consider the energy G corresponding to a fracture rate approaching zero. In addition to van der Waals interactions, two distinct molecular processes resist the opening of a fracture. Chain scission occurs when a bond in a connector chain is ruptured. Chain extraction occurs when a connecting chain is dragged out of the surrounding matrix. Both processes depend on the number of monomers [bar.n] under load along each connector and the density of connectors [sigma] per unit of surface area. For the case of bond scission, Lake and Thomas (72) found

[GRAPHIC OMITTED]

[G.sub.scis] [approximately equal to] [U.sub.x][bar.[sigma]n] (11)

where [U.sub.x] is the order of a typical bond energy. The physical meaning of equation (11) is that in order to rupture a chain between two crosslink points, each monomer in the connector chain must be put under load and brought close to the breakage threshold. For a reactive latex system, one expects [bar.n] [approximately equal to] [N.sub.c].

For the case of chain extraction, Raphael and de Gennes (69) found

[G.sub.extr] [approximately equal to] [U.sub.v][bar.[sigma]n] (12)

where [U.sub.v] is on the order of a van der Waals bond energy. When a chain is extracted, it is exposed to air and becomes extended. The former results in a gain in interfacial energy, and the latter, in a loss of entropy. Both contribute [U.sub.v] per monomer to the overall energy. In a mixed situation in which, under fracture, some chains are extracted and others are broken, Aradian et al. assume that the two types of dissipation are additive. Since [U.sub.x]/[U.sub.v] [approximately equal to] 100, the adhesion energy G is normally dominated by the chemical contribution of bridging chains.

The major finding of this article is that there are three major regimes in [alpha] that contribute to the magnitude of G. (1) When the reaction is much slower than diffusion ([alpha] [much less than] 1), the interface heals completely before significant gel develops. Under these circumstances, the tear strength of the polymer approaches the cohesive strength of the bulk material. (2) When the reaction rate is fast, comparable to the diffusion rate (1 < [alpha] < [[alpha].sub.2]), the interface becomes frozen in a nonequilibrium state. A substantial number of crosslinks develop, and the mean chain length between crosslinks that span the interface is still [N.sub.c]. It is the density of connector chains that span the interface which does not reach equilibrium. (3) When the reaction rate is very fast (1 [much less than] [alpha] = [[alpha].sub.2]), diffusion cannot compete with the reaction. No crosslinked chains bridge the interface, and adhesion is due entirely to pullout of chains that diffuse after the reaction is complete. This is similar to the case of films formed from latex particles that were crosslinked during their synthesis.

When extensive diffusion occurs, the adhesive strength becomes equal to the cohesive strength, which can be calculated by assuming that fracture occurs across an arbitrary plane in the polymer. The number of chains that cross this plane and undergo scission lie inside a Gaussian radius a / [square root of [N.sub.c]] from the fracture, from which they deduce that there are [bar.[sigma]] [approximately equal to] 1 / [square root of [N.sub.c]] [a.sup.2] broken connectors per unit area; a is a length characteristic of the size of a monomer. Thus

G [approximately equal to] [[U.sub.x]/[a.sup.2]] [square root of [N.sub.c]] [approximately equal to] [G.sub.max]([alpha] < 1) (13)

When the reaction rate is more rapid, some chains become trapped. Because of the limited extent of diffusion, the length of the inserted chains are, on average, shorter than in the fast diffusion case. Chain scission is still the dominant contribution to the adhesion energy, and [bar.n] [approximately equal to] [N.sub.c], but the density of chains crossing the interface is not at equilibrium. Under these circumstances, Aradian et al. find

G [approximately equal to] [U.sub.x][N.sub.c][bar.[sigma]] [approximately equal to] [G.sub.max] [1/[[alpha].sup.1/2]] (1 < [alpha] < [[alpha].sub.2]) (14)

The adhesion energy becomes smaller as [alpha] increases.

In the limit of very fast reaction, there is little chance of the system forming crosslinks that are connected by chains that span the interface. Fracture involves pull-out of dangling ends, anchored at one end, that diffuse across the interface. For these chains, [bar.n] [approximately equal to] [N.sub.c], and

G [approximately equal to] [[U.sub.v]/[a.sup.2]][N.sup.1/2] [1/[[alpha].sup.3/4]] ([alpha] > [[alpha].sub.2]) (15)

In this regime, the inverse dependence of G on [alpha] is even more pronounced.

The crossover from the slow reaction to the fast reaction regime occurs at [alpha] = [[alpha].sub.1] = 1. Inside the fast reaction regime, the crossover from chain-scission dominated fracture to fracture dominated by chain pull-out occurs at [alpha] = [[alpha].sub.2], where [[alpha].sub.2] = (N/[N.sub.c])[.sup.2]. Thus, there are three characteristic lengths that affect the magnitude of G: the overall chain length N, the chain length between crosslinks [N.sub.c], and the entanglement length [N.sub.e], [N.sub.e] affects the magnitude of [alpha] through its contribution to [T.sub.dif].

CROSSLINKING CHEMISTRIES

In this section, we describe a variety of different chemistries that have been developed to provide post-application crosslinking in latex films. Our emphasis will be on recent developments and on chemical reactions that have been applied successfully to commercial coatings. We summarize experiments that help us to understand the reaction mechanisms that lead to bond formation, and the competition between bond formation and polymer diffusion in latex films. These principles form the basis for the design of modern coatings that combine high performance with environmental compliance.

Our discussion of these chemistries provides an opportunity to consider the issue of the distribution of crosslinks in the final coating. One commonly distinguishes homogeneous systems, which contain crosslink sites that are statistically distributed throughout and between polymeric particles, from interfacial systems. Interfacial systems contain a high crosslink density at the boundaries of the particles with a crosslink density gradient that decreases in the direction of the core of the particles.

In this context, it is important to recognize that the functional monomers that are introduced into the latex particles to provide post-application cure are normally much more costly than the monomers that make up the base latex. They are also more expensive than monomers such as ethylene glycol dimethacrylate that can be used to introduce crosslinks during emulsion polymerization. As a consequence, most industries have developed strategies to optimize the cost-effectiveness of the post-ambient-cure chemistry. These strategies include the synthesis of core-shell particles with a core crosslinked during the particle synthesis and with the reactive monomers confined as much as possible to the shell. For example, Taylor et al. (73) describes the synthesis of a styrene-acrylate latex containing allyl groups (see below) for post-application oxidative cure. Examples are provided in which allyl methacrylate is introduced into the shell surrounding a crosslinked core representing about 65% of the particle volume. Other strategies include a power-feed to introduce a gradient of monomers into the reaction. In this way, it is possible to produce latex particles with a highly crosslinked soft core with a progressively higher [T.sub.g] polymer composition and lower degree of crosslinking toward the exterior of the particles. (74,75)

Many commercial coatings formulations contain mixtures of latex. These mixtures can consist of latex particles of similar composition but different size, (37) introduced primarily to vary the rheological properties of the formulation. Alternatively, they can contain latex particles of different compositions (as in the case of hard-soft latex blends) that can serve a variety of purposes, such as promoting low temperature film formation (76) or the enhancing block resistance. (77) A more detailed discussion of these strategies is outside the scope of this review.

Melamine and Other Formaldehyde-Based Crosslinking

From the 1950s to today, crosslinking chemistries based on the condensation of amino resins and acrylamide-formaldehyde derivatives have been mainstream technologies for crosslinking hydroxyl-containing waterborne acrylic polymers. (4) Two approaches are used: an additive approach in which melamine-formaldehyde derivatives, urea-formaldehyde, benzoguanamines, or glycourils are added directly to a latex formulation, or a reactive monomer approach where the reactive functionality is attached to the backbone on the polymer. One additive approach that is used commercially employs substantial amounts (typically 20 wt% based on latex solids) of a melamine-formaldehyde (MF) resin, which can include ester or ether derivatives. In the co-monomer approach, acrylamide derivatives such as N-methylolacrylamide, or alkylether-containing monomers such as N-(isobutoxymethyl)acrylamide, at typically 2-4 wt% are introduced during the latex synthesis. Methylol groups readily condense under mildly acidic conditions with hydroxyl, carboxyl, or amide functionality. Methylol groups also combine with each other under acidic or basic conditions. Alkyloxymethyl groups self condense under more acidic conditions. Scheme 1 shows the possible crosslinking reactions.

[FIGURE 13 OMITTED]

Some monomeric melamine-formaldehyde derivatives such as HMMM are water-soluble. The longer chain alkoxy derivatives and oligomeric MF resins have limited water solubility and are introduced into formulations as an emulsion.

In a previous review, it was stated that the addition of "melamine" to a hydroxyl-containing waterborne polymer is an example of interfacial and interstitial crosslinking. (1) From this perspective, it was thought that melamines condense with themselves to form a crosslinked matrix around the particles and with the hydroxyl functionality attached to the waterborne particles. This statement must be qualified. Winnik et al. showed that while HMMM has limited miscibility with simple acrylate resins like PBMA, it is fully miscible with BMA copolymers with 4 mol% each of 2-hydroxyethyl methacrylate (HEMA) and methacrylic acid (MAA). They show that films of P(BMA-HEMA-MAA) containing 10 and 20 wt% HMMM underwent homogeneous crosslinking. (78)

The P(BMA-HEMA-MAA) + HMMM system was the first latex coating in which information was obtained about the relative rates of polymer diffusion and crosslinking during cure. (78) The experiments employed a mixture of donor- and acceptor-labeled latex particles. Polymer diffusion across the intercellular boundaries was monitored using the ET technique. Two important principles emerged from this work. The first can be seen in Figure 11, where we compare a film containing 20 wt% HMMM to a film without HMMM, and plot the extent of mixing against annealing time. At 80[degrees]C, no gel can be detected in either film on the time scale of four hours. One sees that the degree of mixing measured by the increase in the ET quantum efficiency increases to 0.18 in the film without HMMM, but increases to 0.38 in the film containing HMMM. We learn the important fact that HMMM acts as a reactive diluent and a powerful plasticizer. Its presence strongly enhances the rate of polymer diffusion in the coating before it reacts to form a crosslinked matrix. At 120[degrees]C, a significant amount of gel forms as the film is annealed. At early times, the rate of polymer diffusion is enhanced, both for film with and without HMMM. For the film containing HMMM, [f.sub.m] increases rapidly to a value somewhat less than 0.6 and then levels off. Gel formation brings further polymer diffusion to a halt, whereas polymer diffusion continues unimpeded in the film without HMMM.

The second important principle emerges from experiments carried out at different temperatures. While the diffusion rate and the chemical reaction rate increase with an increase in cure temperature, the increase in temperature has a more pronounced influence on the rate of polymer diffusion. In terms of the theoretical description of diffusion and crosslinking rates, raising the annealing temperature lowers the value of the parameter [alpha]. At lower cure temperatures, a smaller fraction of the polymer chains are able to undergo diffusive mixing with polymers in adjacent cells in the film before the cure reaction brings polymer diffusion to a halt.

Melamine derivatives were developed for high temperature applications (>120[degrees]C, >15 min cure time) such as automotive and industrial coatings. While the use of catalysts such as para-toluene sulfonic acid in these systems produces coatings that cure at lower temperatures or faster at higher cure temperatures, additional problems occur. These include a short pot life for the formulation and water sensitivity. The bonds providing the crosslinks in the coating can be hydrolyzed by moisture in the presence of the strong acid catalyst. Still, this crosslinking technology produces industrial coatings with an excellent balance of properties such as chemical resistance, flexibility, and hardness.

N-methylolacrylamide (NMA) is sometimes cited as an example of a reactive monomer that leads to homogeneous crosslinking. Results such as those shown in Figure 4, and concern about the high water solubility of NMA lead to the counter-speculation that NMA is selectively enriched in the shell region of the latex. Under these circumstances, its crossreactions with -OH groups in the latex and its self-condensation would favor interfacial cure. The more hydrophobic N-(isobutoxymethyl)acrylamide (IBMA) has a more uniform distribution in the latex polymer and should lead to more uniform crosslinking. The results of Liu et al. described above on the competition between polymer diffusion and crosslinking in P(BA-MMA-IBMA) latex films were entirely consistent with homogeneous crosslinking. They demonstrated that one can change the value of the control parameter [alpha] for this system in two ways. One can increase the reaction rate without affecting the polymer diffusion rate by increasing the amount of acid catalyst or choosing a stronger acid. Alternatively, one can raise the temperature, which has a larger effect on increasing the polymer diffusion rate than the chemical reaction rate.

[GRAPHIC OMITTED]

Bassett and Sherwin (79) at Union Carbide Corp. explored the mechanism and cure conditions for N-(isobutoxymethyl)acrylamide (IBMA)-based latex films. Their proposed crosslinking mechanism is shown in Scheme 2. A series of latex samples were prepared from a monomer mixture consisting of styrene, ethyl acrylate, 2-hydroxyethyl acrylate, methacrylic acid, and IBMA, and the kinetics of the crosslinking reaction was determined by thermal evolution analysis and thermal gas chromatography. The amounts of isobutyl alcohol evolved from the films during cure were determined as the isobutoxymethyl moieties self-condensed. Using their data, the plots in Figure 12 were constructed. The results show that the addition of methacrylic acid to the composition of the polymer catalyzes the crosslinking reaction. The best result is obtained if no 2-hydroxyethyl acrylate is contained in the composition of the polymer. Grawe and Bufkin (1) discussed other aspects of this work.

Crosslinking latex systems based on acrylamide monomers or glycoluril resins are possible under milder conditions, using latex bound -COOH groups if the latex is first ion-exchanged to convert the acid groups to the -COOH form. This approach was first described by Hahn and Kunz in two patents by Glidden in the 1980s. (80) Winnik and co-workers showed that these conditions are also effective for curing HMMM at temperatures as low as 100[degrees]C in the BMA-HEMA-MAA copolymer latex film described above.

The Klein group at Lehigh (81) has recently shown that latex particles synthesized in the presence of NMA had a substantial gel content (premature gelation), even when prepared at pH values at which no condensation of the N-methylol groups would be expected. Based on their success in suppressing the gel content with carbon tetrabromide as a chain transfer agent, they suggested that crosslinking arose as a consequence of hydrogen abstraction from the N-methylol group during particle synthesis. We would like to point out that the unusual behavior of the latex films seen in Figure 6, which contain NMA groups, may well be due to a substantial gel content in the films that limited the extent of polymer diffusion and interface healing.

[FIGURE 14 OMITTED]

Zinc and Zirconium Crosslinking

In 1965, Rodgers at SC Johnson developed one of the first successful nonformaldehyde crosslinking chemistries. (82) He used zinc salts to crosslink carboxylic-acid containing waterborne dispersions. An example is shown in Scheme 3. A strong coulombic interaction between the zinc and the carboxylate anions leads to tight ion pairs that tend to associate into ionomer-like interactions. Although the scheme shows ammonia evaporating from the coating, the fate of all the ammonia from the amine complex is unknown. The advantage of this technology is that it provides a one-pack stable system that produces ambient curable films with excellent hardness and water resistance. Since ammonia competes with the carboxylates for binding to the zinc ions and reacts with zinc cations, ammonia solutions reverse the crosslinks and solubilize the carboxylate-based coating. Thus, ammonia-based floor cleaners remove zinc crosslinked floor coatings crosslinked with this technology.

[GRAPHIC OMITTED]

[GRAPHIC OMITTED]

Another metal-based crosslinking chemistry is based upon the ability of zirconium salts to crosslink carboxylic acid-containing acrylic coatings and ink films. (83a) In particular, ammonium zirconium carbonate is sold commercially for use in waterborne coatings to improve adhesion to low energy surfaces, and to improve the heat, scrub, water, and solvent resistance of paint and ink films. (83b) Ammonium zirconium carbonate is also effective at altering the rheology of paints. The

aqueous chemistry of this material is complex, but it is known to exist as an oligomeric anionic zirconium hydroxyl carbonate complex in water at pH values between 7.5 and 9.5. Scheme 4 shows that the first step in the crosslinking reaction involves the displacement of the carbonate anion. Subsequent carbonate displacement reactions occur until a crosslinked polymer is obtained.

The effectiveness of Zr salts on crosslinking of coatings varies with the Zr salt concentration and with the pH of the water phase of the formulation. These factors affect the length of the oligomeric ammonium zirconium carbonate complex. Raising the complex concentration and lowering the pH of its water solution reduces its chain length, while lowering its concentration and raising its pH increases the chain length. The crosslinking efficiency of these Zr salts improves as their chain length decreases.

Aziridine Crosslinking

After the Clean Air Act in the U.S. and the oil embargo in the early 1970s, a search began for crosslinking technologies that produced lower formaldehyde emissions than from melamine-based coatings, and provided faster cure at lower temperatures. In the late 1970s, Cordova Chemical Co. introduced aziridine crosslinking technology. (84) The reaction of an aziridine group with an ammonium carboxylate is shown in Scheme 5. XAMA-7, a trifunctional aziridine crosslinker, is prepared by the Michael addition reaction of ethylenimine and trimethyolpropane triacrylate (TMPTA). This ester-based aziridine crosslinker plasticizes latex particles, and gives wide cure profiles. Once XAMA-7 is added to a latex-based formulation, its pot life typically ranges from 2-24 hr depending upon the design of the carboxylic acid-containing acrylic particles and the pH of the formulation. Unfortunately, aziridine crosslinkers do not pass the Ames test and as one of these authors can attest, a certain percentage of the general population develops severe allergic reactions. Still, aziridine-based crosslinkers react in carboxylic acid-containing latex films at room temperature, and this strong attribute results in their current use. Today, most aziridine-based crosslinkers are prepared by reacting propylenimine with multifunctional acrylates. The substitution of propylenimine for ethylenimine results in crosslinkers with lower vapor pressures than those prepared using ethylenimine; thus, these aziridine-based crosslinkers are less hazardous to handle.

[FIGURE 15 OMITTED]

Polycarbodiimide Crosslinking

In the early 1980s, Union Carbide Corp. began the development of crosslinking technologies that provided the benefits of aziridine technology without the associated health risks. This research led to the introduction of polycarbodiimide technology to the coating industry. (27) Eventually, this technology was made water dispersible for easy use. Polycarbodiimide crosslinkers react rapidly with carboxylic acid moieties in acrylic or urethane coatings to form N-acyl ureas, as shown in Scheme 6. Carlson and co-workers (85) at Rohm and Haas Co. also explored the use of this technology and showed its versatility as an ambient curing crosslinker. Today, a polycarbodiimide crosslinker is supplied commercially by GSI EXIM America, Inc. as Carolite[R] and previously by Union Carbide Corp. as UCARLNK XL29SE[R]. The crosslinking technology is used primarily to provide solvent and water resistance as well as to improve the corrosion resistance of maintenance coatings. (86)

During the development of polycarbodiimide crosslinking technology, XAMA-7, an aziridine crosslinker, 3, was compared to a tetrafunctional "linear" polycarbodiimide, 4, and to trifunctional "star" carbodiimides, 5, as seen in Figure 13. (87) To measure the effectiveness of carbodiimide crosslinking, clear films were prepared from formulations of UCAR[R] Vehicle 462 that contained 3, 4, and 5. UCAR Vehicle 462 is a carboxylic acid-containing styrene-acrylic latex. A multifunctional acrylate controls the initial gel content of the latex. Although this description is an over-simplification, it will be a very useful model in discussing film formation and crosslinking mechanisms. To evaluate the crosslinkers 3, 4, and 5, crosslinked films were prepared using the formulation in Table 2.

[GRAPHIC OMITTED]

The aziridine crosslinker, 3, was added neat to the formulated latex. Emulsions of the polycarbodiimide crosslinkers were prepared using a previously published procedure. (88) Emulsions of the carbodiimide crosslinkers, 4 and 5, (27% active in water and amyl acetate) were added to the above partial formulation at five parts of dry carbodiimide crosslinker per 100 parts of dry polymer resin (phr). The final pH of the formulations was 8.4 [+ or -] 0.3. A formulation that contained no crosslinker was used as a control. Small aliquots of the final formulations were placed in Teflon molds. After air-drying the films to remove most of the water in the films, the films were cured at 127[degrees]C for 15 min in a forced-air oven. The dry thickness of the films was approximately 22 mils (0.55 mm). Studies of the cured films by FTIR spectroscopy showed that more than 95% of the carbodiimide functionality reacted during cure. After cure, the films were evaluated for their crosslink density, tensile strength at break, elongation, and solvent resistance.

Since methyl ethyl ketone is used in the coatings industry to determine the solvent resistance of cured coatings, it was chosen as the solvent to study crosslinking efficiency. The crosslink density or molecular weight between crosslinks, [M.sub.c], of the cured films of Table 2, was determined by swelling the films in MEK. The results are shown in Table 3. The control film (film containing no crosslinker) has an [M.sub.c] of 4.1 X [10.sup.4] g/mol, a tensile strength at break of 5.65 MPa, and an elongation-to-break of 830%. The gel from the control film results from the heterogeneous nature of the waterborne particles from UCAR vehicle 462 and is not indicative of cure.

The data show that the polycarbodiimide, 4, and 1,3,6-tri(N-isopropyl-N'-methylene carbodiimide) hexane are the least efficient crosslinkers in obtaining low [M.sub.c] values at 5 phr of crosslinker, whereas XAMA-7, (3), 1,3,6-tri (N-t-butyl-N' methylene carbodiimide) hexane, 1,3,6-tri(N-n-butyl-N'-methylene carbodiimide) hexane, and 1,3,6-tri(N-phenyl-N'-methylene carbodiimide) hexane are the most efficient.

Wet films were cast over Leneta paper, and cured in a forced-air oven. The dry film thickness was 1.2 mils (30 [micro]m). During the course of this work it was shown that the number of MEK double rubs obtained on crosslinked films decreased as the thickness of the films decreased; therefore, smooth films of equivalent thickness were essential in obtaining reproducible results. Each MEK double rub is an average of three measurements. The results show that samples cured at 127[degrees]C for 15 min were resistant to methyl ethyl ketone double rubs. More than 300 MEK double rubs were obtained under these cure conditions. Solvent resistance was also obtained for cure temperatures as low as 60[degrees]C.

Oxirane Crosslinking

Another crosslinking technology that has been investigated is the incorporation of oxirane (epoxide) groups into latex particles. Oxiranes react with carboxylic acids and their ammonium salts. They also react with amines, with each -N[H.sub.2] group reacting with two oxiranes to form two C-N bonds. These reactions are shown in Scheme 7. Glycidyl methacrylate (GMA) has been incorporated into carboxylic acid-containing particles; however, there are many problems with this approach. Typically, the weight-average molecular weight of an acrylic polymer prepared via emulsion polymerization is often in excess of 100,000 g/mol. For an acrylic polymer stabilized with 3% methacrylic acid, there may be more than 30 methacrylic acid repeat units per weight-average polymer chain. When more than one crosslink per weight-average polymer chain occurs, the gel fraction inside waterborne acrylic particles begins to increase. Although the oxirane group introduced by GMA reacts slowly with carboxylic acid groups at ambient temperature, only 3% of the total acid needs to react to begin a gel fraction increase. Another approach is to blend GMA-containing particles with acid- or amine-containing waterborne particles; however, the oxirane group within the particle can still hydrolyze to the diol. For example, in butyl methacrylate waterborne particles that contain GMA, substantial hydrolysis (32%) occurred within one year at ambient temperature. (89) Taylor (90) recently showed that the hydrolysis rate of the oxirane group (as well as acetoacetoxy and carbonate groups) in acrylic waterborne particles is reduced by decreasing the oxygen content of the oxirane-containing waterborne particles and that the hydrolysis rate of oxirane moiety in selected latexes follows first-order kinetics.

[FIGURE 16 OMITTED]

[FIGURE 17 OMITTED]

Tronc et al. (91) described a study of the relative rates of polymer diffusion and crosslinking in films ([T.sub.g] = 8[degrees]C) formed from an MMA-BA-GMA latex with a monomer weight ratio of 52:16:30. The remaining ca. 2 wt% was a dye-bearing monomer, either 1 or 2, as shown in Figure 9. Films of these latexes exhibited the unusual behavior of spontaneous gel formation. For one set of samples with [M.sub.W] [approximately equal to] 90,000, size exclusion chromatography experiments on the particles in the dispersed state showed no gel content. In contrast, films formed from this sample at 23[degrees]C, when they were dry, exhibited a 40% gel content that grew slowly over time, but never reached 100%. Films formed at 4[degrees]C exhibited little gel initially, but the gel content increased over time with a rate that increased with temperature. An example of these results is shown in Figure 14. Energy transfer experiments showed that the polymer diffusion rate was strongly retarded by the increase in gel content in the films. Nevertheless, for films formed at 4[degrees]C and then rapidly heated to 60[degrees]C, polymer diffusion was rapid and led to full randomization of donor and acceptor dyes ([f.sub.m] = 1) in the film. Films formed in the presence of a diamine (epoxy/N[H.sub.2] = 0.75) underwent more rapid gel formation. (92) The extent of gel reached 100%, and at ambient temperature over a period of three days, mechanically tough crosslinked films formed. The crosslinking reaction has a pronounced effect on limiting the extent of intercellular polymer diffusion in these films.

[GRAPHIC OMITTED]

[FIGURE 18 OMITTED]

Oxazoline Crosslinking

In 1991, Nippon Shokubai introduced oxazoline crosslinking technology (Scheme 8) to waterborne coatings. (93) This crosslinking technology was supplied as water-soluble polyoxazolines or as oxazoline-containing waterborne particles. This technology deactivates the carboxylic acid groups attached to the polymer chain during the crosslinking process.

The development of crosslinking technologies to react with the carboxylic acid group of waterborne acrylic particles is no accident. Carboxylic acid-containing monomers such as acrylic, methacrylic, and itaconic acids provide stability for waterborne acrylic particles; however, their presence results in coatings with more water sensitivity than solventborne coatings that do not contain residual carboxylic acid groups attached to the backbone of the polymer. The development of chemistries to deactivate the carboxylic acid moieties during the crosslinking process is still a challenge today for the chemist designing crosslinking chemistries. Unfortunately, chemistries designed to react with carboxylic moieties usually, by their nature, react with water. The result is that one-pack thermosetting acrylic technology has been only marginally successful in fulfilling the stability requirement of a one-pack water-based acrylic thermosetting system. To circumvent this problem, recent technology has examined various approaches to incorporate unsaturation into acrylic latex particles.

Isopropenyl Chemistry

Murray et al. (94) have shown that isopropenyl esters react rapidly under acidic conditions with hydroxyl-containing waterborne particles, as shown in Scheme 9, to form ester crosslinks plus acetone. The reaction between isopropenyl pivalate and 1-pentanol was also studied at 80[degrees]C using toluene as the solvent and p-toluene sulfonic acid as the catalyst. The extent of reaction was followed by monitoring the disappearance of the carbonyl absorption at 1745 [cm.sup.-1]. As shown in Figure 15, the reaction occurs rapidly and completely at relatively low temperatures. The half-life of the reaction is 34 min.

A high-molecular weight styrene-acrylic latex that contained 5 wt% hydroxyethyl acrylate, 5% methacrylic acid, 45 wt% ethyl acrylate, and 45 wt% styrene ([T.sub.g] = 41[degrees]C) was formulated with 4.86 phr of di-isopropenyl adipate and three parts per hundred of p-toluene sulfonic acid. The latex was 45% solids. Films were cast over Leneta paper, and cured for 30 min at the appropriate temperature in a forced-air oven. The dry film thickness of the films was 1.2 mils (30 [micro]m). The results are shown in Table 4. Excellent solvent resistance was obtained at cure temperatures greater than 100[degrees]C. For solventborne coatings, gel fractions (X100) of > 80% were reported for cure temperatures as low as 60[degrees]C. In this study, no additional filming aids were added to lower the [T.sub.g] of films as they cured.

Unsaturation Crosslinking

One of the oldest crosslinking technologies is alkyd chemistry. Alkyd-based coatings cure at ambient temperature and the crosslinking process does not start until the coating is in the presence of oxygen. Hurbert et al. (95) from Akzo Nobel have studied the oxidation reaction of alkyd resins and 3,6-nonadiene in the presence of cobalt ions and organic sensitizers. They concluded from their studies that the oxidative drying process consists of two mechanisms. The first begins with hydrogen abstraction from the activated methylene group ([C.sub.11]) of the linoleic moiety of a fatty acid and leads to hydroperoxide formation through reaction of the carbon-centered radical with oxygen. The second begins with the formation of singlet oxygen, followed by its reaction with an allylic C-H to yield an endo-peroxy moiety. The hydrogen abstraction or autooxidation mechanism does not require any catalyst; however, cobalt ions do catalyze the degradation of hydroperoxides into alkyl, ether, and peroxide radicals, which accelerate the drying process considerably. Recombination of these radicals leads to crosslinking reactions. Cobalt ions also promote the formation of the endo-peroxy moiety, which is best explained by a photooxidation process where singlet oxygen is the reactive species.

[FIGURE 19 OMITTED]

[GRAPHIC OMITTED]

For solvent-based systems, wonderful application properties are obtained. Alkyd crosslinking technology, however, has several disadvantages: alkyd coatings embrittle and yellow as the coating ages. Alkyd chemistry continues to occur within the coating as it ages; thus, aldehydes and ketones are continuously released from the coating. As a result of these drawbacks, there has been extensive exploratory work to design crosslinking chemistry that mimics the cure profiles of alkyd resins, but avoids their inherent deficiencies.

Tillson, (96) using allyl methacrylate (ALMA) as a monomer, first introduced unsaturation into latex particles and showed that allyl-containing latex films undergo oxidative cure. Taylor et al. (73) improved Tillson's approach by adding styrene to an emulsion polymerization process that resulted in more than 95% of the allyl moieties from allyl methacrylate surviving the polymerization process. The copolymerization of ALMA with styrene and a methacrylate ester is presented in Scheme 10.

[FIGURE 20 OMITTED]

As shown in Figure 16, the level of allyl survival increased with the level of methacrylate or styrene in the monomer feed. Styrene was the most effective monomer at promoting allyl survival during the emulsion polymerization process. The allyl survival was also determined via FTIR spectroscopy for a styrene/butyl acrylate-based latex using t-butylaminoethyl methacrylate (t-BAMA). It was shown that t-BAMA inhibits allyl polymerization. The results are shown in Figure 17. The survival of allyl moieties during the emulsion polymerization of allyl methacrylate in the presence of styrene and other monomers is also supported by the reactivity ratio studies of Heathley, Lovell, and McDonald at the University of Manchester. (97)

These technologies, which introduced allyl moieties into waterborne particles, provide cure by reacting with oxygen to produce crosslinked films. (96) As shown in Scheme 11, the first step in the crosslinking reaction likely involves oxygen reacting with the allyl moieties to form allylic hydroperoxides. The reaction pathways after formation of the hydroperoxide moieties have not been studied for the allyl moiety of allyl methacrylate-based latex films, but it is possible that alkyd-type bond making and breaking reactions occur. As discussed earlier, two possible mechanisms exist for the formation of the hydroperoxide: a hydrogen abstraction of the allylic carbon (structure 7 in Scheme 11) or a concerted reaction between singlet oxygen and the double bond that results in a double bond shift (structure 8). (98) More study is needed to determine the oxidation-cure mechanism for allyl methacrylate-based latex films.

Coatings prepared from allyl methacrylate-based emulsions were prepared and monitored at room temperature. Coatings prepared from formulations that contained [Co.sup.2+] as a catalyst increased in gel fraction and decreased in swell ratio as the coating cured at room temperature. For example, for an allyl methacrylate-containing core/shell latex, clear films were cast and cured at room temperature. A decrease in swell ratio was observed (Figure 18).

Thames and co-workers (99) prepared branched acrylate and methacrylate derivatives of castor oil as monomers for emulsion polymerization. These fatty acid monomers contain an isolated double bond that survives free radical emulsion polymerization, and its pendant flexible chain acts as an internal plasticizing moiety. The two ester groups on these monomers apparently increase their hydrophilicity so that they can transport the water phase during emulsion polymerization. The surviving double bond cures under ambient conditions using an alkyd-type mechanism. The homopolymer of the acrylate derivative (CAM) has a [T.sub.g] of -93[degrees]C. Vinyl acetate-based and acrylate-based latexes designed as binders for architectural coatings were prepared and evaluated. When the binder contains 15% CAM, the [T.sub.g] of coatings increases by as much as 12[degrees]C during cure. Because of the low [T.sub.g] imparted to architectural coatings with CAM and the [T.sub.g] swing associated with CAM-based polymers, coatings could be formulated using no coalescing aids. During cure, improvements in the block resistance and dirt pick-up of the coatings occurred. After cure, an increase in the gel fraction of the polymer and the disappearance of unsaturation in the pendant fatty acid chain of the polymer confirmed crosslinking of the coating.

Introduction of Unsaturation Using Aziridine, Oxirane, and Carbodiimide Chemistries

In this section we examine the postfunctionalization of carboxylic acid-containing latex to introduce methacrylate groups. Examples are shown in Scheme 12. Radical initiated polymerization of the pendant methacrylate moieties results in a crosslinking reaction within the latex film. McGinniss, Seidewand, and Robert (100) first used this concept to introduce polymerizable functionality onto the surface of latex particles by reacting 1-(2-aziridinyl)ethyl methacrylate with carboxylic acid-containing waterborne particles. Taylor and co-workers at Union Carbide Corp. laboratories examined this chemistry by nuclear magnetic resonance. The results showed that more than 90% of the theoretical level of methacrylate moieties survives the modification process. (101) Padget (102) reported the results of Robert who introduced methacrylate double bonds into latex particles by reacting GMA with amine-containing waterborne particles. Later, Mylonakis (103) performed similar modifications by reacting GMA with carboxylic acid-containing waterborne particles. Wolfersberger, Schindler, Beckley, and Novak (104) at Rohm and Haas Co. extended Mylonakis's work by reacting GMA with carboxylic acid-containing waterborne particles prepared using a multistaged process. Up to 80% of the glycidyl methacrylate was incorporated onto the surface of their waterborne particles. Rohm and Haas Co. has successfully commercialized a reactive waterborne acrylic polymer prepared from the reaction between GMA and carboxylic acid-containing acrylic particles. This technology has found use as an aqueous UV-curing latex for wood coatings.

[GRAPHIC OMITTED]

COMPARATIVE STUDIES: Taylor and co-workers (101) chose to explore the use of carbodiimide methacrylate for incorporating methacrylate functionality into waterborne particles. The incorporation efficiency and crosslinking efficiency of carbodiimide monomers were compared to 1-(2-aziridinyl)ethyl methacrylate, and epoxy monomers (Scheme 12). Carboxylic acid-containing latex samples were reacted with emulsions of alkyl-carbodiimide ethyl methacrylates and epoxy-based monomers. 1-(2-aziri-dinyl)-ethyl methacrylate was added neat to the latex.

Latex A (STY/MMA/BA/MAA/mercaptoacetic acid, 19.7/44.2/24.6/10/1.5) was neutralized to a pH of 8.5 with a volatile base such as ammonia or triethylamine. The latex was heated at 80[degrees]C until the surface modification was completed, and no carbodiimide absorption was visible by FTIR spectroscopy. Analyses of the resulting latex by nuclear magnetic resonance spectroscopy show that the latex particles contain more than 80% of the theoretical level of pendant methacrylate functionality. The levels of incorporation for the various carbodiimide methacrylates are depicted in Figure 19.

The level of incorporation of 1-(2-aziridinyl)ethyl methacrylate depended upon the pH of latex A. When the initial pH was 1.6, a higher level of incorporation (98%) was obtained; however, if the pH of the latex A was adjusted to 7.5 with triethylamine prior to the addition of 1-(2-aziridinyl)ethyl methacrylate, a lower level of incorporation was obtained (67%). The authors postulated that under acidic conditions, the nitrogen atom on the aziridinyl methacrylate is protonated; thus, the protonated molecule is drawn very near or inside the negatively charged latex particles. The result is that less hydrolysis occurs before the reaction between the carboxylate salt (attached to the latex polymer) and the protonated aziridine moiety. The glycidyl methacrylate data point in the plot (57% [+ or -] 6%; confidence limit, 95%) is an average of eight separate experiments.

[FIGURE 21 OMITTED]

Figure 19 also shows the percent of monomer incorporation as a function of oxygen content. The modification of latex A using 1-(2-aziridinyl)ethyl methacrylate under acidic conditions is not included. The correlation of different reactive species (aziridine, carbodiimide, and oxirane) with oxygen content may be fortuitous, but the authors point out that monomer solubility in water is related to their oxygen content, and that their solubility in water may play a role in the correlation.

CROSSLINKING RESULTS: To demonstrate the crosslinking from waterborne particles containing polymerizable unsaturation, the gel fractions of cured films were obtained. For comparison, latex A was modified with one equivalent of aziridinylethyl methacrylate, or glycidyl methacrylate, or cyclohexylcarbodiimide ethyl methacrylate to one equivalent of the carboxylic acid contained in the latex. Films were cast, and then cured at various temperatures for 30 min in the presence of an organic peroxide. The results are shown in the gel fraction plots in Figure 20. The results show that modification of the latex with the aziridinyl methacrylate produced films with the highest gel content. Cured films produced from latex particles modified with cyclohexylcarbodiimide ethyl methacrylate gave slightly higher gel fractions at cure temperatures greater than 130[degrees]C than cured films produced from latex particles modified with glycidyl methacrylate.

SOLVENT RESISTANCE STUDIES: Latex A, a carboxylic acid-containing acrylic-based latex, was modified with glycidyl methacrylate, cyclohexylcarbodiimide ethyl methacrylate, or t-butylcarbodiimide ethyl methacrylate, then evaluated as clear films over aluminum panels. Films were cured in a forced-air oven or cured under UV light. The results are shown in Table 5. Both glycidyl methacrylate-modified and alkyl carbodiimide methacrylate-modified latexes produce films that cured in the presence of peroxide additives or under UV light. Thermally cured films gave high gel fractions and excellent solvent resistance. For UV-curable latex films, the latex modified with t-butylcarbodiimide ethyl methacrylate gave the best result. Latex B, a vinyl ester-based latex that contained 6% monovinyl adipate, was also modified with cyclohexylcarbodiimide ethyl methacrylate. A waterborne alkyd was added to the formulation as a peroxide generator. The film cured at room temperature to give a high gel fraction and excellent solvent resistance.

[FIGURE 22 OMITTED]

[FIGURE 23 OMITTED]

More recently, Bors and Lavoie (105, 106) at Rohm and Haas Co. have introduced unsaturation into waterborne acrylic particles by reacting ammonia with pendant acetoacetoxy moieties to produce reactive moieties that cure in the presence of peroxide forming materials. Although the chemistry for curing is not revealed, the enamine structure shows that the enamine is an acrylate moiety that is sterically hindered by the amine and methyl moieties (Scheme 13). The methyl moiety also contains allylic hydrogen atoms.

FTIR spectroscopy studies of clear films show that for hydrophilic polymers, the enamine moiety reacts with water from the atmosphere, and slowly disappears (half-life is approximately four days at room temperature), thus reforming the initial acetoacetoxy moiety and ammonia. It is unclear whether the crosslinking reaction involves the enamine, acetoacetoxy moiety, or its "enol" form. As shown in Figure 21, this technology gives cured films with high solvent resistance and low swell ratios. Rohm and Haas Co. has commercialized this technology for light-duty maintenance coatings. ICI and Eastman Chemical Co. have placed enamine and methacrylate functionality onto waterborne acrylic particles by the reaction between acetoacetoxyethyl methacrylate and amine-containing particles. (107, 108) Excellent solvent resistance is reported.

[FIGURE 24 OMITTED]

Acetoacetoxy Chemistry

Since the introduction of acetoacetoxyethyl methacrylate (structure 9 in Figure 22, AAEM), by Eastman Chemical Co. in the late 1980s, (109) acetoacetoxy functional acrylic polymers have been intensively investigated. Geurink et al. (110) have examined the properties of films formed from two-pack crosslinking formulations with acetoacetoxy-containing latex particles. Acetoacetoxy moieties attached to waterborne particles cannot survive the heat stability requirements required for commercialization. The acetoacetoxy moieties in acrylic waterborne particles react with water to form carbon dioxide and acetone forming the repeat unit from hydroxyethyl methacrylate on the backbone on the polymer. As shown by Taylor (90) (Figure 23), the kinetics of hydrolysis for acetoacetoxy moieties within waterborne particles is a function of the oxygen content of the latex polymer. Although acetoacetoxy-containing waterborne particles can be designed to achieve reasonable stability at ambient temperature, the hydrolysis rate constant is temperature dependent and acetoacetoxy-based acrylic latexes should be stored at low temperatures.

A variety of methods have been examined to improve the stability of the acetoacetoxy moieties. It is known that acetoacetoxy moieties react with amines to form enamines; thus, the focus of research has been converting the acetoacetoxy moieties to enamines. As discussed above, Bors and Lavoie used this approach to show that enamines are easily formed by the addition of ammonia to acetoacetoxy-based latex, that enamines are more hydrolytically stable than acetoacetoxy moieties, and that enamines cure via an oxidative process in the presence of fatty acids. (106) The stability of ammonia-based enamines (90) is illustrated in Figure 24.

Feng et al. (111) reported experiments which examined the relative rates of crosslinking and polymer diffusion in films formed from poly(2-ethylhexyl methacrylate) and poly(butyl methacrylate) copolymer latex. Both types of latex were prepared using 10 wt% acetoacetoxyethyl methacrylate as a comonomer. These samples are denoted AA-PEHMA and AA-PBMA respectively. As a crosslinker, 1,6-hexanediamine was added to each dispersion just prior to film formation. The authors showed that the reaction between the acetoacetoxy groups and the diamine occurred on a time scale of minutes. For AA-PEHMA + diamine, the gel content of the newly formed film was greater than 90%. For AA-PBMA + diamine, the crosslinking reaction occurred in the dispersion before the film could dry. The gel content reached 90% in 10 min and 95% in less than one hour. Thus, the chemical reaction is very fast. Nevertheless, the films formed from these latex samples had good integrity and excellent solvent resistance. This type of result, which is very different from one's expectations about films formed from crosslinked latex, is only possible if the reaction between the amine groups and the acetoacetoxy groups is reversible. Polymer diffusion is still possible to create bridging chains which span the intercellular boundaries in the film.

Diacetone Acrylamide Crosslinking

Most reactive moieties designed for crosslinking reactions attached to waterborne particles suffer from instability in water. One newer reactive monomer that has excellent stability in water is (N-(1, 1-dimethyl-3-oxybutyl acrylamide). (112) In the coatings industry, this monomer is referred to by its trade name, diacetone acrylamide (DAAM). It has a melting point of 57[degrees]C, is readily soluble in water (>100g/100g water), is relatively nontoxic (L[D.sub.50], oral, rat: 1-2g/Kg), and it passes the Ames test. Once incorporated into waterborne particles, the polymer films are crosslinked through the ketone moiety using hydrazine-based crosslinkers added to the latex formulation. These crosslinkers are prepared from hydrazine and difunctional organic acids such as adipic acid (ADH). The crosslinking reaction results in the formation of hydrazone moieties that form the crosslink points within the film. The result is that crosslinked films have high gel fractions and excellent solvent resistance. An important advantage of this crosslinking chemistry is that it occurs at such a rate at room temperature that acceptable properties are obtained in crosslinked films within seven days, and the chemistry is stable in a water environment. The crosslinking reaction is illustrated in Scheme 14.

The following example illustrates diacetone acrylamide chemistry. The latex had a pH of 9.2, a particle size of 106 nm, and a [T.sub.g] of 0[degrees]C. An equivalent of N-H moieties from ADH was added to the latex that contained 2 to 3% DAAM. Crosslinking results of films cured at room temperature at 23[degrees]C for seven days are shown in Table 6. Excellent solvent resistance was obtained. After 1000 hr of QUV ultraviolet light exposure, films of the above crosslinked polymer retained excellent gloss retention with essentially no yellowing of the exposed film (yellowness index < 2). Companies such as Akzo Nobel, (113, 114) NeoResins, (115) ICI Industries, (116) and Zeneca (117) have pursued the above cross-linking technology in coatings and inks.

Enamine and Amine Crosslinking

Esser et al. (118) used a combination of ammonia and a diamine in the presence of acetoacetoxy-containing latexes to demonstrate a viable system that had adequate heat stability. Taylor and Collins (119, 120) used similar chemistry to develop a thermosetting latex system.

The primary advantages of the above approaches are the development of thermosetting acrylic-based latexes that cure at ambient temperature. The reaction of an acetoacetoxy enamine with a primary alkyl amine is shown in Scheme 15.

[GRAPHIC OMITTED]

Carbodiimide and Isocyanate-Functional Latex

Carbodiimide groups can be introduced into acrylate latex via emulsion copolymerization using t-butyl-carbodiimidoethyl methacrylate (tBCEMA) or cyclohexyl-carbodiimidoethyl methacrylate as the functional comonomer. (87) One of the ideas associated with the application of this type of latex is to blend the latex with a second latex containing carboxylic acid groups. As discussed above, these groups react to form an N-acyl urea. By packaging the carbodiimide groups in a separate particle, one prevents the reaction with the-COOH groups until a film is formed and the two reactive groups can come into contact. In this way, one may be able to achieve a "two-pack-in-one-pot" formulation. The success of this type of approach depends upon the chemical stability of the reactive groups in the two types of latex particles. The stability of the carbodiimide group depends upon the hydrophobicity of the latex polymer, steric protection of the -N=C=N- group, the pH of the medium, and the storage temperature. (87) Pham and Winnik (121) showed that 91% of these groups in a P(EHMA-co-tBCEMA) latex survived one year at 23[degrees]C. This is not likely to be adequate for a commercial product that might be exposed to higher temperatures during shipment or storage.

Pham (121b) examined films formed from 1:1 mixtures of P(EHMA-co-MAA) and P(EHMA-co-tBCEMA), in which the -N=C=N- groups were introduced as described above. In one example, relatively thick films (0.5 mm) were formed and allowed to age at room temperature for three days. Tensile testing experiments showed that tough elastomeric films were produced with an extension to break of nearly a factor of 7.

In other experiments, Pham and Winnik (122) examined the competition between the reaction and diffusion rates at 60[degrees]C in films prepared from mixtures of P(EHMA-co-MAA) and P(EHMA-co-tBCEMA). Both latex samples contained particles of 120 nm in diameter and polymer molar masses of [M.sub.w] [approximately equal to] 60,000 ([M.sub.w]/[M.sub.n] = 2). P(EHMA-co-MAA) contained 11 mol% -COOH groups and 1 mol% Phe as the donor for ET experiments. P(EHMA-co-tBCEMA) contained 5 mol% -N=C=N- and 1 mol% An as the acceptor dye. Thus, the ratio of (-COOH/-N=C=N-) groups in the system was 2:1. In these experiments, the extent of reaction was monitored using the characteristic -N=C=N- infrared absorption at 2126 [cm.sup.-1]. As the mixture dried to form a film, approximately 20% of the -N=C=N- groups reacted. The right-hand axis of the upper graph in Figure 25 describes the survival of the remaining carbodiimide groups in the dry film. One sees that polymer diffusion was relatively rapid at 60[degrees]C, and that [f.sub.m] reached a value of 0.8 in five hours. At the same time, the gel content had reached 75%, whereas only about 25% of the carbodiimide groups present in the initial film had reacted.

[GRAPHIC OMITTED]

One of the interesting features of this system is that the presence of the -COOH groups in the EHMA copolymer limits the miscibility of the P(EHMA-co-MAA) with PEHMA itself and its copolymers with tBCEMA. (123) It is the reaction between the -COOH groups of P(EHMA-co-MAA) and the carbodiimide groups of P(EHMA-co-tBCEMA) that promotes miscibility, presumably through graft copolymer formation. Full mixing ([f.sub.m] = 1) accompanies complete gelation of the film.

[FIGURE 25 OMITTED]

An alternative strategy for achieving a two-pack-in-one-pot formulation involves the blend of other pairs of reactive latex. In each pair, one latex contains an electrophilic functional group such as acetoacetate or isocyanate, and the other contains a nucleophilic functional group such as a primary amine. For example, Guerts et al. (124) describe the reaction of an acetoacetoxy-functional acrylic latex with a methacrylate latex containing 5-amino-pentyl methacrylate as a comonomer.

El-Aasser and co-workers (125) prepared isocyanate-containing styrene-butyl acrylate latex particles by using 1-(3'isopropenylphenyl)-1-methyl-ethylisocyanate (TMI) as a comonomer. They examined the yield of incorporation of the -N=C=O group into the latex and the stability of this group toward hydrolysis in the dispersion. They obtained latex films with useful mechanical properties by mixing the latex with an aqueous dispersion of [alpha], [omega]-diaminopolyisobutylene. This result is somewhat unexpected, because of the strong immiscibility of polyisobutylene with poly(styrene-co-butyl acrylate). Electron micrograph images showed that the reaction between the groups on the two different polymers created a significant interphase between the polymers.

SUMMARY

Since the late 1970s, there have been significant advances in the development of new chemistries for thermoset latex coatings. Several of these reactions allow crosslinking to take place at room temperature. Noteworthy among these reactions are those between carboxylated latex and water-dispersible carbodiimide or aziridine-containing crosslinking agents. The reaction between the acetoacetoxy group and diamines proceeds rapidly at room temperature, but the shelf life of acetoacetoxy-containing latex is limited by the hydrolysis of the acetoacetoxy group. This group can be converted to an enamine, more stable to hydrolysis, through reaction with ammonia or with a primary amine. The enamine will undergo an exchange reaction with added diamine to form a fully crosslinked network.

One of the most promising strategies for thermoset latex coatings that cure at ambient temperature involves the introduction of unsaturated groups that cure through autooxidation. One approach involves the use of allyl methacrylate in the latex synthesis, employing conditions in which the allyl groups survive the emulsion polymerization reaction. Another strategy involves the postfunctionalization of carboxyl latex with glycidyl methacrylate or other reactive methacrylates to introduce methacrylate groups into the latex particles. With a drier present, these groups undergo free radical polymerization in the dry coating.

To achieve optimum film properties in a thermoset latex film, the rate of polymer diffusion must be greater at the cure temperature than the rate of the crosslinking chemical reaction. In two-pack formulations, the crosslinking agent is mixed with the resin just prior to application of the coating to the substrate. The two-pack-in-one-pot concept is based upon the idea that one could prepare a latex blend in which the two components have functional groups that do not come into contact in the aqueous dispersion. Thus, they cannot react until the dispersion dries to a dense film. Under these circumstances, one has to be aware of the possibility that the two components may be immiscible and react only at the interface between the two types of latex, or the two components may have limited miscibility. Under these circumstances, the chemical reaction can drive mixing between the two components.

Billion lb in U.S.

Special Purpose Coatings:  1.81
OEM Coatings:              2.13
Architectural Coatings:    3.64

Figure 1--Estimated current markets for waterborne resins used in the
U.S.

Table 1--[T.sub.g] (in [degrees]C) of Commercial Filming (Coalescence-
Promoting) Solvents

Commercial Name               [T.sub.g] (JP)  [T.sub.g]  [T.sub.g]
                                              (Hoy) (a)  (CRI) (b)

Exxate 700 Solvent[TM]          -129
Propyl Cellosolve[TM]           -127
Dowanol[TM] PM GE               -125
Dowanol[TM] EB                  -124                       -124.5
Hexyl Cellosolve[TM]            -119
Dowanol[TM] PPh (c)              -96
Dowanol[TM] PnP                 -118
Benzyl Alcohol (c)              -128
Dowanol[TM] PnB                 -116
Butyl Cellosolve[TM] Acetate    -112
Methyl Carbitol[TM]             -111
Hexyl Carbitol[TM]              -108
Butyl Carbitol[TM]              -108
Dowanol[TM] DPM                 -106
Carbitol[TM]                    -105            -103       -104
Dowanol[TM] DPnB                -104
Dowanol[TM] DPnP                -104
Eastasolve EEH                  -104
Propasol Filmer BEP[TM]         -103
TXIB[TM]                        -100
Butyl Carbitol[TM] Acetate      -100            -100
Arcosolv[TM] DPtB                -98
Aromatic A-150[TM]               -98
KP-140[TM]                       -96
t-Butyl acetoacetate             -95
Monomer QM-57T[TM]               -95
Carbitol[TM] Acetate                             -95
Arcosolv[TM] PtB                 -88
Dibutyl phthlate                 -87
Texanol[TM]                      -84             -80        -84.5
Pamolyn[TM] 300                  -80
SER (127)                                                   -78.6
Santicizer[TM] 123               -67

Commercial Name               Chemical Type

Exxate 700 Solvent[TM]        Ester solvent
Propyl Cellosolve[TM]         Ethylene glycol of monopropyl ether
Dowanol[TM] PM GE             Propylene glycol of methyl ether
Dowanol[TM] EB                Ethylene glycol of n-butyl ether
Hexyl Cellosolve[TM]          Ethylene glycol of n-hexyl ether
Dowanol[TM] PPh (c)           Propylene glycol of phenyl ether
Dowanol[TM] PnP               Propylene glycol of n-propyl ether
Benzyl Alcohol (c)            Benzyl alcohol
Dowanol[TM] PnB               Propylene glycol of n-butyl ether
Butyl Cellosolve[TM] Acetate  Butoxyethyl acetate
Methyl Carbitol[TM]           Dipropylene glycol of methyl ether
Hexyl Carbitol[TM]            Diethylene glycol of n-hexyl ether
Butyl Carbitol[TM]            Diethyl glycol of n-butyl ether
Dowanol[TM] DPM               Dipropylene glycol of methyl ether
Carbitol[TM]                  Ethylene glycol of ethyl ether
Dowanol[TM] DPnB              Dipropylene glycol of n-butyl ether
Dowanol[TM] DPnP              Dipropylene glycol of n-propyl ether
Eastasolve EEH                Diethyl glycol of 2-ethylhexyl ether
Propasol Filmer BEP[TM]       Butoxyethoxypropanol
TXIB[TM]                      Diester of 2,2,4-trimethyl-1,3-pentanediol
Butyl Carbitol[TM] Acetate    Butoxyethoxyethyl acetate
Arcosolv[TM] DPtB             Dipropylene glycol of t-butyl ether
Aromatic A-150[TM]            Aromatic hydrocarbon
KP-140[TM]                    Tributoxyethyl phosphate
t-Butyl acetoacetate          t-Butyl acetoacetate
Monomer QM-57T[TM]            Reactive methacrylate monomer
Carbitol[TM] Acetate          Ethoxyethoxyethyl acetate
Arcosolv[TM] PtB              Propylene glycol of t-butyl ether
Dibutyl phthlate              Dibutyl phthlate
Texanol[TM]                   2,2,4-trimethyl-1,3-pentanediol
                                monoisobutyrate
Pamolyn[TM] 300               Conjugated lineleic acid
SER (127)                     Propylene glycol monoversatate
Santicizer[TM] 123            Dibenzyl phthlate

(a) Reference 126.
(b) Reference 127.
(c) [T.sub.g]s were estimated from MFFT measurements.
(d) Dowanol, Carbitol, Proposal, and Cellosolve are trade names of Dow
Chemical Co.; Eastasolve, Texanol, TXIB, and Aromatic A-150 are trade
names of Eastman Chemical Co.; KP-140 is a trade name of FMC; Monomer
QM-57T is a trade name of Rohm and Haas Co.; Arcosolv is a trade name of
Lyondell; Santicizer is a trade name of Solutia; Exxate is a trade name
of Exxon, and Pamolyn is a trade name of Hercules. JP (Johnson Polymer).

Table 2--Formulation Used to Evaluate Multifunctional Carbodiimides 4
and 5, as well as the Triaziridine 3 (XAMA-7)

                                      Mass of Components
Formulation Components                  in Formulation

UCAR[R] Vehicle 462                       100.00
Dimethylethanol amine (50% in water)        0.40
Butyl Cellosolve                            7.42
Water                                       7.42
Crosslinker emulsion (27% in water)     Variable

Table 3--Ultimate Properties of Waterborne Coatings Prepared from
Multifunctional Carbodiimides

                                     Tensile
Compound (a)  [M.sub.c] (g/mole)  Strength (MPa)  Elongation

3                 2961                5.98           510
4                 4706                9.79           430
Isopropyl         4897                9.82           450
Cyclohexyl        3740                6.12           450
n-Butyl           3050                5.97           485
t-Butyl           3008                8.20           510
Phenyl            3088                9.14           560

(a) Level of crosslinker is 5 phr.

Table 4--MEK Double Rub Results for Films Formed from a Hydroxyl-
Containing Latex (a) Treated with Diisopropenyl Adipate (b)

Cure Temperature ([degrees]C)  MEK Double Rubs

100                                 54
120                                120
140                                300
160                                300

(a) Composition by weight: butyl acrylate (45%), styrene (45%),
methacrylic acid (5%), hydroxyethyl acrylate (5%), [T.sub.g] =
41[degrees]C.
(b) Level of crosslinker is 4.86 phr.

Table 5--Crosslinking Results of Cured Films Prepared from an Acrylic
Latex Modified with Reactive Methacrylate Monomers

                                            Reactive   Gel Fraction
Latex Modification                Cure      Diluent      (Panel)

Glycidyl methacrylate          Thermal (a)  None           88
Cyclohexylcarbodiimide ethyl
  methacrylate                 Thermal (a)  None           86
Cyclohexylcarbodiimide ethyl
  methacrylate                 Thermal (a)  TMPTA (c)      88
Glycidyl methacrylate          Photo (b)    PETA (d)       74
Cyclohexylcarbodiimide ethyl
  methacrylate                 Photo (b)    TMPTA (c)      37
t-Butylcarbodiimide ethyl
  methacrylate                 Photo (b)    PETA (d)       73

                                Pencil     Acetone
Latex Modification             Hardness  Double Rubs

Glycidyl methacrylate             3H        +100
Cyclohexylcarbodiimide ethyl
  methacrylate                    4H        +100
Cyclohexylcarbodiimide ethyl
  methacrylate                    3H        +100
Glycidyl methacrylate              H        Cut
Cyclohexylcarbodiimide ethyl
  methacrylate                     H        Cut
t-Butylcarbodiimide ethyl
  methacrylate                    3H        +100

(a) Thermal cure: t-butyl peroxybenzoate (5 phr); 160[degrees]C/30 min.
(b) Photocure: Initiator (Irgacure 651, 4 phr); 300 watts, 14 sec.
(c) Trimethylolpropane triacrylate.
(d) Pentaerythritol tetracrylate.

Table 6--Comparison of Crosslinking Effectiveness for Butyl
Methacrylate-Methyl Methacrylate Copolymer Latex Films With and Without
Diacetone Acrylamide (DAAM) as a Comonomer and Adipic Acid Dihydrazide
as a Crosslinker

Wt% of DAAM                              2.3        None
Wt% adipic acid dihydrazide              1.2        None
Pencil hardness                         H           F
Solvent resistance                      Excellent   POOR
Gel fraction (X100)                     99          None
Water resistance (23[degress]C, 24 hr)  Excellent   Good

References

(1) (a) Bufkin, B.G. and Grawe, J.R., "Survey of the Applications, Properties, and Technology of Crosslinking Emulsions: Part I," JOURNAL OF COATINGS TECHNOLOGY, 50, No. 641, 41 (1978); (b) Grawe, J.R. and Bufkin, B.G., "Survey of the Applications, Properties, and Technology of Crosslinking Emulsions: Part II," JOURNAL OF COATINGS TECHNOLOGY, 50, No. 643, 67 (1978); (c) Bufkin, B.G. and Grawe, J.R., "Survey of the Applications, Properties, and Technology of Crosslinking Emulsions: Part III," JOURNAL OF COATINGS TECHNOLOGY, 50, No. 644, 83 (1978); (d) Grawe, J.R. and Bufkin, B.G., "Survey of the Applications, Properties, and Technology of Crosslinking Emulsions: Part IV," JOURNAL OF COATINGS TECHNOLOGY, 50, No. 645, 70 (1978); (e) Bufkin, B.G. and Grawe, J.R., "Survey of the Applications, Properties, and Technology of Crosslinking Emulsions: Part V," JOURNAL OF COATINGS TECHNOLOGY, 50, No. 647, 65 (1978).

(2) (a) Gottlob, K., German Patent 254, 672 (1912); (b) Gottlob, K., German Patent 255, 129 (1912).

(3) Luther, M. and Hueck, C. (to I.G. Farbenindustrie A.G.), Products of Latex Character and a Process for Producing the Same, U.S. Patent 1,864,078 (1931).

(4) Wicks, Z.W. Jr., Jones, F.N., and Pappas, S.P., "Film Formation, Components, and Appearance," Organic Coatings: Science and Technology, John Wiley and Sons, Inc., New York, pp. 83-119, 1992.

(5) (a) Lee, Do Ik, senior scientist Dow Chemical Co., Private communication, 1999; (b) Thompson, S.J., The S/B Latex Story, Recollections of 'Can Do' at Dow, Dow Chemical Co. (1980).

(6) Watson, D., retired engineer, Rohm and Haas Co., Private communication, 1999.

(7) Miller, W., retired research scientist, UCAR Emulsion Systems, Union Carbide Corp., Private communication, 1999.

(8) Flory, P.J., Principles of Polymer Chemistry, Cornell University Press, London, pp. 359, 495-518, 1953.

(9) (a) Harkins, W.D., "A General Theory of the Mechanism of Emulsion Polymerization," J. Am. Chem. Soc., 69, 1428-1444 (1947); (b) Smith, W.V. and Ewart, R.H., "Kinetics of Emulsion Polymerization," J. Chem. Phys., 16 (6), 592-599 (1948).

(10) Dillion, R.E., Matheson, L.A., and Bradford, E.B., "Sintering of Synthetic Latex Particles," J. Colloid Sci., 6, 108-117 (1951).

(11) Henson, W.A., Taber, D.A., and Bradford, E.B., Ind. Eng. Chem., 45, 735 (1953).

(12) Brown, G.L., "Formation of Films from Polymer Dispersions," J. Polym. Sci., 22, 423-434 (1956).

(13) Vanderhoff, J.W., Tarkowski, H.L., Jenkins, M.C., and Bradford, E.B., "Theoretical Consideration of the Interfacial Forces Involved in the Coalescence of Latex Particles," J. Macromol. Sci., Chem., 1(2), 361-397 (1966).

(14) Sheetz, D.P., "Formation of Films by Drying of Latex," J. Appl. Polym. Sci., 9, 3759-3773 (1965).

(15) (a) Routh, A.F. and Russel, W.B., "Horizontal Drying Fronts in Evaporating Latex Films," AICHE J., 44, 2088 (1998); (b) Routh, A.F. and Russel, W.B., "A Process Model for Altex Film Formation: Limiting Regmes for Individual Driving Forces," Langmuir, 15, 7762-7773 (1999).

(16) Voyutskii, S., Autohesion and Adhesion of High Polymers, Wiley-Interscience, New York, 1963.

(17) Wang, Y., Kats, A., Juhue, D., Winnik, M.A., Shivers, R., and Dinsdale, C., "Freeze Fracture Studies of Latex Films Formed in the Absence and Presence of Surfactant," Langmuir, 8, 1435-1442 (1992).

(18) (a) Wang, Y., Juhue, D., Winnik, M.A., Leung, O.M., and Goh, M.C., "An Atomic Force Microscopy Study of Latex Film Formation," Langmuir, 8, 760-762 (1992); (b) Goh, M.C., Juhue, D., Leung, O., Wang, Y., and Winnik, M.A., "Annealing Effects on the Surface Structure of Latex Films Studied by Atomic Force Microscopy," Langmuir, 9, 1319-1322 (1993).

(19) Hahn, K., Ley, G., Schuller, H., and Oberthur, R., "On Particle Coalescence in Latex Films," Colloid & Polym. Sci., 264, 1092-1096 (1986).

(20) (a) Linne, M.A., Klein, A., Miller, G.A., and Sperling, L.H, "Film Formation from Latex: Hindered Initial Interdiffusion of Constrained Polystyrene Chains Characterized by Small-Angle Neutron Scattering," J. Macromol. Sci., Phys., B27, 217-231 (1988); (b) Yoo, J.N., Sperling, L.H., Glinka, C.J., and Klein, A., "Characterization of Film Formation from Polystyrene Latex Particles via SANS. I. Moderate Molecular Weight," Macromolecules, 23, 3962-3967 (1990).

(21) Zhao, C.L., Wang, Y., Hruska, Z., and Winnik, M.A., "Molecular Aspects of Latex Film Formation. An Energy Transfer Study," Macromolecules, 23, 4082-4087 (1990).

(22) Rubinstein, M. and Colby, R., Polymer Physics, Oxford, New York, 2003.

(23) Flory, P.J. and Rehner, J., "Statistical Mechanics of Crosslinked Polymer Networks," Chem. Phys., 11, 521-526 (1943).

(24) Flory, P.J., "Statistical Mechanics of Swelling of Network Structures," Chem. Phys., 18, 108-111 (1950).

(25) Mark, J.E. and Erman, B., Structure and Properties of Rubberlike Networks, Oxford University, 1995.

(26) Mark, J.E. and Erman, B., Rubberlike Elasticity: A Molecular Primer, Wiley-Interscience, 1988.

(27) Taylor, J.W. and Bassett, D.R., in Technology of Waterborne Coatings, Glass, J.E. (Ed.), ACS Symposium Series 666, American Chemical Society, Washington, D.C., pp. 137-163 (1997).

(28) Billingham, N.C., Molar Mass Measurements in Polymer Science, John Wiley and Sons Inc., New York, 1977.

(29) Flory, P.J., "Molecular Size Distribution in Three-Dimensional Polymers. V. Post-Gelation Relationships," J. Am. Chem. Soc., 62, 30-35 (1947).

(30) Reiser, A. and Pitts, E., "A Photographic Theory of Crosslinking Resins," J. Photogr. Sci., 29, 187-191 (1981).

(31) Reiser, A. and Pitts, E, "Characteristic Curve of Crosslinking Photoresists," Photogr. Sci. and Eng., 20 (5), 225-229 (1976).

(32) Miller, D.R. and Macosko, C.W., "Network Parameters for Crosslinking of Chains with Length and Site Distribution," J. Polym. Sci: Part B: Polym. Phys., 26, 1-51 (1988).

(33) Bauer, D.R. and Budde, G.F., "Crosslinking Chemistry and Network Structure in High-Solids Acrylic-Melamine Coatings," Ind. Eng. Chem, Prod. Res. Dev., 20, 674-679 (1981).

(34) Bauer, D.R. and Dickie, R.A., "Development and Application of Network Structure Models to Optimization of Bake Conditions for Thermoset Coatings," JOURNAL OF COATINGS TECHNOLOGY, 58, No. 738, 41 (1986).

(35) Bauer, D.R., "How to Calculate Crosslink Structure in Coatings," JOURNAL OF COATINGS TECHNOLOGY, 60, No. 758, 53 (1988).

(36) Joanicot, M., Wong, K., Maquet, J., Chevalier, Y., Pichot, C., Graillat, C., Lindner, P., Rios, L., and Cabane, B., "Ordering of Latex Particles During Film Formation," Prog. Colloid Polym. Sci., 81, 175-183 (1990).

(37) Tzitzinou, A., Keddie, J.L., Guerts, J.M., Peters, A.C.I.A., and Satguru, R., "Film Formation of Latex Blends with Bimodal Size Distributions: Consideration of Particle Deformability and Continuity of the Dispersed Phase," Macromolecules, 33, 2695-2708 (2000).

(38) Roulstone, B.J., Wilkinson, M.C., Hearn, J., and Wilson, A.J., "Studies on Polymer Latex Films: I. A Study of Latex Film Morphology," Polym. Int., 21, 87-94 (1991).

(39) Distler, D. and Kanig, G., "Fienstruktur von Polymeren aus Wassriger Dispersion," Colloid Polym. Sci., 256, 1052-1060 (1978).

(40) Chevalier, Y., Pichot, C., Graillat, C., Joanicot, M., Wong, K., Maquet, J., Lindner, P., and Cabane, B., "Film Formation with Latex Particles," Colloid Polym. Sci., 270, 806-821 (1992).

(41) Joanicot, M., Wong, K., Richard, J., Maquet, J., and Cabane. B., "Ripening of Cellular Films," Macromolecules, 26, 3168-3175 (1993).

(42) Sperling, L.H., Klein, A., Sambasivam, M., and Kim, K.D., "Molecular Basis of Healing and Fracture at Polymer Interfaces," Polym. Adv. Technol., 5, 453-472 (1993).

(43) Kim, K.D., Sperling, L.H., Klein, A., and Wignall, G.D., "Characterization of Film Formation from Direct Mini-Emulsified Polystyrene Latex Particles via SANS," Macromolecules, 26, 4624-4631 (1993).

(44) Wool, R.P., Structure and Strength of Polymer Interfaces, Hanser Press, New York, 1995.

(45) Mohammadi, N., Klein, A. and Sperling, L.H., "Polymer Chain Rupture and Fracture Behavior of Glassy Polystyrene. I," Macromolecules, 26, 1019-1026 (1993).

(46) Kim, K.D., Sperling, L.H., Klein, A., and Hammouda, B., "Reptation Time, Temperature, and Cosurfactant Effects on the Molecular Interdiffusion Rate During Polystyrene Latex Film Formation," Macromolecules, 27, 6841-6850 (1994).

(47) Sambasivam, M., Klein, A., and Sperling, L.H., "Energy-Consuming Micromechanisms in the Fracture of Glassy Polymers. 2: Effect of Molecular Weight on the Fracture of Polystyrene," Macromolecules, 28, 152-159 (1995).

(48) Feng, J., Yekta, A., and Winnik, M.A., "Direct Non-Radiative Energy Transfer Across a Sharp Polymer Interface," Chem. Phys. Lett., 260, 296-301 (1996).

(49) Liu, Y.S., Feng, J., and Winnik, M.A., "Study of Polymer Diffusion Across the Interface in Latex Films Through Direct Energy Transfer Experiments," J. Chem. Phys., 101, 9096-9103 (1994).

(50) Farinha, J.P.S., Martinho, J.M.G., Yekta, A., and Winnik, M.A., "Direct Nonradiative Energy Transfer in Polymer Interphases: Fluorescence Decay Functions from Concentration Profiles Generated by Fickian Diffusion," Macromolecules, 28, 6084-6088 (1995).

(51) Boczar, E.M., Dionne, B.C., Fu, Z., Kirk, A.B., Lesko, P.M., and Koller, A.D., "Spectroscopic Studies of Polymer Interdiffusion During Film Formation," Macromolecules, 26, 5772-5781 (1993).

(52) Wang, Y. and Winnik, M.A., "Energy Transfer Study of Polymer Diffusion in Melt-Pressed Films of Poly(methyl methacrylate)," Macromolecules, 26, 3147-3150 (1993).

(53) Winnik, M.A., Wang, Y., and Haley, F., "Latex Film Formation at the Molecular Level: The Effect of Coalescing Aids on Polymer Diffusion," JOURNAL OF COATINGS TECHNOLOGY, 64, No. 811, 51 (1992).

(54) Juhue, D., Wang, Y., and Winnik, M.A., "Influence of a Coalescing Aid on Polymer Diffusion in Poly(butyl methacrylate) Latex Films," Makromol. Chem., Rapid Commun., 14, 345-349 (1993).

(55) Kawaguchi, S., Odrobina, E., and Winnik, M.A., "Non-Ionic Surfactant Effects on Polymer Diffusion in Poly(n-butyl methacrylate) Latex Films," Makromol. Chem., Rapid Commun., 16, 861-868 (1995).

(56) Feng, J., Pham, H., Stoeva, V., and Winnik, M.A. "Polymer Diffusion in Latex Films at Ambient Temperature," J. Polym. Sci.: Part B: Polym. Phys., 36, 1129-1139 (1998).

(57) Kim, H.B. and Winnik, M.A., "Effect of Surface Acid Group Neutralization on Interdiffusion in Latex Films," Macromolecules, 27, 1007-1012 (1994).

(58) Juhue, D. and Lang, J., "Film Formation from Dispersions of Core-Shell Latex Particles," Macromolecules, 28, 1306-1308 (1995).

(59) Rharbi, Y., Boue, F., Joanicot, M., and Cabane, B., "Deformation of Cellular Polymeric Films," Macromolecules, 29, 4346-4359 (1996).

(60) Kim, H.B. and Winnik, M.A., "Factors Affecting Interdiffusion Rates in Films Prepared from Latex Particles with a Surface Rich in Acid Groups and Their Salts," Macromolecules, 28, 2033-2041 (1995).

(61) Taylor, J.W. and Klots, T., "Applied Approach to Film Formation: The Glass Temperature Evolution of Plasticized Latex Films," Eur. Coat. J., 6, 38 (2002).

(62) Taylor, J.W. and Klots, T., "An Applied Approach to Film Formation for Waterborne Coatings," Proc. of the Twenty-Ninth International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 181 (2002).

(63) Zosel, A. and Ley, G., "Influence of Crosslinking on Structure, Mechanical Properties, and Strength of Latex Films," Macromolecules, 26, 2222-2227 (1993).

(64) Tamai, T., Pinenq, P., and Winnik, M.A., "Effect of Crosslinking on Polymer Diffusion in Poly(butyl methacrylate-co-butyl acrylate) Latex Forms," Macromolecules, 32, 6102-6110 (1999).

(65) Winnik, M.A., "Interdiffusion and Crosslinking in Thermoset Latex Films," JOURNAL OF COATINGS TECHNOLOGY, 74, No. 925, 49 (2002).

(66) Liu, R., Winnik, M.A., Di Stefano, F., and Venkatesan, J., "Interdiffusion vs. Crosslinking Rates in Isobutoxyacrylamide-Containing Latex Coatings," Macromolecules, 34, 7306-7314 (2001).

(67) Aradian, A., Raphael, E., and de Gennes, P.G., "Strengthening of a Polymer Interface: Interdiffusion and Crosslinking," Macromolecules, 33, 9444-9451 (2000).

(68) Aradian, A., Raphael, E., and de Gennes, P.G., "Interdiffusion and Crosslinking at Polymer Interfaces," Macromolecules, 35, 4036-4043 (2002).

(69) Raphael, E. and de Gennes, P.G., "Rubber-Rubber Adhesion with Connector Molecules," J. Phys. Chem., 96, 4002-4007 (1992).

(70) Ahagon, A. and Gent, A., "Effect of Interfacial Bonding on the Strength of Adhesion," J. Polym. Sci. Polym. Phys. Ed., 13, 1285-1300 (1995).

(71) Ji, H. and de Gennes, P.G., "Adhesion via Connector Molecules: The Many-Stitch Problem," Macromolecules, 26, 520-525 (1993).

(72) Lake, G.J. and Thomas, A.G., "The Strength of Highly Elastic Materials," Proc. R. Soc. London, Ser A., 300, 108-119 (1967).

(73) Taylor, J.W., Collins, M.J., and Clark, M.D., Waterborne Polymers Using Pendant Allyl Groups, U.S. Patent 5,539,073 (1996).

(74) Bassett, D.R. and Hamielec E.A., Emulsion Polymers and Emulsion Polymerization, ACS Symposium Series 165, American Chemical Society, Washington D.C., pp. 371-387 (1981).

(75) Hoy, K.L., "Effect of Reaction Pathway on Emulsion Polymer Structure," JOURNAL OF COATINGS TECHNOLOGY, 51, No. 651, 27 (1979).

(76) (a) Feng, J., Odrobina, E., and Winnik, M.A., "Effect of Hard Polymer Filler Particles on Polymer Diffusion in a Low-[T.sub.g] Latex Film," Macromolecules, 31, 5290-5299 (1998); (b) Winnik, M.A. and Feng, J., "Latex Blends: An Approach to Zero-VOC Coatings," JOURNAL OF COATINGS TECHNOLOGY, 68, No. 852, 39 (1996).

(77) Eckersley, S.T. and Helmer, B.J., "Mechanistic Consideration of Particle Size Effects on Film Properties of Hard/Soft Latex Blends," JOURNAL OF COATINGS TECHNOLOGY, 69, No. 864, 97 (1997).

(78) Winnik, M.A., Pinenq, P., Kruger, C., Zhang, J., and Yaneff, P.V., "Crosslinking vs. Interdiffusion Rates in Melamine-Formaldehyde Cured Latex Coatings: A Model for Waterborne Automotive Basecoat," JOURNAL OF COATINGS TECHNOLOGY, 71, No. 892, 47 (1999).

(79) Bassett, D.R. and Sherwin, M.A., "Crosslinking Studies of Thermosetting Emulsion Polymers," Waterborne and Higher Solids Coatings Symposium of the Southern Society of Coatings Technology, New Orleans, LA, February 1977.

(80) (a) Hahn, K.G., Thermosetting Acrylic Latexes, U.S. Patent 4,812,491 (1989); (b) Kunz, B.L. and Hahn, K.G., Pigmented Low Cure Emulsion Polymers, U.S. Patent 4,981,883 (1991).

(81) Krishnan, S., Klein, A., El-Aasser, M.S., and Sudol. E.D. "Influence of a Chain Transfer Agent on the Crosslinking of Poly(n-butyl methacrylate-co-methylol Acrylamide) Latex Particles and Films," Macromolecules, 36, 3511-3518 (2003).

(82) Rodgers, J.R., Organic Film Former Compositions of Self Sensitive Systems, U.S. Patent 3,308,078 (1965); (b) Rodgers, J.R., Coating Composition Comprising a Terpolymer, an Alkali Resin, and a Zirconyl-Fugitive Ligand Compound, U.S. Patent 3,320,196 (1965).

(83) (a) Karsa, D.R., Additives for Water-Based Coatings, Royal Chemical Society, London, (1990); (b) Moles, P., MEL Chemical's Data Sheet 117 (2003).

(84) "Polyfunctional Aziridines," Cordova Chemical Co., Commercial Literature (1983).

(85) Trentini, M.C., Gerosa, P., and Carlson, V., "Formaldehyde-Free Waterborne Acrylic Wood Coatings," Eur. Coat. J., 5, 362-367 (1995).

(86) (a) Nissihinbo Industries Literature (2002); (b) "Polycarbodiimide Crosslinkers," Union Carbide Literature.

(87) (a) Taylor, J.W., Monodispersed Multifunctional Carbodiimides, U.S. Patent 5,117,059 (1992); (b) Taylor, J.W., Emulsion and Crosslinkable Compositions Containing Polycarbodiimides, U.S. Patent 5,081,173 (1992); (c) Taylor, J.W., Surface Active Polycarbodiimides and Dispersions, U.S. Patent 5,047,588 (1991).

(88) Taylor, J.W., Surface Active Polycarbodiimides, U.S. Patent 4,820,863 (1989).

(89) Jacobs, J.M., Muijs, P.E., Steven van Es, J.G., and German A.L.J., "Molecular Mass Control in Methacrylic Copolymer Latexes Containing Glycidyl Methacrylate," Appl. Polym. Sci., 61, 9-19 (1996).

(90) (a) Taylor, J.W., Functional Latexes Resistant to Hydrolysis, International Patent Application (PCT), WO 17 698 A1 (1998); (b) Taylor, J.W., Functional Latexes Resistant to Hydrolysis, U.S. Patent 5,962,556 (1999).

(91) Tronc, F., Liu, R., Winnik, M.A., Eckersley, S.T., Rose, G.D., Weishun, J.M., and Meunier, D.M. "Epoxy-Functionalized, Low Glass-Transition Temperature Latex 1. Synthesis, Characterization, and Polymer Interdiffusion," J. Polym. Sci. Part A: Polym. Chem., 40, 2609-2625 (2002).

(92) Tronc, F., Chen, W., Winnik, M.A., Eckersley, S.T., Rose, G.D., Weishun, J.M., and Meunier, D.M. "Epoxy-Functionalized, Low Glass-Transition Temperature Latex II. Interdiffusion vs. Crosslinking in the Presence of a Diamine," J. Polym. Sci. Part A: Polym. Chem., 40, 4098-4116 (2002).

(93) Nippon Shokubia Literature (1994).

(94) Murray, R.E., Eaton, R.F., Upshaw, T.A., Taylor, J.W., and Bassett, D.R., Process Utilizing Alkenyl Carboxylate Crosslinkers, U.S. Patent 5,412,038 (1995).

(95) (a) Hubert, J.C., Venderbosch, R.A.M., Muizebelt, W.J., Klaasen, R.P., and Zabel, K.H., "Singlet Oxygen Drying of Alkyd Resins and Model Compounds," JOURNAL OF COATINGS TECHNOLOGY, 69, No. 869, 59 (1997); (b) Muizebelt, W.J., Donkerbroek, J.J., Nielen, M.W.F., Hussem, J.B., Biemond, M.E.F., Klaasen, R.P., and Zabel, K.H., "Oxidative Crosslinking of Alkyd Resins Studied with Mass Spectrometry and NMR Using Model Compounds," JOURNAL OF COATINGS TECHNOLOGY, 70, No. 876, 83 (1998).

(96) Tillson, H.C., Coating Compositions Comprising a Terpolymer of an Alkyl Acrylate, an Alkyl Methacrylate, and an Ester of a Unsaturated Alcohol with Methacrylic Acid, U.S. Patent 3,219,610 (1965).

(97) Heathley, F.H., Lovell, P.A., and McDonald, J., "NMR Studies of Free-Radical Polymerization and Copolymerization of Monomers and Polymers Containing Allyl Groups," Eur. Polym. J., 29, No.2/3, 255-268 (1993).

(98) (a) Porter, N.A., Lehman, L.S., Weber, B.A., and Smith, K.J., "Unified Mechanism for Polyunsaturated Fatty Acid Autooxidation. Competition of Peroxy Radical Hydrogen Atom Abstraction, [beta]-Scission, and Cyclization," J. Am. Chem. Soc., 103, 6447-6455 (1981); (b) Porter, N.A. and Wujek, D.G., "Autooxidation of Polyunsaturated Fatty Acids, an Expanded Mechanistic Study," J. Am. Chem. Soc., 106, 2626-2629 (1984); (c) Coyle, J.D., Introduction to Organic Photochemistry, John Wiley & Sons, Great Britain (1986).

(99) (a) Brister, E.H., Smith, O.W., and Thames, S.F., "Castor Acrylated Monomer (CAM) in Vinyl-Acrylic Latexes," Twenty-Sixth International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 89 (1999); (b) King, C.L., Smith, O.W., and Thames, S.F., "Acrylic Properties Using CAM Monomer," Proc. Twenty-Sixth International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 100 (1999); (c) King, C.L., Smith, O.W., and Thames, S.F., "Oxidative Curing Profiles of CAM Based Acrylic Latexes," Proc. Twenty-Seventh International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 436 (2000); (d) King, C.L., Smith, O.W., and Thames, S.F., "Oxidative Curing Profiles of CAM Based Acrylic Latexes," Proc. Twenty-Eighth International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 143 (2001); (e) Thames, S.F., Panjnani, K.G., and Fruchey, O.S., Latex Compositions Containing Ethylenically Unsaturated Esters of Long-Chain Alkenols, U.S. Patent 6,001,913 (1999); (f) Thames, S.F., Panjnani, K.G., and Fruchey, O.S., Latex Compositions Containing Ethylenically Unsaturated Esters of Fatty Acids and Applications Thereof, U.S. Patent 6,174,948 (2001); (g) Thames, S.F., Panjnani, K.G., Hariharan, R., and Wang, Z., Internally Plasticizing and Crosslinkable Monomers and Applications thereof, U.S. Patent 6,174,948 (2001).

(100) McGinniss, V.D., Seidewand, J.R., and Robert, J., Ultraviolet Curable Latexes, U.S. Patent, 4,107,013 (1978).

(101) (a) Taylor, J.W. "A Study on the Chemistry of Alkylcarbodiimide Ethyl Methacrylates as Reactive Monomers for Acrylic and Vinyl Ester-based Latexes," Proc. for XXIVth International Conference in Organic Coatings Science and Technology, Athens, Greece (1995); (b) Taylor, J.W., Collins, M.J., and Bassett, D.R., "A Study on the Chemistry of Alkylcarbodiimide Ethyl Methacrylates as Reactive Monomers for Acrylic and Vinyl Ester-based Latexes," Prog. Org. Coat., 35, 215-221 (1999).

(102) Padget, J.C., "Polymers for Water-Based Coatings--A Systematic Overview," JOURNAL OF COATINGS TECHNOLOGY, 66, No. 839, 89 (1994).

(103) Mylonakis, S.G., Air Curable Latex, U.S. Patent 4,244,850 (1981).

(104) Wolfersberger, M.H., Schinder, F.J., Beckely, R.S., and Novak, R.W., Functionalized Multistage Polymers, U.S. Patent 5,306,744 (1994).

(105) Bors, D.A., Air Curing Polymer Compositions, European Patent Application, 492 847 A2 (1991).

(106) Bors, D.A., Lavoie, A.C., and Emmons, W.D., Air Curing Compositions, U.S. Patent 5,484,849 (1996).

(107) Pears, D.A. and Overbeek, G.C., Incorporation of Desired Groups into Polymer, the Polymer so Produced, and Composition Containing Them, European Patent Application, 442 653 A2 (1991).

(108) Taylor, J.W. and Collins, M.J., Waterborne Polymers with Pendant Crosslinkable Groups, International Patent Application (PCT), WO 32 424 A3 (1996).

(109) Del Rector, F., Blount, W.W., and Leonard, D.R., "Application of the Acetoacetyl Functionality in Thermoset Coatings," Waterborne and Higher Solids Coatings Symposium, New Orleans, LA (1988).

(110) Geurink, P.A.J., van Dalen, L., van der Ven, L.G.J., and Lamping, R.R., "Analytical Aspects and Film Properties of Two-Pack Acetoacetate Functional Latexes," Prog. Org. Coat., 27, 73-78 (1996).

(111) (a) Feng, J., Pham, H., MacDonald, P., Winnik, M.A., Guerts, J., Zirkee, H., and van Es, S., German, A.L., "Formation and Cross-linking of Latex Films Through the Reaction of Acetoacetoxy Groups with Diamines Under Ambient Conditions," JOURNAL OF COATINGS TECHNOLOGY, 70, No. 881, 57 (1998); (b) Feng, J., "Molecular and Environmental Aspects of Latex Film Formation," Ph.D. Thesis, University of Toronto, 1996.

(112) Kyowa Hakko U.S.A., Literature bulletin DAA01 (1999).

(113) De Krom, A., Mulder, H., and Mestach, D., Wet Polymer on Nonabsorbing Substrates: Waterborne Acrylate Emulsions for Non-Absorbing Substates with Exceptional Removability, Farbe and Lack, 107, 97-105 (2001).

(114) De Krom, A., Mulder, H., and Mestach, D., "New Developments in Waterborne Printing Inks for Non-Absorbing Substrates," 6th Nurnburg Conference, 173-190 (2001).

(115) Buckmann, F., Overbeek, G.C., and Nabuurs, T., "Self-Crosslinking Surfactant Free Acrylic Dispersion for High Performance Coating Applications," European Congress on Printing Inks, 701-708 (2001).

(116) Overbeek G.C. and Heuts M.P.J., Aqueous Coating Systems, U.S. Patent 4,988,762 (1991).

(117) Overbeek, G.C., Smak, W., Buckman, A.J.P., and Padget, J.C., Production of Aqueous Polymer Compositions, U.S. Patent, 5,859,112 (1999).

(118) Esser, R.J., Devona, J.E., Setzke, D.E., and Wagemans, L., "Waterbased Crosslinkable Surface Coatings," Eur. Coat. J., 10, 732-733 (1998).

(119) Taylor, J.W. and Collins, M.J., Stable Amino-Containing Polymer Latex Blends, U.S. Patent 5,998,543 (1999).

(120) Taylor, J.W. and Collins, M.J., Surfactant-Containing Acetoacetoxy-Functional and Enamine-Functional Polymers, U.S. Patent 6,028,155 (2000).

(121) (a) Pham, H.H. and Winnik, M.A., "Synthesis, Characterization, and Stability of Carbodiimide Groups in Carbodiimide-Functionalized Latex Dispersions and Films," J. Polym. Sci. Part A: Polym. Chem., 38, 855-869 (2000); (b) Pham, H.H., "Polymer Interdiffusion vs. Crosslinking in Carboxylic Acid-Carbodiimide Latex Films," Ph.D. Thesis, University of Toronto, 1999.

(122) Pham, H.H. and Winnik, M.A., "Polymer Interdiffusion vs. Crosslinking in Carboxylic Acid-Carbodiimide Latex Films," Macromolecules, 32, 7692-7695 (1999).

(123) (a) Pham, H.H., Farinha, J.P.S., and Winnik, M.A., "Crosslink-ing, Miscibility, and Interface Structure in Blends of Poly(2-Ethylhexyl Methacrylate) Copolymers. An Energy Transfer Study," Macromolecules, 33, 5850-5862 (2000); (b) Pham, H.H. and Winnik, M.A., "Film Formation from Blends of Carbodiimide and Carboxylic Acid-Functional Latex," in Film Form-ation from Waterborne Dispersions, Provder, T. (Ed.), ACS Books, Washington, D.C., 2000.

(124) (a) Guerts, J.M., "Latices with Intrinsic Crosslink Activity," Ph.D. Thesis, Eindhoven University, 1998; (b) Guerts, J.M., van Es, J.J.G.S., and German, A.L., "Latexes with Intrinsic Crosslink Activity," Prog. Org. Coat., 29, 107-115 (1996).

(125) Mohammed, S., Daniels, E.S., Sperling, L.H., Klein, A., and El-Aasser, M.S., "Isocyanate-Functionalized Latexes: Film Formation and Tensile Properties," J. Appl. Polym. Sci., 66, 1869-1884 (1997).

(126) Hoy, K.L, "Estimating the Effectiveness of Latex Coalescing Aids," JOURNAL OF PAINT TECHNOLOGY, 45, No. 579, 51 (1973).

(127) Toussanit, A., Wilde, M. De, Molenaar, F., and Mulvihill, J., "Calculation of [T.sub.g] and MFFT Depression Due to Added Agents," Prog. Org. Coat., 30, 179 (1997).

James W. Taylor--Johnson Polymer*

Mitchell A. Winnik--University of Toronto ([dagger])

*8310 16th St., P.O. Box 902, MS712, Sturtevant, WI 53177-0902, email: [e-mail].

([dagger]) Dept. of Chemistry, 80 St. George St., Toronto, Ont., Canada, M5S 3H6, email: [e-mail].