410

HEAT STABILIZERS

Vol. 6

HETEROPHASE POLYMERIZATION Introduction Heterophase polymerization is a generic term that describes polymerization reactions under nonhomogeneous conditions with respect to physical and chemical properties of the reaction mixture. This means the existence of gradients such as in density and chemical composition and consequently, the coexistence of different phases. Heterophase polymerizations are best defined in a very general way as processes resulting in polymer dispersions. Polymer dispersion is a state of matter where polymers are finely dispersed in a continuous phase. In this sense these polymerization processes are colloidal in nature and thus bridge two important scientific areas: polymer science and colloid science. Colloid science characterizes systems by two general features, heterogeneity and dispersity, which necessarily occur mutually. Heterogeneity means the coexistence of different phases and is a qualitative feature. It implies that phase boundaries, processes at phase boundaries, mixed interfaces, and in some cases phase formations also are of importance. Dispersity is a quantitative measure that characterizes the degree of division of the heterophase system. It is defined as the reciprocal average characteristic length scale. Besides dispersity the shape, the size, and the interface per mass of the dispersed phase are other characteristic features. In particular, dispersity is responsible for new properties of colloidal systems absent in homogeneous ones, because with increasing dispersity interfacial effects become more and more important. Furthermore, heterogeneity and dispersity together cause an energetic characterization of a dispersion, that is, the existence of an interfacial tension between coexisting phases and an interfacial free energy (interfacial tension times interfacial area). The interfacial free energy contributes substantially to the overall free energy of dispersions, and in many cases it governs the Encyclopedia of Polymer Science and Technology. Copyright John Wiley & Sons, Inc. All rights reserved.

Vol. 6

HETEROPHASE POLYMERIZATION

411

Table 1. Colloidal Dispersions Made of Various Combinations of States of Matter Dispersed phase

Continuous phase

Liquid Solid Gas Liquid Solid Gas Liquid Solid

Gas Gas Liquid Liquid Liquid Solid Solid Solid

Examples

Technical term

Clouds/sprays Smog/dust Lather/meringue Milk/creams Paint/blood Pumice/cork Oil shale/opals Bones/teeth

Aerosol Aerosol Foam Emulsion Suspension Solid foam Porous material Solid suspension

behavior completely. Also, the interface is responsible for all peculiarities, for all pros and cons of heterophase polymerization techniques. The context of heterogeneity and dispersity requires that the material forming the dispersed phase is insoluble/immiscible in the continuous phase. From that follows directly that the interfaces in dispersed systems can be made of all possible combinations of states of matter except the gas–gas combination. Examples of dispersed colloidal systems are given in Table 1. It is interesting to note that indeed all of these combinations can be used to carry out polymerizations. The various kinds of heterophase polymerization techniques are put together in Table 2 in the order of ascending number of components; the components are listed, except the initiating system. In most polymerizations the kind of dispersed system changes in the course of the process owing the fact that a gaseous or liquid monomer is transformed into a solid polymer. Examples are catalytic gas-phase polymerizations on solid catalyst beads (1) or polymerizations in liquid continuous phases of gaseous monomers. Heterophase polymerization can also be carried out in pores of inorganic solid materials such as Zeolites (2–5), or mesoporous MCM-41 and similar silicates (6,7), or inside interlayers of montmorillonite (8). Other special types of heterophase polymerizations in a solid continuous phase are used to modify the pores in solid polymer monoliths (9) or pores in polymeric membranes (10). Table 2. Overview of Heterophase Polymerization Techniques Technique (common name)

Components a

Bulk or mass polymerization Gas-phase polymerization Precipitation polymerization Suspension polymerization Microsuspension polymerization Dispersion polymerization Emulsion polymerization Miniemulsion polymerization Microemulsion polymerization a Only

Monomer, polymer Monomer, polymer Dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer Stabilizer, dispersion medium, monomer, polymer

if the polymer is insoluble in its own monomer such as vinyl chloride/poly (vinyl chloride) or acrylonitrile/polyacrylonitrile.

412

HETEROPHASE POLYMERIZATION

Vol. 6

The dramatic changes in the physico-chemical properties on the way from monomers to polymers leads to changes in the overall reaction conditions during any kind of polymerization. In the case of heterophase polymerizations these changes have a considerable influence on the colloidal properties as well, raising questions with respect to the existence of equilibrium situations of important colloid chemical properties. In this sense the final, stable polymer dispersion is a colloidal system in equilibrium and very well suited as models for colloid chemical investigations. In practice, the term polymer dispersion is used almost exclusively for systems with a liquid continuous phase. In this sense and from the colloid chemical point of view, heterophase polymerizations deal with emulsions and suspensions. A polymer dispersion is nowadays frequently also called a latex, which in Latin means “liquid or “fluid,” and goes back to the Greek word latax, ´ which means “droplet”. The word latex has become a generic term that comprises all kinds of polymer dispersions (11). In this article the terms polymer dispersion, polymer suspension, and latex are used interchangeably. Synonyms for the continuous phase are dispersion medium, homogenous phase, and serum. Any liquid can be a dispersion medium provided it is a nonsolvent for the dispersed material. But in fact, by far the most important industrial dispersion agent is water, for cost, safety, and environmental reasons. In this article, heterophase polymerization and polymer dispersion means an aqueous continuous phase whereas the term inverse, put in front, denotes systems with organic continuous phases. Furthermore, a chaingrowth polymerization by a free-radical initiation mechanism is considered, if not otherwise stated. The aim of this article is to provide a condensed view of heterophase polymerizations in liquid media, ie to point out common features rather than to give comprehensive or detailed views on peculiarities of the different kinds of heterophase polymerizations. The list of references is intended to be useful as a starting point for further literature searches rather than an attempt to provide all references, which would be daunting to author and reader alike. After a brief look at the history and the actual meaning of polymer dispersions, general features will be considered, which thereafter will be applied to the variants of heterophase polymerizations. History. It is historically proven that the Mayas already in 1600 BC were able to use the sap of certain trees to make out of it useful things such as rubber balls, glue, waterproofed clothes, medicine, and much more (12–14). They called this sap caa o-chu “crying wood,” from which the French word for rubber, caoutchouc, was derived. A broader scientific interest in caoutchouc was initiated by the report of Charles de la Condamine and Francois Fresneau on their tour through South America in 1736 and their experiences with the milky white resin and the products thereof (15). There is no doubt that the development of polymer science has been strongly connected with polymer dispersions, ie the investigation of natural rubber (see RUBBER, NATURAL). Michael Faraday determined the exact composition of natural rubber as hydrocarbon in 1826 (16), and precisely described the separation of the rubber and its cleaning—basically with the same techniques that are applied still today. After the discovery of vulcanization by Charles Goodyear in 1839, natural rubber became a raw material of increasing importance. Thirty years later, C. Greville Williams isolated by destructive distillation caoutchouc decomposition products and named the lightest isoprene (17)

Vol. 6

HETEROPHASE POLYMERIZATION

413

and concluded that the caoutchouc hydrocarbon is composed of isoprene polymers. Less than two decades later it was claimed that the reverse process was possible, ie, rubber could be synthesized by polymerizing isoprene (18). Note, the term polymeric was introduced 1832 by J. Berzelius from the Greek π oλνζ , which means “several,” to describe substances possessing the same composition but different properties (19). In the first decade of the 20th century a team headed by F. Hofmann in Germany was successful in developing polymerization of isoprene either in bulk or in benzene solution. The polymerizations were initiated thermally and carried out in autoclaves at temperatures below 250◦ C for 10–150 h. (20). The idea of using an aqueous emulsion of a monomer to carry out polymerization dates back to 1912, when the first patent was filed (21). This patent represents the birth of what we call today heterophase polymerization, and Kurt Gottlob was the pioneer researcher (22,23). In order to copy the conditions employed by Nature during the synthesis of caoutchouc, Gottlob used viscous fluids such as egg albumin, starch, or gelatin as stabilizers. The addition of alkali or organic bases leads to an improvement of the latex stability, and coagulation was possible by decreasing the pH. Polymerization was started thermally with 7 g of egg albumin in 500 g of water and 300 mL of isoprene in an autoclave. The mixture was stirred for several weeks at 60◦ C. Significant progress in the development of heterophase polymerizations was made during the following decades, when typical surfactants such as oleates and alkyl aryl sulfonates were used together with water as well as monomer-soluble peroxides (24–26). An almost comprehensive overview of the development of synthetic rubber and products thereof until the early thirties can be found in References 27 and 28. A description of developments that took place on polymerizations of olefins and diolefins during the following period of time until the end of World War II can be found in References 29 and 30. The pace of development in these early days of heterophase polymerization was mainly determined in industrial research labs, with the benefit that the time span between first laboratory results and commercially available products was short. After World War II more and more research activities were started at universities and independent research institutes all over the world. Summaries of today’s knowledge of heterophase polymerization techniques can be found in References 31–36. Whereas there were only a handful of polymeric products based on heterophase polymerizations during the 1930s in Germany (37), by 1951 there were nearly 200 synthetic latexes made of various polymers commercially available in the USA (38), and today products based on heterophase polymerization techniques determine in almost all fields our living standard, including construction and automobile industry, medical diagnostics, and paper making (39). Economic Importance of Heterophase Polymerizations. Polymer dispersions are an important part of the worldwide polymer business, not only because of their outstanding role in the historic development of polymer chemistry but mainly because of their versatility with respect to both properties and applications. Figure 1 illustrates both the position of polymer dispersions in the hierarchy of science and shows the variety of applications in which they are found. These examples illustrate on the one hand the different branches of industry, that take advantage of polymer dispersions and on the other hand how the daily life is influenced. Heterophase polymerizations are important technologies that can

414

HETEROPHASE POLYMERIZATION

Vol. 6

Paints Adhesives glues Materials science

Rubber products

Textile finishing

Fresh water Pharmacy

Reaction engineering

Biology

Inks toners

Paper wrappings Polymer chemistry

Polymer dispersions

Colloid science

Drug delivery

Carpet backings Medicine

Waste water treatment

Physical chemistry

Physics Additives modifiers

Medical diagnosis

Electronics Nonwoven fabrics Coatings

Fig. 1. Illustration of the importance and meaning of heterophase polymerizations/polymer dispersions. Polymer dispersions as the centerpiece belong to chemistry and have strong interrelations to disciplines of chemistry mentioned in the next level. Also, other scientific areas, which are listed in the third level, use polymer dispersions for research and development. Finally, the outer level contains important application fields of polymers, which were produced by heterophase polymerizations.

produce on a volume scale from only a few milliliters up to 200 m3 high quality polymers with specifically tailored properties. Polymer dispersion is a description for a special state of matter and not connected to a special chemical composition of the polymeric material. Any kind of polymer can be transferred into the state of polymer dispersion. For instance, water-soluble polymers are prepared by inverse heterophase polymerization techniques in organic media. Poly(acrylic acid) or poly(methacrylic acid) polymers and copolymers, quaternary ammonium polymers and copolymers, and polyacrylamide polymers and copolymers belong to this class of important synthetic polymers. The main application areas of these polymers are water and wastewater treatment, detergents, papermaking, coatings, adhesives, textiles, and super-absorbent polymers. The generic term polymer dispersion can be illustrated by means of a “family tree,” as depicted in Figure 2. Primary dispersions are the direct result of heterophase polymerizations and can be subdivided into naturally occurring latexes and synthetic latexes. Secondary dispersions comprise examples where polymers are transferred into polymer dispersions only after their preparation. Artificial latexes are obtained after dispersing solid polymer or emulsifying polymer solution in a proper dispersion medium. The emulsion can be transferred in a second step into a suspension by removing

Vol. 6

HETEROPHASE POLYMERIZATION

415

Polymer dispersions

Primary dispersions

Secondary dispersions

Natural latex Synthetic latex

Artificial latex Block copolymer dispersions

Tim

e

Polymeric colloidal complexes

Fig. 2. Family tree of polymer dispersions. The time axis in this graph denotes the time when the different kinds of polymer dispersions came to awareness of mankind starting with natural latex and ending with recent developments in polymeric colloidal complexes. Some important historic milestones are the following: the biosynthesis of natural rubber in plants takes place on earth since millions of years, the first heterophase polymerization was mentioned in a patent filed in 1912 (21), the first process description to make an artificial latex was published in 1923 (40), and the first block copolymer dispersion made by heterophase polymerization was described in 1952 (41).

the solvent. Artificial latexes have until now never gained much industrial interest although several methods have been developed. There are basically three driving forces for the development of artificial latexes: The first is economic because it is more advantageous to transport polymers as dry solid material instead of as dispersions with about 50% by weight dispersion medium. The second reason has to do with the superior material properties of polymers that cannot be prepared by heterophase polymerization techniques—such as polyurethanes or chlorosulfonated polyolefins—but whose application as polymer dispersion is advantageous or desirable. Combinations of different materials such as mineral substances and polymers in the form of composite polymer dispersions for various potential applications are the third reason for an ongoing research on artificial latexes. Block copolymers in selective solvents for only one block form micelles, which are in fact sterically stabilized polymer particles and thus belong to the class of polymer dispersions (42–44). Moreover, heterophase polymerization techniques can be used to prepare bock copolymers in a very efficient way (45–48). One of the oldest attempts, if not the oldest, to prepare block copolymers by radical polymerization was undertaken via aqueous heterophase polymerization (41). Polymeric colloidal complexes are the result of manipulating polymer solutions

416

HETEROPHASE POLYMERIZATION

Vol. 6

by inducing interactions that lead to dispersions. These interactions can be either hydrophobic or electrostatic in nature. Examples are the formation of hollow spheres via layer-by-layer assembly of oppositely charged polyelectrolytes onto various templates (49–52). But also a single kind of block copolymer can form polymeric colloidal complexes via self-assembly, such as poly(ethylene glycol)-bpoly(N-isopropyl acrylamide) copolymers, where the poly(N-isopropyl acrylamide) block becomes insoluble at temperatures above 32◦ C and particles are formed via hydrophobic interactions. These particles consist of a poly(N-isopropyl acrylamide) core and a stabilizing poly(ethylene glycol) shell. Such particles can easily be prepared by heterophase polymerization started with poly(ethylene glycol)– ceric ion redox system (53). Other examples of such dispersions are complexes between amphiphilic block copolymers with one polyelectrolyte block and either oppositely charged surfactants (48,54–56) or other amphiphilic block copolymers with oppositely charged polyelectrolyte block (48,57). Besides these recent developments in research on preparation of polymer dispersions, they also play an important role (1) as model systems for soft matter physics, especially in understanding interactions in colloids (58–60), (2) as materials to trigger new characterization methods (61), (3) as objects to observe directly nucleation and growth of crystals (62), (4) as components in advanced material science research (63), and (5) also as object of single particles’ handling and modification (64). The economic importance of heterophase polymerization products is not easy to quantify. There is on the one hand the direct value of the polymer products and on the other hand the value added to products, which are made involving polymer dispersions or which are even made only on the base of very specialized polymer dispersions. Table 3 presents the breakdown of worldwide annually produced amounts of different kinds of polymers/polymer dispersions by aqueous heterophase polymerization techniques. According to these numbers about 25% of the total amount of synthetic polymers, that is at the moment some 2 × 108 t per annum worldwide (39), is prepared by heterophase polymerizations. Except some very minor amount of copolymers for paints, poly(vinyl chloride) is exclusively produced by heterophase polymerization techniques. About 80% is produced by suspension polymerization, followed with about 15% by emulsion and microsuspension polymerization, and the rest is by mass or bulk polymerization (65,66,70). Contrary to polystyrene, poly(vinyl chloride) is insoluble in its own monomer. But vinyl chloride still swells its polymer to about 35% at saturation [the Flory–Huggins interaction parameter is between 0.9 and 1.0 (71)] and hence, the vinyl chloride bulk polymerization is a precipitation polymerization where a swollen polymeric phase is formed after a

Table 3. Worldwide Aqueous Heterophase Polymerization Products (Solids) Product Poly(vinyl chloride) Synthetic rubber Other latexes Natural rubber Polystyrene

Amount, 106 t

Reference

25 10 8 7 2

65,66 67 68 66 69

Vol. 6

HETEROPHASE POLYMERIZATION

417

few percents of conversion. The 2 × 108 t of polymers produced worldwide represent a financial value of more than euro 200 billion (39) and consequently, as a rough estimate, heterophase polymerization techniques contribute at least with one quarter. This is very likely a lower limit because the about 8 × 106 t of emulsion polymers represent a value of more than euro 20 billion (72). Inverse heterophase polymerization techniques are employed on an industrial scale to produce about 1.5 × 106 t of various synthetic water-soluble polymers. The global consumption of such polymers in the year 2000 was about 2.2 × 106 t representing a value of about euro 3.5 billion, where roughly two-thirds is estimated to be realized by polymers produced via inverse heterophase polymerization techniques (73).

Peculiarities of Heterophase Polymerization The peculiarities of heterophase polymerizations are directly connected with the dispersed state. Figure 3 shows schematically the main difference between 1 g of polystyrene as bulk material and as a dispersion. The bulk material is a single particle with a diameter (D) of 1.22 cm and a total surface area (AP ) of 4.676 cm2 , whereas the same amount of material in the dispersed state with an average particle size of 50 nm is subdivided or compartmentalized into 1.455 × 1016 particles (N), with a total surface of 1.143 × 106 cm2 . This increase in both N and AP causes both the advantages and problems of heterophase polymerizations and polymer dispersions compared to homogeneous counterparts. The prices mentioned in Figure 3 for that 1 g of polystyrene in different states show interestingly that the increase in the total surface area almost corresponds to the increase in the

Macroscopic

Dispersed

D = 1.22 cm

D = 50 × 10−7 cm

N=1

N = 1.455 × 1016

A P = 4.767 cm2

A P = 1.143 × 106 cm2

Price: ~

1 × 10−3 g−1

Price: ~

3.1 × 102 g−1

Fig. 3. Illustration of the physical effect of dispersion (not to scale); 1 g of polystyrene with a volume V p of 0.952 cm3 .

HETEROPHASE POLYMERIZATION

Vol. 6

1011

1028 1027 1026 1025 1024 1023 1022 1021 1020 1019 1018 1017 1016 1015 1014 1013 1012 1011

1010

A spec, m2/md3

109 108 107 106 105 104 10−1

100

101

102

103

104

105

N , md−3

418

106

D, nm

Fig. 4. Specific surface area and particle number for the model dispersion in dependence on the average particle size.

price. The price for the dispersion is that of carboxylated polystyrene spheres with a diameter of 50 nm, which are sold as size-calibration standards in 10-ml batches with a solids content of 2.5% (74). Some of the most important consequences of compartmentalization are illustrated by means of a model dispersion, which is composed of equal volumes of water and organic phase in dependence on average particle size (D in nm). These features are the specific surface area (Aspec ) and the number of particles (N), both calculated per dispersion volume in md 3 (cf Fig. 4), the average distance between dispersed particles (dpp ) and the amount of surfactant relative to the mass of total organic phase (W surf,t ) (cf Fig. 5), the number 104

106 105

103 102

103 102

101

Wsurf,t, %

dpp, nm

104

101 100

100 10−1 10−1

100

101

102

103

104

105

10−1 106

D, nm

Fig. 5. Average distances between the particles and amount of surfactant to cover the whole interface with a saturated layer in dependence on average particle size for the model dispersion.

Vol. 6

HETEROPHASE POLYMERIZATION

419

1000 100 100

10 1 0.1

φSE, %

n S, %

10

1

0.01 0.001 10−1

100

101

102

103

104

105

0.1 106

D, nm

Fig. 6. Fraction of interface molecules per particle and fraction of interfacial free energy in dependence on average particle size for the model dispersion.

of molecules in the surface layer of the particle relative to the total number of molecules per particle (nS ) and the fraction of free surface energy relative to the total free energy of the droplet (φ SE ) (cf Fig. 6). 6 9 10 D

Aspec = N=

ϕdp 6 1027 π D−3 √

−3

dpp = 102 MS Wsurf,t = ϕdp ρdp  ns = 102

(2)

N

(3)




Aspec cmc + NA as



D − 2Ddp 1− D3

φSE =

(1)

σ Aspec σ Aspec + Pd v

 (4)

3  (5)

(6)

Figures 4–6 were generated by means of the following assumptions. The volume fraction ϕ dp of the dispersed phase is 0.5, its density ρ dp is 0.9 g/cm3 , the surfactant is sodium dodecylsulfate with a molecular weight M S = 288.4 g/mol, a surface coberage as = 0.50 nm2 per molecule, and a critical micelle concentration cmc = 10 − 2 M, the volume of a single particle is v and the pressure inside is Pd = 105 N/m2 , the interfacial tension between dispersed phase and continuous

420

HETEROPHASE POLYMERIZATION

Vol. 6

fluid phase is σ = 10 − 2 N/m, and the size of a molecule of the dispersed phase is Ddp = 0.7 nm. Note that the terms particle and droplet are used interchangeably, to describe the dispersed phase without taking into account at the moment a particular state of matter of this phase. σ Aspec corresponds to the amount of work required to expand the interfacial area or surface area. The data depicted in Figures 3–6 clearly reveal the colloidal nature of heterophase polymerization systems. In particular, model dispersions with an average particle size of 50 nm and 50 µm consist of 7.643 × 1021 md − 3 and 7.643 × 1012 md − 3 particles, have specific surfaces of 1.2 × 108 m2 /md − 3 and 1.2 × 105 m2 /md − 3 , an average distance between particle surfaces of 50.8 and 50,800 nm, an amount of 8.167% and 8.4 × 10 − 3 % of surface molecules, need 26.15% and 0.665% of sodium dodecylsulfate to cover the interface completely, and the surface free energy contributes with 92.3 and 1.18% to the overall free energy, respectively. Just for comparison, a soccer field is about 104 m2 large and a ball for that game has a fraction of 8.75 × 10 − 7 % surface molecules, provided it consists of bulk rubber. The high surface free energy and hence the higher total free energy of a dispersion in comparison with a homogeneous system is the reason that permanently acting forces reduce the total free energy, ie they lower the interfacial area by coalescence or coagulation processes. Thus, working with heterophase polymerization techniques also means facing the problem of colloidal stability. From a technical point of view there are three main advantages compared with homogeneous systems because of the compartmentalization. First, the viscosity of polymer dispersions is independent of the degree of polymerization and can be kept low up to high solids content. Second, the heat of polymerization caused by the enthalpy of the chain growth reaction, which for common monomers is between −50 and −100 kJ/mol is removed easily in tiny portions coming from the individual particles through the continuous phase. Third, the particle size in the micrometer and submicrometer range allows morphological variations on length scales, which are accessible with difficulty by means of other procedures. Another important relation of these rather general considerations is with respect to the initiator concentration (CI ), which for radical polymerizations is typically in the order of 10 − 3 M. Interesting quantities are the (formal) number of initiator molecules per particle (nI ) and the number of radicals in the steady state (nR,equ ) per particle calculated according to equation 7 assuming a termination rate constant kt of 108 L/(mol · s) and an initiator decomposition rate constant 2fkd of 8.5 × 10 − 5 s − 1 . 1 nR,equ = N




2 f kd CI kt

1/2 (7)

Curves for nI and nR,equ are plotted in Figure 7 for two initiator concentrations CI = 10 − 3 M (curves a1 , b1 ) and CI = 1 M (curves a2 , b2 ). The horizontal gray line meets the curves at values for nI and nR,equ equal to 1. The corresponding particle or droplet sizes represent limiting values above which enough initiator molecules are present and radicals are generated to have one initiator molecule per particle and to start polymerization in each compartment, respectively. Note that for both initiator concentrations, even for the unrealistically high value of

1013 1012 1011 1010 1089 10 1067 10 105 104 1032 10 101 100 10−1 10−2 10−3 10−4 10−5 −6 10−7 10 −8 10−9 10 −1 10

a2 a1 b2 b1 100

101

102 103 D, nm

104

105

10

1013 1012 1011 1010 109 108 107 1065 104 10 1032 10 101 100 10−1 10−2 10−3 10−4 10−5 10−6 10−7 10−8 10−9 6

421

n R,equ

HETEROPHASE POLYMERIZATION

nI

Vol. 6

Fig. 7. Average number of initiator molecules per particle (a1 , a2 , left axis) and steadystate radical concentration per particle (b1 , b2 , right axis) in dependence on the average particle size for the model dispersion.

CI = 1 M, the necessary particle sizes to contain at least one growing radical per particle are above 100 nm. Even if this is a formal and crude appraisal, it leads to interesting consequences with respect to polymerizations in preformed monomer drops. The various heterophase polymerization techniques mentioned in Table 2 can be distinguished with respect to the average particle sizes at the end of the polymerization. Figure 8 compares these values with characteristic length scales of other colloidal systems and the wavelength of electromagnetic radiations, which are frequently used to investigate polymer dispersions. This diagram clarifies that heterophase polymerization techniques cover the whole size range of colloidal systems. The larger the size of the final latex, the broader is the size range the particles pass through during their genesis and consequently, the larger is the changes in colloidal properties and effects influencing particle morphology. For instance, microemulsion polymerization, where the final particle size is clearly below 100 nm and in many cases even below 50 nm, the particles change their size only few tens of nanometers during the process. In contrast, during precipitation polymerization the average particle size changes over several orders of magnitude, say from less than 50 nm to more than 500 µm in some cases. Second, joint consideration of Figures 5 and 8 reveals that the amount of stabilizer necessary for a particular type of heterophase polymerization strongly depends on the average particle size. This is important with respect to end use properties of the polymeric product because stabilizers are auxiliary materials that may influence the desired polymer properties badly. Microemulsion polymerizations require amounts of stabilizers in the order of the monomer mass, whereas suspension polymerization at the other end of the size scale manage with stabilizer concentrations less than 0.1% relative to monomer mass. The interrelation between polymer and colloid science in the case of heterophase polymerizations is nicely expressed by the following consideration.

422

HETEROPHASE POLYMERIZATION Emulsion polymerization

Vol. 6

Suspension polymerization Other techniques

Microsuspension Miniemulsion Microemulsion

Viruses Fume

Dust Spray

Mist

100

UV

101

102

Rain

Clouds & fog

Smog

X-rays

Human hair

Bacteria

Visible

103

Near-infrared

Far-infrared

104

105

106

D, nm

Fig. 8. Characterization of colloidal systems by characteristic length scales.

Heterophase polymerizations may be considered as compartmentalized homogeneous polymerizations, with a lot of interactions between the compartments (particles, drops) on the one hand and between compartments and the continuous phase on the other. In almost all cases, a polymer particle consists of more than one macromolecule. For instance, if the polystyrene of the example given in Figure 3 has a molecular weight of 6.022 × 105 g/mol (average degree of polymerization Pn of about 6000) the single bulk particle consists of 1018 macromolecules, whereas each of the colloidal particles is composed of 69 polymer molecules. Each polymer chain can be regarded as a permanent record of all the reactions by which it was produced. For a free-radical polymerization mechanism, which is by far the most important one for heterophase polymerizations, the duration of these reactions (τ pol ) can be estimated by means of equation 8 to be in the order of seconds or even less for the above polystyrene example, depending on the polymerization temperature (monomer concentration CM ≈ 6 M, propagation rate constant kp = 237, 341, 480 L/(mol · s) for 50, 60, 70◦ C, respectively (75). τpol =

Pn 2 = kp CM (2 f kd CI kt,d )1/2

(8)

In contrast to single polymer molecules, the particles grow during the entire reaction and hence, they can be regarded as a permanent record of the overall polymerization reaction. The duration of heterophase polymerizations is typically in the order of hours. In this sense, recording molecular weight and particle size data should allow an almost perfect reaction control on different timescales.

Vol. 6

HETEROPHASE POLYMERIZATION

423

Another peculiarity of heterophase polymerization is that almost all kinds of hydrophilic and hydrophobic monomers, stabilizers, initiators, chain-transfer agents, and other auxiliary materials can be applied together in one polymerization system. Of course, this can lead to a pretty muddled scenario where a variety of interactions and reactions between both phases have to be taken into account. Furthermore, a clear demarcation between the different kinds of heterophase polymerizations is often not possible. The particle size is not suited to classify heterophase polymerizations unambiguously, because between all techniques regions exist that more or less overlap, and the particle size at the end of the polymerization depends on particular conditions. Also the names of heterophase polymerizations appear somewhat illogical rather than rational. For example, according to the general definitions in Table 1, the names emulsion and suspension polymerization describe the situations at the beginning and at the end of the process, respectively, whereas the name dispersion polymerization is based on the generic term in colloid science. However, this situation points out the importance of whether a liquid (emulsion) or a solid (dispersion) forms the dispersed phase, which may change in the course of the polymerization. The main difference between a liquid and a solid interface with respect to the surface tension is given. In either case, the surface tension is a region where the stress differs from that in the bulk of the droplets. In case of a liquid, the surface tension is the same at all points and for all directions in the surface. It is a force that stretches the interface isotropically in all directions. But the surface tension of a solid surface need not be isotropic and hence, its value can be different for different directions in the surface. The impacts of this fundamental difference between an emulsion and a suspension on the mechanism of heterophase polymerizations are poorly understood. Despite all the problems mentioned, an attempt at demarcation is made in Table 4 by means of the absolutely necessary prerequisites to carry out a particular type of heterophase polymerization. The minimum requirements for carrying out emulsion polymerization are a certain concentration of monomer dissolved in the continuous phase and an initiator. The concentration of the monomer must be high enough to ensure that both the chain length and the concentration of oligomers or polymers formed during the initial period in the continuous phase are high enough to form polymer particles. The same minimum requirements are Table 4. Absolutely Necessary Prerequisites for Various Kinds of Heterophase Polymerizations in Liquid Dispersion Media Polymerization Bulk Precipitation Dispersion Suspension Microsuspension Emulsion Miniemulsion Microemulsion

Stabilizer

Free or dispersed monomer phase

No No No Yes Yes No Yes Yes

Yes No No Yes Yes No Yes Yes

Solubility of initiator Monomer phase Either phase Either phase Monomer phase Monomer phase Either phase Either phase Either phase

424

HETEROPHASE POLYMERIZATION

Vol. 6

necessary for bulk, precipitation, and dispersion polymerizations. Note that in a bulk polymerization the monomer also forms the continuous phase for its own insoluble polymer. With this it is possible to make a clear demarcation between suspension, microsuspension, miniemulsion, and microemulsion polymerizations, where the presence of a free, dispersed monomer phase (monomer droplets) is a prerequisite by definition. Another criterion to distinguish between bulk, suspension, and microsuspension polymerization and the different techniques is the initiator type: whereas the former techniques need an initiator that is soluble in the monomer phase, the initiator type plays practically no role for the other polymerization techniques listed in Table 4.

Kinetics of Heterophase Polymerizations In general, all reactions according to the particular polymerization mechanism can take place in either phase of a two-phase system. For free-radical polymerization this means initiation, propagation, termination, and chain transfer can take place in both the continuous and the dispersed phase. The kinetics inside the dispersed phase is of special interest, as the goal of heterophase polymerization is to take advantage of the compartmentalized reaction system, which is only possible if the major part of the monomer conversion takes place inside the particles. If this condition is fulfilled, the overall polymerization rate (rP ) can be expressed by equation 9. The product kp CM P is the expression for a homogeneous reaction system, where kp , CM , and P are the polymerization rate constant, the monomer concentration, and the radical concentration, respectively. The further equality on the right-hand side describes the overall rate of a compartmentalized polymerization, where CM,p is the monomer concentration inside the compartments and the radical concentration is expressed by means of the average number of radicals per particle (n), the total number of particles (N), and Avogadro’s number (N A ). rP = kP CM P = kp CM,p

nN NA

(9)

All of the four unknowns in equation 9 (kp , CM,p , n, and N) can be determined experimentally. Values for kp are nowadays also determined by pulse polymerization techniques in combination with size-exclusion chromatography under conditions of heterophase polymerizations. Briefly, high radiation pulses generate radicals, where a few of them terminate but most of them will start to propagate. After a distinct time interval a second pulse generates again a lot of small radicals, which terminate almost all of the growing radicals. If termination between two pulses is neglected, the propagation rate constant can now be estimated by kp = DP/(CM
t). DP is the degree of polymerization determined by size-exclusion chromatography,
t is the time between two pulses, and CM is the monomer concentration, which in case of neat droplets is CM = 1/vmon where vmon is the molar volume of the monomer. N is calculated via the relation N=

6FG  π 100 − FG ρ2 D3 

Vol. 6

HETEROPHASE POLYMERIZATION

425

where ρ 2 is the particle density. The calculation of N requires the determination of both solids content and average particle size. In contrast, the determination of CM,p is much more lavish because it requires the determination of the amount of unreacted monomer that can be present either in the form of neat monomer droplets or inside monomer-swollen polymer particles. For a monomer, which is liquid at room temperature, CM,p can be determined in the following way. A certain amount of latex with known solids content is centrifuged just with a speed that is high enough to separate the monomer, which is present in neat droplets. Then, to the latex, which still contains monomer-swollen particles, initiator is added and a polymerization is started. After determination of the solids content of the treated latex, the amount of monomer in the swollen particles can be calculated from the increase in solids content. Finally, n can be calculated from equation 9 and the heterophase polymerization is completely characterized by experimental means. From the viewpoint of the mechanism of heterophase polymerization both n and N have inspired researchers since the early days of heterophase polymerization research. N is strongly connected with the particle nucleation mechanism and n with the distribution of active centers among the phases and hence the overall kinetics.

The Formation of Polymer Particles The common feature among heterophase polymerization techniques with respect to particle formation is the formation of a new phase, ie the polymeric phase. Besides phase formation the terms generation or nucleation of particles can be used synonymously to describe the process, which finally leads to the manifestation of the polymer as new phase. This phase formation may, in some cases where the reaction system is at the beginning a homogenous single phase, coincide with the development of the heterogeneous nature of the polymerizing system as it is, according to Table 4, the case for dispersion, precipitation, and heterogeneous bulk and emulsion polymerizations. In such cases the polymerization reaction starts in the homogeneous phase and the generation of the polymer phase exhibits all characteristics of a first-order phase transformation, ie thermodynamic functions and parameters such as molar volume and free energy show abrupt changes at the phase transition point. In contrast, if the reaction system is at the beginning heterogeneous, the polymerization reaction can start in either phase. If the polymerization happens predominantly inside the monomer droplets the kind of phase transformation depends on the Flory–Huggins interaction parameter (χ m,p ). The criteria for solubility or insolubility within the framework of the Flory– Huggins theory are for high molecular weight polymers over the whole composition range χ m,p ≤ 0.5 and χ m,p ≥ 0.5, respectively. The larger the χ m,p , above 0.5, the more insoluble is the polymer in its own monomer or the less favored is the interaction between monomer and polymer. For example, for the systems styrene/polystyrene, vinyl chloride/poly(vinyl chloride), and acrylonitrile/ polyacrylonitrile, χ m,p values are ∼0.5, > 0.9, and >1, respectively. These values suggest that polystyrene is completely soluble in styrene, and poly(vinyl chloride) is only swollen by its monomer, but polyacrylonitrile is insoluble in acrylonitrile. Hence, the consequences for formation of the polymer phase during polymerization

426

HETEROPHASE POLYMERIZATION

Vol. 6

inside monomer droplets are as follows. For styrene/polystyrene, the transition from monomer droplets to polymer particles proceeds very smoothly via a polymer in monomer solution. The polymer concentration of this solution increases correspondingly to the overall conversion as long as at complete conversion solid polystyrene particles remain. Vinyl chloride is not a solvent for its polymer and hence, phase separation inside the monomer phase takes place at low conversion. But as the polymer is swollen by its monomer, the transition into the glassy (solid) state starts only at a conversion above 70%. The situation is still a little more different for acrylonitrile, which swells its polymer only to a negligible extent, as in that case χ m,p ≥ 1 (76). Consequently, a sharp phase transition (first-order) takes place at very low conversion but in contrast to the other two systems the polymerization proceeds in an outer layer by precipitation of growing polymer chains onto existing polymer particles. Another consequence of the absence of an interaction between monomer and polymer is the formation of nonspherical particles, as the glass-transition temperature of polyacrylonitrile (T g ∼ 100◦ C) is higher than polymerization temperature and hence the polymer inside the particles is almost immobile, and the spherical equilibrium shape, which is obtained if the particle is able to minimize its interfacial free energy, cannot be attained. Note that for a given interfacial tension the interfacial free energy is determined by the interfacial area, and the area per unit volume is the lowest for a sphere compared with any other body (eg the surface area per unit volume is 4.83 for sphere, 5.56 for a right cylinder, 6 for a cube or cuboid, and 6.83 for a cone). In summary, if formation of the polymer phase takes place inside the monomer droplets (a process that is sometimes referred to as droplet nucleation) the phase transformation itself is rather unspectacular. The transition from a solution to a solid is, if it ever will takes place during the polymerization under the particular conditions, finished only at high conversion, when the T g of the polymer/monomer mixture in the particles is above the polymerization temperature. Thus, the more interesting case is the formation of polymer particles in the continuous phase, such as during dispersion and emulsion polymerization. Under these conditions, a clear first-order phase transition is observed because the system changes in a very short period of time from homogeneous to heterogeneous. Nucleation in emulsion polymerization has been a matter of controversial discussion because for more than five decades no direct experimental data on the nucleation process were available. Discussions were mainly centered on the question of the role of emulsifying agents, in particular on the role of micelles as precursors of polymer particles. “Micellar nucleation theory,” as it was expressed by Smith–Ewart (77) on the baseis of ideas developed by William Harkins (78–80), states: When a free radical from the surrounding water phase enters a micelle, it initiates polymerization of the monomer in the micelle. When this polymerization starts, monomer from the surrounding medium diffuses into the polymerizing region until shortly this region is no longer identified as a micelle, but is now considered a growing polymer particle. . . . In other words, a polymer particle is formed ad hoc when a radical enters a micelle. That this is an oversimplification is revealed in Harkin’s original papers.

Vol. 6

HETEROPHASE POLYMERIZATION

427

He considered three possible loci for “the initiation of polymer particle nuclei” (79), based on his studies during 1942–1943 in connection with the Rubber Reserve Company (Synthetic Rubber Program of the United States Government) (1) the soap micelles, if present, (2) the aqueous phase, excluding soap micelles and polymer–monomer particles, and (3) the monomer droplets. He considered the monomer droplets as . . . relatively unimportant in initiating new polymer particles, since thus far no evidence has been obtained to show that a monomer droplet can do more than change into a polymer particle of smaller size. With regard to the micelles as locus of polymer particle initiation, Harkins described very interesting results of X-ray scattering investigations (78,80,81). It was found that styrene at saturation concentration increases the diameter of fatty acid micelles by 1.2 nm (styrene is present only as dissolved in water and solubilized in micelles but not as droplets) but upon polymerization initiated with peroxodisulfate the size of the micelles decreased to its initial value. This process could be repeated several times by consecutive swelling and polymerization until the surface of the polymer particles was grown so large that the soap concentration decreased as a result of the adsorption on the particles below the cmc. These results led to the important conclusion that a growing polystyrene chain is incompatible with the alkyl chains in the interior of micelles and hence, polymer chains grow out of micelles. The incompatibility between polystyrene and alkyl chains was repeatedly proved more than 50 years later by attempts to polymerize the monomer inside double layers of dioctadecyldimethyl ammonium bromide vesicles. Transmission electron microscopy pictures reveal that phase separation takes place during the polymerization in a way that the polystyrene molecules gather at a particular place, whereas the monomer was uniformly distributed over the whole bilayer (82). A remaining question of importance is at which chain length of the growing polymer molecule the phase separation occurs. If the further fate of escaped radicals is considered, two other mechanisms of particle formation are possible. besides micellar nucleation. The first of these nonmicellar theories of particle formation is known as homogeneous nucleation and its development and quantification owes entirely to Robert Fitch (83–86). This theory, originally developed for more water-soluble monomers such as methyl methacrylate, considers initiation in the aqueous phase, capture of growing radicals by existing particles, and precipitation of a single chain if it becomes insoluble in water as individual steps. The second possibility is called aggregative nucleation, according to which nucleation occurs when a critical supersaturation of growing or dead oligomers in the continuous phase is reached and this solution becomes unstable and separates into a polymer phase and a less concentrated continuous phase (87,88). In contrast to so-called homogeneous nucleation the aggregative nucleation is a multichain process, which is governed by thermodynamic rules known from classical nucleation theory (89) or spinodal decomposition (90). It is interesting to note that Harkins came to a conclusion, which completely agrees with aggregative nucleation mechanism when he concluded (81), that in the absence of monomer droplets, “It is not improbable that several polymer molecules may meet each other and form a small aggregate as the source of a single particle.” Experimental techniques have

428

HETEROPHASE POLYMERIZATION

Vol. 6

70

1017

60

1016

50

1015

40

1014

30

1013

20

1012

10 10−7

10−6

10−5

10−4

10−3 10−2 C S, M

10−1

100

N, cm−3

γ, mN/m

been developed that allow a direct observation of the nucleation step in emulsion polymerization (88,91–93), The results obtained so far clearly support the aggregative nucleation mechanism. Moreover, the results confirm that nucleation in heterophase polymerization can be considered and treated within the framework of general nucleation theories as bubble formation, or condensation, or crystallization processes. According to these theories nucleation requires supersaturation of the nucleating species, which are oligomers born in the continuous phase. In this sense another nucleation criterion for heterophase polymerizations is to bring the system, which is in this situation a solution of oligomers in the continuous phase, into a thermodynamically unstable intermediate state. Role of Emulsifiers. There is no doubt that all kinds of surface-active compounds have a strong influence on particle formation as they influence the interfacial tension of the nuclei, which is an essential parameter. On the other hand, emulsion polymerization can be carried out at emulsifier concentrations below the cmc or even in the absence of emulsifiers. In the latter case one has to take care that polymer end groups arising from primary radicals or water-soluble comonomers contribute to stability. However, as a general rule for all kinds of heterophase polymerizations, for a given recipe the particle number depends strongly on the stabilizer concentration as is shown in Figure 9. Similar curves were also obtained for styrene, with ammonium peroxodisulfate as initiator as well as for different acrylic and methacrylic acid ester monomers (94). The most striking feature of the curve in Figure 9 is the S shape, which suggests regions of different stabilizer efficiency with respect to particle concentration if it would be expressed as d logN/d logCs (cf Fig. 10). Another measure for the efficiency of a surfactant is its surface tension–concentration plot. A comparison of both plots (as is also shown

1011 101

Fig. 9. Particle concentration in dependence on SDS concentration for batch emulsion polymerization of styrene (filled circles) and interfacial tension concentration plot for SDS (black line, further details see text). (Recipe: 40 g of water, 10 g of styrene, 0.32 g of potassium peroxodisulfate, variable amounts of SDS, temperature 80◦ C, particle concentration calculated from solids content and particle size at final conversion, particle size: intensity average from dynamic light scattering.)

Vol. 6

HETEROPHASE POLYMERIZATION

429

C S, M 10−5

10−4

10−3

10−2

1017

10−1

100

101 1021

cmc

1020

N, cm−3

1015 1014

1019

1013

dN/dC S, mol−1

1016

1012 1018 100

1011 10 γ, mN/m

Fig. 10. Plots of the final particle number of styrene emulsion polymerization versus interfacial tension of a SDS solution with the same concentration (filled circles), final particle number versus SDS concentration (filled squares), and the efficiency of the surfactant with regard to the final particle number in a styrene emulsion polymerization (gray line) (recipe: cf. caption of Fig. 9).

in Figure 9) reveals that the regions with the steepest change in both curves coincide. The surface tension–concentration plot was measured with a bubble pressure tensiometer using the bubble pressure difference method as described in Reference 95. This method is especially designed not to measure exact surface tension values but rather changes under various conditions even inside stirred reactors at elevated temperatures. To draw conclusions regarding the role of emulsifiers in heterophase polymerizations the surface tension measurements should be conducted under conditions that correspond as closely as possible to those during the polymerizations. This is of special importance because it is known that the cmc of sodium dodecylsulfate increases with temperature but decreases with increasing ionic strength (96). The surface tension–concentration plot depicted in Figure 9 was obtained at polymerization temperature (80◦ C) in presence of potassium sulfate at a concentration corresponding to the potassium peroxodisulfate concentration employed for polymerization and the water was saturated with toluene to mimic the influence of styrene. Experimental conditions have been controlled so that the curves in Figure 9 can be directly compared. The steep decrease in γ and the steep increase in N originate from the effect of increasing surfactant adsorption at a given surface area. In the case of γ measurement this is the nitrogen bubble interface to water and during emulsion polymerization it is the particle water interface. The strong dependence between γ and N at surfactant concentrations below the cmc shows in the logN versus logγ plot in Figure 10. Furthermore, this plot proves that the S shape of the logN–logCS curve is a consequence of surfactant adsorption. If this is true, then the good correlation between N and γ suggests that the interface area of the nuclei formed during the particle nucleation period does not depend on the surfactant concentration but is rather

430

HETEROPHASE POLYMERIZATION

Vol. 6

constant for a given rate of initiation. Consequently, the surfactant concentration and the adsorption behavior of the surfactant determines the particle surface area, which can be stabilized and hence the number of final latex particles. This argument leads to the conclusion that it could happen that the number or the total surface area of nuclei formed at the end of the particle nucleation period is at low surfactant concentration larger than the amount that can be stabilized by the particular surfactant concentration. Consequently, the number of particles decreases by coagulation or coalescence processes until the particle area matches the area of surfactant necessary for stabilization, which is as CS,C , where as is the area covered by 1 mol of surfactant and CS,C is the critical surfactant concentration necessary for stabilization of all nuclei formed during the nucleation period. Increasing the surfactant concentrations leads to a point where the surface area of the nuclei matches as CS,C and the number of particles remains constant after nucleation. The higher the surfactant concentration, the longer is the duration of the nucleation period. On the other hand, increasing surfactant concentration means that the period of time between the start of the polymerization and the appearance of the first particles (ie the duration of the prenucleation period) decreases, as it was concluded from model calculations (88). All these effects have also been observed experimentally (88,91,97). The data summarized in Figures 9 and 10 make clear that the cmc is no special point with regard to the final particle concentration as it falls in a region where the efficiency (d logN/d logCs ) decreases in dependence on the sodium dodecylsulfate concentration. The decreasing efficiency with increasing surfactant concentration can be explained with the increasing duration of the nucleation period. The longer the nucleation period, the larger the earlier-formed particles can grow and the more they compete with the newly formed ones for the surfactant. This interpretation of the role of emulsifier in emulsion polymerization agrees with experimental results, which were obtained in a series of very well designed experiments by Dunn and Al-Shahib (98–100). They used a series of sodium alkyl sulfate surfactants from C8 to C18 at equal concentrations of 60 mM in styrene emulsion polymerization and were able to show that the number of micelles initially present does not govern the final particle number. A reexamination of these results as it was done in Reference 101 revealed that the number of particles correlates with the cmc (the lower the cmc, the higher is the number) and the adsorption energies of the surfactants (the higher the adsorption energy, the higher is the number). These results clearly support the idea that the stabilizing power or the adsorption strength of the surfactants, which are the stronger the larger the hydrophobic chains, determines the final particle number. This is the same effect as expressed in Figure 10 by the correlation between the surface tension and the particle number for a particular surfactant. For heterophase polymerizations, which start in the continuous phase, such as emulsion polymerization (in the absence of monomer droplets) or dispersion polymerization (cf Table 4), a question of importance might be at which degree of conversion or, more general, at which solids content the first particles appear. This question can be answered by sensitive in-line or on-line characterization techniques such as conductivity and turbidity measurements or dynamic light scattering as described for emulsion polymerization in References 91 and 93 or dispersion polymerization in Reference 102. These results show that nucleation takes place after a distinct period of time where polymerization proceeds under

Vol. 6

HETEROPHASE POLYMERIZATION

431

homogeneous conditions (the prenucleation period) at very low solids content. For instance, the duration of the prenucleation period for an emulsifier-free styrene emulsion polymerization started with potassium peroxodisulfate (2.5 mM) at 60◦ C was found to be 431 s. Within 1 s a huge number of particles is formed (1.76 × 1013 cm − 3 ), with an average diameter of 13 nm (92). From these numbers it can be concluded that each particle consists of about 7000 styrene units at maximum, and the minimum amount of polystyrene formed up to this moment is 2.13 × 10 − 5 g/cm3 . For dispersion polymerization of methyl methacrylate in methanol (5 wt%) at 55◦ C with poly(vinylpyrrolidone) as stabilizer and 2,2 azobisisobutyronitrile as initiator, the first particles were detected with dynamic light scattering 3–4 min after starting the reaction (102). Both sets of experiments reveal another common feature, which is typical for phase transformation or nucleation processes (103). The experimental data are afflicted with large scatter and to get trustworthy results requires enormous experimental efforts and a large number of repeats. The reason for this is of general nature, ie, any firstorder phase transition is characterized by the necessity to surmount an energy barrier via fluctuations. A free energy of activation for such processes exists, which strongly depends on all experimental conditions such as temperature, pressure, and composition. Even minor changes in the experimental conditions can cause huge effects, because the rate of the phase transition depends exponentially on the activation free energy. This strong influence on heterophase polymerizations means, for example, that the kind of reactor material or even the replacement of a stirrer can change reaction rate and product properties. In a comprehensive study it was shown that emulsion polymerization of methyl methacrylate is strongly influenced by the kind of reactor material (stainless steel, Teflon, glass) especially at low emulsifier concentrations (104). In many technical emulsion polymerizations the problems connected with particle nucleation are avoided by using so-called seed latexes. In this case polymerization is started heterogeneously in presence of preformed polymer particles (seed particles). Polymerization recipe and procedure are adjusted (sometimes) carefully in such a way that a high radical absorption rate (cf eq. 12) is realized to capture all radicals from the continuous (aqueous) phase by seed particles in order to avoid both supersaturation of oligomers in the continuous phase and subsequently nucleation of new particles (105). Although the particular mechanism may differ, the common feature of particle nucleation in heterophase polymerizations is the formation of the polymeric phase. Furthermore, the final number of polymeric particles depends on the stabilization, which is employed in the particular heterophase polymerization process. Again, the larger the average particle size at a given solids content, the lower is the interfacial area, which has to be stabilized and hence, the lower the amount of stabilizing groups needed. In aqueous emulsion polymerization the micellar nucleation mechanism is very popular, although clear experimental evidence has not been found at any time. There might be mainly two reasons for this popularity. First, it is the simplicity of this approach together with the similarity to seed polymerization techniques where polymerization starts inside the seed particles by entry of radicals from the continuous phase. Although seed particles differ from micelles in many respects (dynamic behavior, size, and composition) a clear demarcation is complicated as several common features exist, such as swelling with monomer and ability of radical absorption

432

HETEROPHASE POLYMERIZATION

Vol. 6

from the continuous phase. The situation is even more complicated in the case of micelles made of polymeric surfactant, where the gap to a seed particle is still narrower because dynamic behavior and composition can be almost identical to seed particles (106,107). Second, the lack of any direct experimental data on the nucleation step during aqueous heterophase polymerization may have hindered the acceptance of nucleation models based on general nucleation theories, although for dispersion polymerization in organic media the classical nucleation theory has been considered as a possible mechanism at least since the mid-1970s (108).

Compartmentalized Polymerization Kinetics Average Number of Radicals per Particle. The centerpiece of equation 9 and of all heterophase polymerization kinetics is a formula to calculate the concentration of growing radicals inside each compartment. The situation sketched in Figure 11 can be used to derive a suitable general expression. If the number of radicals per particles is denoted with n the particles can be classified as N n , which means the number of particles containing n radicals. The formation of radicals by initiator decomposition and subsequent initiation of the chain growth is allowed to take place in either phase (ri,w and ri,d is rate of initiation in continuous and dispersed phase, respectively). Furthermore, the number of radicals in the dispersed phase increases by absorption or entry (rabs denotes the rate) and decreases by exit or desorption and termination. In these equations kd,w and kd,d are the rate constants of initiator decomposition in the continuous and in the dispersed phases, respectively, and f is the efficiency factor, kabs,i is the rate constant j,e of primary radical absorption by the droplets, kabs is the absorption rate constant of propagating radicals (Pj,e,w ) in dependence on chain length (j) and kind of end group (e), kt,d is the termination rate constant in the dispersed phase, and kdes is the desorption rate constant. If in steady-state the rate of formation of N n particles

r i,w = 2fk d,wC I,w Initiation

rabs dn = = k abs,i R i,w + dt N dn = r i,d = 2fkd,dC I,d dt

ΣΣ z

j= 0

j,e

k abs P j,e,w

e

Absorption, entry

Initiation (n − 1) dn = −2kt,d n v dt dn = −kdesn dt

Termination

Desorption

Fig. 11. Illustration of most important reactions in heterophase polymerization systems.

Vol. 6

HETEROPHASE POLYMERIZATION

433

equals the rate of their disappearance the recursion formula given by equation 10 is obtained.   (n + 2)(n + 1)Nn + 2 + δ(n + 1)Nn + 1 + α Nn − 1 + βd Nn − 2 = Nn n(n − 1) + δn + α + βd Nn (10)

A balance of the radicals in the continuous phase leads to equation 11, where j,e kt,w is the termination rate constant of propagating radicals (Pj,e,w ) in dependence on chain length (j) and kind of end group (e), termination rate constant of primary radicals and Ri,w is the concentration of primary radicals. rabs = ri,w + kdes

∞

nN − 2

n= 1

z j =0

j,e

2 2 kt,w Pj,e,w − 2kt,w Ri,w

(11)

e

Rearrangement, and defining, the coefficient by combination of rate constants into equations 12, lead to equation 13, which together with equation 10 can be used to calculate n. The average number of radicals per particles is defined by equation 14.

δ=

rabs vNA2 ri,d vNA2 ri,d vNA2 kdes vNA ; βd = ; α= ; βd = kt,d Nkt,d kt,d N kt,d N

rabs = kabs,i Ri,w +

z j =0

e

2Nkt,d kt,w j,e kabs Pj,e,w ∼ = kabs Crad,w Y = 2 kabs vNA2

(12)

α = βw + δ n − Yα 2

(13)

nNn n= N

with

N=



Nn

(14)

Various researches have contributed over decades to our present knowledge. Wendell Smith and Roswell Ewart started in 1948 (77) by writing down their famous recursion formula for emulsion polymerization, which takes into account absorption, desorption, and termination of radicals, where initiator decomposition is assumed to take place exclusively in the continuous phase. Two months after the Smith–Ewart paper appeared, anarticle—nowadays virtually forgotten—was submitted by Haward dealing with polymerization in a system of discrete particles (109). He took into account generation of either single radicals or two at once inside particles and termination. Either consideration (77,109) led to the identical recursion formula if the formation of single radicals inside the particles is considered to be equivalent to radical entry from the continuous phase. As an analytical solution of the recursion formula is impossible, approximations have to

434

HETEROPHASE POLYMERIZATION

Vol. 6

be considered. Assuming that there is only radical entry or generation of single radicals inside particles, and termination equation 10 leads for n = 0, 1, and 2 to equations 15, 16, 18, α N0 2

(15)

α ( N1 − N0 ) 6

(16)

α α N0 − ( N1 − N0 ) 12 2

(17)

n = 0 ⇒ N2 = n = 1 ⇒ N3 = n = 2 ⇒ N4 =

Now Haward’s argumentation was as follows. The higher the dispersity (ie smaller the v or larger the N), the smaller the α, and for v → 0 or N → ∞, this results in N 2 = N 3 = N 4 = 0. From equations 16 and 17, it follows that N 1 = N 0 and hence that n = 12 , which is an important limiting case for small particles. This result is identical to the famous Smith–Ewart’s case 2, where they argued that n = 12 when the particles are so small that entry of a radical into a particle N 1 results in immediate termination. Thus a particle contains either none or only one radical on average. The argumentation that v → 0 or N → ∞ is obviously more appropriate, as equations 10–14 suggest that n strongly depends on the degree of dispersity. However, the treatment of Smith–Ewart is more general than that of Haward as it allows radical exchange between both phases via radical absorption into the particles from the continuous phase and radical exit out of the particles into the continuous phase. This leads to an analogous form of equation 15 if radical exit, entry, and termination are considered: n = 0 ⇒N1 =

 1 1 α N0 − 2N2 or N2 = (α N0 − δ N1 ) δ 2

Following the arguments of Smith and Ewart (77), two more limiting cases with respect to n are possible. Case 1: if δ  α, it follows that N 1 → 0 and N → N 0 , and hence n 1. Case 3: if radical absorption (or generation inside particles) is much larger than desorption (δ α), it follows that N 2 =⇒ N 0 , which leads, only if the particle size is large enough, to n  12 and hence to an approximation of bulk kinetics. In the next two decades, Stockmayer (110) and O’Toole (111) dealt with exact solutions of the recursion formula by means of modified Bessel functions. In particular, O’Toole presented explicit expressions for N n and showed that the solution obtained by Stockmayer is physically questionable for small but finite rates of radical desorption. Although these contributions represent a crucial step in the direction of a better understanding of heterophase polymerization kinetics, they still leave important questions unanswered, especially regarding the interplay between kinetic events inside and outside the particles. John Ugelstad and his group in two benchmark papers in 1967 and 1976 have treated in particular the role of continuous phase kinetics (112,113). They introduced radical balances in the continuous phase as described by equations 11 and 13. They solved equation 13 taking the procedure described by O’Toole with Bessel functions. Employing continued fractions to describe the ratio of Bessel functions, they finally obtained

Vol. 6

HETEROPHASE POLYMERIZATION

435

log n

Y S/E 3

δ=0

S/E 2

n = 0.5 δ

S/E 1

δ>1

log βw

Fig. 12. Illustration of an Ugelstad plot.

a rather simple expression in form of a series: n=

2α 2α 1 2α + + +··· 2 δ δ+1 δ+2

which converges fast for all α ≥ 0 and δ ≥ 0 (112). n is now computed in dependence on α for given values of δ, and finally equation 13 is solved with these values of n and appropriate values of Y to get corresponding values of β w . The results of this procedure are the famous Ugelstad plots, where log n is plotted versus logβ w , with δ and Y as parameters. Such a plot is schematically shown in Figure 12. The three limiting cases of the Smith–Ewart theory can be clearly seen. The gray arrows indicate the shift of the solution behavior if one of the parameters δ or Y is changed. For instance, on the one hand decreasing δ values (which means decreasing desorption or increasing termination) lead to the appearance of a typical Smith–Ewart case 2 behavior characterized by n = 0.5. On the other hand increasing Y from zero (Y = 0 means, according to equation 11, that no termination takes place in the continuous phase) leads to a spreading of the curves, with δ as the parameter along β w -axis. This means that the linear relation log n versus logβ w , which is obtained for δ > 1, is shifted to larger β w values. Shinzu Omi has pointed out that a recursion formula, which considers only radical formation inside the particles and desorption (only β d and δ are considered in eq. 10) could be used to describe polymerization in large monomer droplets (114). As α (ie the droplet size) increases, n converges to a straight line, which corresponds to the solution of bulk kinetics [n = (β d )1/2 ]. The average number of radicals per compartment is indeed the centerpiece of heterophase polymerization kinetics. For theoretical descriptions of n, detailed knowledge not only about kinetic events in either phase but also regarding exchange processes between particles and dispersion medium are required.

HETEROPHASE POLYMERIZATION

n1−4

436

10 9 8 7 6 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 −7 −8 −9

Vol. 6

n2 n1

n3 n4 10−1

100

101

102

103

104

105

106

D, nm

Fig. 13. Calculation of with various approximations of n the recursion formula equation 10 (cf text).

The Influence of Dispersity. For different types of heterophase polymerization and in context with dispersity as a characteristic feature, it is necessary to emphasize that n strongly depends on particle size. To describe this behavior, several approximations based on equation 10 can be found (113,114). In Figure 13, some n approximations are compared regarding their dependence on the particle size, which is given by the definitions in equations 12. The first approximation n1 = (α/2)1/2 is based on Stockmayer’s solution for the Smith–Ewart case 3 (n  1). This can be directly compared with the approximation for bulk kinetics proposed by Omi (114), n4 = (β d )1/2 , and also with the hypothetical number of radicals per particle in equilibrium (nR,equ ), as depicted in Figure 7. Another approximation derived by Ugelstad and Hansen for Smith–Ewart case 3 kinetics is n3 = (β w /4Y)1/2 , with consideration of water phase termination (Y = 0). Finally an approximation for Smith–Ewart case 1 derived by Ugelstad and Hansen (113) is considered (n2 ), which however is valid for the whole range of n as long as the following relations are fulfilled: δ n  β w Yα 2 and Yα 2  α− δn. Equation 9 can be used to calculate n from experimental polymerization rate data provided CM,p , kp and N are known. To do so is really extensive and tedious work, as one has to determine solids content, particle size, and monomer concentration inside the particles in the course of the polymerization. Figure 14 shows an example of such investigations for a batch styrene emulsion polymerization in dependence on both conversion (X) and average particle size (115). The results of this procedure are influenced by some experimental problems such as the estimation of rP by numerical differentiation of conversion–time curves, the experimental determination of CM,p , and the proper choice of the particle size average value to calculate N. Despite all these problems the results depicted in Figure 14 lead to important conclusions. Batch emulsion polymerization of styrene, even if 2,2 -azobisisobutyronitrile is used as initiator,

Vol. 6

HETEROPHASE POLYMERIZATION

437

D, nm 70

80

90

100

110

120

130

1.4

120

1.2

110

1.0

100

0.8

90

0.6

80

0.4

70

n

130

0.2

D, nm

60 1.6

60 0

10

20

30

40 X, %

50

60

70

80

Fig. 14. Average number of radicals per particles in dependence on both conversion and particle size and particle size in dependence on conversion for a batch styrene emulsion polymerization; gray squares n versus conversion; open squares n versus particle size; gray circles’ D versus conversion. (Recipe: 1000 g of water, 200 g of styrene, 4.268 g of SDS, and 1.74 g of 2,2 -azobisisobutyronitrile as initiator, temperature 60◦ C, for details see Ref. 115.

which is soluble in both phases [about 10% of AIBN is dissolved in water at 25◦ C (116)] obeys Smith–Ewart case 2 nicely as n is between 0.4 and 0.6, with a slight tendency to increase between 5 and 60% of conversion and to increase much steeply thereafter. This increase, even if only slight at conversions below 65%, indicates that there are still other conditions influencing n. One of these possible influences is the change in the average particle size during polymerization, which may cause a change in n as discussed above. The particle size data depicted in Figure 14 (gray circles, bottom x-axis and right y-axis) show an almost linear increase with conversion over the whole range. Indeed, the change in D can explain the slight increase in n at low conversions. The steep increase in n, however, cannot be explained with the increase in D, which is very smooth (cf data in Fig. 13). Another possible explanation of the increase in n is, according to equation 9, a change in CM,p . Figure 15 shows that CM,p also changes over the whole conversion range (115,117,118). CM,p is expressed as the volume fraction of monomer in the swollen particles φ M , which relates to CM,p with the following relation: φ M = CM,p v V mon . This experimental result contradicts an important assumption of emulsion polymerization theory, where constant CM,p values are assumed as long as a free monomer phase is present. For styrene, this should be a conversion of up to about 50%. But the experimental results are trustworthy because they were obtained at different times, in different laboratories, and by different researchers. Consequently, the decrease in CM,p may contribute over the whole conversion range to increasing n values. Waning monomer concentration inside the droplets V/particles leads to an increase in viscosity, which at the same time causes decreasing mobility and radical termination inside the droplets. This gel or Trommsdorff effect, although known

438

HETEROPHASE POLYMERIZATION

Vol. 6

0.9 0.8

φM

0.7 0.6 0.5 0.4 0

10

20

30 X, %

40

50

60

Fig. 15. Change of monomer volume fraction in latex particles during batch styrene emulInitiator: potassium peroxodisulfate (117); sion polymerization with SDS as emulsifier. Initiator: 2,2 -azobisisobutyronitrile (118); Initiator: 2,2 -azobisisobutyronitrile (115).

to be not very pronounced in styrene emulsion polymerization (70), finds its expression under conditions of emulsion polymerization in a steep increase in n. In contrast, a drastic increase in the number of radicals per particle by more than an order of magnitude was directly observed in emulsion polymerizations of methyl methacrylate (MMA) with electron spin resonance spectroscopy (119). A Smith–Ewart case 1 behavior (n 1) can be observed during emulsion polymerization of monomers with a high chain-transfer rate constant such as vinyl chloride because the monomer radicals have a high tendency to escape from the particles. Especially for vinyl chloride emulsion polymerization, Ugelstad and Hansen (113) derived the following equation 18 to calculate n, which predicts dependence, n ∝ D3/2 .  1/2      2/3  ri,w /2kdes n = N − 1/2 k1 D3 + k2 D2 k1 = 6/π ri,w /2kt,w k2 = 6/π (18) It was shown that n values calculated with this equation agree well with estimated ones by means of equation 9 up to a conversion of some 60% in batch emulsion polymerization of vinyl chloride, leading to particle sizes below 80 nm (120). Under these conditions the n values for vinyl chloride, are between 10 − 4 and 10 − 3 . However, for technically important continuous emulsion polymerizations of vinyl chloride, it was found that under steady-state conditions n depends on the square of D, as described by equation 19 (120). 

n = k D2

(19)

The data depicted in Figure 16 show that for a continuous emulsion polymerization of vinyl chloride leading to a very broad particle size distribution (10 nm < D < 2000 nm) n changes over almost 5 orders of magnitude from 10 − 4 up to 10.

Vol. 6

HETEROPHASE POLYMERIZATION

439

101 100

n

10−1 10−2 10−3 10−4 101

102 D, nm

103

Fig. 16. Calculated average number of radicals per particle during a continuous emulsion polymerization of vinyl chloride (120).

This is a peculiar situation because at the same time particles do coexist, showing all features of compartmentalization effect (n < 1) together with particles obeying bulk or solution kinetics (n > 1). In summary of this part, the average number of radicals per compartment is the centerpiece of heterophase polymerization kinetics. Its value depends mainly on the rate with which active, propagating centers appear inside the particles either by decomposition of monomer-soluble initiators or by entry from the continuous phase. Furthermore, particle size and overall concentration of compartments influence n in such a way that it increases with both increasing D and decreasing N. If the viscosity inside the particles is so high that termination by radical recombination is hindered, n increases as an expression of the gel effect in compartmentalized polymerization systems. Competitive Growth of Particles with Different Size. Particle growth kinetics is an important aspect especially if at the end of polymerization a certain particle size or size distribution is required. In this context the relative growth of particles with different diameters is of importance (so-called competitive growth) and consequently, how a given particle size distribution changes during polymerization (if nucleation of new particles is excluded by taking appropriate action). The volume growth of a particle is given by equation 20, where K is assumed to be independent of particle size. dv = K Dx dt

(20)

For the size growth follows equation 21, with volume v = π6 D3 , which leads under the assumption that both K and K  independent of time (ie K and K  only depend on polymerization conditions) after integration to equation 21. Note that the particle size is always the unswollen size; thus using equation 21 means the size dependence of the monomer concentration is neglected, which is reasonable for surfactant-stabilized particles with sizes larger than 50 nm (121). D and D0

440

HETEROPHASE POLYMERIZATION

Vol. 6

are the sizes at the beginning and the end of polymerization with duration
t, respectively. dD  = K Dx − 2 dt

(21) 

D3 − x − D03 − x = K
t

(22)

A closer inspection of equations 21 and 22 reveals that if x < 3, smaller particles will grow faster and the particle size distribution is self-sharpening. On the contrary, if x > 3, the distribution becomes broader as larger particles grow faster than smaller ones. However, if x = 3, particle growth does not depend on particle size. There is experimental evidence that competitive growth during seeded emulsion polymerization slightly favors smaller particles (122). Other experimental and theoretical studies confirmed in the case of water-soluble peroxodisulfate initiator growth exponents of x = 2.5 for particle sizes larger than about 150 nm, which decreased toward zero when the particle size was decreased below 150 nm. However, for oil-soluble initiators, x was equal to 2.5 for a larger size range until the particles reached a critical size needed to sustain two growing radicals (123). This means that for larger particles, ie pseudo-bulk systems, particle growth does not depend on particle size and hence, the particle size distribution evolves uniformly in time. In a series of vinyl chloride emulsion polymerizations an exponent of x between 2 and 3 was determined, which slightly decreases with increasing initiator concentration (peroxodisulfate) as predicted by a theory considering desorption and reabsorption of radicals (124). Model calculations on the kinetics of competitive growth in aqueous polymerization of bidisperse seed systems with oil-soluble initiators showed that x is strongly dependent on rate of initiator decomposition, desorption rate, and the diameter ratio. If only 1% of the initiator is soluble in water, x approaches a value of 3 for low and high decomposition rates, independent of desorption rate. For intermediate values of initiator decomposition rate, x < 3 is obtained, indicating self-sharpening of the particle size distribution. In contrast, with a completely water-insoluble initiator the order is found to be constantly equal to 3, which indicates size-independent growth rate (125). Also, more sophisticated model considerations with regard to particle growth come to similar conclusions; for example, one author stated that as particle size distribution evolves it usually becomes narrower, and a more monodisperse distribution can be obtained provided formation of new particles can be avoided (126). An essential advantage of heterophase polymerization, especially emulsion polymerization and related procedures, is that it allows simultaneously high polymerization rates and high molecular weights owing to its compartmentalized nature. In this context a question arises with regard to the relation between particle size and molecular weight. There are at least two contradictory effects necessary to consider: the dependence of the average number of radicals on the particle size (cf above) and the dependence of the monomer concentration inside the particles (cf below). As both values get larger with increasing particle size and on the one hand the molecular weight increases with increasing monomer concentration but decreases with higher n, a theoretical prediction is very hard. Figure 17 shows for

Vol. 6

HETEROPHASE POLYMERIZATION

441

Molecular weight, g/mol

106

105

104

102

103

D, nm

Fig. 17. Average molecular weights in dependence on average particle sizes for emulsion polymerizations of styrene, where changing emulsifier concentration varied the particle size (conditions cf. caption of Fig. 9); lines are for guiding the eye. Mw ; Mn .

the same polymerization as in Figures 9 and 10 the experimentally determined dependence of the weight- and number-average molecular weights on the average particle size for the final latexes. For this particular set of experiments the average molecular weight increases with decreasing average particle size. The slope in this log–log plot changes for particle sizes below 100 nm. For smaller particles the molecular weight increase is much steeper than for larger particles. The plots in Figure 18 reveal that in dependence on the emulsifier concentration

80

60

M n, g/mol

105

50 40 30

Surface tension, mN/m

70

20 10−7

10−6

10−5

10−4

10−3 S, M

10−2

10−1

100

10 101

Fig. 18. Experimentally obtained relation between number-average molecular weight (M n ), left ordinate, and surface tension, right ordinate, and emulsifier concentration (for polymerization conditions, see caption of Fig. 9); lines are for guiding the eye. surface tension; Mn .

442

HETEROPHASE POLYMERIZATION

Vol. 6

the increase in the molecular weight starts in the vicinity of the critical micelle concentration. At this point a large number of particles can be stabilized and the effect of compartmentalization really starts to act.

Technical Realizations The easiest way to carry out heterophase polymerization is to mix water and a monomer, say styrene, which is able to undergo thermal polymerization, in a vessel at elevated temperatures. After a couple of hours. (the time depends on the temperature) the mixture becomes turbid because of the formation of polystyrene particles. Figure 19 shows a transmission electron microscopy picture of polystyrene particles obtained in this simple way. Although the amount of polystyrene formed was low and the reproducibility of this procedure very bad, a heterophase polymerization took place, and this offers a nice example of how easily heterophase polymerization can be carried out. Recipes. The minimum number of chemically different components for heterophase polymerizations is two. This is the case if a monomer, which is a nonsolvent for its own polymer is polymerized in bulk. Light-induced bulk polymerization of acrylonitrile and vinyl chloride might be examples for this procedure. However, well-defined products especially with regard to colloidal properties, require more efforts. In order to carry out reproducible heterophase polymerization up to high conversions with highly controlled polymeric and colloidal properties, at

Fig. 19. Transmission electron microscopy pictures of polystyrene particles obtained by thermal polymerization at 90◦ C in an all-Teflon (DuPont) reactor after duration of 7 h; the bar indicates 3.5 µm.

Vol. 6

HETEROPHASE POLYMERIZATION

443

least three more components may be necessary, that is, an initiator, a dispersion medium or continuous phase, and a stabilizer. The kind of chain-growth mechanism (condensation or addition) together with the desired polymeric product determines whether a catalyst or an initiator, and which kind, is employed. The thermal properties of the monomers and polymers determine the polymerization temperature and hence the properties of the initiating system. The Continuous Phase. As a general rule any polymerization recipe that works well under homogeneous conditions can be transformed into a heterogeneous system by the choice of a corresponding continuous phase that should be liquid at least at polymerization temperature. The continuous phase has to fulfill three basic requirements: it should be chemically inert, a nonsolvent for the polymer, and possess good heat-transfer properties. If however, the continuous phase interferes with recipe components, the transfer from homogeneous to heterogeneous polymerization conditions requires additional efforts and if needed, changes of one or the other recipe component. Table 5 gives examples of the types of polymerization chemistry applied in heterogeneous polymerization with fluid continuous phases. A variety of different media have been used as continuous phases, for example, aqueous solutions with high ionic strength (130), polymeric melts (47), or supercritical CO2 (132). Other Components. The smaller the particle size, at a given phase ratio, the more difficult it is to ensure colloidal stability (cf Fig. 5). This means that for aqueous heterophase polymerizations in the order suspension < microsuspension < emulsion ≤ miniemulsion < microemulsion, the stabilizer concentration increases. Contrary to the simple polymerization of styrene in water, polymerization recipes for industrially important polymer dispersions comprise up to six monomers, frequently more than two emulsifiers, more than one initiating system, and a few other aids like biocides, defoaming agents, plasticizers for supporting film formation (39). The monomer-to-water ratio is adjusted in such a way that a solid content results typically between 40 and 60% or even higher. The amounts of surfactants and initiator (mainly peroxodisulfate) are typically between 0.5 and 2% (w/w) relative to the monomers and 0.5% (w/w) relative to water, respectively. A peculiarity of suspension polymerization relates somewhat to recipes and more so to reactor operation. Industrial suspension polymerizations are carried out batchwise in reactors with a volume of up to 200 m3 . These reactors are designed with a maximum cooling capacity corresponding to the maximum heat release. However, the maximum heat release normally occurs only for a limited period of time (gel effect) and hence the average heat release during the entire polymerization is much lower. In order to increase the reactor output, so-called initiator cocktails are used that allow polymerization as long as possible at the maximum rate (148,149). Another possibility to avoid a substantial loss in reactor productivity at conversions where the maximum rate of polymerization is not attained is the use of temperature-programmed reaction. An example of such a procedure in order to increase the productivity of vinyl chloride suspension polymerization reactors is given in Reference 150. Reactors and Operation Modes. A large variety of reactors and mixing systems can be used for heterophase polymerizations. The enzymatic biosynthesis of natural rubber in Hevea brasiliensis trees or of poly-3-(hyroxyalkylalkanoates) in bacteria cells takes place in living organisms. For instance, the enzymatic

444

HETEROPHASE POLYMERIZATION

Vol. 6

Table 5. Types of Polymerization Chemistry Applied in Heterophase Polymerization Polymerization chemistry Enzymatic biosynthesis

Example

(1) Synthesis of cis-1,4-polyisoprene by enzymatic condensation of iso-pentenyl pyrophosphate in plants (2) Synthesis of poly-3-(hydroxyalkylalkanoates) in bacteria even an industrial scale Condensation (1) Preparation of polyurethanes via miniemulsion from isophorone diisocyanate and bisphenol A or dodecane diol in aqueous media (2) Tetraethyl silicate reaction in alcoholic solution in presence of water under alkaline conditions Ring opening (1) Polymerization of cyclic esters such as (pseudo-anionic) ε-caprolactone in 1,4-dioxane and heptane with diethyl aluminum ethoxide or sodium trimethyl silanoate as initiator (2) Copolymerization of D,L-lactide and glycolide in supercritical CO2 with stannous octoate as catalyst Ionic polymerization (1) Cationic polymerization of permethyl cyclosiloxanes with tetraalkylammonium chloride surfactants in alkaline aqueous solutions (2) Anionic polymerization of permethyl cyclosiloxanes with dodecylbenzene sulfonic acid as both catalyst and surfactants in acid aqueous solutions (3) Styrene polymerization in hexane with sec-butyl lithium as initiator Noble metal–olefin (1) Olefin complexes of rhodium(I) chloride for complexes aqueous emulsion polymerization of 1,3-butadiene to get crystalline trans-1,4-polybutadiene (2) Binuclear nickel ylide catalysts to get high density polyethylene in aqueous emulsion or miniemulsion (3) Monocyclopentadienyl complex of trimethoxy titanium with a borate activator as metallocene catalyst to polymerize styrene in aqueous emulsions in presence of surfactants Radical, conventional Most prominent way to initiate heterophase polymerization in organic and aqueous continuous phases Radical, controlled Various controlled radical techniques: reversible addition fragmentation chain transfer Atom-transfer radical polymerization nitroxide mediated

Reference 127

128 129

130 131

132

133

134

135 136

137

138

36,139

140–142,144 145–147

Vol. 6

HETEROPHASE POLYMERIZATION

445

condensation of dimethylallylpyrophosphate to cis-poly-1,4-isoprene takes place in tubular vessels in the tissues of the tree, which might be considered as a continuous tube reactor at least during the harvest. There are basically two different types of reactors for synthetic heterophase polymerization: the well-mixed stirred tank reactors or the tube flow reactors. The former may operate batchwise, semibatchwise, or continuously whereas the latter is restricted to continuous operations. Both types of reactors, that is a single continuous stirred-tank reactor and a tube flow reactor, differ significantly in their residence time distribution. This is of particular importance because the particle size distributions, provided particle growth processes mainly govern the size distribution, reflect the residence time distributions. Remember, particles grow during the entire reaction time in contrast to the case of polymer molecules in radical polymerizations. Compared with homogeneous condensation or bulk polymerizations, the technical realization of emulsion polymerizations in both laboratory and on an industrial scale is easy. The minimum equipment needed is a container that can be closed and that is chemically resistant to water and the monomers. In the case of gaseous monomers, the container must additionally withstand the corresponding vapor pressure. This type of reactor, for instance beverage bottles of different styles with crown caps or caps with a septum to allow sampling, was employed for many investigations during the early days of emulsion polymerization (151). These so-called bottle reactors were fixed to a rotating shaft and immersed in a thermostated bath. Nowadays rotation thermostats with metal block heaters are available and frequently in use (152). With a large number of reaction vials it is possible to investigate many recipe variations simultaneously or to acquire conversion time data with sampling volumes large enough for the necessary latex and polymer characterizations. Moreover, for laboratory investigations, stirred vessels with heating and cooling jackets are widely used, although, especially if larger sampling volumes are desired, the bottle reactors still have some advantages. In industrial processes stirred reactors with numerous technical variations are mainly employed. Besides jacket cooling, the heat of polymerization can be removed by various techniques as, for instance, hot cooling or internal fittings inside the reactor. Reactors for commercial processes have a size between 10 and 200 m3 and are pressure-proof up to 200 MPa (2000 atm) if necessary (39). The vessels normally have a cylindrical shape with dished tops and bottoms. The mixing system depends on the particular heterophase polymerization technique, as it has to fulfill different requests. A detailed discussion is beyond the scope of this article and the reader is referred to chemical engineering text books or reviews dealing with special heterophase polymerization techniques such as for emulsion polymerization (153,154) and for bulk and suspension polymerization (155,156). Currently, commercially pure batch processes play a major role for suspension and bulk polymerization but only a minor role for emulsion polymerizations. The most important procedure for effecting polymer dispersions by emulsion polymerization on a technical scale is semibatch or feed processes, which are very flexible regarding product properties. Depending on the required properties with respect to particle size distribution, molecular weight distribution, chemical composition in the case of copolymerization, and particle morphology, numerous feeding policies have been developed. Almost all kinds of consecutive

446

HETEROPHASE POLYMERIZATION

Vol. 6

shell morphologies can be prepared nowadays by means of computer-controlled monomer feed streams. This is a tremendous improvement compared to the early attempts to prepare heterogeneous particle morphology by changing the monomer feed compositions with the aid of so–called near tanks and far tanks, as described in Reference 157. A variety of emulsion polymerizations [for example poly(vinyl chloride) and rubber production] are carried out continuously, mainly in continuous stirred-tank reactors (CSTRs) or in a series consisting of up to 8 CSTRs. In the latter case intermediate feed streams are frequently applied to control copolymer composition and particle morphology. A specific and very exciting feature of continuous processes in stirred tank reactors is the occurrence of damped or sustained oscillation of conversion and also of latex and polymer properties (particle size, molecular weight, and the corresponding distributions) even at the steady state for different monomers (158–163). Although continuous emulsion polymerization technology is more than 60 years old it is still attracting researchers investigating the dynamic behavior in order to achieve control (164) and also developing procedures that allow transfer of products from batch or semibatch processes to continuous reactors in order to increase productivity (165). In large-scale technical reactors, the hydrodynamic conditions are much more important than in laboratory-scale reactors. Sufficient emulsification is essential for a proper control of the polymerization process and product properties. If the stirring conditions are not optimized, problems can arise regarding the homogeneity of the reaction mixture as well as heat removal. In contrast, if the shear forces are too high the latex may become unstable to coagulation [cf below and (39)]. It is generally assumed that agitation has no effect on the kinetics and this topic is ignored in almost all textbooks. However, a few results have been published indicating an influence. For example, kinetics may be affected if the inert gas for purging contains traces of oxygen, if a chain-transfer agent has to diffuse out of the monomer droplets into the particles (166), if the monomer diffusion to the reaction loci is influenced (167), or if the emulsifier distribution between the interfaces is changed considerably (168). Nowadays, the influence of the stirrer speed receives more and more attention (169–172), and increasing the stirrer speed, starting from well-mixed conditions, has been carried out in all these investigations. A drastic change in both the kinetics and the polymer and particle properties has been reported if the stirrer speed is increased starting from values so low that a bulky free monomer phase on top of the reactor still exists (104). Increasing the stirrer speed leads to an increase in the polymerization rate, the particle size, and the molecular weight. Besides stirred tank reactors, in which the composition per volume unit is the same over the entire reactor volume, tubular reactors have aroused increasing interest over the last few years. The references in Table 6 seem to support this trend. Even for suspension polymerization, more and more continuous procedures based on tube reactors are being investigated. For instance, a 5-m3 batch tank reactor can be replaced by a 0.05-m3 continuous loop reactor (175). In summary, many different technical realizations exist for preparing polymer dispersions, ranging from biosynthesis in many plants and bacteria over bottle reactors to high-tech and computer-controlled production systems. Emulsion polymerizations have even been carried out in space under the conditions

Vol. 6

HETEROPHASE POLYMERIZATION

447

Table 6. Developments in Reactor Design for Heterophase Polymerizations Reactor type

Design element

Tubular reactor

Development of a pulsation operation mode to eliminate reactor fouling and plugging; comparison between batch and continuous stirred-tank reactors; review on tubular reactors Styrene emulsion polymerization Experimental investigations and modeling of emulsion copolymerization in continuous loop reactors Suspension polymerization to prepare large porous polymer beads in a tubular reactor; less polydisperse particle size distribution than in batch stirred-tank reactor Emulsion polymerization; 100-m-length tube, with an inner diameter of 1 cm; high cooling capacity; residence time distribution comparable with a cascade of 600–1500 continuous stirred-tank reactors Inverse suspension polymerization of acrylamide in a batch oscillatory baffled reactor 1 m in length, with a diameter of 5 cm); up to 9 baffles oscillate with a frequency between 1 and 5 Hz Continuous emulsion polymerization of styrene in a Couette–Taylor vortex flow reactor; comprehensive kinetic study in dependence on reactor operation parameters Continuous emulsion polymerization of Styrene, influence of flow conditions Continuous seeded copolymerization of styrene and methyl methacrylate

Pulsed tubular reactor Loop process

Tubular reactor

Wicker tube reactor

Tubular reactor

Couette–Taylor

Taylor vortex flow Pulsed-packed column

Reference 173

174 175

176

177

178

179

180 181

of zero gravity (182) under the supervision of the Lehigh Emulsion Polymers Institute. Procedures. From a technological point of view heterophase polymerizations can be carried out either batchwise, semicontinuous (or semibatch), or continuous. In the batchwise case the reactor is filled with all ingredients before the polymerization is started and the reactor content is removed at the end of the polymerization. In a semibatch procedure, at the start of the polymerization the reactor is filled only partially and a stream of either neat monomers or monomer emulsion with constant or deliberately changed composition is fed continuously until the reactor is filled. After a final post-feeding batch reaction period, the reactor is emptied. A continuous procedure means that all necessary ingredients are fed and final latex is removed continuously. In all three cases the polymerization can be carried out in the absence or presence of preformed particles (so-called seed particles). Reactions in the absence of seed particles are frequently called ab initio polymerizations and require that particle nucleation takes place. Table 7 is an

Table 7. Heterophase Polymerization Techniques with Continuous Fluid Phases Common name

448

Initiator

Stabilizer

Precipitation Suspension

Lyophobic, Lyophilic Lyophobic

Dispersion Microsuspension or minisuspension Emulsion

Lyophilic Lyophobic Lyophobic, Lyophilic

None Polymeric or protective colloid Polymeric Polymeric plus surfactant All kinds or none

Miniemulsion

Lyophilic, Lyophobic

All kinds

Microemulsion

Lyophilic, Lyophobic

All kinds

Procedures

Particle size

Reference

Batch Mainly Batch

mm range 10–500 µm

183–186 155,187–191

Batch Batch; high shear

1–20 µm 1–10 µm

192–196) 197–199

Batch; semi-batch; continuous; seed Batch, semibatch; continuous Batch, semibatch

5 nm–10 µm

101,153,200–204

50–500 nm

205–211

10–100 nm

212–222

Vol. 6

HETEROPHASE POLYMERIZATION

449

attempt at a condensed view of the great variety of procedures for carrying out heterophase polymerization. A few typical characteristics of various techniques are put together there. The references given in Table 7, except those for dispersion polymerization, refer mainly to water as continuous phase. As already pointed out, the corresponding processes with nonaqueous continuous phases are denoted with the prefix inverse (223). In inverse heterophase polymerization systems the dispersed phase is hydrophilic and in cases where the monomer is a solid and insoluble in the continuous phase, aqueous solutions of the monomers are used in order to support emulsification. In this sense all procedures except dispersion polymerization have inverse mirror images. The interested reader can find detailed information with regard to inverse precipitation and suspension polymerization in Reference 224, inverse emulsion and suspension polymerization in Reference 225, all types of inverse polymerization in Reference 139, and inverse emulsion and microemulsion polymerization in References 226 and 227. On closer examination of this information, together with that in Tables 2 and 4, one may come to the conclusion that a classification with regard to the final particle size is much more meaningful than with regard to separate recipe components as is frequently done. The final polymer dispersion is the result of a complicated interplay between all recipe components and the polymerization temperature that governs thermodynamics. This is especially true for technical polymerizations owing to the increased number of components. Initiators and Stabilizers. The combinations of initiator and stabilizer play a crucial role. Studies of the influence of the kind of initiator on heterophase polymerization and latex properties have been published. Some efforts have been made to investigate the influence of water-soluble and oil-soluble initiators on emulsion polymerization (213,228–230) and microemulsion polymerization (116,231). In systematic kinetic studies it turned out that the general behavior of emulsion polymerization is the same for water-soluble and oil-soluble initiators (232). The reason for the similar behavior is the desorption of initiator radicals from the polymer particles, which governed the kinetics under the particular conditions, rather than the contribution of the fraction of oil-soluble initiators dissolved in the dispersed phase. Note that this can only be valid for recipes containing surfactants because the polymerization of lyophobic monomers with lyophobic initiators in the absence of stabilizers is a precipitation polymerization without any control of colloidal stability. There are not only similarities but also differences between lyophobic and lyophilic initiators. For instance in Reference 229, the authors found that use of dibenzoyl peroxide leads to lower average molecular weight (about a factor of 5) of the polystyrene than use of potassium peroxodisulfate (emulsifier: hexadecanoic acid) because of chain transfer to the initiator. But for both initiators the authors claimed a rapid exchange of low molecular weight radicals between latex particles and water phase. The type of oil-soluble peroxide, especially its polarity, influenced the results (233). The authors found that higher the polarity, higher was the initiation rate in emulsion polymerization (emulsifier: Nekal, which is of the short-chain alkyl naphthalene sulfonate type) compared to bulk polymerization of styrene. Thus sec-butylhydroperoxide had a 14000 times higher initiation rate in emulsion polymerization whereas that of dibenzoyl peroxide was the same in both cases. The application of a mixture of potassium peroxodisulfate and dibenzoyl peroxide in emulsion polymerization styrene (emulsifier:

450

HETEROPHASE POLYMERIZATION

Vol. 6

alkyl sulfonate, E30) is described in Reference 234. The authors found that the rate of polymerization depends on potassium peroxodisulfate concentration but not on dibenzoyl peroxide concentration. In contrast, the molecular weight depends on dibenzoyl peroxide concentration and not on peroxodisulfate concentration under the particular polymerization conditions. However, these results are contradictory to those of Reference 229 and reveal the necessity to compare experimental data obtained under almost identical conditions. Also, for heterophase polymerizations there is the sensitivity of radical polymerization kinetics to changes in colloidal conditions that must be considered. A systematic study of emulsion polymerization of styrene with different initiator–emulsifier systems (IES) at temperatures between ambient and 100◦ C was carried out (235). This study included sodium dodecylsulfate and sodium C15 -alkylsulfonate (E30, average carbon chain length of 15) as emulsifiers and potassium peroxodisulfate, 2,2 -azobisisobutyronitrile, and various symmetrical poly(ethylene glycol)–azo compounds as initiators (PEGA initiators where the added number denotes the molecular weight of the poly(ethylene glycol) chain). Note, PEGA200 is soluble in both phases owing to the bizarre solubility pattern of poly(ethylene glycol) (45,47 and references therein) and PEGAS200 (the sulfate of PEGA200) is a surface-active initiator mainly located at the interface. With this selection all possible loci of initiator decomposition were covered: the water phase, the oil phase, and the interface. The experiments were evaluated in the form of Arrhenius plots for the average degree of polymerization and the particle number. Both data sets revealed a strong influence of the IES on the experimentally observed dependencies in the case of styrene emulsion polymerization. For instance, the activation energy of the mean degree of polymerization mainly depends on the initiator decomposition and on the locus of radical formation. The activation energy changes in the following order: potassium peroxodisulfate (EA ∼ −35 kJ/mol), 2,2 -azobisisobutyronitrile and PEGA200 (EA ∼ −50 kJ/mol), and PEGAS200 (EA ∼ −80 kJ/mol), independent of the kind of emulsifier and its concentration. The foregoing results for a variety of lyophilic and lyophobic initiators were obtained with low molecular surfactants/emulsifiers as is typical of emulsion or microemulsion recipes. For suspension polymerizations polymeric stabilizers such poly(vinyl alcohol) are typically employed. In a study of suspension polymerization of vinyl chloride [stabilizer: poly(vinyl alcohol) having a degree of hydrolysis of 72.5%] it was shown that the locus of initiator decomposition has an influence on the final particle size distribution. This study is special in the sense that the oil-soluble initiator [bis(4-t-butylcyclohexyl)peroxodicarbonate] was either dispersed in the continuous phase or predissolved in the monomer (236). When the initiator was predissolved in the monomer, the polymerization occurred uniformly in all drops (ideal suspension polymerization behavior), whereas in the other case, additionally to suspension polymerization an emulsion polymerization took place leading to a fraction of smaller-sized particles. The following recipe was used to investigate the influence of the kind of initiator (water- or oil-soluble or both) on the suspension polymerization of styrene at 80◦ C: 40 g of water, 10 g of styrene, 10 mg of poly(vinyl alcohol) (type W25/140 from Wacker Polymer Systems with a degree of hydrolysis of 84–89%), 1.184 × 10 − 3 M initiator (ie 0.32 g of potassium peroxodisulfate or 0.2867 g of dibenzoyl peroxide or 0.673 g PEGA200). The final dispersions at complete conversion have a different appearance, as illustrated by the photograph in Figure 20. From potassium peroxodisulfate to

Vol. 6

HETEROPHASE POLYMERIZATION

C

D

451

I

Fig. 20. Appearance of polystyrene dispersions stabilized with poly(vinyl alcohol) prepared with different initiators. C: initiator soluble in the continuous phase, potassium peroxodisulfate; D: initiator soluble in the disperse phase, dibenzoyl peroxide; I: initiator located mainly at the interface, PEGA200.

dibenzoyl peroxide to PEGA200, the main loci of initiator decomposition is shifted from the aqueous phase via the oil phase to the interface. The dispersion appears coarser and coarser. In case of PEGA complete phase separation occurred, with large aggregates floating around. The difference between dibenzoyl peroxide and potassium peroxodisulfate is less pronounced—but clearly revealed by the transmission electron microscopy pictures shown in Figure 21. The particles obtained in presence of potassium peroxodisulfate look like normal latex particles (spherical solid particles), whereas the particles prepared with the oil-soluble initiator resemble much softer or bleeded structures. However, the scanning electron microscopy pictures of the main fractions of both dispersions (cf Fig. 22) at a glance hardly show any difference. If there is a difference, then the particle size distribution of dispersion (D with dibenzoyl peroxide) appears to be a little broader. Furthermore, in case of initiation in the styrene phase with dibenzoyl peroxide, a few larger particles are also visible (cf Fig. 23), representing either polymerized single-monomer droplets or coalescence structures. These results on the one hand confirm the results of Brooks (236) and on the other reveal that the overall particle size distribution for poly(vinyl alcohol) as polymeric stabilizer depends strongly on the kind of initiator, in contradiction to the above results with low molecular weight emulsifiers. Besides the distinctions in morphology, differences with regard to the molecular properties also exist, as shown by means of molecular weight distributions (cf Fig. 24) and average molecular weights (cf Table 8). Looking at these data, one has to consider that, although the initiator concentration was the same on a molar basis, the decomposition rates, and hence the initiation rates, were different. The decomposition rate constants at 80◦ C for potassium peroxodisulfate,

452

HETEROPHASE POLYMERIZATION

Vol. 6

(a)

(b)

Fig. 21. Transmission electron microscopy pictures of the small particle size fraction of the dispersion shown in Figure 20a. Sample C (potassium peroxodisulfate); b Sample D (dibenzoyl peroxide); the bars indicate 1 µm.

dibenzoyl peroxide, and PEGA 200 are 3.5 × 10 − 4 s − 1 (237), 8.1 × 10 − 4 s − 1 (238), and 6.0 × 10 − 4 s − 1 (47), respectively. A comparison with the data in Table 8 shows, that obviously effects other than the initiator decomposition rate determine the molecular weight (potassium peroxodisulfate should lead to the lowest molecular

Vol. 6

HETEROPHASE POLYMERIZATION

2µm

EHT = 3.00 kV

WD = 4 mm

Signal A = SE2

Date :5 Feb 2003

Signal A = SE2

Date :5 Feb 2003

453

(a)

2µm

EHT = 3.00 kV

WD = 3 mm

(b)

Fig. 22. Scanning electron microscopy pictures of the main fraction of polymer dispersions C and D of Figure 20a. Sample C (potassium peroxodisulfate); b. Sample D (dibenzoyl peroxide); the bars indicate 2 µm.

weight and dibenzoyl peroxide to the highest). One possibility to explain the low molecular weights in case of dibenzoyl peroxide is the chain-transfer reaction as mentioned in Reference 229. This effect might contribute but cannot explain the whole situation. Moreover, there are distinct differences not only in the average molecular weights but also in the shape of the molecular weight distribution. Polymerization I leads to the polymer with the highest molecular weight, followed by sample C and finally sample D. The molecular weight distribution in all three cases is characterized by shoulders or even separated peaks in a range between 2 and 3 × 104 g/mol and about 105 g/mol. Additionally the polymer obtained with PEGA200 has a shoulder at 106 g/mol, which is typical for this kind of initiator (152). The distributions of Figure 24 show that sample D covers the lower molecular weight range whereas sample I covers the higher molecular weight range and interestingly, sample C extends over the whole range. Note that all the molecular weight data are reproducible, as indicated by the average values given in Table 8. In the sense that the molecular weight distribution is a permanent record of the polymerization reaction, an explanation of these data (Fig. 24 and Table 8) might

454

HETEROPHASE POLYMERIZATION

200µm

EHT = 3.00 kV

WD = 4 mm

Vol. 6

Signal A = SE2

Date :5 Feb 2003

Fig. 23. Scanning electron microscopy picture of the large particle size fraction of the dispersion D of Figure 20 (dibenzoyl peroxide). 120 100

UV intensity, au

80 60 40 20 0 102

103

104

105

106

107

Molecular weight, g/mol

Fig. 24. Molecular weight distribution dependence on the locus of initiator decomposition (cf. Fig. 20). — C; - - - D; · · · · I.

be attributed to various reaction loci (continuous phase, dispersed phase, and interface), but a real proof needs further investigation. Consequently, an important point in heterophase polymerization is the control of the particle size or particle size distribution, especially in cases where the dispersed state, besides the polymer properties, is important for application.

Vol. 6

HETEROPHASE POLYMERIZATION

455

Table 8. Weight- and Number-Average Molecular Weights (M w , M n ) of the Polymers Prepared with Different Initiatorsa Sample

M w , g/mol

M n , g/mol

M w /M n

C

2.13 × 10 1.76 × 105 3.33 × 104 3.97 × 104 2.73 × 105 2.78 × 105

3.13 × 10 3.38 × 104 1.29 × 104 1.57 × 104 6.42 × 104 6.84 × 104

6.87 5.21 2.58 2.52 4.25 2.78

5

D I a cf

4

Fig. 20 and text.

Control of Particle Size. In principle, particle size control requires the control of the formation of the dispersed phase either by control of the nucleation process or by restricting the polymerization to a preformed dispersed state. Because the control of nucleation is complicated, the second possibility is the method of choice for most of the industrial applications of heterophase polymerization. A more or less well-defined dispersed state as a starting point for polymerization is realized either by the application of seed particles or by emulsification processes. The former method means that preformed particles are used, which serve as reaction loci. The seed particles can either be swollen with the monomer mixture (seeded batch process) or the monomer mixture can be continuously fed into the reactor (so-called seed and feed procedures, or semibatch or semicontinuous process). The polymerization can be restricted to the seed particles by the application of proper means (33) such as the application of radical scavengers (which are soluble in the continuous phase) in the case of radical polymerizations. The second method requires the preparation of monomer droplets of average size a little larger than the desired final particle size because of the shrinkage taking place during polymerization. The two principal techniques for the preparation of emulsions are comminution and condensation (cf (239) and references therein). Comminution techniques require the input of mechanical energy to crush the bulk monomer phase into smaller drops, whereas condensation techniques are thermodynamically driven phase transitions leading to droplet formation. Four main groups of comminution techniques can be distinguished: rotor/stator systems (including all type of stirrers), high pressure homogenizers, ultrasound devices, and membranes. Except for membrane techniques these methods are based on the input of mechanical energy, in the course of which eddies are formed, transferring the mechanical energy to the droplets. The emulsification proceeds in such a way that as soon as the shear exerted by the turbulent microeddies on the droplet interface exceeds the cohesive forces of the liquids in the drops, they split up to smaller units. The cleavage occurs as long as a balance between the external stress and the internal stress is reached. At this steady state the droplet size distribution becomes time-independent. The final droplet size is influenced by various parameters such as the volume–phase ratio, the viscosity of both phases, the mutual solubility of both phases, the kind and concentration of surfactants, the power input, the stirrer as well as the vessel geometry, the diminution energy, and thermodynamic changes during the emulsification process (reactions, temperature). Contrary to

456

HETEROPHASE POLYMERIZATION

Vol. 6

comminution techniques, the preparation of emulsions by condensation does not require mechanical energy, except gentle stirring sometimes to avoid creaming or settling as a result of density differences between both phases. Condensation processes are mainly determined by thermodynamic principles. There are basically two different types of condensation methods: droplet nucleation and swelling of dispersed phases; but the combination of both also has some meaning, especially for heterophase polymerizations (cf Swelling and Nucleation). Microemulsions are the most prominent example of purely thermodynamically driven emulsion formation, which can be considered as swelling of surfactant solutions. Preparation of emulsions and their use in various heterophase polymerization processes has been reviewed (239). An emulsion made by comminution techniques is thermodynamically not stable and hence it is exposed to degradation processes. Degradation of Emulsions. Emulsions can degrade via different mechanisms: phase separation, Ostwald ripening, aggregation (as generic term for flocculation or coagulation and coalescence), and phase inversion. Additionally, changes in the thermodynamic conditions (composition and temperature) compared to the preparation conditions may have a devastating influence on emulsion stability. Phase separation means that an emulsion may separate into two phases: one phase enriched with the droplets and the other enriched with the continuous phase. Depending on the density difference between the continuous and dispersed phase these kinds of separation are called creaming or sedimentation depending on whether the upper or the lower phase is formed by the droplets, respectively. Aggregation processes (flocculation, coagulation, coalescence) can take place when the average distance between emulsion droplets is so close that attractive forces become dominant. The stability of emulsions can be controlled very effectively by the addition of proper stabilizers in sufficient concentrations so that generally in practice the aggregation of emulsions is not a serious problem. Phase inversion is a process where for a given stabilizer the continuous phase becomes the dispersed one and vice versa. This is mainly observed in the case of polymeric surfactants with a stabilizing moiety possessing a critical solution temperature. Prominent examples are surfactants with poly(ethylene glycol) units. Increasing the temperature leads to an increase in the hydrophobicity of the surfactant and it may subsequently promote the stabilization of water in oil instead of oil in water emulsion (cf Bancrofts’s rule). Furthermore, whether phase inversion occurs or not depends on the polarity of the oil phase; the kind of electrolyte and its concentration; other additives as for example organic, water-soluble solvents increasing the oil solubility in water; and on the oil volume fraction. Ostwald ripening as a degradation mechanism is a direct consequence of the polydispersity of the droplet size distribution after comminution. Thermal fluctuation in the size or in the curvature may produce polydispersity as well. It is easy to show (cf for instance (239)) that the chemical potential of a dispersed phase compared to that of the same bulk material is increased by 4σ V mon /Dd . Consequently, for the concentration of the dispersed phase just at the interface C(D) (where σ holds) follows equation 23, with C0 being the solubility of the dispersed phase in the continuous medium.

kB T lnC ( D) =

4σ vmon + kB T lnC0 Dd

(23)

Vol. 6

HETEROPHASE POLYMERIZATION

457

From equation 23 it can be seen that smaller the objects, higher is the C(D). As Dd goes to infinity C(D) equals C0 . This means that smaller objects have a tendency to dissolve whereas larger objects grow by the uptake of matter that is released from the small ones. Two points are important for heterophase polymerization: (1) For Ostwald ripening to occur, a certain water solubility is required (this is given for many monomers in Table 9) and (2) a direct contact between the droplets is not necessary as molecular diffusion through the continuous phase is sufficient. On the basis of these considerations it was supposed that it might be possible to stabilize emulsions against Ostwald ripening by addition of small amounts of a lyophobic substance. This was done by Higuchi and Misra (240) and experimentally verified by stabilization of carbon tetrachloride droplets in water with hexadecane. Kabalnov investigated Ostwald ripening theoretically under such conditions (241). Two cases are of special interest for polymerization. First, that the solubility of the lyophob in the continuous phase is zero (CLC0 = 0), which means that the total number of particles remains constant. However, the distribution of the dispersed phase from small to large particles changes the composition of the particles. The volume fraction of the lyophob (φ L ) in the smaller particles increases and in the larger ones decreases. Indeed, investigations on Ostwald ripening with sedimentation field-flow fractionation in fluorocarbon emulsions confirmed this theoretical prediction (242). This redistribution of the dispersed phase comes to an end when the capillary and the concentration forces (Raoult’s law) are in equilibrium. The equilibrium condition requires that the chemical potential of the monomer is the same in all particles. Equation 24 illustrates the situation at equilibrium (subscript e) regarding the lyophob for two particles [superscripts (1) and (2)] of different diameters, where vm is the molar volume of the monomer.

1 φL,e −

4σ vmon 1 4σ vmon 1 2 = φL,e − (1) (2) kB T Dd,e kB T Dd,e

(24)

Compensating for the Ostwald ripening requires φ L values depending on the droplet size, in the sense that smaller drops require a higher φ L and vice versa. However, this is practically impossible to control during the initial emulsification process and consequently, Ostwlad ripening inevitably takes place during and after any emulsification process and leads to a change in droplet size distribution during a certain maturation time. An equilibrium size distribution can be reached and Ostwald ripening will stop if between the average drop size (Dd,i ) and φ L at the end of the comminution process (subscript i) the inequality given by equation 25 holds. A detailed discussion and derivation of the equations can be found in References 239 and 241. φL,i >

4σ vmon 3kB T Dd,i

(25)

If this condition is not met, Ostwald ripening will further increase the differences in the chemical potential for particles with different sizes. This means that

458

HETEROPHASE POLYMERIZATION

Vol. 6

Table 9. Water Solubility of Monomers Monomer n-Octyl acrylate Dimethylstyrene Vinyl neodecanoate 2-Ethyl hexyl acrylate 2-Ethyl hexyl acrylate Vinyl neononanoate Vinyl 2-ethylhexanoate Vinyl toluene n-Hexyl acrylate Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene n-Butyl methacrylate n-Butyl methacrylate Hydroxyoctyl methacrylate Styrene Butadiene Vinyl pivalate n-Butyl acrylate n-Butyl acrylate Butadiene n-Butyl acrylate Chloroprene Butadiene Butadiene n-Butyl acrylate n-butyl acrylate Butadiene Butadiene Hydroxyhexyl methacrylate Butadiene Propyl methacrylate Ethyl methacrylate Propyl acrylate Ethyl methacrylate Vinylidene chloride Methyl methacrylate Ethyl acrylate Methyl methacrylate Ethyl acrylate Methyl methacrylate Methyl methacrylate Methyl methacrylate Vinyl chloride

Solubility, M 0.00034 0.00045 0.000504 0.0005389 0.000543 0.000543 0.000587 0.001 0.0012 0.00192 0.002256 0.00261 0.002976 0.003 0.00308 0.0035 0.00368 0.00423 0.004389 0.005 0.005568 0.00628 0.0078 0.0109 0.011 0.011 0.0129 0.013 0.015 0.0152 0.0161 0.0167 0.0229 0.024 0.037 0.0389 0.0447 0.0454 0.05 0.05 0.066 0.1498 0.15 0.15 0.1504 0.1504 0.159 0.1598 0.1672

Temperature ◦

25–50 C 25–50◦ C 20◦ C 20◦ C 20◦ C 20◦ C 25–50◦ C 25–50◦ C 30◦ C 25◦ C 25◦ C 20◦ C 40◦ C 25–50◦ C 50◦ C 20◦ C 50◦ C 30◦ C 50◦ C, 760 mm Hg 20◦ C 20–50◦ C 30◦ C, 760 mm Hg 25–50◦ C 25–50◦ C 25◦ C, 760 mm Hg 20◦ C 20◦ C 15◦ C, 760 mm Hg 15◦ C, 798 mm Hg 50◦ C 0◦ C, 760 mm Hg 20◦ C 20◦ C 25–50◦ C 30◦ C 25–50◦ C 25–50◦ C 20◦ C 20◦ C 25◦ C 20◦ C

Source 253 253 254 255 94 254 254 253 253 94 256 257 256 255 257 353 257 94 255 259 256 257 254 94 253 257 256 253 253 257 254 255 257 256 258 257 94 94 255 255 253 256 253 253 255 255 94 254 255

Vol. 6

HETEROPHASE POLYMERIZATION

459

Table 9. (Continued) Monomer

Solubility, M

Vinyl chloride Hydroxybutyl methacrylate Ethyl acrylate Ethyl acrylate Vinyl acetate Vinyl acetate Vinyl acetate Vinyl acetate Hydroxypropyl methacrylate Ethylene Methyl acrylate Methyl acrylate Methyl acrylate Methyl acrylate Acrylonitrile Acrylonitrile Acrylonitrile Acrylonitrile Acrolein Methacrylic acid Hydroxyethyl methacrylate

Temperature ◦

25 −50 C 50◦ C

0.17 0.710 0.179 0.184 0.232 0.29 0.29 0.296 0.382 0.4 0.604 0.616 0.635 0.65 1.358 1.376 1.395 1.6 3.1 11.6 ∞

20◦ C 25–50◦ C 28◦ C 20◦ C 50◦ C 25–50◦ C 20◦ C 25–50◦ C 0◦ C 20◦ C 25◦ C 25–50◦ C 25–50◦ C 25◦ C 50◦ C

Source 253 258 256 94 254 253 256 255 258 253 256 94 255 253 256 256 256 253 253 254 258

the addition of a completely water-insoluble hydrophobic compound during the emulsification also does not per se prevent Ostwald ripening. Ostwald ripening takes place in any case, except for an exactly monodisperse droplet size distribution, which is hard to realize in actual practice. Hexadecane, which is frequently used as lyophob in miniemulsion polymerization, has a water solubility of about 4 × 10 − 11 M at room temperature (243), which is low but not zero. In this case the analysis of Ostwald ripening (241) shows that there exists a critical drop size (Dd,c ) in the droplet size distribution for which the flux of the lyophob out of the drops is zero. The critical drop size is given by equation 26, where CLC is the concentration of the lyophob in the continuous phase. Dd,c =

4vmon σ   kB Tln CLC /CLC0 φL (Dd,c )

(26)

Note that for Dd > Dd,c and Dd < Dd,c , the flux is positive and negative, respectively. The characteristic time of the exchange of monomer from droplets during Ostwald ripening (τ ex ) is given by relation 27, where φ MC is the dimensionless solubility of the monomer in the continuous phase (φ MC0 = CMC0 V mon ) and DMC is the monomer diffusion coefficient in the continuous phase. τex ∝

Dd2 φMC0 DMC

(27)

460

HETEROPHASE POLYMERIZATION

Vol. 6

To preserve the droplets as reaction loci, the polymerization has to compete with the bleeding of the emulsion droplets (239). A consequence of these considerations is that the best lyophob for the preparation of emulsion droplets for subsequent polymerization are polymers in form of seed particles. Swelling is a thermodynamically controlled process and consequently, swollen seed particles are not subjected to degradation processes to the same extent as emulsion droplets made by comminution processes. Furthermore, in most cases the chemistry of the seed particles does not interfere with the final properties of the dispersion. For example a seed particle with a diameter of 20 nm represents in final particles of about 100 nm diameter only less than 1% of the total volume. However, seed processes with suppressed nucleation of new particles are not always the method of choice. High Solids Polymer Dispersions. Polymer dispersions with higher solids content, clearly above 50 wt%, gain increasing interest for at least two reasons: increasing reactor capacity or productivity (cost reduction) and improved product properties (244–246). From the practical point of view the problem of high solids polymer dispersions is directly connected with rheology, as the viscosity of dispersions goes to infinity at certain solids content. The central problem is very general, that is, consideration of filling space with spheres goes back to Johannes Kepler. The stacking of cannon balls or oranges is intuitively done in a way that uses the available space in the most effective way. In two dimensions, exactly six spheres fit around a single sphere in the center, and six fit around each of the outer spheres. The next layer of spheres fits into the hollows of the first layer and again there is a hexagonal pattern. Kepler suspected that this filling pattern (about 75%) is the tightest way to pack spheres, but he was unable to prove it. This is known as Kepler’s conjecture and is still today a matter of research (247). For latex spheres close packing depends on the potential that imparts stability to them. Soft potentials interact over long distances and prevent closer contacts between the particles (cf stabilization). In this sense it is not the mass fraction (solids content) but the effective volume fraction (including a hydrodynamic layer thickness) that determines viscosity. To get latexes with solids content of about 70% or even higher requires polymodal particle size distributions with larger and smaller spheres (the smaller spheres have to fit in the interstitial space between the larger ones). This principle was successfully applied to get high solids polymer latexes, with up to 70% solids content and low viscosity (245). It was found that there is a critical ratio between the larger and smaller sphere diameters, which should be above 10 in order to get both low viscosities and high solids content. The preparation of polymodal latexes requires the occurrence of multiple nucleation events in the course of the polymerization, which can be realized either by emulsifier feeding policies (244,246) and/or by proper choice of recipe components (comonomers) (245). In summary, the application of seed particles in heterophase polymerizations is very promising and indeed, for many industrial polymerizations state-of-the-art (39). Furthermore, particles with special properties such as particle size, monodispersity, functionality, and porosity are prepared by the application of seed particles and activated swelling procedures [cf below and (248–250,252)].

Vol. 6

HETEROPHASE POLYMERIZATION

461

Solubility and Solubilization in the Continuous Phase Monomers. One of the most important properties of a particular monomer or monomer mixture that is subjected to heterophase polymerization is its solubility in the continuous phase. Table 9 summarizes the water solubility of monomers, which are frequently used in heterophase polymerizations. Monomers with different water solubility of course behave differently during the polymerization process in several ways. Although the data in Table 9 for a particular monomer show some scatter they confirm that there is indeed no limit for monomers with regard to their water solubility to be applied in heterophase polymerization. Even if the monomers are so hydrophilic that their polymers are water-soluble they can be applied as comonomers together with more hydrophobic monomers. Of course, the water solubility of monomers depends on temperature, and for gaseous monomers also on pressure, as indicated in Table 9. However, the water-solubility is not always a linear function of temperature, as is demonstrated in Figure 25 for styrene and methyl methacrylate. Besides the range of the solubility, the shape of the curves is also distinctly different for both monomers. Whereas the water solubility for styrene increases over the entire temperature range, the curve for methyl methacrylate goes through a shallow minimum at about 45◦ C. At temperatures 100

Solubilty, M

10−1

10−2

10−3

10−4

10−5

0

20

40 T, °C

60

80

100

Fig. 25. Solubility values for styrene and methyl methacrylate in aqueous emulsions.  water in styrene, determined with either Karl–Fischer titration or cloud point (261); strene in water, determined with formaldehyde–sulfuric acid reagent (261); styrene in water determined via cloud point (261); MMA in water (260); (262).

462

HETEROPHASE POLYMERIZATION

Vol. 6

Table 10. Interfacial Tensions between Monomer Droplets (σ m,w ) and the Corresponding Polymer Particles (σ p,w ) and Water, and Surface Area of Sodium Dodecyl Sulfate Occupied at Saturation (aS ) Monomer/polymer Methyl acrylate Vinyl acetate Ethyl acrylate Propyl acrylate n-Butyl acrylate Styrene Vinyl chloride Methyl methacrylate n-Butyl methacrylate n-Octyl methacrylate

σ m,w , mN/m

σ p,w , mN/ma

13–14a 18–19a 21–22a 26–27a 30–31a 40–43a 20.2b 20.8b 38.7b 51.6b

1.75 23.5

32.7 37.8 26.0 36.7

aS a nm2 1.1 0.90 0.69 0.60 0.50 0.40 0.79 0.54

a Values

from Ref. 255. calculated from the regression σ m,w = −11.667 × logCm,w + 11.179, which was determined from data given in Ref. 255. b Values

below 25◦ C the values for styrene obtained with different methods do not agree closely, but at temperatures above 40◦ C the agreement is almost perfect. Furthermore, Figure 25 reveals that it is always necessary to consider mutual solubilities for a particular monomer/water combination as not only monomers are soluble in water but water also is soluble in monomers. In the case of some hydrophobic monomers like styrene, the solubility of water in the organic phase is even higher than that of the monomer in water. This is the case for many alkanes and alcohols as well (259). The consequence of this fact is of some importance as together with water, hydrophilic radicals may enter the monomer phase and start polymerization. Moreover, as the forces between the molecules of the liquids forming an emulsion (the cohesive and adhesive forces) determine not only the interfacial tension but also the mutual solubilities there is a relation between the interfacial tension and the miscibility (the higher the miscibility, the lower is the interfacial tension). Table 10 summarizes values for the interfacial tension between monomer droplets and the corresponding polymer particles and water, respectively, which clearly show this effect. The interfacial tensions of polymer– water interfaces are not directly accessible, but have been calculated mainly on the basis of polymer polarity data (255). Furthermore, Table 10 shows that the sodium dodecylsulfate adsorption on the polymer particles is also influenced by the polarity of the polymer surface. There is a general rule: the higher the polarity of the interface, the larger is the area covered by one sodium dodecylsulfate molecule in a saturated adsorption layer. Consequently, the lower the number of stabilizing groups per unit area, the higher is the polarity of the interface. Thus, the hydrophilicity of a monomer and the polarity of a polymer have direct influence on all colloid chemical properties during heterophase polymerization, as the interfacial tension influences nucleation, swelling of the particles with monomer, surfactant adsorption, and stability of the droplets/particles. Also, the polymerization kinetics in both phases is influenced because the rate of polymerization is directly proportional to the monomer concentration (cf above)..

Vol. 6

HETEROPHASE POLYMERIZATION

463

Table 11. Solubility and Solubilization of Organic Molecules (Monomers and n-Heptyl Mercaptan) in Water and Surfactant Solutions, Respectivelya Monomer

Surfactant

Solubility/solubilization, M

Butadiene Butadiene Styrene Styrene Styrene Styrene MMA MMA MMA n-Hm f n-HM

None 2.8% SFb None 0.093 M KP None 0.093 M KPc None 1% CTABe 3% CTAB None 0.3 M SDS

0.0351,50◦ C, saturation pressure 0.196,50◦ C, saturation pressure 0.00308, 40◦ C 0.096,40◦ C 0.00368, 50◦ C 0.139,50◦ C 0.1598, 25◦ Cd 0.2 0.32 0.00007 0.0925

a From

Ref. 264. is a commercial fatty acid surfactant. c KP is potassium palmitate. d Value from Ref. 254. e DCTAB is cetyl trimethyl ammonium bromide. f n-HM is n-heptyl mercaptan. b SF

In the context of solubility it is necessary to briefly mention the phenomena of solubilization. This terminus technicus was introduced by J. W. McBain for a particular process of increasing the solution concentration of substances in a given medium (263). Solubilization can be understood as a sorption phenomenon as it leads to a decrease in the number of molecules of a substance in a reservoir outside the solution but in equilibrium with the solution, by adding an adsorbent to the solution (see ADSORPTION). For heterophase polymerizations the adsorbents are surfactant molecules, which decrease the number of molecules outside the continuous phase (eg in monomer droplets). The amount of the adsorbed or solubilized material increases steeply with increasing stabilizer concentration and levels in the vicinity of the cmc. Solubilization has nothing to do with a true solution but is an effect connected with colloidal systems, which can be formed either by micelles or other particles able to interact with the adsorbate. For aqueous heterophase polymerizations this means that the solubilizing power of micelles, or in other words the solubility of a monomer in a surfactant solution is much greater than in water (cf selected data in Table 11). As solubilization is mediated by micelles and hence colloidal in nature, it depends on the ionic strength in aqueous solution and on temperature. Dissolved Gases. Oxygen is known to interfere with radical polymerization in different ways. On the one hand its presence leads to detrimental effects such as induction periods, retardation, and side products via peroxide formation. On the other hand in vinyl chloride emulsion polymerization, a certain level of oxygen leads in the presence of basic buffers to an acceleration of the polymerization as a result of decomposition of in situ–formed peroxides (265). Either at higher concentrations of oxygen or in the absence of buffer, retardation of the polymerization and long inhibition periods are observed. Studies of the influence of oxygen have been carried out since the beginning of heterophase polymerization

464

HETEROPHASE POLYMERIZATION

Vol. 6

research, when oxygen was, even if unintentionally, the sole initiator (21,24). Under conditions of heterophase polymerization, it is important to consider oxygen partition between all phases. In principle, there are three places where oxygen can be found: in the headspace, the continuous phase, and the dispersed phase. In aqueous emulsion polymerization, most of the oxygen is concentrated in the headspace and smaller amounts only in organic phases and water. However, it is necessary to consider that the solubility of oxygen is higher in many organic solvents than in water (266). In a calorimetric study of both seeded styrene and seeded styrene/butyl acrylate emulsion polymerization, the presence of oxygen leads to an inhibition period and in due course the redistribution of oxygen from the headspace causes a retardation of the reaction (267). The action of oxygen as both a water–soluble and a monomer–soluble gas was proven in a series of styrene emulsion polymerizations at 60◦ C with sodium dodecyl sulfate as emulsifier and potassium peroxodisulfate as initiator (268). The authors determined the saturation concentration of oxygen in watery solutions of various compositions at polymerization temperature. The saturation value for pure water, water plus emulsifier, and water plus emulsifier plus styrene was found to be 5.40, 5.58, and 6.53 ppm, respectively. A saturated continuous phase under polymerization conditions (water plus emulsifier plus styrene) sparged with nitrogen of high purity would decrease in dissolved oxygen to a level lower than 0.05 ppm in about 10 min. The authors observed a significant inhibition period only if oxygen is present in concentrations higher than 50% of the saturation concentration. Furthermore, at the end of the inhibition period the oxygen level in the aqueous phase was found to be about 3 ppm independent of the starting values. This experimental finding suggests that high enough initiation rates can outweigh the inhibition action of dissolved oxygen. Another action of dissolved gas, not only oxygen but also nitrogen or other inert gases, is its interference with the nucleation process, as was shown in Reference 91. The reproducibility during the nucleation stage of an ab initio emulsion polymerization was drastically improved if the water and the monomer were carefully degassed instead of purged with nitrogen prior to polymerization and if the water was added to the reactor at a temperature higher than polymerization temperature in order to avoid bubble generation.

Swelling of Polymer Particles Swelling means that polymer particles in a dispersed state even in the presence of an excess amount of solvent for the polymer do not dissolve but imbibe only a limited amount of that solvent. The importance of this phenomenon for heterophase polymerizations becomes clear by the fact that most of the monomers (acrylates, methacrylates, styrene) are good solvents note only for their own but also other polymers. Thus, during heterophase polymerization monomer droplets and polymer particles can coexist side by side. Swelling is of special practical importance because the colloidal particles consisting of both polymer and monomer are the main reaction loci. The monomer concentration in swollen polymer particles is governed by equilibrating the free energy of mixing between polymer and monomer with the interfacial free energy of the particles and the swelling

Vol. 6

HETEROPHASE POLYMERIZATION

465

pressure (cf equation 28) (121). The term on the right-hand side of equation 28 is the free energy of mixing between polymer and monomer, where the Flory– Huggins–Rehner expression for cross-linked polymers is used. In equation 28, σ is the interfacial tension between the swollen particles and the continuous phase, PSW is the swelling pressure, χ mp is the polymer monomer interaction parameter, φ 2 is the polymer volume fraction in the swollen particle, RT is the thermal energy, j is the average degree of polymerization of the polymer molecules, ρ 2 is the polymer density, and Mc is the average molecular weight between two cross-links in the network. If the latex particles are not cross-linked (Mc =⇒ ∞) and if PSW 0, equation 28 becomes identical with the Morton–Kaizerman–Altier equation (269) also (270). Between φ 1 , the monomer volume fraction, and φ 2 , the relation φ 1 + φ 2 = 1 exists. The swollen (r) and unswollen particle radii (r0 ) are connected with V , φ 2 via (r/r0 )3 = 1/φ 2 , whereas φ 1 is related to CM,p via ϕ1 = CM,p Vmon = Vm,p m,p + Vp,p with V m,p and V p,p being the monomer volume and the polymer volume inside the swollen particles, respectively.


2σ + PSW r





 vmon 1 vmon ρ2 1/3 φ2 + χm,p φ22 + (φ2 − φ2 /2) = − ln(1 − φ2 ) + 1 − RT j MC (28)

The right-hand side of equation 28, except the last term, for which one needs to consider only for cross-linked particles, promotes swelling whereas the terms on the left-hand side counteract swelling. The first term on the left-hand side is the interfacial free energy, which counteracts swelling as a result of an increase in the particle interface. The second term on the left-hand side originates from a volume work due to attractive forces between polymer chains in concentrated solutions, as it has been concluded from osmotic modulus measurements (271). The pressure in equation 28 can be considered as swelling pressure of colloidal particles, comparable to that known from the swelling of macroscopic gels, contributing together with the partial molar free interfacial energy to the equilibrium with the chemical potential of the swelling agent (
µ1 ). It can be in the order of up to 10 MPa, depending on the decoration of the particle interface (272). An illustration of the dependence of the equilibrium monomer volume fraction (φ 1 ) on the particle size is shown in Figure 26. The higher the φ 1 , the larger are the particles, the lower is the interfacial tension, and the higher is the temperature, but it decreases with increasing degree of cross-linking and with increasing Flory– Huggins interaction parameter. The two limiting cases of equation 28 for r =⇒ ∞ and φ 2 =⇒ o is of special meaning. In the case of a bulk polymer (r =⇒ ∞) the resulting equation is the osmotic or swelling pressure relation given by the equation
µ1 = −s V mon , with PSW = −π s . If the polymer concentration is vanishing (φ 2 =⇒ 0), the result is the Young–Laplace equation (
Pd = 2σ /r) for a colloidal droplet or bubble embedded in a fluid phase. This behavior can be regarded as expression of the situation that the swollen latex particles represent a polymer gel or solution but with colloidal dimensions. Equation 28 considers the influence of surfactants on swelling only via the interfacial tension. But this may be insufficient as entropy effects are completely neglected. The above results

466

HETEROPHASE POLYMERIZATION

φ1

Vol. 6

D T 0.5 σ cross-linking χ > 0.5

100 nm

D

Fig. 26. Illustration of the dependence of φ 2 on polymer and particle parameters.

with regard to the behavior of swollen micelles, which are stable as long as the monomer is not polymerized and the surfactant tails can keep their mobility, demonstrate the influence of entropic effects on surfactant–monomer interaction. For monomer partition inside swollen latex particles, which are stabilized with adsorbed surfactants these effects favor the formation of a monomer rich shell rather than a uniform distribution throughout the whole particle volume. This effect is very similar to adsolubilization, which describes the solubilization of organic compounds inside adsorbed stabilizer layers on inorganic particles in order to modify them by admicellar polymerization (273). There is clear experimental evidence by small-angle X-ray scattering that in poly(methyl methacrylate) particles swollen with methyl methacrylate the monomer is concentrated inside an outer shell of approximately 2-nm thickness. The particles were stabilized with two kinds of surfactants: sodium dodecylsulfate, which was present already during the polymerization, and sodium dodecylbenzenesulfate, which was added during the swelling experiments to avoid coagulation (274). Swollen polymer particles are high concentration polymer solutions (about 50% by weight) and hence, the swelling pressure counteracts swelling due to attractive forces between entangled polymer chains and might contribute to formation of the monomer-rich shell. This gradient with regard to monomer concentration inside swollen particles should cause during polymerizations, especially in seeded polymerizations, a morphology gradient with regard to composition. In this sense, the morphology of polymer particles in heterophase polymerization, especially emulsion polymerization, has

Vol. 6

HETEROPHASE POLYMERIZATION

467

been a matter of ongoing discussion for several decades (275–279). The monomer concentration inside the particles as main reaction loci (ie swelling of latex particles) determines rate of polymerization, particle morphology, molecular weight, and in copolymerizations the chemical composition also of the copolymers. Consequently, methods either to measure or to predict the monomer concentration inside particles are a key aspect in heterophase polymerization research and are of permanent importance (280–282). John Ugelstad’s theoretical and experimental work to development new methods of monomer emulsion preparation (283) makes good use of one of the pillars of colloid science, the Kelvin equation. The Kelvin equation says that larger objects, in coexistence with smaller ones, will grow in size at the expense of the smaller objects, which have a tendency to dissolve. This effect is known in colloid science as Ostwald ripening (discussed earlier). Equation 29 shows the Kelvin equation for liquid droplets surrounded by vapor where p is the vapor pressure outside the drop, p0 is the bulk vapor pressure, v is the molar volume of the liquid, and RT is the thermal energy (284). The smaller the radius of the drops (r), the larger is p and hence, the higher the tendency to degrade. 

p 2σ v ln = p0 r RT

(29)

However, as discussed above, the dissolution can be retarded or even prevented if the drops contain a substance that is insoluble in the continuous phase (compound 2 or lyophob). In this case the decrease in size increases the chemical potential of the lyophob inside the smaller drops and generates a force counteracting Ostwald ripening. A detailed discussion of emulsion stability with regard to subsequent polymerization can be found in Reference 239. In addition to the stabilizing effect of compound 2, Ugelstad was able to show that compound 2 has an additional effect on swelling, as the entropy term in the swelling equation (eq. 28), is (1 − 1/j)φ 2 , becomes more and more important as long as compound 2 has a low molecular weight (or suppose it is an oligomer where j is in the order of about 5). On the basis of these two effects, Ugelstad and co-workers developed an activated two-step swelling procedure, which enabled them to prepare and to commercialize large monodisperse particles for various applications as high value specialty polymer colloids (251,285). In brief, in the first step an emulsion of the low molecular and highly water-insoluble compound 2, which may contain a solvent that is water-soluble, is added to a suspension of polymer particles (seed). The solvent promotes the transport of compound 2 through the aqueous phase and allows swelling of the polymer particles, which is further facilitated if the emulsion droplets are smaller than the seed particles (action of the Kelvin equation). Then, in a second swelling step the monomer emulsion is added. Because of the high entropy gain caused by mixing of compound 2, with the monomer inside the particles the swelling ability of the seed is activated compared to particles without compound 2. For instance, using dioctyladipate as compound 2, it was possible to swell polystyrene particles with a diameter of 1.55 µm with chlorobenzene as model compound to a diameter of about 30 µm, ie the seed particles imbibed more than a 7000-fold amount of their own volume. During subsequent polymerization,

468

HETEROPHASE POLYMERIZATION

Vol. 6

if instead of chlorobenzene a monomer is used in the second swelling step, precautions are necessary to avoid nucleation of new particles. The interaction between monomers and polymers during heterophase polymerizations is very complex and determines essential features of the particular heterophase polymerization technique such as polymerization rate and particle morphology. A control of swelling enables the preparation of polymer particles with specially tailored properties.

Shrinkage of Monomer-Swollen Polymer Particles Any polymerization reaction causes a decrease in the average distance between the acting molecules, ie the average distance between individual monomer molecules is larger than that between the repeating units in the polymer. This causes an increase in density although the length of a carbon–carbon double bond is shorter than that of single bond. Table 12 summarizes densities of selected monomers and polymers, which play an important role in heterophase polymerization. The densities of the monomers depend much more on the temperature than those for the polymers, as shown in Figure 27. To illustrate the shrinkage, the volume mon mol and vru , respectively, can per monomer molecule and per repeating unit vmol be calculated by means of equation 30, where M mon is the molecular weight of the monomer, N A the Avogadro’s number, and ρ 1 and ρ 2 are the monomer and polymer densities, respectively..

Table 12. Densities of Selected Monomers (ρ m ) and Corresponding Polymers (ρ p ) at Various Temperatures Monomer T◦ C Methyl methacrylate

Styrene

Vinyl chloride

Vinyl acetate

a Values b Values

taken from Ref. 286. taken from Ref. 287.

Polymer

ρma g/cm3

T◦ C

ρp,b g/cm3

20 40 60 80 20 40 60 80 20

0.9428 0.9203 0.8971 0.8730 0.9049 0.8869 0.8686 0.8498 0.9115

0 20 25

1.195 1.190 1.180

40 60 80 20 40 60 80

0.8760 0.8375 0.7950 0.9323 0.9075 0.8816 0.8544

ρ p - 1.04–1.065; Average ρ p = 1.05

ρ p - 1.391–1.431; depending on polymerization temperature, average ρ p - 1.40

20 25 50 1120

1.191 1.190 1.17 1.110

Vol. 6

HETEROPHASE POLYMERIZATION

469

1.3

Density, g/cm3

1.2

1.1

1.0

0.9

0.8

0

20

40

60

80

100

120

T , °C

Fig. 27. The dependence of density of selected monomers and polymers on temperature Vinyl acetate; poly(vinly acetate); methyl methacrylate; (cf also Table 12). poly(metnyl methacrylate).

mon vmol =

Mmon ru Mmon vmol = NA ρ1 NA ρ2

(30)

With the data given in Table 12 the corresponding values for styrene are, at mon = 2.04 × 80◦ C when the average polystyrene density of 1.05 g/cm3 is used, vmol ru 10 − 22 cm3 and vmol = 1.65 × 10 − 22 cm3 . These values mean that styrene drops, under the assumption that there is no mass transfer, would shrink from the start until the end of the polymerization at 80◦ C by about 19.12%. This is an enormous number and has in the case of heterophase polymerizations much more important consequences than for homogeneous polymerizations, where the reaction volume shrinks evenly. The sketch in Figure 28 elucidates these consequences. Case A is the ideal behavior where a polymerization proceeds smoothly within a droplet beginning from the center and ending with an evenly shrunk particle with homogeneous density throughout the whole particle volume. Case B represents the other extreme where polymerization starts at the particle–water interface and thus, a shell is formed that might prevent shrinkage of the polymerizing drop (indicated by the stacking of the triangles). In that case, as polymerization proceeds, monomer moves from the center to the shell as a result of swelling and leaves just one big hole in the middle. Because the size of the final particle is that of the starting drop, this process is called constraint shrinkage. Case C is intended to illustrate the real situation where polymerization begins at many sites inside a drop. In that case shrinkage is in between both extremes, and a void structure on the nanometer-size scale characterizes the particle morphology. Equations 31 can be used to quantify the droplet shrinkage and to estimate the magnitude of the

470

HETEROPHASE POLYMERIZATION

Vol. 6

A

B

C

Fig. 28. Illustration of effects caused by volume shrinkage during polymerization in discrete particles.

effects. Dp and Dm denote the polymer particle and monomer droplet diameter, respectively,
D is the decrease in diameter due to shrinkage according to case A, and dh is the size of the single hole formed as described in case B (cf Fig. 28).


Dp = Dm

ρm ρp

1/3


D = Dm 1 −




ρm ρp

1/3 


 ρm 1/3 dh = Dm 1 − ρp

(31)

The data plotted in Figure 29 show the development of
D and dh for styrene and vinyl chloride polymerizations in dependence on temperature, where for the polymers the average density value as given in Table 8 were used. These model calculations show that shrinkage is an important factor during heterophase polymerization, which decisively influences particle morphology and should not be neglected, especially not in cases where the polymerization takes place mainly inside monomer-swollen droplets. Shrinkage may lead to porous particles and can also be utilized to prepare hollow particles.

Vol. 6

HETEROPHASE POLYMERIZATION

471

20

d h /D m, %; ∆D /D m, %

18 16 14

Vinyl chloride

12 10 8 6 Styrene 4 2

0

20

40

60

80

100

120

T, °C

Fig. 29. Shrinkage and hole diameter with respect to the droplet size for polymerization of styrene and vinyl chloride in dependence on polymerization temperature (according to equations 31 and case A and B in Fig. 28). —
D; · · · · dh .

Stabilization and Stabilizers The excess interfacial energy of polymer dispersions, compared with bulk or solution systems, is the driving force for an inherent tendency to decrease the interfacial area and hence, to instability. Stabilization of particles and droplets was recognized as a key issue in heterophase polymerization from the very beginning, when Kurt Gottlob described in the first patent (21) the utilization of natural polymers such as starch, egg albumin, or gelatin as stabilizers. Some excellent contributions placing special emphasis on colloidal stabilization in connection with heterophase polymerization can be found in References 288–291. The most important principle of stabilization is that the stabilizer or the stabilizing groups must be soluble in or strongly interacting with the continuous phase. This principle is known as Bancroft’s rule and was formulated in 1912 by Wilder Bancroft in a paper on titled The Theory of Emulsification, when he generalized the results of emulsification experiments known at that time: “When water is the dispersing phase, the emulsifying agent should be a hydrophile colloid. . ., the emulsifying agent should be an oleophile colloid in case the emulsion is to contain water in drops” (292). The same principle is behind the empirical HLB (hydrophilic–lipophilic balance) concept, which was developed mid-20th century in order to make the choice of the most effective emulsifier for a desired emulsion easier (293). Hydrophilic emulsifiers, which favor the formation of oil-in-water emulsions have HLB numbers greater than 10, whereas emulsifiers that are more soluble in oil phases have lower HLB numbers and stabilize water–in–oil emulsions more easily. A scientific basis for the HLB concept can be found in the cohesive energy ratio of emulsions that was developed by Beerbower and Hill (294) and which was successfully applied to formulate for instance polymerizable microemulsions (295). Stability is achieved if the Gibbs free energy of mixing is negative. In this sense colloids can be of two basic types, lyophobic and lyophilic. Lyophilic colloids interact strongly

472

HETEROPHASE POLYMERIZATION

Vol. 6

whereas lyophobic colloids have only a very restricted interaction with the continuous phase. Polymer dispersions can be of both types. Examples of lyophilic polymer dispersions, which can also be absolutely stable in a stricter thermodynamic sense are block copolymers in selected solvents or microgel particles such as cross-linked polystyrene in toluene or cross-linked poly(styrene sulfonate) particles in water. In contrast lyophobic polymer colloids in a bare state are thermodynamically unstable and need, in order to keep dispersed for a longer period of time, to be stabilized by third substances. Attractive Forces. The origin of attractive forces, which are summarized as van der Waals forces, between adjacent particles is the interaction between the molecules and atoms of which the particles are made and is basically electromagnetic in nature. There are three sources of forces attracting uncharged molecules: First, forces produced by aligned permanent dipoles in separate molecules (Keesom forces). Second, forces between permanent dipoles and induced dipoles (Debye forces). Both forces account mainly for binary interactions, because two dipoles, which are in optimum alignment for attractive interaction, cannot be at the same time in the optimum position for attractive interaction with a third molecule. The third kind of forces are of quantum-mechanical nature (London forces) and caused by interaction due to the coupling of random fluctuations of the electrical fields of adjacent molecules, which reduces the free energy and hence are attractive. The London attractive forces are fundamentally different from Keesom or Debye forces, as a molecule can interact by means of London forces with all molecules in its neighborhood at the same time. The attractive potential energy (V A ) decreases with increasing separation distance (dpp ) according to equation 32, where kL is the London interaction constant, Z is the number of electrons in the outer shell, ν c is the frequency of the fluctuations, h is Planck’s constant, and α 0 the static polarizability of the molecules.

VA = −

kL kL = 0.75Z0.5 hνc α02 6 dpp

(32)

The calculation of the attractive potential between condensed particles separated by vacuum goes back to Hamaker (296–298). He obtained, under the assumption that the pairwise interaction as used by London for gas molecules can be applied to condensed phases, from an integration over all possible interactions between attracting elements of two particles the following expression (eq. 33), where A is the Hamaker constant, which has the dimension of energy, q is the concentration of interacting centers per unit volume, and H a geometrical function of the particles. An approximate expression of H is given for two spheres of radius r at distance dpp r. Equation 33 shows that the energy of attraction varies with the reciprocal of the separation distance. Model calculations have shown that it might be of importance over a distance of tens of nanometers rather than within the close-contact region below 1 nm.

VA = − AH A= π 2 q2 kL H =

D 1 · 2 dpp·12

(33)

Vol. 6

HETEROPHASE POLYMERIZATION

473

There are several possibilities to refine these calculations such as the consideration of correlation frequencies over the whole spectral range (London considered only ultraviolet frequencies) and the use of a continuum electrodynamic model rather than the assumption of pairwise interaction. Notwithstanding these problems, the Hamaker model can also be applied to the interaction of condensed particles in liquid media to gain knowledge of the basic principles. As the electromagnetic coupling of fluctuating dipoles needs a certain time the fluctuations can move out of phase in dependence on the distance of the interacting particles, which leads to decreasing attractive potential (retardation effect). The properties of the intervening medium between interacting bodies contribute in specific ways. The primary medium effect is due to the medium influence on the transmission of the electromagnetic field between interacting particles. The strength of the radiation of the electric field depends inversely on the dielectric constant, and the optical path length between interacting particles increases in proportional to the square root of the dielectric constant of the medium. The latter influences the retardation effect and reduces attraction. The secondary medium effect considers the finite attraction between particles and molecules of the continuous phase. Under such circumstances the attractive energy is smaller than that between the particles in vacuum, and for the energy balance one has to consider that particle– molecule contacts of the medium will be replaced by particle–particle contacts. This means that chemically identical or similar particles will attract each other, but that this must not necessarily be the case for chemically different particles, where under particular circumstances the dispersed state may be energetically favored. For attracting particles the secondary medium effect finds its expression in a so–called effective Hamaker constant, as defined by equation 34, where Ap is the Hamaker constant of the polymeric particles and Am that of the medium in vacuum. 2  1/2 − A Aeff = A1/2 p m

(34)

Some values of Hamaker constants for various polymers and liquids that can be applied as continuous phase in heterophase polymerizations are put together in Tables 13 and 14. These data reveal that effective Hamaker constants are clearly smaller than the values for the corresponding polymers and continuous phases. However, the data collected in Table 13 show that Hamaker constants of various polymers are not that much different but the values obtained with different methods for the same polymer differ much more. Consequently, before using values of Hamaker constants it is always necessary to check their origin in order to be sure to compare values obtained with same approximations. According to equation 34 it is possible that the effective Hamaker constant, and hence the London attraction forces also, can be zero. Such situations are virtually obtained in the case of lyophilic microgel particles, which are highly swollen with dispersion medium. To counteract the attractive forces in lyophobic colloids repulsive forces must operate and consequently, the question regarding an overall interaction potential between colloidal particles arises. Equation 31 leads for separation distances close

474

HETEROPHASE POLYMERIZATION

Vol. 6

Table 13. Hamaker Constants for Different Polymers (A in 10 − 20 J) Calculated with Different Approximate Proceduresa Polymer Polystyrene Poly(ethylene terephthalate) Polyethylene Polypropylene Polyamide-6 Polyacrylonitrile Regenerated cellulose

1b

2c

3d

4e

5f

6g

4.62 3.94 3.97 3.87 3.33 3.47 4.19

7.7 8.4 8.25 8.25 8.1 8.0 —

7.7 6.4 6.2 6.2 6.1 6.0 6.5

— 4.0 — — 4.3 2.8 3.9

5.7 5.2 4.4 3.6 4.7 — 3.8

3.46 3.05 3.08 3.06 2.59 2.96 —

a From

Ref. 299. spectroscopic data with ionization energy IE , A = 27 IE (−1)2 /[64( + 2)2 ] c Same equation in footnote b but from refractive index measurements. d Approximate formula, A = 4.02 × 10 − 19 ((ε − 1)/(ε + 1))2 . e Approximate formula, A = 6π r2 γ d , where γ d is the surface tension dispersion part, and 6π r2 is s s the size of interacting surface volume element. f Same approximate formula as in footnote e but different methods to determine γ d . s 2 g Approximate formula, A 11 = 0.75hvc α0 {1 − [3πqα0 (ε − 1)/(ε + 1) − 1.5]}. b From

Table 14. Effective Hamaker Constants for Different Combinations of Materials Aeff × 10 − 20 J Material a

Water Pentane Hexane Decane Tetradecane Hexadecane Paraffin waxb Poly(methyl methacrylate)c Poly(vinyl chloride)c Polystyrenec Polystyrenec Polyisoprenec Polytetrafluoroethylenec

Air

H2 O

3.7 3.8 4.1 4.8 5.1 5.2

0 0.3 0.4 0.5 0.5 0.5 0.02 1.05 1.30 0.95 0.911 0.437 0.333

7.11 7.78 6.58 6.37 5.99 3.8

a Values

taken from Ref. 300. taken from Ref. 301 (cf. footnote d in Table 13). c Values taken From Ref. 302. b Values

to zero to an infinite attraction. However, this result contradicts the practical experience that any material can be comminuted into colloidal objects. Repulsive Forces. For an overall balance of interaction forces, one has to consider the repulsion that occurs if surfaces come as close as an atomic diameter together. This strong repulsion is known as Born repulsion (V R,B ) and changes −x , where x is a number of order larger than 10 (290). As the in proportion to dpp Born repulsive potential acts only at separation distances of atomic diameters, it does not contribute to the stability of colloidal polymer dispersions. There are

Vol. 6

HETEROPHASE POLYMERIZATION

475

other forces counteracting already the inherent attraction at larger separation distances. Charge Stabilization. The contact between materials with different dielectric constants leads according to Coehn’s rule (303,304) to electric charges, as the material with the lower permittivity charges negatively. For water as continuous phase, with a permittivity of about 80, which is larger than that of polymers and many inorganic materials, this means that interfaces in contact will charge negatively. Coehn’s rule means furthermore that charge stabilization is inherently present in colloidal systems. For practical applications in polymer dispersions (ie also during heterophase polymerizations) the effect of charge stabilization due to permanent charges, either by ionic stabilizers, comonomers, or initiator fragments, is much more important. The repulsive potential between two equally charged particles (V R,es ) separated at a center-to-center distance of (dpp + D) is given by equation 35, where Q is the charge on each particle and α is the permittivity of the continuous phase.

VR,es =

Q2 (dpp + D)ε

(35)

Equation 35 can be rewritten by using the capacitance of a conducting sphere (CDε/2) and the surface potential ( = Q/C) in the form of equation 36.  2  2 D/2 ε  VR,es =  dpp + D

(36)

A closer inspection of equation 36 reveals that (1) particles in water have both larger capacitance and repulsive potential than those in organic continuous media, (2) the repulsive potential falls with increasing separation between the particles, and (3) at separation distances much smaller than the particle size, larger particles are easier to stabilize than smaller ones because V R,es = (D/4)ε × ζ and the ratio V R,es V A is independent of D. For electrostatically stabilized particles the interaction energy between two approaching particles can be calculated by means of a theory, which was developed independently during the 1940s by Derjagin and Landau in the former Soviet Union and by Verwey and Overbeek in the Netherlands. This so–called DLVO theory assumes that long-range interactions determine colloidal stability and hence, the interaction potential (V int ) is calculated as the sum of attractive van der Waals–London forces and repulsive electrostatic forces based on the Goy–Chapman theory of charge distribution near a surface (305). Equations 37a–c can be used to calculate V int in dependence on the separation distance in order to get insight into the principles of the interaction between charge-stabilized colloids. In these equations Aeff is the effective Hamaker constant of particles interacting through the continuous phase, V salt is the ionic strength, κ − 1 is the Debye screening length, ε and ε 0 are the permittivities in the continuous phase and in vacuum, z is the stoichiometric valency of the electrolyte,

476

HETEROPHASE POLYMERIZATION

Vol. 6

and  is the surface potential of the particles.

 64kB TCsalt 02 Aeff 1 + exp( − κd ) Vint (dPP ) = πa − PP 12π dPP κ2

(37a)

0.5



 εε0 kB T  κ −1 = 

 (zi e)2 Csalt

(37b)

i


0 = tanh

ze 4kB T

 (37c)

Figures 30a and 30b illustrate the interaction between charged particles in dependence on particle size and ionic strength in the continuous phase. Both parameters (D and V salt ) have a strong influence on the interaction potential. The maximum at small separation distances (usually dpp 0. Flocculation is thermodynamically impossible if
H f is positive and
Sf is negative. This is called combined enthalpic–entropic stabilization and is observed in dioxane–water and methanol–water mixtures as the continuous phase for dispersions stabilized with poly(vinyl alcohol) and poly(ethylene glycol), respectively. If the magnitude of T
Sf is larger than that of
H f and both
Sf and
H f are negative, it is called entropic stabilization because the large negative entropy causes stability. However, entropically stabilized systems may flocculate upon cooling as
Gf might become smaller than zero. In the case of an enthalpically stabilized system, which might become unstable upon heating, the magnitude of
H f is larger than that of T
S, and both
H f and
Sf are greater than zero. The stabilization of dispersions in toluene or n-heptane with polystyrene and poly(12-hydroxy stearic acid), respectively, is entropic,. whereas poly(vinyl alcohol) in 2 M sodium chloride aqueous solution and polyisobutylene in 2-methylbutane act as enthalpic stabilizers. Note that poly(ethylene glycol) can act in entropic, enthalpic, and combined stabilization in methanol, 0.48 M aqueous magnesium sulfate solution, and methanol– water mixture, respectively. Depletion Interactions. A polymer need not necessarily be adsorbed to impart stability to a colloidal dispersion. This kind of interaction, is described as depletion interaction, emphasizing the nonadsorbed (or the depleted particle– water interface) state of the lyophilic polymers. The interaction between dispersed particles and dissolved polymers can be described by means of phase diagrams showing regions of stability and instability depending on the concentration and size of both species. Again a simple but vivid picture can be used to explain the principle of this kind of interaction. The action of dissolved polymers on dispersed particles depends on two characteristic lengths: the size of the dissolved polymer (Dee ) and the average distance between the particles (dpp , cf above). If dpp  Dee , that is a stable situation as the particle density is low and both particles and polymers move freely in the continuous phase. If the particle concentration increases and dpp decreases such that dpp > Dee , a situation may possibly be reached where the local concentration of the polymer in the vicinity of two particles is higher than that in a region of lower particle density. Thus, an osmotic force arises, and in order to dilute this region, continuous phase fluid soaks in and pushes the particles apart. However, if Dee > dpp and the concentration of dissolved polymer is larger outside an interstitial volume between particles, the flux of continuous phase fluid, which tries to dilute the local high polymer concentration, presses the particles together and hence, leads to coagulation of the dispersion. An analogous situation to depletion interaction, purely entropically driven, might also exist in neat colloidal suspensions of hard monodisperse spheres, that is, in good

480

HETEROPHASE POLYMERIZATION

Vol. 6

approximation in purely nonionically stabilized particles possessing cutoff or hard repulsive potential. In the case of such particles, a higher degree of order must not necessarily mean a loss of entropy. In random close packing at space filling of 64%, direct contact between the particles occurs. In contrast, in a crystalline hexagonal arrangement at maximum space filling of about 75%, free volume exists between the particles and hence, each sphere has freedom of movement (mainly rotation). This entropy of the free volume outweighs the configurational entropy (with regard to the sphere center) at 54.5% space filling and crystallization starts purely driven by the gain of configurational entropy. In a mixture of a few large and many small spheres (a situation that is comparable with the above particle plus dissolved polymers) both entropy-driven attraction and repulsion can occur. Around each large sphere is an excluded volume for the smaller particles with a thickness of their size. If the distance between the larger spheres is below the size of the smaller spheres, the excluded volume regions can overlap and hence, the effective volume for the smaller spheres is increased by
vc , which causes an entropy gain
Ssl = kB N ss ln[(vc +
vc )/vc ], where vc is the volume of continuous phase and N ss the number of small particles. This entropy gain causes a decrease in the free energy and the existence of an attractive force for the larger particles. In the reverse case, if the distance between the larger drops is greater, then an entropic repulsive force exists. This is because before another particle fits in the gap the volume of continuous phase has to increase and then it might be larger per particle within the gap than outside. This again causes a gain in entropy and a loss in free energy (307). Obviously, dissolved polymers with large Dee can lead to so–called bridging flocculation if they are able to strongly interact with colloidal particles. Such processes are carried out commercially in breweries or in wastewater treatment facilities with water-soluble polymers prepared by means of heterophase polymerization in organic media. Combinations of Electrostatic and Steric Stabilization. From the application point of view in heterophase polymerizations, steric stabilization does not require the presence of high molecular weight polymers and can also be realized by low molecular weight nonionic stabilizer based on poly(ethylene glycol) as hydrophilic groups (low molecular weight means a molar mass below 103 –104 g/mol). The simplest case, which indeed is frequently applied, is the joint application of ionic and nonionic low molecular weight surfactants. Advantage can be taken from the fact that the combination of electrostatic and steric stabilization gives on the one hand additional stability at low electrolyte concentrations and on the other hand the dispersion remains sterically stabilized at higher electrolyte concentrations (290). The effect of steric stabilization caused by adsorbed poly(ethylene glycol) chains has been clearly demonstrated by Napper and co-workers (308,309). The remarkable results of these studies are as follows: First, the critical coagulation concentration of sterically stabilized poly(vinyl acetate) dispersions against magnesium sulfate is increased by two orders of magnitude compared with purely electrostatic stabilized dispersions (3.9 × 10 − 1 versus 6.2 × 10 − 3 M MgSO4 at 298 K). Second, for poly(vinyl acetate) latexes with diameters between 30 and 230 nm stabilized with poly(ethylene glycol) moieties (chain length of about 220), the point of incipient flocculation is independent of the particle size. Third, the electrolyte stability of particles stabilized jointly by electrostatic and steric components is governed by the steric contribution. Moreover, the flocculation conditions

Vol. 6

HETEROPHASE POLYMERIZATION

481

are independent of the poly(ethylene glycol) chain length. Fourth, if once flocculated sterically stabilized particles can easily be redispersed if the conditions that have caused flocculation are removed, provided the stabilizer is strongly adsorbed. In other words, dispersions with only poorly anchored steric stabilizers may be flocculated more easily and may not be redispersed. Fifth, increasing the molecular weight of the steric stabilizer can kinetically retard the displacement of weakly anchored stabilizer. Sixth, both nonaqueous and aqueous dispersions stabilized with polymers require for stability that the continuous phase liquid is better than a  solvent for the stabilizing chains. This is the application of Bancroft’s rule and later refinements (cf above) to polymeric stabilizers. Compared with low molecular weight surfactants, polymers possess more degrees of freedom regarding the modification of their properties. There are many possibilities, such as the molecular weight or molecular weight distribution and in the case of random copolymers their composition and the distribution of chemically different groups along the chain. Block copolymers can offer additional possibilities with regard to various architectures such as linear block copolymers, or star-shaped block copolymers, or graft block copolymers. Furthermore, polymers can stabilize dispersions via all stabilization mechanisms (electrostatic and steric stabilization) and offer additional possibilities by a combination of various mechanisms in a single molecule. In the sense of combined stabilization, block copolymers made of one polyelectrolyte and one strong anchor block offer another possibility of so– called electro–steric stabilization absent in the above simply joint application of both ionic and nonionic stabilizers. A great deal has been done during the last few years with regard to both theoretical and experimental investigations of the behavior of polyelectrolytes on surfaces in dependence on various parameters such as concentration, chain length, geometry of the surface, and the ionic strength (Csalt ) in the continuous phase. The use of such block copolymers with strong acid groups (such as sulfonates) as stabilizers in aqueous heterophase polymerization has, compared with poly(carboxylic acid) blocks, the advantage that the stability is practically independent of the pH. Pincus derived, for strong polyelectrolyte tethered to spherical particles and forming a corona of thickness
R, a scaling relation that predicts a dependence of
R ∝ Cα salt , with α = −1/5 (310). This relation predicts an extraordinary electrolyte stability compared with low molecular weight stabilizers, which show an exponential dependence on the ionic strength (cf equations 37a–37c). Indeed, the extraordinary stability against electrolytes was experimentally observed for various types of polyelectrolyte block copolymers such as poly(alkyl methacrylate-b-sulfonated glycidyl methacrylate) (311) but also for poly(ethyl ethylene-b-styrene sulfonate) (312) as stabilizers for polymer dispersions. Moreover, in the latter case a Pincus brush behavior (
R ∝ Cα salt ) was observed but with clear experimental evidence that the exponent α depends on the ratio of corona thickness to particle size. If the ratio 2
R/D is lower than 0.7, α is about −0.1, but at larger corona thickness in relation to particle size, the Pincus behavior (α = −0.2) is observed (313). A shrinkage of the corona with over several orders of magnitude increase in ionic strength was also observed for triblock copolymer particles composed of poly(styrene sulfonate-bN-isopropyl acrylamide-b-styrene), which were directly prepared by heterophase polymerization as described in Reference 48. Furthermore, experimental evidence is given that the shrinkage of the corona is directly related to the polyelectrolyte

482

HETEROPHASE POLYMERIZATION

Vol. 6

behavior of the poly(styrene sulfonate) hairs because in case of nonionic sterically stabilized particles of poly(ethylene glycol-b-N-isopropyl acrylamide-bmethyl methacrylate), the corona thickness remains unchanged under the same conditions.

Optimum Stabilization and Heterophase Polymerization Technique. In summary, effective stabilization of particles during heterophase polymerization requires, besides some theoretical knowledge in physical and polymer chemistry, a lot of experience also and is always dependent on to the particular polymerization process. Aqueous and nonaqueous heterophase polymerizations require not only different stabilization mechanisms but also different stabilizers. Another expression of Bancroft’s rule of thumb for emulsions that are for many heterophase polymerizations the starting point is that the liquid with the highest solubility for the stabilizer forms the continuous phase wherein the other liquid is dispersed. In principle, there are no limitations for the choice of a stabilizer for heterophase polymerizations if only the above general rules are complied—anionic, cationic, amphoteric, or nonionic, monomeric or polymeric, organic or inorganic, linear or branched, comb-like or star-like all, reactive or nonreactive, single or mixed systems, monofunctional or multifunctional, naturally occurring or synthetic—and even colloidal solid spheres have been applied as stabilizers. In the case of colloidal solid spheres, so–called Pickering stabilizers, which have practical meaning for suspension polymerizations, the starting emulsion is a three-phase rather than a two-phase system. The interaction between a Pickering stabilizer and the continuous phase is expressed by means of the contact angle. Droplets of a second liquid are stabilized in a continuous phase if the contact angle between the continuous phase and the sphere is below 90◦ , whereas if the contact angle is above 90◦ the inverse emulsion is preferably formed (314,315). During heterophase polymerizations, as well as during storage of the dispersions, stabilizers are needed to maintain the colloidal state and to guarantee the maximum possible benefit from that state. However, during the final application when the colloidal state is no longer needed, they may lead to problems as they are in general not compatible with the polymer produced. For example, in case of hydrophobic coatings made from aqueous polymer dispersions, low molecular weight stabilizers have a tendency to assemble in hydrophilic domains and finally can cause failure of the coating. One strategy to solve this problem is the use of stabilizers that are fixed in place. This is possible either by the use of polymeric or reactive surfactant that can participate in the polymerization reaction. A summary of recent developments in this field where a great deal has been done during the last decade can be found in Reference 316. Recently, the possibility of tailoring polymeric surfactants has been demonstrated as poly(ethylene glycol-b) (partly sulfonated olefin) have been synthesized and successfully applied as stabilizers in aqueous heterophase polymerization. These polymeric stabilizers combine all kinds of stabilization mechanisms in one molecule, that is, electrostatic, steric, and electrosteric stabilization together with reactivity in radical polymerization (152). The development of stabilizers for heterophase polymerization is an ongoing topic. But, for any new surfactant or stabilizer one faces with the problem that it is only an auxiliary material that has to meet strict cost requirements and has to improve existing solutions in two, frequently contradictory, fields, that is, during polymerization and during application of the solid products.

Vol. 6

HETEROPHASE POLYMERIZATION

483

Copolymerization The majority of polymers prepared by heterophase polymerizations are copolymers. Copolymerization is an effective way to tailor polymer properties for a variety of applications with only a limited construction set of monomers. The hierarchical or modular assembly of polymer dispersions, which goes in descending order with regard to size from dispersion via particles to polymer molecules and finally to monomer units offers almost unlimited possibilities to tailor required properties. For general considerations concerning copolymerization mechanism, the reader is referred to standard textbooks such as References 317–319 (see also BLOCK COPOLYMERS and GRAFT COPOLYMERS). In general, the composition of the copolymer chain is a function of the reactivity ratios and of the instantaneous concentrations of the monomers at the reaction loci. Considering a binary copolymerization, which means that there are two kinds of growing chains, one ended with a radical of monomer 1 (P1 ) and the other ended with a radical of monomer 2 (P2 ), equations 37b as defined below are frequently used. In these equations f 1 and f 2 are the mole fractions of monomers 1 and 2, respectively, at any time in the reaction loci (f 1 + f 2 = 1), F 1 and F 2 are the mole fractions of monomer 1 and 2, respectively, in the copolymer formed at that time (F 1 + F 2 = 1), M 1 and M 2 are the monomer concentrations of both monomers in the reaction loci at any time, r1 and r2 are the reactivity ratios, k11 , k22 , k12 , and k21 are the rate constants for the possible four propagation reactions P1 + M 1 , P2 + M 2 , P1 + M 2 , and P2 + M 1 , respectively.   r1 − 1 f12 + f1 k11 k22   2  F1 =  r1 = r2 = k12 k21 r1 + r2 − 2 f1 + 2 1 − r2 f1 + r2

(41)

F 1 and F 2 are the instantaneous copolymer compositions and depend on the particular conditions at a given time in a given reaction loci in the heterogeneous reaction system. The overall or cumulative copolymer composition is the sum over all times and reaction loci, and may be different from that obtained in a homogeneous reaction system. The application of equations 41 for heterophase polymerization requires the knowledge of the relative concentrations of both monomers in the reaction loci or their partitioning behavior between the different phases. An additional complication arises from the fact that heterophase copolymerization is carried out with monomers that differ markedly in their solubility in the different phases. For instance, in many industrially important emulsion polymerizations, acrylic or methacrylic acid is used as comonomer to impart useful properties to latex particles during polymerization and application. This means that the composition of the monomer mixture in water and in the monomer-swollen polymer particles differs considerably. Thus, the composition of the copolymers formed in both phases differs as well. Wang and Poehlein investigated the composition of cooligomers formed during styrene–acrylic acid batch emulsion copolymerization (318). The cooligomers were isolated by precipitating a dioxane solution of dried latex samples in an excess of water (1/15 v/v). After separating the precipitated copolymer the water phase was dried and the residual solids (a mixture

484

HETEROPHASE POLYMERIZATION

Vol. 6

Table 15. Characterization of Styrene–Acrylic Acid Oligomers Formed during a Batch Emulsion Polymerizationa Seed No No No No Yes Yes Yes Yes

f1 0.861 0.861 0.725 0.725 0.861 0.861 0.725 0.725

Conc., % 9.19% 16.99 16.33 26.32 7.00 21.16 9.11 22.00

F1

MW

n1

n2

0.235 0.372 0.063 0.092 0.334 0.341 0.086 0.064

599 511 1046 1000 787 767 1113 1175

2 2 1 1 3 3 1 1

6 4 14 13 9 9 15 16

at 50◦ C with SDS as stabilizer and potassium peroxodisulfate as initiator 320; 1: styrene; 2: acrylic acid; MW: molecular weight without end groups, in g/mol; n1 , n2 : number of styrene and acrylic acid units per cooligomer.

a Polymerizations

of cooligomers, inorganic salts, and emulsifier) were dissolved in ethanol in order to isolate the cooligomers. They found that the composition of these products depends, besides on the starting monomer ratio, mainly on the presence or absence of seed particles. The data are summarized in Table 15 with the same abbreviations as used for equation 41. The values for f 1 and f 2 refer to the overall concentrations in the reactor and not to the reaction locus, which is the aqueous phase. These data confirm on the one hand the importance of identifying and characterizing the particular reaction loci in heterophase copolymerizations and on the other hand the necessity to carry out future work in order to understand the role of reactions in the continuous phase that are crucial although only a minor part of all monomers is converted there. Despite these problems, it is proved in many examples that the main reaction locus in heterophase polymerization is the monomer-swollen dispersed phase, and if the concentration of monomers in the whole reaction system is allowed to adjust according to the partition coefficients and f 1 /f 2 can be estimated in the reaction locus, then there is good agreement between experimental and calculated compositions using equation 41 and r values derived from homogeneous kinetics. This is the case even for monomers with huge differences in their water solubility, such as styrene and acrylonitrile, as long as monomer droplets are present (321). For further information on emulsion copolymerization, the reader is referred to References 322–324. It is the general case for batch polymerizations that F 1 = f 1 , that is, both monomers are not consumed with equal rates, which means that the less reactive monomer will be accumulated in the remaining monomer mixture as the reaction proceeds. In the case of an azeotropic point in the F 1 –f 1 diagram, the mutual reactivity of both monomers changes, that is, if F 1 > f 1 the monomer 1 is more reactive, whereas if F 1 < f 1 the monomer 2 is more reactive. Thus, the composition of the copolymer is shifted, or a composition drift occurs, during copolymerization in accordance with equation 41. This composition drift cannot be avoided during batch or discontinuous heterophase copolymerizations where the monomers are fed into the reactor only at the beginning. However, the shift in the copolymer composition can be compensated if proper feeding strategies are applied. If the goal is constant copolymer composition throughout the whole reaction, the monomers have to feed into the reactor according to their consumption. In order to do so it is very

Vol. 6

HETEROPHASE POLYMERIZATION

485

helpful to know the instantaneous copolymer composition or the f values in the reaction loci. For instance, the f values can be determined by gas chromatography (321), and the cumulative copolymer composition by nuclear magnetic resonance spectroscopy (325), from which the instantaneous copolymer composition can be deduced. Both methods need much longer times to deliver results than the characteristic growth time of a copolymer molecule, which is on the order of a minute (cf above). Notwithstanding these difficulties a lot of experimental data is a prerequisite for successful modeling. An example is the work of Jean Guillot on emulsion copolymerization (326). Starting from monomer-swollen particles as the main reaction loci he examined two possibilities to calculate f values. The first approach is based on experimental partition coefficients and the second on more general consideration with regard to thermodynamics of polymer solutions. Another way to control heterophase polymerization reactors is based on mathematical modeling in connection with reaction calorimetry, which means in the simplest case monitoring reactor temperature. To combine copolymer composition with reactor temperature control is of special benefit for large-scale, industrial reactors owing due to the exothermic polymerization reaction. A summary of possibilities to control emulsion polymerization reactors can be found in References 327–330. Composition control is by far more complicated in heterophase copolymerizations because of the more compartmentalized nature than in homogeneous copolymerizations. This situation may tempt researchers to the question whether it will ever be possible to model heterophase copolymerizations. Some of the problems toward satisfying models of emulsion copolymers, especially with respect to obtaining reliable kinetic constants and other model parameters, have been discussed in an overview paper by van Herk and German (331). As accurate, detailed models to solve the problem are not available today a possible way out is the use of hierarchical fuzzy-logic model based controllers as described for styrene–butadiene emulsion copolymerization in Reference 332 or neural networks as developed for a batch polymerization of methyl methacrylate (333). From the point of view of materials science, heterophase copolymerization is extremely important because it is the method of choice to tailor application properties of polymer dispersions and that of the products made thereof. Mechanical properties of the polymers such as glass-transition temperature or film formation temperature, elasticity, durability, tensile strength, water uptake, adhesion, light resistance, or solvent resistance, and also colloidal properties such as stability against electrolyte or temperature changes (freeze–thaw stability), surfactant adsorption, or interaction with pigments and other fillers are influenced by copolymer composition in the particle interior and on the surface, respectively. The diversity of monomers allows the preparation of polymers with desired glass-transition temperatures, somewhere in between 150 and −100◦ C by radical heterophase polymerization techniques (cf T g values of homopolymers in Table Table 16). Provided the copolymerization behavior is known, the comonomer mixture to realize a particular glass-transition temperature can be approximated fairly well from the T g ’s of the homopolymers according to equation 42, where the w’s are weight fractions of the monomers in the random copolymer. wi 1 = Tg Tg,i i

(42)

486

HETEROPHASE POLYMERIZATION

Vol. 6

Table 16. Glass-Transition Temperatures (T g ) for Some Homopolymersa Homopolymer Acrylamide Methacrylic acid o-Vinyl toluene Phenyl methacrylate t-Butyl methacrylate Acrylonitrile Acrylic acid Methyl methacrylate Cyclohexyl methacrylate p-Vinyl toluene Styrene i-Propyl methacrylate Vinyl chloride Vinyl pivalatea m-Vinyl toluene Ethyl methacrylate sec-Butyl methacrylate 2-Hydroxyethyl methacrylate t-Butyl acrylate n-Propyl methacrylate n-Hexadecyl acrylate Vinyl acetateb Vinyl acetate 2-Hydroxypropyl methacrylate n-Butyl methacrylate Tetradecyl acrylate Cyclohexyl acrylate Vinyl propionate Vinyl propionateb Methyl acrylate Vinyl versatate Vinyl isopropylether n-Dodecyl acrylateb n-Hexyl methacrylate i-Propyl acrylate Tetradecyl methacrylate i-Butylvinyl ether n-Octyl methacrylate Ethyl acrylate Vinylidene chloride Vinyl 2-ethylhexanoateb n-Propyl acrylatec Ethyl vinyl ether n-Propyl acrylate n-Butyl acrylate Butyl vinyl ether n-Hexyl acrylateb Vinyl caproate

Tg ◦ C 153 130 115–125 110 107 96–106 106 105 104 101 100 81 80 70 72–82 65 60 55 41 35 35 33 32 26 20 20 16 12 10 5–8 −3 −3 −3 −5 −5 −9 −19 −20 −20 to −27 −23 −36 −37 −42 −52 −52 to −57 −55 −57 −60 to −85

Vol. 6

HETEROPHASE POLYMERIZATION

487

Table 16. (Continued) Homopolymer 2-ethylhexyl acrylate n-octyl acrylateb n-dodecyl methacrylate n-decyl methacrylate Ethylene Butadiene Octadecyl methacrylate

Tg ◦ C −60 to −77 −65 −65 −70 −70 to −77 −87 −100

a Data

taken from Ref. 334, except those indicated otherwise. b Ref. 254. c Ref. 33.

Equation 42 is an empirical relation, which is based on the experimental observation that the T g values of a binary copolymer (i = 2) vary monotonically with composition. T g values are important because they determine many application properties of polymers and in relation to the application temperature whether it is rubbery or glassy. In case of film formation and coatings applications of polymer dispersions, the T g values determine at which temperature interdiffusion of polymer molecules across particle boundaries can start. Besides this application as building blocks for the main polymer (which is a random copolymer) comonomers are also applied to act with the aim either to contribute to colloidal properties or to allow subsequent chemical reactions. Monomers frequently employed in radical heterophase polymerizations are put together in Table 17. The monomers listed in Tables 9, 10, 16, and 17 suggest that there is an almost infinite number of combinations that can be applied in radical heterophase copolymerizations. The individual comonomers used in a particular comonomer mixture employed in commercial heterophase polymerization have to fulfill different tasks. The main monomers, ie the monomers of greater molar portion in the copolymers are responsible for the mechanical properties such as glass-transition temperature, film formation temperature, mechanical strength, elasticity, and so on. Functional monomers, which are only a minor part of the monomer mixture, have nevertheless to fulfill important tasks. They are in most cases more lyophilic than the main monomers and hence may contribute to latex stability. The most important examples in this sense, for aqueous continuous phases, are carboxylic acid monomers, with several benefits. These groups can exist in both ionized and un-ionized form; depending on pH, their copolymers contribute in ionized form to colloidal stabilization as electrosteric stabilizers; they can be used to adjust dispersion viscosity in dependence on pH; in case of latex paints they improve adhesion as well as incorporation of pigments; and finally carboxylic acid groups are available as reactive sites for further reactions. Possibilities to carry out reaction either between particles or polymer molecules are crucial for a variety of applications. One of the technically important fields for post-polymerization reactions is cross-linking of the polymer during film formation in order to increase the properties of coatings (259,332–336). With regard to chemical diversity there are virtually no limitations in heterophase polymerization for reactive groups.

488

HETEROPHASE POLYMERIZATION Table 17. Classification of Monomers for Radical Heterophase Polymerizations Acryl monomers Ethyl acrylate Butyl acrylate 2-Ethylhexyl acrylate Methyl methacrylate Butyl methacrylate Diethyl amino ethyl methacrylate Dimethyl amino ethyl methacrylate Acrylonitrile Acrylamide Methacrylamide Vinyl monomers Ethylene Styrene Vinyl chloride Vinyl acetate Vinyl propionate Vinyl 2-ethylhexanoate Vinyl neononanoate Vinyl neodecanoate Vinyl sulfate Diene monomers Butadiene Chloroprene Isoprene Monomers with multiple, nonconjugated double bonds Diallyldimethyl ammonium chloride Diisopropenyl benzene Divinyl benzene Ethyleneglycol dimethacrylate Trimethylolpropane triacrylate Penterythritol teraacrylate Specialty monomers with reactive groups Carboxyl acid Acrylic acid Methacrylic acid Maleic acid Itaconic acid Hydroxyl Hydroxyethyl acrylate Hydroxyethyl methacrylate Hydroxypropyl acrylate Hydroxypropyl methacrylate N-Methylol acrylamide N-Methylol methacrylamide Hydroxymethylated diacetone acrylamide Epoxy Glycidyl acrylate Glycidyl methacrylate

Vol. 6

Vol. 6

HETEROPHASE POLYMERIZATION

489

Table 17. (Continued) Aldehyde Acrolein Alkoxy n-Isobutoxymethyl acrylamide n-Butoxymethyl methacrylamide n-Methoxymethyl acrylamide Allyl n-Methylol carbamate Hydroxymethylated N-formyl-N  -acryloylmethylenediamine Peroxy t-Butyl peroxoacrylate 2-t-Butylperoxyethyl acrylate Silane Vinyl trimethoxy silane γ -Methacryloxypropyl trimethoxysilane β-Methacryloxyethoxytrimethoxysilane

Some of the most important reactive groups used as cross-linking sites in latex paints are listed in Table 18, together with some additional information regarding reaction conditions (cf Table 17 also for functional comonomers).

Particle Morphology The development of inhomogeneous particle morphology is certain to happen in all kinds of heterophase polymerizations, not only with regard to the molecular weight of the polymers formed at different times during the polymerization but also with regard to chemical composition. This radial distribution of properties from the core to the periphery of the particles is a direct consequence of the heterogeneous nature of the reaction system and cannot be avoided. Driving forces are differences in the compatibility caused by enthalpic or entropic effects of various species, which differ either with regard to size or composition. In this sense hydrophilic and hydrophobic groups tend to assemble at the particle water interphase and in the interior of the particles, respectively. Consequently, the question arises to what extent a particular particle morphology represents an equilibrium structure. This is mainly governed by the mobility of the chains and hence, on the particular polymerization conditions such as T g of the polymer, polymerization temperature, degree of swelling, degree of polymerization, and whether cross-linking occurs or not. In other words, the particle morphology is basically determined by competition between thermodynamic and kinetic factors during the particular heterophase polymerization process. Modeling morphology development during particle growth is complicated mainly by difficulties in treating the mobility of the polymer chains inside the particles in a proper way. During homopolymerization the situation is comparatively easy, as on the one hand low molecular weight species try to stay in place in less dense parts of the particle where they can gain more configurational entropy, and on the other hand hydrophilic end groups try to reach the aqueous phase in order to minimize

490

HETEROPHASE POLYMERIZATION

Vol. 6

Table 18. Combinations of Reactive Group for Cross-Linking Reactions in Latexes Functional group O C

C

Examples

Partner

Acrylic acid Methacrylic acid

Heat Acidic Polyvalent metal Acidic Basic Polyhydric alcohols Acidic Heat Methylol compounds Epoxides Amino groups Self-cross-linking Acidic Acidic Other C-methylol Acidic Ester interchange Amino groups

O−

C

C

C

OH

C

Hydroxyalkyl acrylates Hydroxyalkyl methacrylates Carboxylic acid

Catalyst

Glycidyl acrylate

Carboxylic acid

Acidic

Glycidyl methacrylate N-Methylol acrylamide

Amino groups Self-cross-linking

Basic Acidic

N-Methylol methacrylamide

Amino groups

Acidic

Epoxides Self-cross-linking

Acidic Acidic

Amino groups

Acidic

Alkyd resins

Acidic

C O

C

C

N CH2 CH2 OH

O

C

C

N CH2 O C

i-Butoxymethyl acrylamide

O

i-Butoxymethyl methacrylamide

interaction energy. The situation is different for composite particles, where the interaction energy must be additionally considered; this is mainly the interfacial energy between chemically different components. Torza and Mason (339) pioneered this type of investigation when they determined the final morphology of a three-component heterophase emulsion (three mutually immiscible liquids, where mobility effects play no major role) by minimizing the interfacial free energy given by equation 43.
GI = σ1,2 A1,2 + σ1,3 A1,3 + σ2,3 A2,3

(43)

Vol. 6

HETEROPHASE POLYMERIZATION

491

Case 1 S1 = σ2,3 − σ1,2 − σ1,3 S1 ≥ 0 "liquid 1" will spread over "liquid 2" 3

S1 < 0 "liquid 2" and "liquid 3" will form largest interface

1 2

Case 2 S2 = σ1,3 − σ1,2 − σ2,3 S2 ≥ 0 "liquid 2" will spread over "liquid 1"

3 2

S2 < 0 "liquid 3" and "liquid 1" will form largest interface 1 Case 3 S3 = σ1,2 − σ1,3 − σ2,3 S3 ≥ 0 "liquid 3" will spread over "liquid 1"

3

S3 < 0 "liquid 2" and "liquid 1" will form largest interface

2

1

Fig. 31. Illustration of the spreading-coefficients approach to determine the thermodynamically favored morphology by means of the wetting picture, where a drop (gray) sits on a surface (patterned) both surrounded by a third component (white).

The σ ’s are the interfacial tension and the A’s are the interfacial area between the liquids 1, 2, and 3. The conventions were made that phase 1 is chosen so that σ 1,2 > σ 2,3 (which means that S1 < 0), and that liquid 2 forms the continuous phase in which drops of both liquid 1 and liquid 3 are present. The analysis has been carried out by means of spreading coefficients for each liquid (S), which are defined by equation 44 and illustrated in Figure 31 by means of the wellknown phenomenon of wetting. In this sense spreading coefficients greater than zero mean the drop wets the surface and forms a thin film, while for negative S’s the liquid forms a lens, and interfacial contact between all three phases is thermodynamically favored. Spontaneous spreading or wetting means that the contact angle between the corresponding phases is zero and that the spreading molecules adhere to the substrate molecules more strongly than they cohere with one another. S1 = σ2,3 − σ1,2 − σ1,3

S2 = σ1,3 − σ1,2 − σ2,3

S3 = σ1,2 − σ1,3 − σ2,3

(44)

Knowing the spreading coefficients it is now possible to predict thermodynamically favored morphologies. The predicted morphologies in Table 19 are easy to imagine with the information in Figure 31. According to the above convention that liquid 2 forms the continuous phase and S1 > 0 is excluded, there are only three equilibrium morphologies possible, that is, core shell, single drops, and partial engulfing. This analysis revealed that the interfacial tensions between the phases are key factors that govern the final morphology. For practical applications, equilibrium interfacial tensions, for which one nneds to consider mutual

492

HETEROPHASE POLYMERIZATION

Vol. 6

Table 19. Effect of Spreading Coefficients (Interfacial Tensions) on Mixed Particle Morphologya Spreading coefficient S1 < 0 S2 < 0 S3 > 0 S1 < 0 S2 < 0 S3 < 0 S1 < 0 S2 > 0 S3 < 0 S1 < 0 S2 > 0 S3 > 0

Largest interface between liquids

Spreading

Resulting morphology

2 and 3 3 and 1

b

b

3 over 1

2 and 3 3 and 1 2 and 1 2 and 3

b

b

2 over 1

2 and 1 2 and 3

b

b

2 over 1 3 over 1

Core: liquid 1 Shell: liquid 3 Continuous phase: liquid 2 No core shell All liquids to contact Partial engulfing No core shell No contact between 3/1 Single drops in liquid 2 Not realizable System cannot decide No equilibrium

b

b

b b b

b

a Spreading b Either

coefficients as defined by equation 42 and in Figure 31. no interface is formed or no spreading occurs.

solubilities and swellabilities of all components, must be used because initial spreading can be very different from equilibrium spreading. According to equation 43, it is necessary to also consider the interfacial areas with regard to minimizing the free energy. In other words, the final morphology also depends on the relative amounts of the components. Methods to estimate the morphology in dependence on both the interfacial tensions and the volume fractions of the dispersed phases were developed and applied by several groups such as Sundberg (340–344), T¨ornell (345,346), El-Aasser (347,348), and Waters (349) to tailor particle morphology. The results can be summarized as follows. In order to determine the thermodynamically favored morphology, which is that with the lowest free energy, it is necessary to compare
G1 for as many as possible morphologies. To show the principle, two simple cases are considered in context with Figure 31. The continuous phase is formed by liquid 2; lyophobic liquid 3 is present in the form of seed particles, with size D3 ; and lyophilic liquid 1 is added to form composed particles of size Dt as a result of interaction with the seed particles. First, the change in free interfacial energy for the formation of normal core shell particles (
G1 cs ) with the lyophilic liquid 1 forming the shell is given by equation 45. This is the normal or expected core shell morphology because the lyophilic phase forms the shell. The reference or starting free energy state is that for the lyophobic seed particles, which also forms the core in the final particles. 2 2 2
Gcs I = σ1,3 π D3 + σ1,2 π Dt − σ2,3 π D3

(45)

It is advantageous to reduce
G1 cs by  × D3 2 and to ascribe the change in the free energy per unit area of seed particles to interfacial tension. This leads to equation 46.

Vol. 6

HETEROPHASE POLYMERIZATION
Gcs I π D32

=
σ cs = σ1,3 + σ1,2

Dt2 − σ2,3 D32

493

(46)

Second, the reduced free energy change for the formation of inverted core shell particles (α) with the lyophobic liquid 3 and the lyophilic liquid 1 in the shell and the core, respectively, leads to equation 47. Note that the same reference state is assumed as above.
σ ics = σ1,3

D12 D32

+ σ2,3

Dt2 − σ2,3 D32

(47)

Equations 46 and 47 clearly show the influence of the amount of material forming a particular phase on the thermodynamically favored morphology. It is possible to consider simple morphologies other than these and examples can be found in References 340–349. The sizes of the phases can be expressed as volume fractions, leading to equations, that are independent of particle size. The crucial point to use this type of equations in predicting particle morphology is to get reliable data for the various surface tensions. In general, these results mean for aqueous two-stage heterophase polymerizations that the more hydrophilic polymer that has the lower interfacial tension to water tends to form the particle shell (normal core–shell structure). This is absolutely true from the thermodynamic point of view but the influence of the particular polymerization conditions can lead to a variety of morphologies in between such as raspberry-, sandwich-, or acorn-like structures, or occlusions (“salami”-like structures), and even inverted core-shell structures are possible. The influence of the polymerization procedure was demonstrated impressively in a series of experiments carried out by Lee and Rudin (350). The authors showed that if styrene is polymerized onto poly(methyl methacrylate) seed particles under conditions maintaining a high core viscosity or low mobility, that is, with a redox initiator at low temperatures with a slow monomer feed (no swelling), composite particles with an inverted core-shell morphology are formed, as the more hydrophilic poly(methyl methacrylate) should, from a thermodynamic point of view, form the shell. In contrast, if the polymerization is conducted in a way that favors higher mobility in the seed particles (no monomer feed but batchwise monomer addition with swelling of the seed particles before initiation with potassium peroxodisulfate at 60◦ C), the result is a salami-like morphology showing occlusions of polystyrene subparticles and that large parts of the poly(methyl methacrylate) core have been moved at the water interface. The morphology formation in the course of heterophase polymerization in seed particles swollen with a second-stage monomer is extremely interesting, as phase separation and nucleation processes take place either consecutively or in parallel. These processes are influenced strongly by polymerization kinetics and thermodynamics of phase separation and hence, the final morphology is determined by the chain mobility inside the particles, which to some extent is the result of a competition between thermodynamics and polymerization kinetics. Note that, whereas the influence of polymerization kinetics ends sometime or other, thermodynamics acts as long as the particles exist. Indeed some authors describe change in particle morphology

494

HETEROPHASE POLYMERIZATION

Vol. 6

during storage of polymer dispersions (344). An interesting approach to model this behavior is described by Gonzales-Ortiz and Asua in a series of papers (351– 353). They treated the polymerization of monomer 1 in seed particles of polymer 2. Phase separation leads to the formation of clusters. The motion of the clusters inside the particles is determined by a balance between van der Waals and viscous forces. Furthermore, cluster coalescence is allowed and polymerization is assumed to take place in both domains inside the particles. The authors simulated methyl methacrylate polymerization in polystyrene seed particles and found that the final particle morphology depends strongly on the polymerization kinetics and that the slower the rate of polymerization, the closer the final morphology is to the equilibrium structure. The formation of particles with a particular morphology is not only impossible to avoid but is moreover an advantage of heterophase polymerization processes, because by means of controlling morphology formation application properties can be tailored to very specific needs. The following examples of practical importance should illustrate these possibilities for aqueous heterophase polymerizations. In this case it is practically impossible to avoid the concentration of hydrophilic groups at the particle–water interface, which means that it is difficult not to build up particles with a core–shell structure. Hydrophilic groups can arise from initiators or comonomers, or sidereactions such as hydrolysis of ester groups in monomers [acrylic or methacrylic acid ester monomers or vinyl acetate cf (93)] or oxidation reaction in the aqueous phase in connection with initiator decomposition (92,237,354–356). The first example is comonomers with functional groups, which are mostly of polar, hydrophilic character (cf Table 18) and which are employed in concentrations of a few percent relative to the main monomers in order to prepare polymer molecules with reactive sides for post-polymerization modifications such as cross-linking to increase mechanical properties of coatings. It might be of interest to control the distribution of these comonomers over the entire particle volume rather than to have them accumulated at the particle–water interface. There are two general and one specific possibility to do so in dependence on the properties of the hydrophilic comonomer (cf (337,357,358). The first of the general possibilities is the use of main monomers with a higher solubility in water (ethyl acrylate, instead of butyl acrylate, for instance, cf Table 9) leads to the formation of more hydrophobic copolymer with these reactive comonomers and favors a more equal distribution in the particle volume. The second general possibility is the control of the hydrophilic comonomer content in the monomer mixture during feed procedures. The amount of hydrophilic comonomer units at the particle–water interface is the lowest if the whole amount of this comonomer is added at the beginning of the feed process; it is in a middle range if it is equally distributed over the entire feed duration; and it is highest if it is added only at the end of the monomer feed period. The third principle can be additionally applied if the hydrophilicity of the comonomers can be easily controlled, such as in case of carboxylic acid comonomers by changing the pH value. These comonomers are much more hydrophobic in the acid form than in the carboxylated form and hence, at high pH-values they will preferably be accumulated at the particle–water interface. Note that emulsifier adsorption depends strongly on the polarity of the particle-water interface and is the lower with regard to the adsorbed amount the higher the polarity (cf above, Solubility and Solubilization) (255,336,357).

Vol. 6

HETEROPHASE POLYMERIZATION

495

Another example of desired core-shell morphologies are polymer particles composed of regions with different mechanical or optical properties. For instance, impact modifiers for hard and brittle plastic materials are core-shell latex particles with a cross-linked rubbery core and a hard shell. The shell polymer should be identical to the hard material and must be grafted to the rubber core in order to prevent phase separation. Only in this case the modifier particles can effectively absorb tensile stresses and impart additional impact strength. Impact modifiers are used to improve mechanical properties of various commodity polymers such as polystyrene, poly(vinyl chloride), and poly(methyl methacrylate). The problem of impact modification is more complex if the optical properties such as transparency, clarity, and glass-like appearance of the polymeric material should not be changed. This necessitates an adaptation of the refractive indices between the bulk material and the modifier particles. However, the problem is not that easy to solve, as the refractive index is strongly temperature-dependent, especially for particle sizes with good stress-absorbing properties. Making the particles smaller may solve the problem with regard to the temperature dependence of the refractive index but at the expense of a decrease in mechanical properties. The solution of this problem is multilayered particles, where the core and the outer shell are composed of the hard polymer and in between both is the cross-linked rubber layer, which is on the one hand thick enough to impart the required impact modification and on the other hand thin enough to retain unchanged the optical properties over a larger range of temperatures. The morphology can be specifically tailored by multiple arrangements of rubber layers followed by hard polymer layers, where it is unimportant what the core material consists of provided the thickness of the rubber layers is smaller than the wavelength of visible light to ensure only a slight dependence of the refractive index on temperature. However, this example elucidates that it is important to have strategies to prepare well-defined core– ¨ shell transitions. In extension of Sutterlin (237), the following conditions can be formulated in order to get well-defined or less-defined shells (Table 20). Note that the preparation of core-shell morphologies requires a heterophase polymerization technique that allows the application of consecutive monomer addition. A specific core-shell morphology is given when the core is a void (hollow particles), that is, the particle is a porous material. In the dried state these particles

Table 20. Conditions to Prepare Polymer Particles with Shells of Different Properties Sharp core shell transition Polymerization rate  monomer feed rate Avoid swelling of core polymer Cross-linked core More lyophobic core on more lyophilic shell More incompatible core and shell Lower stabilizer concentration More lyophilic emulsifier Lower polymerization temperature More lyophilic initiator Higher initiator concentration

Diffuse core shell transition Monomer feed rate  polymerization rate Allow swelling of core polymer Non-cross-linked core More lyophilic core on more lyophobic shell More compatible core ad shell Higher stabilizer concentration Less lyophilic emulsifier Higher polymerization temperature Less lyophilic initiator Lower initiator concentration

496

HETEROPHASE POLYMERIZATION

Vol. 6

can be considered as polymer gas (mostly air) composites. From the synthetic point of view it is necessary to distinguish between particles with a single hole and really porous particles consisting of a network of pores and channels. Single hollow particles find for instance application as synthetic low density pigments (due to the large difference of refractive index between polymer and air) in the paint and paper industry or as UV protection filters in cosmetic industry, whereas porous particles find application as the stationary phase in separation processes (chromatography). There are several ways to make hollow particles. The first way (cf 337,359) is a multistep process, which starts with the preparation of particles with a high content of carboxylic acid groups. These hydrophilic particles are used as seed in a subsequent polymerization and covered with a more hydrophobic shell made of polymer with T g above the desired service temperature of the composite material. In a next step the pH is increased at a temperature above the T g of the shell, which causes the particles to swell considerably and to fill with water. This structure can be frozen in by decreasing the temperature again and after drying and dewatering, the carboxylic acid rich polymer in the core collapses onto the shell and air diffuses into the particles. Another possibility of making hollow particles, which is interesting from the general viewpoint of the various possibilities offered by heterophase polymerization techniques, requires the exchange of the continuous phase. Core–shell particles composed of a low molecular weight polymer in the core and a slightly cross-linked shell are transferred from a continuous phase in which both the core polymer and the shell polymer are insoluble into a medium that is a good solvent for the core and a swelling agent for the shell. Under these conditions the core polymer diffuses into the continuous phase if the cross-link density of the shell is properly adjusted and by ultrafiltration the core can be drained (360,361). A summary of the work done over the last decades on preparation of voided particles by encapsulation of an organic liquid, which is a nonsolvent for the polymer, can be found in Reference 362. It is a two-stage process, where the first stage is a batch process leading to encapsulation of a mixed liquid drop with a polymer layer and the second stage is stabilization of the drops by further polymerization in the shell with a monomer feed process containing cross-linking monomers. The formation of the polymer shell surrounding the liquid drops requires low molecular weight polymers and hence the polymerization is carried out in the presence of mercaptans as chain-transfer agents. Shell formation takes place at a certain degree of conversion when the nonsolvent (isooctane) monomer (styrene) mixture becomes a progressively poorer solvent for the polymer (polystyrene). Phase separation takes place predominantly at the drop–water interphase owing to energetic reasons, as the polystyrene–water interfacial tension (32.7 mN/m, cf Table 10) is lower than that for isooctane–water [the value for n-octane–water is 51.5 mN/m (363)]. The authors performed some thermodynamic calculations (equilibrium model based on Flory–Huggins theory) that were successful in characterizing the initial and final stages of the process. The process conditions can be adjusted in a way that particles either with a single void or with diffusive microvoids are obtained. A procedure to prepare multihollow particles by a stepwise treatment of polymer particles consisting of carboxylic acid copolymers was developed by Okubo and co-workers. The process of hole formation is similar to the one above based on core-shell particles. Increasing the pH

Vol. 6

HETEROPHASE POLYMERIZATION

497

at elevated temperatures leads to a swelling of the particles with water and the subsequent decrease in pH leads to a collapse of the chains, followed by void formation. Deceasing the temperature below T g causes the structure to freeze, as explained in Reference 364. The authors showed that this principle is virtually universal as they were able to obtain similar morphologies also with copolymer particles containing dimethyl 2-amino ethyl methacrylate instead of methacrylic acid upon the reverse order of pH changes (365). Multihollow particles can also be prepared by so–called double emulsion, that is, an emulsion droplet is itself an emulsion and hence a host for smaller drops. In a water-in-oil-in-water emulsion (w/o/w), drops that are water-in-oil emulsions are dispersed in an aqueous continuous phase. This type of emulsion needs two sets of emulsifiers: hydrophobic ones (smaller HLB values) designed to stabilize the interface of the internal water-in-oil emulsion, and emulsifiers with a higher HLB value to stabilize the external interface of the oil-in-water emulsion. If the oil in the water-in-oil emulsion is a monomer, hollow particles should result after polymerization and drying where the water has acted practically as porogen. To control void concentration and droplet stability is not that easy because stability has to be controlled for both kinds of dispersed droplets. It was found that stability is improved if a urethane macromonomer as viscosity enhancer was incorporated in the monomer mixture (methyl methacrylate and ethylene glycol dimethacrylate as cross-linker). The results clearly show that increasing the amount of urethane macromonomer up to 15 W% relative to the monomer mixture increases both the yield of particles with holes and the number of holes per particle (366). Another way of making porous particles is the application of so–called porogens, which are liquids that are essentially immiscible with the continuous phase and nonsolvent for the polymer but solvents for the monomers. This procedure leads exclusively to porous (multihollow) particles and requires, as in the case of encapsulation of liquid drops, the polymerization to start from an emulsion or highly swollen seed particles. This technique is the application of the principle of macroscopic gel formation to heterogeneous polymerization techniques. Porous polymeric materials, either macroscopic or beads, are prepared by copolymerization in the presence of a cross-linker such as radical polymerization of styrene in presence of divinylbenzene, which is the very “classical system.” Materials of this kind are porous only upon swelling. To control pore size and porosity an inert substance is added, which is only responsible for pore formation and which will be recovered before the material is used. The former procedure leads to materials that are porous only in the swollen state, whereas the latter results in permanent porosity (367). The inert component can either be a nonsolvent for the polymer or a linear polymer, preferably of the same kind as the network polymer but uncross-linked, where its molecular weight controls the pore properties. In general, the pore size is controlled by both the concentration of cross-linker and inert component. The larger the amount of inert component for a given amount of crosslinker, the larger is the pore size, whereas the pore size decreases with increasing cross-linker concentration at given ratio of inert component. Furthermore, the poorer the interaction between porogen and polymer, the easier the phase separation occurs and hence, the larger the pores. These principles are more than 50 years old, but remarkable developments have taken place during the 1990s. The “classical” poly(styrene-co-divinylbenzene) beads are very hydrophobic, but

498

HETEROPHASE POLYMERIZATION

Vol. 6

as separation tasks are more and more shifted toward aqueous systems more hydrophilic beads are needed. One example is permanently porous hydrogel beads made of poly(vinylpyrrolidinone-co-ethylene dimethacrylate), which were prepared by aqueous suspension polymerization (vinyl pyrrolidinone/ethylene dimethacrylate = 1.5 w/w) with a mixture of cyclohexanol and dodecan-1-ol (4:l w/w) as inert diluent (150 w-% relative to the amount of monomers). In order to carry out the heterophase polymerization, the composition of the aqueous phase and the stabilizing system employed were rather unusual. The aqueous phase was a 20 w-% solution of sodium chloride containing 7 w-% magnesium chloride. This high ionic strength is needed to avoid polymerization in the continuous phase. At polymerization temperature, before inserting the organic phase, 1 M sodium hydroxide was added drop wise under stirring, thus forming the stabilizer, which is colloidal magnesium hydroxide. Then the organic phase, containing 2,2-azobisisobutyronitrile, was added and the polymerization allowed to proceed until final conversion. At the end of the polymerization the magnesium hydroxide stabilizer was dissolved by adding HCl (368). Another modern development in the field of polymer beads as stationary phases for separation techniques is their equipment with chiral properties, thus enabling the separation of enantiomers. Examples are described in References 369 and 370 in which the authors do not use suspension polymerization, but rather highly swollen polymer particles, prepared via Ugelstad’s activated swelling method. In brief, the process is as follows: A monodisperse polystyrene seed (prepared by emulsifier-free emulsion polymerization) was activated by swelling with dibutyl phthalate, which was added in the form of an aqueous emulsion stabilized with sodium dodecylsulfate. The monomer mixture containing the chiral monomer, initiator, cross-linker, and toluene was sonicated in an aqueous sodium dodecyl sulfate solution and added to the activated seed particles. After the emulsion drops were sucked up by the activated particles, poly(vinyl alcohol) was added as polymeric stabilizer and the polymerization started by heating. After washing and drying, the particles were treated with toluene to extract the linear polystyrene. The polystyrene acts in this case as both seed polymer and porogen. Another new development during the last decade is the use of supercritical CO2 as porogen, which has the advantage that macroporous polymer beads can be synthesized in aqueous suspension polymerization in the absence of any organic solvents. The results show that the pressure, which is applied to adjust the solvency of the supercritical CO2 together with the stirring speed, controls the pore size (371). All these processes for hollow or porous particle preparation are multistep procedures, but there are also results indicating that hollow particles can be obtained in single-step polymerizations. An interesting possibility, which is also of general importance for heterophase polymerizations, is described in Reference 372. The authors investigated suspension polymerization of styrene in the presence of divinylbenzene as cross-linker, with gelatin and poly(diallyl dimethyl ammonium chloride) as stabilizer, and either 2,2 -azobisisobutyronitrile or dibenzoyl peroxide as initiator. The reaction was started after emulsification at 40◦ C by increasing the temperature to its final value (between 60 and 80◦ C). If the polymerization reaction takes place predominantly in the outer shell of the droplets, then porous particles with a single hole in the center are obtained. The copolymers produced early during the polymerization have a higher degree of cross-linking

Vol. 6

HETEROPHASE POLYMERIZATION

499

due to the reactivity ratios and form a dense shell at the droplet water interface. For geometric reasons this cross-linked shell cannot shrink as much as required by the density change from monomer to polymer. As polymerization proceeds the swelling of the shell with monomer will change gradually in dependence on crosslinking (cf eq. 48) and consequently the core depletes in monomer until finally a hole is formed (cf Shrinkage and Fig. 28). Hole formation was observed at high polymerization rates, that is, at higher temperatures and higher initiator concentrations. Phase separation that takes place inside the confined space of latex particles can also cause the formation of hollow or voided particles. For instance, hollow particle formation, besides other morphologies, was observed in styrene emulsion polymerizations either initiated with poly(ethylene glycol)–azo initiators (312) or in the presence of poly(ethylene glycol) macromonomers (373). Also, ionic groups, if they are present at high concentrations, can cause phase separation and the formation of multihollow morphologies. Such behavior was observed in the case of low molecular weight polystyrene latex particles and is discussed in detail in References 92 and 374. Also, for pure homopolymers particle morphology can be an issue, such as in the case of vinyl chloride polymerization. Poly(vinyl chloride) particles are by no means homogeneous, but rather possess a complicated morphology, which is determined by various parameters. The properties of poly(vinyl chloride) grains prepared by suspension polymerization such as size, density, and porosity are primarily determined by the nature and concentration of the stabilizer and initiator and by the hydrodynamic conditions in the reactor. These grains can have a size up to a few hundred micrometers and are in fact aggregates of a number of droplets with a size of some ten micrometers. Each of these particles is composed of smaller ones with a size of only a few micrometers, which are aggregated into a porous network. These smaller particles consist of still smaller domains (69,156). But these smaller particles also have an internal structure and so on. The smallest substructures have a size of about 10 nm and were identified by of electron microscopy of particles, which were etched with gas molecules (375–377). All investigations regarding the morphology of poly(vinyl chloride) particles, regardless of the polymerization procedure, are in conformity with a hierarchical structure composed of domains with various sizes. The morphology of poly(vinyl chloride) grains is not only determined by the polymerization process but for dispersions prepared by emulsion and microsuspension polymerization also by the conditions during spray-drying, which is used to isolate the polymer (69,156). Control of the grain morphology is very important as for many applications the grains are dispersed in plasticizers that, before further processing can occur, have to swell the grains.

Controlled Aggregation Besides the methods of morphology formation during the polymerization process, the morphology of polymer colloids can also be changed after polymerization by controlled aggregation processes, as already briefly mentioned above. Such processes require close contact between latex particles and hence, they always face the problem of stability control because at the end there should still be a colloidal

500

HETEROPHASE POLYMERIZATION

Vol. 6

polymer dispersion and not a completely coagulated system. From the thermodynamic stability point of view the process of decreasing the interfacial area is always favored. One possibility to carry out such processes is heterocoagulation in aqueous dispersions (water is phase 2 according to Fig. 31) of small particles (phase 3) on oppositely charged larger ones (phase 1) as described by Ottewill and Waters (378–380). Combined stability with ionic and nonionic stabilizers is advantageous in order to ensure stability during the whole process. The application of above principles of minimization of free energy (the reference state is characterized by both kinds of particles being separated) and introducing volume fractions (φ 1 , φ 3 ) in equation 48 defines conditions for engulfment of smaller particles (phase 3) by larger particle (phase 1).

2/3

σ2,3 − σ1,3 1 − φ1 > 2/3 σ1,2 φ

(48)

3

Equation 48 reveals the influence of several variables on the thermodynamics of the engulfment. The process is favored the more hydrophobic the smaller particles the more hydrophilic the larger particles, the less the interfacial tension between both polymers, and the larger the volume fraction of the smaller particles that are forming the core in the composite particles. The applicability of these kinds of predictions as well as of the whole process to prepare composite latex particles has been successfully demonstrated (290,378–380). In a comprehensive study the possibilities of controlled aggregation processes to enlarge the particle size of polymer dispersions were investigated (381–383). Such processes might also be of technical importance as it is possible to produce aggregates with a narrowly distributed size of a few micrometers (up to 15 µm). The aggregation process of negatively charged latexes was induced by the addition of a cationic surfactant under stirring. If the concentration of the cationic surfactant is in a proper range, which causes only partial charge neutralization, the final aggregates are stable and still possess a negative surface potential. This indicates that the surfactant binding to the primary particles is a cooperative process, and besides being controlled by electrostatic interaction is also controlled by hydrophobic interaction of the surfactant tails. The whole aggregation is a three-step process, starting with formation of a gel-like network, which is broken up by shear forces and finally leads to smaller but compacter primary aggregates. These can further agglomerate into larger, secondary aggregates. Their size is mainly determined by the stirring speed and the surface chemistry of the particles (stabilization). Narrow size distribution of the secondary aggregates was only observed for latexes prepared in the presence of acrylic acid, which causes additional stabilization because of the formation of an electrosteric layer. Another kind of controlled aggregation is described in Reference 384. The authors describe the pH-induced coagulation of two rubbery latexes, poly(butyl acrylate) and poly(styrene-co-butadiene), both stabilized by stearic acid. Controlling the decrease in the pH leads to coagulation of the poly(butyl acrylate) particles onto the more hydrophilic poly(styrene-cobutadiene) particles.

Vol. 6

HETEROPHASE POLYMERIZATION

501

Particle Morphology and Application. The morphology of the particles is the main key controlling their properties and application fields. Although in many practical cases the particle morphology, which builds up during the polymerization process, is not known in detail, it is the specific particle morphology that makes a dispersion suited for a particular application. Table 21 gives a few general examples of how certain particle morphologies can contribute to solution of application problems. Hybrid Polymer Dispersions. Particle morphology is not only an issue in the case of heterophase copolymerization or seeded polymerizations but also so– called hybrid systems are arousing increasing interest. Hybrid systems mean in the widest sense the combination of materials arising from different manufacturing processes, such as a combination of more than one generic class of polymers (polyurethanes–acrylics or epoxy–acrylics) or also a combination of heterophase polymerizations with inorganic materials. The former class of hybrid systems plays an important role in the preparation of paints (336). The most widely utilized technique is the application of seed polymers, which are not prepared by heterophase polymerizations but by other techniques, in aqueous emulsion polymerizations. The seed can be either directly water–dispersable owing to hydrophilic functional groups or can be emulsified in water using a combination of surfactants and high shear; if necessary or possible, it can also be dissolved in the monomer mixture prior to emulsification. This kind of sequential polymerization can lead to superior products because totally different properties can be combined on nanometer size scales first in the particles and after application in the final coating. An example of this kind of hybrid heterophase polymerization is given in Reference 385. The colloidal size scale makes heterophase polymerizations a promising tool to prepare hybrid materials with inorganic components. Table 22 lists some recently published examples prepared by polymerization of the particular monomer or monomer mixture in the presence of inorganic material. Composites offer unusual combinations of mechanical properties (stiffness, strength, weight and cost reduction) that are hard to attain from the individual components alone. Heterophase polymerization in this context means not only the encapsulation of inorganic material (as most of the examples given in Table 22) but also the polymerization inside layered structures in the inorganic material, such as silicates (montmorillonites). This so–called intercalation polymerization leads to modification on size scales of a few nanometers and to additional benefits compared to microcomposites or reinforced plastics on millimeter length scales (401). Preparation of nanocomposites also requires specific interactions between the polymers and the inorganic material in order to avoid phase separation (with regard to short–term and long–term storage or use) and to profit form modifications on nanometer length scales. Nonclassical or Novel Developments In the kind of heterophase polymerizations considered so far, except microemulsion polymerization, the stabilizers are used to impart colloidal stability to the disperse phases, which are either monomer droplets or polymer particles in spherical shape. In addition to these “classical” heterophase polymerizations there are also

Table 21. Selection of Core Shell Particle Morphologies for Particular Applications Application

Monomers

Adhesive

Styrene/(meth)acrylates

Pressure-sensitive adhesives Thermoset coatings

(Meth)acrylates/styrene

Paper coating

Various plus reactive methylol groups

White finish

Styrene butadiene plus polar monomers Hydrophobic plus (meth)acrylic acid Methyl methacrylate plus (meth)acrylic acid Hydrophobic monomer plus reactive monomer Hard and soft monomers

Glue

Polyurethanes polyester polyols

Mortar

Vinyl acetate plus various comonomers Various amphiphilic block copolymers

Hydrophilic filler 502

Galenical polymer Diagnostic material

Research products

Morphology Core: high T g Shell: low T g Core: (meth)acrylates Core: T pol > 60◦ C

Benefit Core: cohesion Shell: adhesion Retention of adhesive strength; warp resistance Cross-linking

Shell: T pol < 60◦ C Core: styrene butadiene Shell: acrylic acid, Increased water resistance and acrylamide, acrylonitrile higher gloss Core: hydrophobic Shell: plus acrylic acid Water retention in plastics Core: methyl methacrylate; Shell: (meth)acrylic acid Core: high; T g Shell: reactive, activated

Redispersable Powder

Composite morphology, no statistical copolymer Secondary dispersions in water

High elasticity and low caking

High stability coupling of proteins

Poly(vinyl alcohol) as protective colloid

Crystalline parts give high strength Redispersable powder

Core: hydrophobic Shell: hydrophilic

Redispersable pure solids

Vol. 6

HETEROPHASE POLYMERIZATION

503

Table 22. Polymeric–inorganic Hybrid Composites Prepared with Heterophase Polymerization Techniques Inorganic material Magnetite Pfizers magnetite Na–montmorillonite LiClO4 , fumed silica CaCO3 CaCO3 TiO2 SiO2 , Al2 O3 , TiO2 Carbon black Magnetite Na–montmorillonite SiO2 Montmorillonite Layered silicate

Heterophase polymerization

Reference

Suspension, styrene/divinyl benzene Suspension, monomer mixture Emulsion, methyl methacrylate Poly(ethylene glycol)dimethacrylate Emulsion, styrene Miniemulsion, styrene Miniemulsion, styrene (Mini)emulsion, n-butyl acrylate Miniemulsion, styrene Inverse emulsion, hydroxyethyl methacrylate, methacrylic acid Emulsion, aggregation, styrene Emulsion, 4-vinyl pyridine plus n-butyl acrylate or n-butyl methacrylate Emulsion, acrylonitrile plus butadiene plus styrene Emulsion or suspension, Methyl methacrylate

386 387 388 389 390 391 392 393,394 395 396 397 398 399 400

heterophase polymerizations where the stabilizer determines the morphology and the shape of the disperse phase. To illustrate this, Figures 32 and 33 show illustrations of binary and ternary surfactant phase diagrams, respectively (402). The surfactant concentrations for all the “classical” heterophase polymerizations is in the left part of both phase diagrams, that is, in regions where surfactants form

mi ce lla

T

se ha rp

lam

cub

cub2 hex

0

c

S

1

Fig. 32. Illustration of a binary phase diagram of a surfactant solution. S: weight fraction, surfactant; T: temperature; cub: cubic phase built up of discrete globular micelles; hex: hexagonal phase; cub2: bicontinuous cubic phase; lam: lamellar phase; c: crystalline phase.

504

HETEROPHASE POLYMERIZATION

Vol. 6

D

I

II

C

S

Fig. 33. Schematic drawing of a ternary phase diagram for heterophase polymerizations (part of the ternary phase prism at constant temperature). I: range for “normal” heterophase polymerizations; II: range for heterophase polymerizations in lyotropic phases.

only micelles. Searches have been inspired by the question as to what will happen if polymerizations are carried out in cubic, hexagonal, or lamellar phases, that is, at much higher surfactant concentrations. Basically, there are two possibilities: first, the use of normal monomers as third phase, which swell the surfactant phase (ternary case), and second, the use of reactive, that is polymerizable surfactants (binary case). In either case the interesting question is whether or not it will be possible to replicate the surfactant phase by polymerization. There is no general answer for either case. For the binary case, it was shown in a comprehensive study that for retention of surfactant lyotropic liquid crystalline phases during polymerization, the best position of the polymerizable group is neither at the end of the hydrophobic chain nor within the hydrophilic head group but rather a position where on the one hand chain mobility is reduced and on the other hand head group interactions are not able to prevent polymerization (403). The latter point means that in the case of ammonium surfactants, methacryloyloxyethyl groups attached to the nitrogen behave more readily than allyl groups. Similar results were obtained for (2-methacryloyloxyethyl)dodecyldimethyl ammonium bromide in comparison with 11-(methacryloylundecyl)trimethyl ammonium bromide. Whereas the former polymerizes readily in all phases (cubic, hexagonal, and lamellar) the polymerization of the latter one ceases in the hexagonal phase at about 60%. But both surfactants polymerize under retention of the macroscopic order (404). The situation for the ternary case is more complicated. Here the interesting question is whether or not it might be possible to use the swollen lyotropic phases as templates for the final polymer phase. It has been shown that the very first formed polymer changes the bicontinuous phase structure. Although a direct templating does not seem possible, the kind of lyotropic starting phase influences the final structure (405).

Vol. 6

HETEROPHASE POLYMERIZATION

505

A review describing the state of the art using templates (including lyotropic surfactant phases) to prepare organic polymeric networks has recently been published (406). The monomer solubility in the continuous phase plays in heterophase polymerization, a crucial role as long as diffusion processes through the continuous phase are necessary. This is for instance the case during emulsion polymerization. where the monomer has to diffuse from the reservoir (either droplets or tank in case of batch or feed procedures, respectively) to the particles. However, if the goal is to restrict the polymerization to droplets formed by diminution techniques (for instance suspension or miniemulsion polymerization) a zero solubility in the continuous phase is desirable because Ostwald ripening and bleeding of the drops is impossible (cf above). For some applications very hydrophobic polymers are needed. It is difficult to bring highly hydrophobic monomers such as lauryl methacrylate or stearyl acrylate through the aqueous phase into the particles as the main reaction loci during technically important emulsion polymerization. Recently it was found that β-cyclodextrins, which are cyclic oligosaccharides made up of seven glucopyranose units, can form inclusion complexes with hydrophobic substances. One example is the application of cyclodextrin instead of surfactants for emulsion polymerization of moderately hydrophobic monomers such as butyl methacrylate (407) or emulsion homopolymerization of very hydrophobic monomers such as dodecyl and octadecyl methacrylate, also in the absence of emulsifying agents (408). These hydrophobic monomers led to massive coagulation and incomplete conversion during emulsion polymerization in the presence of common sulfonate surfactants. Another example is the application of cyclodextrins as “transport shuttle” for hydrophobic monomers in emulsion copolymerization (409,410). This application is of interest also in order to increase the hydrophobicity of polymeric materials. Process Models. The demands in industry on increasing product quality and constancy, on most efficient use of reaction space, and on cost reduction require increasing efforts with regard to process and product quality control. Much progress has been made in this field during the last decade. One way to meet industrial requirements is the development and application of process models, which are often applied together with on-line data acquisition systems. In order to completely master a polymerization process a multidisciplinary approach that considers kinetic and polymer property models (rate of polymerization, copolymer composition, particle size, and molecular weight distributions), hardware and software sensors, nonlinear observers for the data interpretation, and easily tuneable, robust controllers (411,412) are needed. The complexity of heterophase polymerization processes makes them difficult to model in detail, especially because the estimation of model parameters is the crucial point. There are basically two ways out of this dilemma in order to fulfill the demand of reactor and product control: First, the development of on-line techniques, that allow a fast determination of important product properties in time frames fast enough to allow remediation when problems occur during the polymerization. Second, the development of mathematical algorithms that provide accurate estimates of the properties of the reactor content. An example of the state of the art, and how to design such an observer, can be found in References 413. The authors developed a continuous–discrete observer that accurately estimates parameters for emulsion copolymerization and

506

HETEROPHASE POLYMERIZATION

Vol. 6

that has been applied for batch copolymerization of styrene and butyl acrylate as well as for semibatch copolymerization of methyl methacrylate and butyl acrylate. The control of batch heterophase polymerizations is a difficult task owing to the gel effect and the absence of lasting stable operation conditions. Frequently in these cases an adapting, self-tuning controller is applied. An example of this for a styrene batch suspension polymerization can be found in Reference 414. The application of a nonlinear predictive controller for a batch vinyl chloride suspension polymerization is given in Reference 415. All these controllers and software techniques require input data on the polymerization reactions. An important application property of a polymer is its molecular weight and hence, its control during polymerization is of special importance. Because of the lack of on-line methods (hardware sensors) a software sensor was developed that calculates the feed streams of monomer and chain-transfer agent required to get a desired molecular weight distribution (416). The control strategy uses a nonlinear-model–based controller and experimental data of both unreacted monomer and chain-transfer agent obtained with gas chromatography. Also, another strategy was developed, which is based on reaction calorimetry data (417). Recently it was shown that it is possible to carry out on-line monitoring of molecular masses during polymerization reactions. To perform this a diluted reactant stream from a polymerization reactor was made to flow through a series of detectors comprising refractive index, ultraviolet absorbance, and time-dependent static light scattering. This combination of detectors allows the determination of the absolute weight-averaged molecular weight of polymers (418). In order to get viscosity data it is also possible to include single-capillary viscometer in the series of detectors (419). The methodology was further improved to allow the finding of instantaneous weight-averaged molecular mass distributions during polymerization reactions and hence, it allows the on-line monitoring of polydispersity (420). Using this technique it was possible to analyze acrylamide polymerization in aqueous solution and to deduce from on-line data important kinetic parameters such as reaction orders and initiator decomposition rate constant (421). Reviews of on-line techniques to get required data can be found in References 422–424). Reaction Calorimetry. The easiest way to follow a polymerization is to monitor the temperature of the reaction mixture or to use reaction calorimetry (cf above). Landau gives an excellent review on reaction calorimetry, its principles and application in chemical research in Reference 425. Table 23 summarizes papers dealing with reaction calorimetry of heterophase polymerizations. Reaction calorimetry gives on-line information with regard to the overall heat released during the reaction. This heat release is directly connected with the monomer conversion (eq. 49). dM dVP = ∝ HF ≡ rp = − dt dt




3 dN

D

dt

2 dD

+ ND

dt

 (49)

Equation 49 can be considered as the basic relation in the case of on-line monitoring of compartmentalized polymerizations. H F is the heat flow, rp is the rate of polymerization, M is the overall monomer concentration, t is the time, V P is the total polymer volume, D is the average particle diameter, and N is the particle

Vol. 6

HETEROPHASE POLYMERIZATION

507

Table 23. Reaction Calorimetry in Heterophase Polymerizations Characteristics Feed policies to maximize productivity, principles, modeling semi-batch emulsion polymerization Adiabatic calorimetry, radical heterophase polymerization, data acquisition, mass balances, parameter estimation Kinetics styrene emulsion polymerization, varying emulsifier and initiator concentration (sodium dodecyl sulfate and potassium peroxodisulfate) Suspension polymerization of styrene, thermokinetic parameters for scale-up Heterophase polymerization, different monomers, initiators and emulsifiers, reaction rate profiles Semi-batch acrylate emulsion polymerization, isoperibolic calorimetry, increase reactor performance Isoperibolic calorimetric investigation of styrene heterophase polymerization with polyurethane emulsifiers

Reference 426 427 428–431 432 433 434 435

concentration. Equation 49 leads immediately to two important conclusions: First, the heat flow practically contains all information about the polymerization process. Second, for a compartmentalized polymerization this information can be extracted only if an additional on-line method is applied. The second method should preferably be a method for the analysis of the particle size or better yet the particle size distribution (PSD). Such a combination is described in References 436 and 437. A small aliquot of dispersion was withdrawn from the reactor with a sampling loop and passed through an on-line density meter (to determine the monomer conversion) and capillary hydrodynamic fractionation apparatus (to measure the PSD). The methods described so far with the terminus technicus “on-line” require the removal of an aliquot of the reaction to carry out the analysis. The only exception is reaction calorimetry, where the temperature is measured either inside the reactor or through its walls. There is another group of techniques available, so–called in-line techniques, which neither require the removal of reaction mixture nor its destruction or dilution. The majority of these techniques is based on various kinds of radiation and hence mainly comprise spectroscopic methods. Owing to the rapid progress in the recent years with regard to optical fiber techniques, and computer techniques these techniques are ready to use inside closed reactors. A summary of papers dealing with in-line techniques to investigate heterophase polymerizations is given in Table 24. These selections suggest that there are a variety of in-line techniques under development or already available to monitor almost all characteristic quantities of heterophase polymerizations. Most of the techniques are only useful to investigate special features in laboratory experiments at low solids content (high dilution) such as transmission measurements, fluorescence spectroscopy, or surface tension. For some others the equipment is very expensive and the gain of additional information frequently does not justify the enormous price and running costs of these methods such as nuclear magnetic resonance or electron spin resonance spectroscopy. In contrast, other methods such as infrared and Raman spectroscopy or ultrasound are very promising.

508

HETEROPHASE POLYMERIZATION

Vol. 6

Table 24. Selection of In-Line Methods to Analyze Heterophase Polymerizations Method Near-infrared spectroscopy (short wavelength 700–1100 nm) Near-infrared spectroscopy (wave-length 1500–1800) Near-infrared spectroscopy

Raman spectroscopy

Raman spectroscopy Raman spectroscopy

Ultrasound Ultrasound

Ultrasound velocity and attenuation Electron spin resonance (ESR)

Time–resolved fluorescence

Solid–state nuclear magnetic resonance spectroscopy Surface tension Combination of turbidity and conductivity, surface tension, Fourier-transformed infrared spectroscopy

Application

Reference

Styrene emulsion polymerization conversion monitoring

438,439

suspension polymerization evaluation of average particle size Emulsion polymerization, detection of monomer droplets upon addition of butyl acrylate and methyl methacrylate individually and in mixtures Emulsion polymerization of methyl methacrylate and copolymerization of styrene/butyl acrylate Styrene suspension polymerization (signal evaluation) Emulsion copolymerization of styrene and butadiene, determination of styrene monomer in the reactor Emulsion homo- and copolymerization principles of applications, conversion Conversion (from velocity) and particle size distribution (from attenuation) for poly(vinyl chloride) and polyterafluoroethylene dispersions All kinds of dispersions, to determine particle size distribution and zeta potential Radical concentrations during emulsion polymerization, lifetime of radicals in reaction compartments Emulsion polymerization of styrene spectroscopy pyrene–labeled macroinitiators; dynamics in polymerizing particles Emulsion polymerization of butyl acrylate in D2 O, conversion Emulsion polymerization of styrene and methyl methacrylate, emulsifier-free Nucleation during emulsion polymerization of styrene, vinyl acetate, and methyl methacrylate; continuous emulsion polymerization of methyl methacrylate, phase transition of polystyrene oligomers with Fourier-transformed infrared spectroscopy

440,441 442

443

444 445

446,447 448

449–453

454–456

457,458

459 93,95 93

Vol. 6

HETEROPHASE POLYMERIZATION

509

From Batch to Continuous Operation. A last point to mention here is that over the last few years technological developments took place in order to transfer classical batch heterophase polymerization processes such as suspension, miniemulsion, and microemulsion polymerization into semibatch or even continuous operation. In the case of suspension (cf Table 6) and miniemulsion polymerization (460,461) this is done in order to increase reactor productivity. The polymerization inside preformed drops requires for continuous operation more than one continuous stirred-tank reactor or plug flow reactors in order to realize residence time distributions, which are as much as possible similar to that of batch reactors. This is necessary in order to achieve an almost equal conversion in all droplets. Continuous stirred-tank reactors possess an exponential residence time distribution whereas the residence time distribution in batch reactors is an orthogonal function. For microemulsion polymerization the goal to develop semibatch procedures is primarily to improve the yield with regard to emulsifier usage as at the end of batch operations polymer particles coexist with empty micelles. The high emulsifier concentration at the start, which was necessary to form the microemulsion, is at the end of the process much too high to ensure particle stability. It is now well established to start monomer feed and if necessary some initiator at the end of the batch period in order to increase the solids content and hence also the emulsifier yield (217,259,270,462). The average particle size stays small (a small increase in average particle is observed, because besides new nucleation and growth seed polymerization also takes place as a result of swelling of the first generation of particles), because there is enough emulsifier available to stabilize newly formed particles. If the fed monomer differs from the first-stage monomer, a special kind of composite particle is obtained that possesses interesting mechanical properties, compared with products from common emulsion polymerization (274) owing to the smaller particle sizes. In final conclusion, heterophase polymerization is a generic term that describes a complex field of polymer science, that spans preparation of both organic and inorganic polymer and mixtures thereof and that is not restricted to a certain polymerization mechanism or technology. It is among the oldest areas in polymer chemistry but it is developing with ever-increasing speed and comprises nowadays more than “classical” emulsion and suspension polymerization techniques. BIBLIOGRAPHY “Emulsion Polymerization” in EPST 1st ed., Vol. 5, pp. 801–859, by Edward W. Duck, The International Synthetic Rubber Company; “Emulsion Polymerization” in EPSE 2nd ed., Vol. 6, pp. 1–51, by Gary W. Poehlein, Georgia Institute of Technology; “Suspension Polymerization” in EPST 1st ed., Vol. 13, pp. 552–571, by Elliot Farber, Tenneco Chemicals, Inc.; “Suspension Polymerization” in EPSE 2nd ed., Vol. 16, pp. 443–473, by Eric A. Grulke, Michigan State University. ´ F. Stepanek, ´ 1. J. Kosek, Z. Grof, A. Novak, and M. Marek, Chem. Eng. Sci. 56, 3951– 3977 (2001). 2. H. L. Frisch, Br. Polym. J. 17, 149–153 (1985). 3. P. Enzel and T. Bein, J. Phys. Chem. 93, 6270–6272 (1989). 4. C.-G. Wu and T. Bein, Science 264, 1757–1759 (1994). 5. A. Matsumoto, T. Kitajima, and K. Tsutsumi, Langmuir 15, 7626–7631 (1999).

510

HETEROPHASE POLYMERIZATION

Vol. 6

6. Y. S. Ko and S. I. Woo, Macromol. Chem. Phys. 202, 739–744 (2001). ¨ 7. S. Spange, A. Graser, A. Huwe, F. Kremer, C. Tintemann, and P. Behrens, Chem. Eur. J. 7, 3722–3728 (2001). 8. M. Biswas and S. S. Ray, Polymer 39, 6423–6428 (1998). 9. U. Meyer, F. Svec, J. M. J. Frechet, C. J. Hawker, and K. Irgum, Macromolecules 33, 7769–7775 (2000). 10. T. Yamaguchi, T. Suzuki, T. Kai, and S.-I. Nakao, J. Membr. Sci. 194, 217–228 (2001). 11. D. C. Blackley, Polymer Lattices, Vol. 1, 2nd ed., Chapman & Hall, London, 1997, pp. 1–2. 12. D. Hosler, S. L. Burkett, and M. J. Tarkanian, Science 284, 1988–1991 (1999). 13. J. Falbe and M. Regitz, eds., R¨ompps Chemie Lexikon, 9th ed., Georg Thieme Verlag Stuttgart, New York, 1993, p. 2945. 14. H.-G. Elias, Große Molekule, ¨ Springer Verlag, Berlin, 1986, pp. 5–7. 15. C. de la Condamine and F. Fresneau, Mem. Acad. R. Sci. 488–510 (1751). 16. M. Faraday, Q. J. Sci. Liter. Art 21, 19–28 (1826). 17. C. G. Williams, Philos. Trans. R. Soc. London 150, 241–255 (1860). 18. M. G. Bouchardat, Compt. Rend. 89, 1117–1120 (1879). 19. J. Berzelius, Jahresbericht Physische Wissenschaften 12, 64 (1833). 20. Ger. Pat. DRP 250,690 (Sep. 6, 1909) (to Farbenfabriken, former F. Bayer & Co.). 21. Ger. Pat. DRP 254,672 (Dec. 12, 1912) (to Farbenfabriken, former F. Bayer & Co.). 22. K. Gottlob, Gummi-Ztg. 33, 508–509 (1919). 23. K. Gottlob, Gummi-Ztg. 33, 534–535 (1919). 24. Ger. Pat. DRP 558,890 (Jan. 9, 1927) (to I. G. Farbenindustrie Akt.-Ges.). 25. U.S. Pat. 1,732,795 (Oct. 22, 1929), R. P. Dinsmore (to Goodyear title a rubber company of Aleion, Ohio). 26. H. Fikentscher, H. Gerrens, and H. Schuller, Angew. Chem. 72, 856–867 (1960). 27. G. S. Whitby and M. Katz, Ind. Eng. Chem. 25, 1204–1211 (1933). 28. G. S. Whitby and M. Katz, Ind. Eng. Chem. 25, 1338–1348 (1933). 29. W. P. Hohenstein and H. Mark, J. Polym. Sci. 1, 127–145 (1946). 30. W. P. Hohenstein and H. Mark, J. Polym. Sci. 1, 549–580 (1946). 31. F. A. Bovey, I. M. Kolthoff, A. I. Medalia, and E. J. Meehan, Emulsion Polymerization, Wiley-Interscience, New York, 1955. 32. D. C. Blackley, Emulsion Polymerisation Theory and Practice, Applied Science Publishers Ltd., London, 1975. 33. R. G. Gilbert, Emulsion Polymerization, A Mechanistic Approach, Academic Press, Inc., London, 1995. 34. R. M. Fitch, Polymer Colloids: A Comprehensive Introduction, Academic Press, Inc., San Diego, 1997. 35. J. M. Asua, ed., Polymeric Dispersions: Principles and Applications, Kluwer Academic Publishers, Dordrecht, 1997. 36. P. A. Lovell, M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997. 37. H. Fikentscher, Angew. Chem. 51, 433 (1938). 38. F. A. Bovey, I. M. Kolthoff, A. I. Medalia, and E. J. Meehan, Emulsion Polymerization, Wiley-Interscience, New York, 1955, pp. 1–25. 39. D. Distler, Waßrige ¨ Polymerdispersionen: Synthese, Eigenschaften, Anwendungen, Wiley-VCH, Weinheim, 1999. 40. J. B. Tuttle, India Rubber World, 488–490 (1923). 41. A. S. Dunn and H. W. Melville, Nature 169, 699–700 (1952). 42. S. F¨orster and M. Antonietti, Adv. Mater. 10, 195–217 (1998). 43. K. Mortensen, Polym. Adv. Technol. 12, 2–22 (2001). 44. A. Milchev, A. Bhattacharya, and K. Binder, Macromolecules 34, 1881–1893 (2001).

Vol. 6

HETEROPHASE POLYMERIZATION

511

45. K. Tauer, Polym. Adv. Technol. 6, 435–440 (1995). 46. L. Rosengarten and K. Tauer, Ber. Bunsen-Ges. Phys. Chem. 100, 734–737 (1996). ¨ 47. K. Tauer, M. Antonietti, L. Rosengarten, and H. Muller, Macromol. Chem. Phys. 199, 897–908 (1998). 48. K. Tauer and V. Khrenov, Macromol. Symp. 179, 27–52 (2002). 49. F. Caruso, R. A. Caruso, and H. M¨ohwald, Science 282, 1111–1114 (1998). 50. E. Donath, G. B. Sukhorukov, F. Carauso, S. A. Davis, and H. M¨ohwald, Angew. Chem, Int. Ed. 37, 2201–2205 (1998). 51. F. Caruso and H. M¨ohwald, J. Am. Chem. Soc. 121, 6039–6046 (1999). 52. F. Caruso, Chem. Eur. J. 6, 413–419 (2000). 53. M. D. C. Topp, I. H. Leunen, P. J. Dijkstra, K. Tauer, C. Schellenberg, and J. Feijen, Macromolecules 33, 4986–4988 (2000). 54. T. K. Bronich, A. V. Kabanov, V. A. Kabanov, K. Yu, and A. Eisenberg, Macromolecules 30, 3519–3525 (1997). 55. E. A. Lysenko, T. K. Bronich, A. Eisenberg, V. A. Kabanov, and A. V. Kabanov, Macromolecules 31, 4511–4515 (1998). 56. E. A. Lysenko, T. K. Bronich, A. Eisenberg, V. A. Kabanov, and A. V. Kabanov, Macromolecules 31, 4516–4519 (1998). 57. A. Zintchenko, H. Dautzenberg, K. Tauer, and V. Khrenov, Langmuir 18, 1386–1393 (2002). ¨ 58. R. Klein and G. Nagele, Curr. Opin. Colloid Interface Sci. 1, 4–10 (1996). 59. T. Okubo and K. Kiriyama, Ber. Bunsen-Ges. Phys. Chem. 100, 849–856 (1996). 60. T. Palberg, J. Phys.: Condens. Matter 11, R323–R360 (1999). 61. C. Sinn, R. Niehuser, E. Overbeck, and T. Palberg, Particle Particle Syst. Charact. 16, 95–101 (1999). 62. U. Gasser, E. R. Weeks, A. Schofield, P. N. Pusey, and D. A. Weitz, Science 292, 258–262 (2001). 63. Y. Xia, B. Gates, Y. Yin, and Y. Lu, Adv. Mater. 12, 693–713 (2000). 64. H. Misawa and S. Juodkazis, Prog. Polym. Sci. 24, 665–697 (1999). 65. M. Engelmann, Klassische Polymere im Aufwind, in (Gendorf) Meeting of the GDCh Fachgruppe Makromolekulare Chemie, TU Dresden, Germany, Mar. 25–26, 1996. 66. D. R. Bassett, in Polymeric Dispersions: Principles and Applications, Elizondo, Spain, June 23–July 5, 1996, NATO Advanced Study Institute. 67. H. Mooibroek, K. Cornish, Appl. Microbiol. Biotechnol. 53, 355–365 (2000). 68. D. Urban, K. Takamura, eds., Polymer Dispersions and Their Industrial Applications, Wiley-VCH, New York, 2002. 69. J. M. Sosa and K. P. Blackmon, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, CRC Press, Inc., Boca Raton, 1996, pp. 8032–8042. 70. P. V. Smallwood, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 17, 2nd ed., John Wiley & Sons, Inc., New York, 1989, pp. 295–297. 71. H. Gerrens, Dechema Monographien 49, 53–73 (1964). 72. Freedonia, Study 1457, Sept. 2001. 73. Hewin International, Inc., in “Water-Soluble Polymers: Potential Applications and Markets”, Trends, Developments, and Forecasts into the 21st Century, Amsterdam, the Netherlands, 1996. 74. Polyscience, Inc., Monomers and Polymers Catalog 2000, p. 70. 75. A. M. van Herk, in J. M. Asua, ed., Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 17–30. 76. C. Hagiopol, Copolymerization: Toward a Systematic Approach, Kluwer Academic/Plenum Publishers, New York, 1999, p. 201. 77. W. V. Smith and R. H. Ewart, J. Chem Phys. 16, 592–599 (1948). 78. W. D. Harkins, J. Chem. Phys. 13, 381–382 (1945).

512 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93.

94. 95. 96.

97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107.

HETEROPHASE POLYMERIZATION

Vol. 6

W. D. Harkins, J. Chem. Phys. 14, 47–48 (1946). W. D. Harkins, J. Am. Chem. Soc. 69, 1428–1444 (1947). W. D. Harkins, J. Polym. Sci. V 217–251 (1950). M. Jung, I. van Casteren, M. J. Monteiro, A. M. van Herk, and A. L. German, Macromolecules 33, 2320–3629 (2000). R. M. Fitch, C. H. Tsai, Polym. Lett. 8, 703–710 (1970). R. M. Fitch, C. H. Tsai, in R. M. Fitch, ed., Polymer Colloids, Plenum Press, New York, 1971, pp. 73–102. R. M. Fitch, Br. Polym. J. 5, 467–483 (1973). R. M. Fitch, Off. Dig., J. Paint Technol. Eng. 37, 32–48 (1965). ¨ K. Tauer and I. Kuhn, Macromolecules 28, 2236–2239 (1995). ¨ K. Tauer, I. Kuhn, in J. M. Asua, ed, Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 49–65. D. Kashiev, Nucleation: Basic Theory with Applications, Butterworth-Heinemann, Oxford, 2000. K. Binder, in P. Haasen, ed., Material Science and Technology, Vol. 5: Phase Transformations in Materials, VCH, Weinheim, 1991, pp. 405–471. ¨ I. Kuhn and K. Tauer, Macromolecules 28, 8122–8128 (1995). ¨ K. Tauer, R. Deckwer, I. Kuhn, and C. Schellenberg, Colloid Polym. Sci. 277, 607–626 (1999). K. Tauer, K. Padtberg, and C. Dessy, in E. S. Daniels, E. D. Sudol, and M. S. El Aasser, eds., Polymer Colloids, Science and Technology of Latex Systems (American Chemical Society Symposium Series 801), American Chemical Society, Washington, D.C., 2001, pp. 93–112. ¨ N. Sutterlin, in R. M. Fitch, ed., Polymer Colloids II, Plenum Press, New York, 1980, pp. 583–597. K. Tauer, C. Dessy, S. Corkery, and K.-D. Bures, Colloid Polym. Sci. 277, 805–811 (1999). P. Mukerjee and K. J. Mysels, Critical Micelle Concentrations of Aqueous Surfactant Systems, Nat. Stand. Ref. Data Ser., (Nat. Bur. Stand. (U.S.), NSRDS-NBS 36) Washington, D.C., 1971, pp. 51–52. ¨ K. Tauer, I. Kuhn, and H. Kaspar, Prog. Colloid Polym. Sci. 101, 30–37 (1996). W. A. G. R. Al-Shahib, and A. S. Dunn, J. Polym. Sci., Polym. Chem. Ed. 16, 677–684 (1978). A. S. Dunn and W. A. Al-Shahib, in R. M. Fitch, ed., Polymer Colloids, Plenum Press, New York, 1971, pp. 619–628. A. S. Dunn, in I. Piirma, ed., Emulsion Polymerization, Academic Press, Inc., New York, 1982, pp. 221–245. K. Tauer, in J. Texter, ed., Reactions and Synthesis in Surfactant Systems, Marcel Dekker, Inc., New York, 2001, pp. 429–453. S. Shen, E. D. Sudol, and M. S. El-Aasser, J. Polym. Sci., Part A: Polym. Chem. 32, 1087–1100 (1994). A. Laaksonen, V. Talanquer, and D. W. Oxtoby, Annu. Rev. Phys. Chem. 46, 489–524 (1995). K. Tauer, C. Schellenberg, and A. Zimmermann, Macromol. Symp. 150, 1–12 (2000). D. Distler, Waßrige ¨ Polymerdispersionen: Synthese, Eigenschaften, Anwendungen, Wiley-VCH, Weinheim, 1999, pp. 16–19. ¨ T. Rager, W. H. Meyer, G. Wegner, K. Mathauer, W. Machtle, W. Schrof, and D. Urban, Macromol. Chem. Phys. 200, 1681–1691 (1999). M. Gerst, H. Schuch, and D. Urban, in J. E. Glass, ed., Associative Polymers in Aquepus Media (ACS Symposium Seres 765), American Chemical Society, Washington, D.C., 2000, pp. 37–51.

Vol. 6

HETEROPHASE POLYMERIZATION

513

108. K. E. J. Barrett, H. R. Thomas, in K. E. J. Barrett, ed., Dispersion Polymerization in Organic Media, John Wiley & Sons, Inc., New York, 1975, pp. 151–108. 109. R. N. Haward, J. Polym. Sci. 4, 273–287 (1949). 110. W. H. Stockmayer, J. Polym. Sci. XXIV, 314–317 (1957). 111. J. T. O’Toole, J. Appl. Polym. Sci. 9, 1291–1297 (1965). 112. J. Ugelstad, P. C. Mork, and J. O. Aasen, J. Polym. Sci., Part A1 5, 2381–2388 (1967). 113. J. Ugelstad and F. K. Hansen, Rubber Chem. Technol. 49, 536–609 (1976). 114. S. Omi, H. Yuyama, and G.-H Ma, J. Polym. Sci., Part A: Polym. Chem. 36, 367–369 (1998). ¨ 115. K. Stahler, Ph.D. dissertation, Max Planck Institute of Colloids and Interfaces, Gohn, and University of Potsdam, Germany, 1994. 116. K. Suzuki and M. Nomura, Macromol. Symp. 179, 1–12 (2002). 117. B. M. E. van der Hoff, Adv. Chem. Ser. 34, 6–31 (1962). 118. M. S. Ryabova, S. M. Sautin, and N. I. Smirnov, Zh. Prikl. Khim. 48, 1577–1582 (1975). 119. H.-Y. Parker, D. G. Westmoreland, and H.-R. Chang, Macromolecules 29, 5119–5127 (1996). ¨ 120. K. Tauer, G. Reinisch, H. Gajewski, and I. Muller, J. Macromol. Sci. Chem. A 28, 431–460 (1991). 121. K. Tauer, H. Kaspar, and M. Antonietti, Colloid Polym. Sci. 278, 814–820 (2000). 122. J. W. Vanderhoff, J. F. Vitkuske, E. B. Bradford, and T. Alfrey Jr., J. Polym. Sci. XX, 225–234 (1956). 123. G. W. Poehlein and J. W. Vanderhoff, J. Polym. Sci., Polym. Chem. Ed. 11, 447–452 (1973). 124. J. Ugelstad, H. Flogstad, F. K. Hansen, and T. Ellingsen, J. Polym. Sci., Polym. Symp. 42, 473–485 (1973). 125. P. C. Mork, J. Polym. Sci. Part A: Polym. Chem. 33, 2305–2316 (1995). 126. Ref. 33, pp. 236–244. 127. R. A. Backhaus, Isr. J. Botany 34, 283–293 (1985). 128. G. Q. Chen, G. Zhang, S. J. Park, and S. Y. Lee, Appl. Microbiol. Biotechnol. 57, 50–55 (2001). 129. F. Tiarks, K. Landfester, and M. Antonietti, J. Polym. Sci., Part A: Polym. Chem. 39, 2510–2524 (2001). 130. W. St¨ober, A. Fink, and E. Bohn, J. Colloid Interface Sci. 26, 62–69 (1968). 131. S. Slomkowski, S. Sosnowski, and M. Gadzinowski, Colloid Surf., A 153, 111–118 (1999). 132. D. D. Hile and M. V. Pishko, J. Polym. Sci., Part A: Polym. Chem. 39, 562–570 (2001). 133. U. S. Pat. 2,891,920 (1959), J. F. Hyde, and J. R. Wehrly (to Dow Corning Corp.). 134. D. R. Weyenberg, D. E. Findlay, J. Cekada Jr., and A. E. Bey, J. Polym. Sci., Part C 27, 27–34 (1969). 135. M. A. Awan, V. L. Dimonie, and M. S. El-Aasser, J. Polym. Sci., Part A: Polym. Chem. 34, 2633–2649 (1996). 136. R. E. Rinehart, J. Polym. Sci., Part C 27, 7–25 (1969). 137. A. Tomov, J.-P. Broyer, and R. Spitz, Macromol. Symp. 150, 53–58 (2000). 138. B. Manders, L. Sciandrone, G. Hauck, and M. O. Kristen, Angew. Chem. Int. Ed. 40, 4006–4007 (2001). 139. K. Landfester and H.-P. Hentze, in J. Texter, ed., Reactions and Synthesis in Surfactant Systems, Marcel Dekker, Inc., New York, 2001, pp. 471–499. 140. J. Qiu, B. Charleux, and K. Matyjaszewski, Prog. Polym. Sci. 26, 2083–2134 (2001). 141. M. F. Cunningham, Prog. Polym. Sci. 27, 1039–1067 (2002). 142. A. Butte, G. Storti, and M. Morbidelli, Macromolecules 34, 5885–5896 (2001). 143. J. G. Tsavalas, F. J. Schork, H. de Brouwer, and M. J. Monteiro, Macromolecules 34, 3938–3946 (2001).

514

HETEROPHASE POLYMERIZATION

Vol. 6

144. G. Moad, J. Chiefari, Y. K. Chomg, J. Krstina, R. T. A. Mayadunne, A. Postma, E. Rizzardo, and S. H. Thang, Polym. Int. 49, 993–1001 (2000). 145. S. Jousset, J. Qiu, and K. Matyjaszewski, Macromolecules 34, 6641–6648 (2001). 146. J. Qiu, T. Pintauer, S. G. Gaynor, K. Matyjaszewski, B. Charleux, and J. P. Vairon, Macromolecules 33, 7310–7320 (2000). 147. G. Scmidt-Naake, M. Drache, and C. Taube, Angew. Makromol. Chem. 265, 62–68 (1999). 148. A. E. Kanakov, A. G. Kronman, Yu. D. Semchikov, G. L. Groshev, E. F. Sitnikova, and T. V. Vodop’yanova, Russ. J. Appl. Chem. 70, 1308–1312 (1997). 149. J. M. Pinto and R. Giudici, Chem. Eng. Sci. 56, 1021–1028 (2001). 150. G. D. Longeway and D. E. Witenhafer, J. Vinyl Additive Techn. 6, 100–103 (2000). 151. Ref. 31, pp. 271–324. 152. K. Tauer, A. Zimmermann, and H. Schlaad, Macromol. Chem. Phys. 203, 319–327 (2002). 153. G. W. Poehlein, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 6, 2nd ed., John Wiley & Sons, Inc., New York, 1986, pp. 1–51. 154. G. W. Poehlein, in J. M. Asua, ed., Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 305–331. 155. E. A. Grulke, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 16, 2nd ed., John Wiley & Sons, Inc., New York, 1989, pp. 443–473. 156. P. V. Smallwood and M. J. Bunten, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 17 2nd ed., John Wiley & Sons, Inc., New York, 1989, pp. 295–376. 157. D. R. Bassett and K. L. Hoy, in D. R. Bassett, and A. E. Hamielec, eds., Emulsion Polymers and Emulsion Polymerization (ACS Symposium Series 165), American Chemical Society, Washington D.C., 1981, pp. 371–387. 158. B. Jacobi, Angew. Chem. 64, 539–543 (1952). 159. G. W. Poehlein and D. J. Dougherty, Rubber Chem. Technol. 50, 601–638 (1977). 160. C. Kiparissides, J. F. McGregor, and A. E. Hamielec, Can. J. Chem. Eng. 58, 48–55 (1980). 161. A. W. DeGraff and G. W. Poehlein, J. Polym. Sci., Part A-2 9, 1955–1976 (1971). 162. G. W. Poehlein, Polym. Intern. 30, 243–251 (1993). 163. F. J. Schork, in C. McGreavy, ed., Polymer Reactor Engineering, Chapman and Hall, London, 1994, pp. 148–202. 164. N. Ohmura, K. Kitamoto, T. Yano, and K. Kataoka, Ind. Eng. Res. 40, 5177–5183 (2001). 165. H. N. Hipps, G. W. Poehlein, and F. J. Schork, Polym. React. Eng. 9, 135–160 (2001). 166. H. Gerrens, Fortschr. Hochpol. Forsch. 1, 234–328 (1959). 167. K. Arai, M. Arai, S. Iwasaki, and S. Saito, J. Polym. Sci., Polym. Chem. Ed. 19, 1203– 1215 (1981). 168. M. Nomura, M. Harada, W. Eguchi, and S. Nagata, J. Appl. Polym. Sci. 16, 835–84 (1972). 169. M. F. Kemmere, J. Meuldijk, A. A. Drinkenburg, and A. L. German, J. Appl. Polym. Sci. 74, 3225–3241 (1999). 170. M. Zubitur and J. M. Asua, J. Appl. Polym. Sci. 80, 841–851 (2001). 171. M. F. Kemmere, J. Meuldijk, A. A. Drinkenburg, and A. L. German, J. Appl. Polym. Sci. 79, 944–957 (2001). 172. S. Oprea and T. Dodita, Prog. Org. Coat. 42, 192–201 (2001). 173. D. A. Paquet Jr. and W. H. Ray, AIChE J. 40, 73–87 (1994). 174. C. Sayer and R. Giudici, Braz. J. Chem. Eng. 19, 89–103 (2002). 175. C. Abad, J. C. de la Cal, and J. M. Asua, in J. M. Asua, ed., Polymeric Dispersions: Principles and Applications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 333– 347.

Vol. 6

HETEROPHASE POLYMERIZATION

515

176. P. J. Dowding, J. W. Goodwin, and B. Vincent, Colloid Polym. Sci. 278, 346–351 (2000). 177. M. Apostel, W. Pauer, H.-U. Moritz, J. Kermesk¨otter, and K.-D. Hungenberg, Chimia 55, 229–230 (2001). 178. X. Ni, D. C. Bennett, K. C. Symes, and B. D. Grey, J. Appl. Polym. Sci. 76, 1669–1676 (2000). 179. X. Wei, H. Takahashi, S. Sato, and M. Nomura, J. Appl. Polym. Sci. 80, 1931–1942 (2001). 180. K. Kataoka, N. Ohmura, M. Kouzu, Y. Simamura, and M. Okubo, Chem. Eng. Sci. 50, 1409–1416 (1995). 181. F. H. A. M. van den Bommen, J. Meuldijk, and D. Thoenes, Chem. Eng. Sci. 54, 3283– 3290 (1999). 182. J. W. Vanderhoff, M. S. El-Aasser, F. J. Micale, E. D. Sudol, C. M. Tseng, A. Silwanowizc, D. M. Kornfeld, and F. A. Vicente, J. Dispersion Sci., Technol. 5, 231–246 (1984). 183. G. Vidotto, A. Crosato-Arnaldi, and G. Talamini, Makromol. Chem. 122, 91–104 (1969). 184. M. Carenza and G. Palma, Eur. Polym. J. 21, 41–47 (1985). 185. A. Guyot, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 9, CRC Press, Boca Raton, 1996, pp. 7228–7237. 186. H. D. H. St¨over, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 9, CRC Press, Boca Raton, 1996, pp. 7237–7238. 187. H. Bieringer, K. Flatau, and D. Reese, Angew. Makromol. Chem. 123/124, 307–334 (1984). 188. H. G. Yuan, G. Kalfas, and W. H. Ray, J. Macromol. Sci.—Rev. Macromol. Chem. Phys. C 31, 215–299 (1991). 189. J. E. Puig and E. Mendizabal, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 8215–8220. 190. E. Vivaldo-Lima, P. E. Wood, A. E. Hamielec, and A. Penlidis, Ind. Eng. Chem. 36, 939–965 (1997). 191. P. J. Dowding and B. Vincent, Colloids Surf., A 161, 259–269 (2000). 192. D. J. Walbridge, in G. C. Eastmond, A. Ledwith, S. Russo and P. Sigwalt, eds., Chain Polymerization II, Vol. 4, Pergamon Press, Oxford, 1989, Chapt. “15”, pp. 243–260. 193. H. D. H. St¨over, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 3, CRC Press, Boca Raton, 1996, pp. 1900–1905. 194. J. L. Cawse, in P. A. Lovell, and M. S. El-Aasser eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997, pp. 743–761. 195. W.-H. Li, K. Li, and H. D. H. St¨over, J. Polym. Sci., Part A: Polym. Chem. 37, 2295–2303 (1999). 196. K. E. J. Barrett, ed., Dispersion Polymerization in Organic Media, John Wiley & Sons, Inc., London, 1975. 197. M. F. Cunningham, Polym. React. Eng. 7, 231–257 (1999). 198. P. C. Mork, B. Saethre, and J. Ugelstad, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 8559–8566. 199. M. J. Bunten, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 17, 2nd ed., John Wiley & Sons, Inc., New York, 1989, pp. 329–376. 200. G. Markert, Angew. Makromol. Chem. 123/124, 285–306 (1998). 201. Q. Wang, S. Fu, and T. Yu, Prog. Polym. Sci. 19, 703–753 (1994). 202. M. S. El-Aasser, E. D. Sudol, in P. A. Lovell and M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997, pp. 37–58. 203. S. Sajjadi and B. W. Brooks, Chem. Eng. Sci. 55, 4757–4781 (2000). 204. H. Kawaguchi, in T. Sugimoto, ed., Fine Particles: Synthesis, Characterization and Mechanism of Growth, Marcel Dekker Inc., New York, 2000, pp. 592–608. 205. B. Saethre, P. C. Mork, and J. Ugelstad, J. Polym. Sci., Part A: Polym. Chem. 33, 2951–2959 (1995).

516

HETEROPHASE POLYMERIZATION

Vol. 6

206. I. Aizpurua, J. Amalvy, M. J. Barandiaran, J. C. de La Cal, and J. M. Asua, Macromol. Symp. 111, 121–131 (1996). 207. E. D. Sudol, M. S. El-Aasser, in P. A. Lovell, and M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997, pp. 699–722. 208. Y. Luo, J. Tsavala, and F. J. Schork, Macromolecules 34, 5501–5507 (2001). 209. I. Capek and C. S. Chevn, Adv. Polym. Sci. 155, 101–165 (2001). 210. M. Antonietti and K. Landfester Prog. Polym. Sci. 27, 689–757 (2002). 211. J. M. Asua Prog. Polym. Sci. 27, 1283–1346 (2002). 212. F. Candau, in J. I. Kroschwitz, ed., Encyclopedia Polymer Science and Engineering, Vol. 9, 2nd ed., John Wiley & Sons, Inc., New York, 1989, pp. 718–724. 213. I. Capek and P. Potisk, Eur. Polym. J. 31, 1269–1277 (1995). 214. J. Santhanalakshmi, and K. Anandhi, J. Appl. Polym. Sci. 60, 293–304 (1996). 215. L. M. Gan and C. H. Chew, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 4321–4331. 216. M. Antonietti, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 4331–4333. 217. M. Rabelero, M. Zacarias, E. Mendizabal, J. E. Puig, J. M. Dominguez, and I. Katime, Polym. Bull. 38, 695–700 (1997). 218. N. Girard, Th. F. Tadros, and A. I. Bailey, Colloid Polym. Sci. 276, 999–1009 (1998). 219. F. Candau, M. Pabon, and J.-Y. Anquetil, Colloids Surf., A 153, 47–59 (1999). 220. N. Sosa, R. D. Peralta, R. G. Lopez, L. F. Ramos, I. Katime, C. Cesteros, E. Mendizabal, and J. E. Puig, Polymer 42, 6923–6928 (2001). 221. C. C. Co, P. Cotts, S. Burauner, R. de Vries, and E. W. Kaler, Macromolecules 34, 3245–3254 (2001). 222. F. Reynolds, K. Jun, and Y. Li, Macromolecules 34, 165–170 (2001). 223. D. Hunkeler, F. Candau, C. Pichot, A. E. Hamielec, T. Y. Xie, J. Barton, V. Vaskova, J. Guillot, M. V. Dimonie, and K.-H. Reichert, Adv. Polym. Sci. 112, 115–133 (1994). 224. V. F. Kurenkover and V. A. Myagchenkov, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 47–54. 225. D. J. Hunkeler and J. Hernandez-Barajas, in J. C. Salamone, ed., Polymeric Materials Encyclopedia, Vol. 10, CRC Press, Boca Raton, 1996, pp. 3322–3333. 226. F. Candau, in P. A. Lovell and M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997, pp. 723–741. 227. Ref. 34, pp. 64–68. 228. H. Edelhauser and J. W. Breitenbach, J. Polym. Sci. XXXV, 423–428 (1959). 229. J. W. Breitenbach and H. Edelhauser, Makromol. Chem. 44–46, 196–211 (1961). 230. I. Capek, Adv. Colloid Interface Sci. 91, 295–334 (2001). ´ 231. J. E. Puig, V. H. P´erez-Luna, M. P´erez-Gonzalez, E. R. Mac´ıac, B. E. Rodr´ıguez, and E. W. Kaler, Colloid Polym. Sci. 271, 114–123 (1993). 232. J. M. Asua, V. S. Rodriguez, and E. D. Sudol, M. S. El-Aasser, J. Polym. Sci., Part A: Polym. Chem. 27, 3569–3587 (1989). 233. S. S. Ivanchev, N. I. Solomko, V. V. Konovalnko, and V. A. Jurshenko, Dokl. Akad. Nauk SSSR 191, 593–595 (1970). 234. V. N. Pavljuchenko, N. N. Lesnikova, N. A. Birdina, G. A. Otradina, and A. P. Agapitov, Vysokomol. Soedin. 35, 240–243 (1993). 235. K. Tauer and S. Kosmella, Polym. Int. 30, 253–258 (1993). 236. M. Zerfa and B. W. Brooks, DECHEMA Monogr. 131, 249–257 (1995). 237. I. M. Kolthoff and I. K. Miller, J. Am. Chem. Soc. 73, 3055–3059 (1951). 238. G. Odian, Principles of Polymerization, Wiley-Interscience, New York, 1991, p. 215. 239. K. Tauer, in K. Holmberg, ed., Handbook of Applied Surface and Colloid Chemistry, John Wiley & Sons, Inc., New York, 2001, pp. 8175–200.

Vol. 6 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251.

252. 253. 254.

255. 256. 257. 258.

259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274.

HETEROPHASE POLYMERIZATION

517

W. I. Higuchi and J. Misra, J. Pharm. Sci. 51, 459–466 (1962). A. S. Kabalnov, A. V. Pertzov, and E. D. Shchukin, Colloids Surf. 24, 19–32 (1987). J. G. Weers, and R. A. Arlauskas, Langmuir 11, 474–477 (1995). K. S. Birdi, in K. S. Birdi, ed., Handbook of Surface and Colloid Chemistry, CRC Press, Boca Raton, 1997, p. 100. F. Chu, C. Graillat, and A. Guyot, J. Appl. Polym. Sci. 70, 2667–2677 (1998). R. Becker, A. Hashemzadeh, and H. Zecha, Macromol. Symp. 151, 567–573 (2000). C. Tang and F. Chu, J. Appl. Polym. Sci. 82, 2352–2356 (2001). J. C. Lagarias, Discrete Comput. Geom. 27, 165–193 (2002). D. Horak, Acta Polym. 47, 20–28 (1996). M. Okubo, M. Shiozaki, M. Tsujihiro, and Y. Tsukuda, Colloid Polym. Sci. 269, 222– 226 (1991). M. Okubo and T. Nakagawa, Colloid Polym. Sci. 270, 853–858 (1992). J. Ugelstad, Mfutakamba, P. C. Mork, T. Ellingsen, A. Berge, R. Schmid, L. Holm, A. Jorgendahl, F. K. Hansen, and K. Nustad, J. Polym. Sci., Polym. Symp. 72, 225–240 (1985). J. Ugelstad, P. Stenstad, L. Kilaas, W. S. Prestvik, A. Rian, K. Nustad, R. Herje, and A. Berge, Macromol. Symp. 101, 491–500 (1996). J. W. Vanderhoff, Chem. Eng. Sci. 48, 203–217 (1993). G. A. Vandezande, O. W. Smith, and D. R. Bassett, in P. A. Lovell and M. S. ElAasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., Chichester, 1997, pp. 563–587. B. R. Vijayendran, in R. M. Fitch, ed., Polymer Colloids II, Plenum Press, New York, 1980, pp. 209–224. E. R. Blout and H. Mark, Monomers, Interscience Publishers, Inc., New York, 1951. Ref. 31, pp. 156–158. J. J. G. S. van Es, J. M. Geurts, J. M. Verstegen, and A. L. German, in J. M. Asua, ed., Polymeric Dispersions: Principles and Applications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 451–462. A. S. de Buruaga, J. C. de la Cal, and J. M. Asua, Polymer 41, 1269–1276 (2000). A. L. Warg, W. J. Roberts, in E. R. Blout, and H. Mark, eds., Monomers, Interscience Publishers, Inc., New York, 1951, p. 51. W. H. Lane, Ind. Eng. Chem. 18, 295–296 (1946). F. Wenzel and K. Lehmann, Lexikon der technischen Chemie, Vol. 16, 4th ed., VCH, Weinheim, 1978, p. 610. M. E. L. McBain and E. Hutchinson, Solubilization and Related Phenomena, Academic Press, Inc., New York, 1955. Ref. 31, pp. 114–163. P. C. Mork and J. Ugelstad, Eur. Polym. J. 5, 483–492 (1969). ¨ Grundstoffindustrie, Tabellenbuch Chemie, 7th ed., VEB Deutscher Verlag fur Leipzig, 1975, p. 298. L. Lopez de Arbina, L. M. Gugliotta, M. J. Barandiaran, and J. M. Asua, Polymer 39, 4047–4055 (1998). M. F. Cunningham, K. Geramita, and J. W. Ma, Polymer 41, 5385–5392 (2000). M. Morton, S. Kaizerman, and M. W. Altier, J. Colloid Sci. 9, 300–312 (1954). O. Braun, J. Selband, and F. Candau, Polymer 42, 8499–8510 (2001). W. Burchard, Prog. Colloid Polym. Sci. 78, 63–67 (1988). M. Antonietti, H. Kaspar and K. Tauer, Langmuir 12, 6211–6217 (1999). J. O’Haver, B. Grady, J. H. Harvell, and E. A. O’Rear, in J. Texter, ed., Reactions and Synthesis in Surfactant Systems, Marcel Dekker, Inc., New York, 2001, pp. 537–545. A. Aguiar, S. Gonzalez-Villegas, M. Rabelero, E. Mendizabal, J. E. Puig, J. M. Dominguez and I. Katime, Macromolecules 32, 6767–6771 (1999).

518

HETEROPHASE POLYMERIZATION

Vol. 6

275. S. S. Medvedev, in IUPAC International Symposium on Macromolecular Chemistry “Kinetics and Mechanism of Polyreactions,” Budapest, Hungary, 1969, Akad´emiai Kiad´o, Budapest 1971, pp. 39–63. 276. M. R. Grancio and D. J. Williams, J. Polym. Sci., Part A1 8, 2617–2629 (1970). 277. D. H. Napper, J. Polym. Sci., Part A1 9, 2089–2091 (1971). 278. P. Keusch and D. J. Williams, J. Polym. Sci., Polym. Chem. Ed. 11, 143–162 (1973). 279. J. L. Gardon, J. Polym. Sci., Polym. Chem. Ed. 11, 241–251 (1973). 280. R. Tognacci, G. Storti, and A. Bertucco, J. Appl. Polym. Sci. 62, 2341–2349 (1996). 281. A. L. German, B. G. Manders, H. F. Zirkzee, B. Klumperman, and A. M. van Herk, Macromol. Symp. 111, 107–120 (1996). ¨ 282. H. Bodug¨oz and O. Guven, J. Appl. Polym. Sci. 83, 349–356 (2002). 283. J. Ugelstad, P. C. Mork, K. H. Kaggerud, T. Ellingsenand, and A. Berge, Adv. Colloid Interface Sci. 13, 101–140 (1980). 284. R. Defay, I. Prigogine, A. Bellemans, and D. H. Everett, Surface Tension and Adsorption, Longmans, Green & Co. Ltd., London, 1966, 217–222. 285. J. Ugelstad, P. Stenstad, L. Kilaas, W. S. Prestvik, A. Rian, K. Nustad, R. Herje, and A. Berge, Macromol. Symp. 101, 491–500 (1996). 286. T. E. Daubert, R. P. Danner, H. M. Sibul, and C. C. Stebbins, Physical and Thermodynamic Properties of Pure Chemicals Data Compilation, Part 1-5, Taylor and Francis, Washington, D.C., 1998. 287. J. Brandrup and E. H. Immergut, Polymer Handbook, 3rd ed., John Wiley & Sons, Inc., New York, 1989, Section V. 288. D. H. Napper, Polymeric Stabilization of Colloidal Dispersions, Academic Press, Inc., London, 1983. 289. D. W. J. Osmond, F. A. Waite, in K. E. J. Barrett, ed., Dispersion Polymerization in Organic Media, John Wiley & Sons, Inc., London, 1975, pp. 9–44. 290. R. H. Ottewill, in P. A. Lovell, and M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, John Wiley & Sons, Inc., New York, 1997, pp. 59–121. 291. I. Piirma, Polymeric Surfactants (Surfactant Science Series 42), Marcel Dekker, Inc., New York, pp. 1–16. 292. W. Bancroft, J. Phys. Chem. 16, 177–233 (1912). 293. W. C. Griffin, J. Soc. Cosmetic Chemists 1, 311–326 (1949). 294. A. Beerbower and M. W. Hill, Detergents & Emulsifiers 1971 Annual, pp. 223–236. 295. C. Holtzscherer and F. Candau, Colloid Surf. 29, 411–423 (1988). 296. H. C. Hamaker, Rec. Trav. Chim. 55, 1015–1026 (1936). 297. H. C. Hamaker, Rec. Trav. Chim. 56, 3–25 (1937). 298. H. C. Hamaker, Rec. Trav. Chim. 56, 727–748 (1937). 299. H.-J. Jacobasch and K.-H. Freitag, Acta Polym. 30, 453–469 (1979). 300. D. F. Evans and H. Wennerstr¨om, The Colloidal Domain, VCH, Weinheim, 1994, p. 221. 301. F. W. Fowkes, Ind. Eng. Chem. 56, 40–52 (1964). 302. D. B. Hough and L. R. White, Adv. Colloid Interface Sci. 14, 3–41 (1980). 303. G. Jung, Z. Elektrochem. Angew. Phys. Chem. 35, 1–2 (1929). 304. A. Kuhn, W¨orterbuch der Kolloidchemie, Verlag von Theodor Steinkopff, Dresden, 1932, pp. 25–26, 100. 305. D. F. Evans and H. Wennerstr¨om, The Colloidal Domain, VCH, Weinheim, 1994, pp. 110–127, and pp. 333–338. 306. D. H. Napper, Polymeric Stabilization of Colloidal Dispersions, Academic Press, Inc., London, 1983, pp. 144–159. ¨ 307. C. Bechinger, H.-H. von Grunberg, and P. Leiderer, Phys. Bl. 55, 53–56 (1999). 308. D. H. Napper and A. Netschey, J. Colloid Interface Sci. 37, 528–535 (1971).

Vol. 6

HETEROPHASE POLYMERIZATION

519

309. D. H. Napper, J. Colloid Interface Sci. 32, 106–114 (1970). 310. P. Pincus, Macromolecules 24, 2912–2919 (1991). 311. L. Leemans, R. Fayt, Ph. Teyssie, and N. C. de Jaeger, Macromolecules 24, 5922–5925 (1991). ¨ 312. K. Tauer, H. Muller, L. Rosengarten, and K. Riedelsberger, Colloids Surf., A 153, 75–88 (1999). ¨ 313. H. Muller, PhD thesis, Max Planck Institute of Colloids and Interfaces, Golm, and University of Potsdam, Germany, Cuvillier Verlag G¨ottingen, 1997, pp. 82–85. 314. G. Lagaly, O. Schulz, and R. Ziemehl, Dispersionen und Emulsionen: eine Einfuhrung ¨ in die Kolloidik feinverteilter Stoffe einschließlich der Tonminerale, Dr. Dietrich Steinkopff Verlag, Darmstadt, 1997, pp. 254–262. 315. S. Abend, N. Bonnke, U. Gutschner, and G. Lagaly, Colloid Polym. Sci. 276, 730–737 (1998). 316. A. Guyot and K. Tauer, in J. Texter, ed., Reactions and Synthesis in Surfactant Systems, Marcel Dekker, Inc., New York, 2001, pp. 547–575. 317. G. Odian, Principles of Polymerization, 3rd ed., John Wiley & Sons, Inc., New York, 1991. 318. S. L. Rosen, Fundamental Principles of Polymeric Materials, 2nd ed., John Wiley & Sons, Inc., New York, 1993, 200–213. 319. C. Hagiopol, Copolymerization: Toward a Systematic Approach, Kluwer Academic/Plenum Publishers, New York, 1999. 320. S.-T. Wang and G. W. Poehlein, J. Appl. Polym. Sci. 50, 2173–2183 (1993). 321. A. Guyot, J. Guillot, C. Pichot, and L. Rios Guerrero, in D. R. Bassett, and A. E. Hamielec, eds., Emulsion Polymers and Emulsion Polymerization (ACS, Symposium Series 165), American Chemical Society, Washington, D.C., 1981, pp. 415–436. ¨ 322. J. Guillot and C. Pichot, eds., Emulsion Copolymerization, Huthig & Wepf, Basel, 1985; Reprinted from Makromol. Chem., Suppl. 10/11, 1–536 (1985). 323. D. C. Blackley, Polymer Lattices Science and Technology Vol. 2: Types of Lattices, Chapman & Hall, London, 1997, pp. 206–232. 324. Ref. 319, pp. 192–206. 325. F. A. Bovey and P. A. Mirau, NMR of Polymers, Academic Press, Inc., San Diego, 1996, pp. 212–220. 326. J. Guillot, Makromol. Chem. Suppl. 10/11, 235–264 (1985). 327. J. Dimitratos, G. Elicabe, and C. Georgakis, AICheE J. 40, 1993–2021 (1994). 328. M. Embirucu, E. L. Lima, and J. C. Pinto, Polym. Eng. Sci. 36, 433–447 (1996). 329. R. Giudici, Latin Am. Appl. Res. 30, 351–356 (2000). 330. C. Kiparissides, Chem. Eng. Sci. 51, 1637–1659 (1996). 331. A. M. van Herk and A. L. German, Macromol. Theory Simul. 7, 537–565 (1998). 332. J. M. Asua, Polym. React. Eng. 9, 37–67 (2001). 333. J. Zhang, A. J. Morris, E. B. Martin, and C. Kiparissides, Chem. Eng. J. 69, 135–143 (1998). 334. J. Snuparek, Prog. Org. Coat. 29, 225–233 (1996). 335. Ref. 323, pp. 355. 336. J. C. Padget, J. Coat. Technol. 66, 89–105 (1994). ¨ 337. N. Sutterlin, Makromol. Chem. Suppl. 10/11, 403–418 (1985). 338. Ref. 323, pp. 358–368. 339. S. Torza and S. G. Mason, J. Colloid Interface Sci. 33, 67–83 (1970). 340. D. C. Sundberg, A. P. Casassa, J. Pantazopoulos, and M. R. Muscato, J. Appl. Polym. Sci. 41, 1425–1442 (1990). 341. C. L. Winzor and D. C. Sundberg, Polymer 33, 4269–2279 (1992). 342. E. J. Sundberg and D. C. Sundberg, J. Appl. Polym. Sci. 47, 1277–1294 (1993).

520

HETEROPHASE POLYMERIZATION

Vol. 6

343. Y. G. Durant and D. C. Sundberg, J. Appl. Polym. Sci. 58, 1607–1618 (1995). 344. D. C. Sundberg and Y. G. Durant, in J. M. asua, ed., Polymeric Dispersions: Principles and Applications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 177–188. 345. J.-E. L. J¨onsson, H. Hassander, L. H. Jansson, and B. T¨ornell, Macromolecules 24, 126–131 (1991). 346. J.-E. L. J¨onsson, H. Hassander, and B. T¨ornell, Macromolecules 27, 1932–1937 (1994). 347. Y.-C. Chen, V. L. Dimonie, and M. S. El-Aasser, Macromolecules 24, 3779–3787 (1991). 348. V. L. Dimonie, E. S. Daniels, O. L. Shaffer, and M. S. El-Aasser, in P. A. Lovell, and M. S. El-Aasser, eds., Emulsion Polymerization and Emulsion Polymers, J. Wiley & Sons, Inc., New York, 1997, pp. 293–326. 349. J. A. Waters, Colloids Surf., A 83, 167–174 (1994). 350. S. Lee and A. Rudin, J. Polym. Sci., Part A: Polym. Chem. 30, 2211–2216 (1992). 351. L. J. Gonzales-Ortiz and J. M. Asua, Macromolecules 28, 3135–3145 (1995). 352. L. J. Gonzales-Ortiz and J. M. Asua, Macromolecules 29, 383–389 (1996). 353. L. J. Gonzales-Ortiz and J. M. Asua, Macromolecules 29, 4520–4527 (1996). 354. D. A. House, Chem. Rev. 62, 185–203 (1962). 355. E. J. Behrman and J. O. Edwards, Rev. Inorg. Chem. 2, 179–206 (1980). 356. K. Tauer and R. Deckwer, Acta Polymerica 49, 411–416 (1998). 357. B. R. Vijayendran, Makromol. Chem. Suppl. 10/11, 271–295 (1985). 358. D. I. Lee, Makromol. Chem., Macromol. Suppl. 33, 117–131 (1990). 359. V. N. Pavlyuchenko, O. V. Sorochinskaya, S. S. Ivanchev, V. V. Klubin, G. S. Kreichman, V. P. Budtov, M. Skrifvars, E. Halme, and J. Koskinen, J. Polym. Sci., Part A: Polym. Chem. 39, 1435–1449 (2001). 360. C. Schellenberg, S. Akari, M. Regenbrecht, K. Tauer, F. M. Petrat, and M. Antonietti, Langmuir 15, 1283–1290 (1999). 361. C. Schellenberg Ph.D. thesis, Max Planck Institute of Colloids and Interfaces, Golm, Germany and University of Potsdam, Germany, Shaker Verlag, Germany 2000. 362. C. J. McDonald, K. J. Bouck, A. B. Chaput, and C. J. Stevens, Macromolecules 33, 1593–1605 (2000). 363. K. S. Birdi, in K. S. Birdi, ed., Handbook of Surface and Colloid Chemistry, CRC Press, Boca Raton, 1997, pp. 71–118. 364. M. Okubo, K. Ichikawa and M. Fujimura, Colloid Polym. Sci. 269, 1257–1262 (1991). 365. M. Okubo and H. Mori, Colloid Polym. Sci. 275, 634–639 (1997). 366. J.-W. Kim, Y.-G. Joe, and K.-D. Suh, Colloid Polym. Sci. 277, 252–256 (1999). 367. J. Seidl, J. Malinsky, K. Dusek, and W. Heitz, Adv. Polym. Sci. 5, 113–213 (1967). 368. D. Horak and M. J. Benes, React. Function. Poly. 45, 189–195 (2000). 369. K. Lewandowski, F. Svec, and J. M. J. Frechet, Chem., Mater. 10, 385–391 (1998). 370. M. Xu, E. Brahmachary, M. Janco, F. H. Ling, F. Svec, and J. M. J. Frechet, J. Chromatogr. A 928, 25–40 (2001). 371. C. D. Wood and A. I. Cooper, Macromolecules 34, 5–8 (2001). 372. Y.-D. Jo, K.-S. Park, J.-H. Ahn, and S.-I. Ihm, Eur. Polym. J. 32, 967–972 (1996). 373. S. Kawaguchi, K. Tano, M. Maniruzzman, and K. Ito, Macromol. Symp. 150, 101–108 (2000). 374. K. Tauer, K. Riedelsberger, R. Deckwer, A. Zimmermann, H. Dautzenberg, and J. Thieme, Macromol. Symp. 155, 95–104 (2000). 375. H. Zecha, I. Zenke, H. J. Purz, and C. Bischof, Angew. Makromol. Chem. 104, 169–188 (1982). ¨ 376. A. Buchtemann, E. Schulz, and K. Tauer, Acta Polym. 36, 654–657 (1985). ¨ 377. K. Tauer, E. Schulz, A. Buchtemann, G. Kurjakova, and W. Jaeger, Plaste Kautschuk 33, 51–53 (1986). 378. R. H. Ottewil, A. B. Schofield, and J. A. Waters, Colloid Polym. Sci. 274, 763–771 (1996).

Vol. 6

HETEROPHASE POLYMERIZATION

521

379. R. H. Ottewill, A. B. Schofield, J. A. Waters, and N. St. J. Williams, Colloid Polym. Sci. 275, 274–283 (1997). 380. R. H. Ottewill, A. B. Schofield, and J. A. Waters, J. Dispersion Sci., Technol. 19, 1151– 1162 (1998). 381. L. Lazar, S. A. M. Hesp, and G. E. Kmiecik-Lawrynowicz, Part. Sci., Technol. 18, 103–120 (2000). 382. L. Lazar and S. A. M. Hesp, Part. Sci., Technol. 18, 121–142 (2000). 383. L. Lazar and S. A. M. Hesp, Part. Sci., Technol. 18, 143–162 (2000). 384. K. Maruyama, M. Kawaguchi, and T. Kato, Colloids Surf., A 189, 211–223 (2001). 385. F. J. Schork, G. W. Poehlein, S. Wang, J. Reimers, J. Rodrigues, and C. Samer, Colloid Surf., A 153, 39–45 (1999). 386. M. Ochiai, M. Masui, M. Asanae, M. Tokunaga, and T. Iimura, J. Imag. Sci., Technol. 38, 415–420 (1994). 387. H. K. Mahabadi, T. H. Ng, and H. S. Tan, J. Microencaps. 13, 559–573 (1996). 388. D. C. Lee and L. W. Jang, J. Appl. Polym. Sci. 61, 1117–1122 (1996). 389. J. Hou and G. L. Baker, Chem. Mater. 10, 3311–3318 (1998). 390. J. Yu, J. Yu, Z.-X. Guo, and Y.-F. Gao, Macromol. Rapid Commun. 22, 1261–1264 (2001). 391. N. Bechthold, F. Tiarks, M. Willert, K. Landfester, and M. Antonietti, Macromol. Symp. 151, 549–555 (2000). 392. B. Erdem, E. D. Sudol, V. L. Dimonie, and M. S. El-Aasser, Macromol. Symp. 155, 181–198 (2000). 393. Q. Wang, H. Xia, and C. Zhang, J. Appl, Polym. Sci. 80, 1478–1488 (2001). 394. H. Xia, C. Zhang and Q. Wang, J. Appl. Polym. Sci. 80, 1130–1139 (2001). 395. F. Tiarks, K. Landfester, and M. Antonietti, Macromol. Chem., Phys. 202, 51–60 (2001). 396. K. Wormuth, J. Colloid Interface Sci. 241, 366–377 (2002). 397. M. B. Ko and J. Y. Jho, Polym. Bull. 46, 315–322 (2001). 398. J. I. Amalvy, M. J. Percy, and S. P. Armes, Langmuir 17, 4770–4778 (2001). 399. L. W. Jang, C. M. Kang, and D. C. Lee, J. Polym. Sci., B: Polym. Phys. 39, 719–727 (2001). 400. X. Huang and W. J. Brittain, Macromolecules 34, 3255–3260 (2001). 401. E. P. Giannelis, Adv. Mater. 8, 29–35 (1996). 402. B. J¨onsson, B. Lindman, K. Holmberg, and B. Kronberg, Surfactants and Polymers in Aqueous Solutions, John Wiley & Sons, Inc., Chichester 1998, pp. 74–82. 403. K. M. McGrath and C. J. Drummond, Colloid Polym. Sci. 274, 612–621 (1996). 404. D. Pawlowski, A. Haibel, and B. Tieke, Ber. Bunsen-Ges. Phys. Chem. 102, 1865–1869 (1998). 405. M. Antonietti and H.-P. Hentze, Colloid Polym. Sci. 274, 696–702 (1996). 406. H.-P. Hentze and M. Antonietti, Curr. Opin. Sol. State Mat. Sci. 5, 343–353 (2001). 407. S. Rimmer and P. I. Tattersall, Polymer 40, 6673–6677 (1999). 408. S. Rimmer and P. I. Tattersall, Polymer 40, 5729–5731 (1999). 409. E. E. Meyer, M. F. Islam, W. Lau, and H. D. Qu-Yang, Langmuir 16, 5519–5522 (2000). ¨ 410. R. J. Leyrer and W. Machtle, Macromol. Chem., Phys. 201, 1235–1243 (2000). 411. T. F. McKenna, S. Othman, G. Fevotte, A. M. Santos, and H. Hammouri, Polym. React. Eng. 8, 1–38 (2000). 412. P. Guinot, N. Othman, G. Fevotte, and T. F. McKenna, Polym. React. Eng. 8, 115–134 (2000). 413. C.-M. Astorga, N. Othman, H. Hammouri, and T. McKenna, Contr. Eng. Practice 10, 3–13 (2002). 414. R. F. Machado and A. Bolzan, Chem., Eng. J. 70, 1–8 (1998). 415. Z. Nagy and S. Agachi, Comp. Chem., Eng. 21, 571–591 (1997).

522

HETEROPHASE POLYMERIZATION

Vol. 6

416. A. Echevarria, J. R. Leiza, J. C. de la Cal, and J. M. Asua, AIChE J. 44, 1667–1679 (1998). 417. M. Vicente, S. BenAmor, L. M. Gugliotta, J. R. Leiza, and J. M. Asua, Ind. Eng. Chem. 40, 218–227 (2001). 418. F. H. Florenzano, R. Strelitzki and W. F. Reed, Macromolecules 31, 7226–7238 (1998). 419. R. Strelitzki, W. F. Reed, J. Appl. Polym. Sci. 73, 2359–2368 (1999). 420. W. F. Reed, Macromoelcules 33, 7165–7172 (2000). 421. A. Giz, H. Catalgil-Giz, A. Alb, J.-L. Brousseau, and W. F. Reed, Macromoelcules 34, 1180–1191 (2001). 422. W.-D. Hergeth, in J. M. Asua, ed., Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 267–288. 423. W.-D. Hergeth, in J. M. Asua, ed., Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 243–256. 424. O. Kammona, E. G. Chatzi, and C. Kiparissides, J. Macrmol. Sci., Rev.—Macromol. Chem., Phys. C 39, 57–134 (1999). 425. R. N. Landau, Thermochim. Acta 289, 101–126 (1996). 426. P. E. Gloor, R. J. Warner, Thermochim. Acta 289, 243–265 (1996). 427. M. Mosebach and K.-H. Reichert, J. Appl. Polym. Sci. 66, 673–681 (1997). 428. L. Varela de la Rosa, E. D. Sudol, M. S. El-Aasser, and A. Klein, J. Polym. Sci., Part A: Polym. Chem. 34, 461–473 (1996). 429. L. Varela de la Rosa, E. D. Sudol, M. S. El-Aasser, and A. Klein, J. Polym. Sci., Part A: Polym. Chem. 37, 4054–4065 (1999). 430. L. Varela de la Rosa, E. D. Sudol, M. S. El-Aasser, and A. Klein, J. Polym. Sci., Part A: Polym. Chem. 37, 4066–4072 (1999). 431. L. Varela de la Rosa, E. D. Sudol, M. S. El-Aasser, and A. Klein, J. Polym. Sci., Part A: Polym. Chem. 37, 4073–4089 (1999). 432. G. Maschio, I. Ferrara, C. Bassani, and H. Nieman, Chem., Eng. Sci. 54, 3273–3282 (1999). ¨ 433. K. Tauer, H. Muller, C. Schellenberg, and L. Rosengarten, Colloid Surf. A. 153, 143– 151 (1999). 434. S. Erwin, K. Schulz, H.-U. Moritz, C. Schwede, and H. Kerber, Chem., Eng. Technol. 24, 305–311 (2001). 435. K. Jahny, H.-J. P. Adler, and H.-U. Moritz, Macromol. Chem. Phys. 202, 2915–2920 (2001). 436. V. Liotta, C. Georgakis, E. D. Sudol, and M. S. El-Aasser, Ind. Eng. Chem., Res. 36, 3252–3263 (1997). 437. V. Liotta, E. D. Sudol, M. S. El-Aasser, and C. Georgakis, J. Polym. Sci., Polym. Chem. 36, 1553–1571 (1998). 438. C. Wu, J. D. S. Danielson, J. B. Callis, M. Eaton, and N. L. Ricker, Prog. Control Qual. 8, 1–23 (1996). 439. C. Wu, J. D. S. Danielson, J. B. Callis, M. Eaton, and N. L. Ricker, Prog. Control Qual. 8, 25–40 (1996). 440. A. F. Santos, E. L. Lima, and J. C. Pinto, J. Appl. Polym. Sci. 70, 1737–1745 (1998). 441. A. F. Santos, E. L. Lima, and J. C. Pinto, J. Appl. Polym. Sci. 77, 453–462 (2000). 442. R. A. M. Viera, C. Sayer, E. L. Lima, and J. C. Pinto, Polymer 42, 8901–8906 (2001). 443. A. Al-Khanbashi, M. Dhamdhere, and M. Hansen, Appl. Spectr. Rev. 33, 115–131 (1998). 444. T. J. Vickers, D. R. Lombardi, B. Sun, H. Wang, and C. K. Mann, Appl. Spectr. 51, 1251–1253 (1997). 445. C. Bauer, B. Amram, M. Agnely, D. Charmot, J. Sawatzki, N. Dupuy, and J.-P. Huvenne, Appl. Spectr. 54, 528–535 (2000).

Vol. 6

HETEROPHASE POLYMERIZATION

523

446. M. Morbidell, G. Storti and A. Siani, in J. M. Asua, ed., Polymeric Dispersions: Principles and Apllications, Kluwer Academic Publishers, Dordrecht, 1997, pp. 257–266. 447. A. Siani, G. Storti, and M. Morbidelli, J. Appl. Polym. Sci. 72, 1451–1477 (1999). 448. G. Storti, A. K. Hipp, and M. Morbidelli, Polym. React. Eng. 8, 77–94 (2000). 449. A. S. Dukhin, P. J. Goetz, and V. Hackley, Colloid, Surf., A 138, 1–9 (1998). 450. T. H. Wines, A. S. Dukhin, and P. Somasundaran, J. Colloid Interface Sci. 216, 303–308 (1999). 451. A. S. Dukhin, P. J. Goetz, T. H. Wines, and P. Somasundaran, Colloid, Surf., A 173, 127–158 (2000). 452. A. S. Dukhin and P. J. Goetz, Colloid Surf. A. Physicochem. Eng. Asp. 192, 267–306 (2001). 453. A. S. Dukhin and P. J. Goetz, Adv. Colloid Interface Sci. 92, 73–132 (2001). 454. S. S. Kim and D. G. Westmoreland, J. Polym. Sci., Part A: Polym. Chem. 32, 3031–3037 (1994). 455. H.-Y. Parker, D. G. Westmoreland, and H.-R. Chang, Macromolecules 29, 5119–5127 (1996). 456. H. R. Chang, W. Lau, H.-Y. Parker, and D. G. Westmoreland, Macromol. Symp. 111, 253–263 (1996). 457. S. Rudschuk, J. Adams, and J. Fuhrmann, Macromol. Chem., Phys. 199, 771–776 (1998). 458. S. Rudschuk, J. Adams, and J. Fuhrmann, Nucl. Instrum Methods Phys. Res., Sect. B 151, 341–345 (1999). 459. K. Landfester, S. Spiegel, R. Born, and H. W. Spiess, Colloid Polym. Sci. 276, 356–361 (1998). 460. C. J. Samer and F. J. Schork, Polym. React. Eng. 5, 85–124 (1997). 461. G. W. Poehlein and J. Schork TRIP 1, 298–302 (1998). 462. S. Roy and S. Devi, Polymer 38, 3325–3331 (1997).

KLAUS TAUER Max Planck Institute of Colloids and Interfaces

Abbreviations A: Hamaker constant (general) A1,2 , A1,3 , A2,3 : interfacial areas between phases 1–2, 1–3, and 2–3, respectively Aeff : effective Hamaker constant of interacting particles Am : Hamaker constant of the medium between interacting particles Ap : total surface/interface of dispersed phase material Ap : Hamaker constant of interacting particles as : area covered by a surfactant molecule Aspec : specific surface/interface of dispersed phase either per unit mass or volume of either dispersed phase or water (eq. 1) C: capacity of a conducting sphere C0 : solubility of the dispersed phase material in the continuous phase CI : initiator concentration CLC : concentration of the lyophob in the continuous phase CLC0 : solubility of the lyophob in the continuous phase CM : monomer concentration

524

HETEROPHASE POLYMERIZATION

Vol. 6

CM,p : monomer concentration inside dispersed phase particles cmc: critical micelle concentration of surfactant/stabilizer CMC0 : solubility of the monomer in the continuous phase CS : surfactant/stabilizer concentration CS,C : critical surfactant concentration for stabilization of particles after nucleation CS,L : concentration of stabilizing polymer (lyophilic polymer) in the volume of the hydrodynamic layer Csalt : ionic strength in the continuous phase D: diameter of dispersed phase material D0 : diameter of seed particles Dd : droplet diameter Dd,c : critical droplet diameter for which Ostwald ripening stops (eq. 25) Dd,e : droplet diameter at equilibrium Ddp : size of dispersed phase molecule Dee : characteristic length of a dissolved polymer molecule dh : diameter of a single hole Dm : monomer droplet diameter DMC : diffusion coefficient of monomer in the continuous phase Dp : monomer particle diameter dpp : interparticle distance (eq. 3) Dt : diameter of final composite particles e: type of end group Ekin : kinetic energy of particles in a shear field (approximation) f : radical efficiency for starting polymerization f 1 , f 2 : mole fractions of monomer 1and 2 in a comonomer mixture F1 , F2 : mole fractions of monomer 1 and 2 build in the copolymer FG: solids content in parts per 100 parts dispersion h: Planck’s constant H: geometrical function of interacting particles j: chain length k11 , k22 , k21 , k12 : propagation rate constants of M 1 with P1 , M 2 with P2 , M 1 with P2 , and M 2 with P1 , respectively kabs,i : rate constant of primary radical absorption by particles of dispersed phase kB : Boltzman’s constant kd,d : rate constant of initiator decomposition in the dispersed phase kd,w : rate constant of initiator decomposition in the continuous phase kdes : rate constant of radical desorption kL : London interaction constant kp : propagation rate constant kt,d : termination rate constant in dispersed phase kt,w : rate constant of radical termination in the continuous phase M 1 , M 2 : concentration of both comonomers M mon : molecular weight of the monomer M S : molecular weight of surfactant/stabilizer N: number of particles of dispersed phase material (eq. 2) n: number of growing centers per particle of dispersed phase N A : Avogadro’s number

Vol. 6

HETEROPHASE POLYMERIZATION

525

nI : number of initiator molecules per particle of dispersed phase N n: number of particles of dispersed phase with n growing centers nR,equ : number of radicals per particle of dispersed phase at steady state (eq. 7) nS : number of molecule of disperse phase at the surface/interface (eq. 5) N ss : number of small spheres in the mixture with large spheres P: concentration of propagating radicals p : vapor pressure outside the drop (eq. 29) p0 : bulk vapor pressure of the liquid forming a drop (eq. 29) P1 , P2 : growing chain ended with a radical of monomer 1 and growing chain ended with a radical of monomer 2, respectively Pd : pressure inside dispersed phase particle P j,e,w : concentration of radicals of chain length j and end group e in the continuous phase Pn : average degree of polymerization PSW : swelling pressure Q: concentration of interacting centers per unit volume R: gas constant r: radius of swollen particles r0 : radius of unswollen particles r1 , r2 reactivity ratios (eq. 41) rabs : absorption rate of radicals from the continuous phase into particles ri,d : rate of initiation in the dispersed phase ri,w : rate of initiation in the continuous phase rp : overall rate of polymerization (eq. 9) S1 , S2 , S3 : spreading coefficients of phase 1, 2, and 3, respectively (eq. 44) T g : glass-transition temperature T g,i : glass-transition temperature of a homopolymer of monomer i v: particle volume v : molar volume of the liquid forming a drop (eq. 29) V A : attractive potential (eq. 32) vc : volume of continuous phase vc,p : molar volume of the molecules forming the continuous phase V int : interaction potential between charge–stabilized spheres (eq. 37a) vm,p : monomer volume in swollen particles vmon : molar volume of the monomer vp,p : polymer volume in swollen particles V R,B : Born’s repulsion potential V R,es : repulsion potential between equally charged spheres vrel : relative velocity of two approaching particles in a shear field V st : steric repulsion potential wi : weight fraction of monomer i in a copolymer W surf,t : amount of surfactant relative to the total mass of organic phase (or dispersed phase) (eq. 4) Z: number of electrons in the outer shell of a molecule z: stoichiometric valency of the electrolyte dissolved in the continuous phase ∆D: decrease in diameter caused by shrinkage during polymerization ∆Gf : free energy of the flocculation process ∆GI : interfacial free energy (eq. 43)

526

HETEROPHASE POLYMERIZATION

Vol. 6

∆H f : change in enthalpy during flocculation ∆Pd : Young–Laplace pressure ∆R: hydrodynamic stabilizer layer thickness ∆Sf : change in entropy during flocculation ∆Ssl : entropy gain between interacting small and large hard spheres ∆vc : increase in the effective volume for the smaller spheres due to overlap of the excluded volume regions of the larger spheres ∆µ1 : chemical potential of swelling agent (monomer) Θ: theta condition of a dilute polymer solution (in the limit of infinite molecular weight, the situation where
H = T
S) Ψ: surface potential α: exponent in the scaling relation between the hydrodynamic layer thickness (
R) and the ionic strength (Csalt ) α0 : static polarizability of molecules χm,p : Flory–Huggins interaction parameter between monomer and polymer χS,cp : Flory–Huggins interaction parameter between the stabilizing polymer and the continuous phase ε: permittivity φ1 : monomer volume fraction inside swollen particles. φ1 , φ 3 : volume fractions of phase 1 and 3, respectively (eq. 48) φ2 : polymer volume fraction inside swollen particles φL : volume fraction of the lyophob inside dispersed phase droplets φL,e : volume fraction of the lyophob inside dispersed phase droplets in equilibrium φL,i : volume fraction of the lyophob inside dispersed phase droplets just after comminution φMC0 : molar solubility fraction of monomer in the continuous phase φSE : fraction of free surface energy relative to total free energy of a particle of dispersed phase (eq. 6) γ: interfacial tension liquid vapor ψ dp : volume fraction of dispersed phase κ−1 : Debye’s screening length (eq. 37b) ν c : frequency of fluctuations π s : osmotic pressure of a bulk gel ρdp : density dispersed phase ρm1 : monomer density ρp1 : polymer density σ: interfacial tension dispersed fluid continuous fluid phase σ 1,2 , σ 1,3 σ 2,3 : interfacial tension between phases 1–2, 1–13, and 2–3, respectively τ ex : exchange time of a droplet during Ostwald ripening τ pol : duration to form an average polymer molecule (eq. 8) n¯ : average number of radicals per particle j,e kabs : rate constant of absorption of propagating radicals with chain length j and end group e t,e : rate constant of radical termination in the continuous phase with chain kabs length j and end group e D¯ d,i : average droplet diameter just after comminution

Vol. 6

HETEROPHASE POLYMERIZATION

527

M¯ C : average molecular weight between two cross-links v mon mol : volume per monomer molecule v ru mol : volume per repeating unit of the monomer in the polymer ∆G Ics : change in interfacial free energy for the formation of core–shell particles (eq. 45) ∆σ cs i: change in reduced interfacial free energy for the formation of core–shell particles or inserted core-shell particles (eqs. 46 and 47) 2fkd : initiator decomposition rate constant, including efficiency for initiating polymerization (f )

HIGH DENSITY POLYETHYLENE. HYDROGELS.

See ETHYLENE POLYMERS, HDPE.

See Volume 2.

β-HYDROXYALKANOATES. See POLY3-(HYDROXYALKANOATES).

HYPERBRANCHED POLYMERS.

See Volume 2.