Advances in Colloid and Interface Science 85 2000.

1 33

Temperature-sensitive aqueous microgels

Robert PeltonU
McMaster Centre for Pulp and Paper Research, Department of Chemical Engineering, McMaster Uni ersity, Hamilton, Ontario, Canada L8S 4L7

Abstract An account of the preparation and characterization of temperature-sensitive aqueous microgels based on poly N-isopropylacrylamide. was first published in 1986. Since then there has been a steady increase in the number of publications describing preparation, characterization and applications of temperature-sensitive microgels. This paper reviews the important developments in the area of temperature-sensitive aqueous microgels over the last decade. Although most of the work involves gels based on poly N-isopropylacrylamide., other polymers are also considered. Core shell latex particles exhibiting temperature-sensitive properties are also described. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Thermal sensitive microgels; Temperature-sensitive microgels; Latexes; Colloids; Swelling; Polymer surfactant interactions

Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2. Microgel synthesis . . . . . . . . . . . . . . . . . . . . . 2.1. Homogeneous gels . . . . . . . . . . . . . . . . . 2.1.1 Novel synthetic methods . . . . . . . . . 2.2. Core shell microgels . . . . . . . . . . . . . . . 2.3. Non-NIPAM temperature-sensitive microgels 3. Microgel properties . . . . . . . . . . . . . . . . . . . . 3.1. Gel structure . . . . . . . . . . . . . . . . . . . . 3.2. Gel swelling . . . . . . . . . . . . . . . . . . . . .
U

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. 2 . 5 . 5 . 8 . 9 . 10 . 11 . 11 . 13

Tel.: q1-905-525-9140 ext. 27045; fax: q1-905-528-5114. E-mail address: peltonrh@mcmaster.ca R. Pelton.

0001-8686r00r$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 1 - 8 6 8 6 9 9 . 0 0 0 2 3 - 8

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

3.2.1. Swelling theory . . . . . . 3.2.2. Swelling measurements . 3.2.3. Swelling in other solvents 3.3. Surface activity . . . . . . . . . . . 3.4. Rheological properties . . . . . . 3.5. Electrical properties . . . . . . . . 3.6. Gel surfactant interactions . . . 3.7. Colloidal stability . . . . . . . . . . 4. Microgel applications . . . . . . . . . . . 4.1. Biotechnology . . . . . . . . . . . . 4.2. Controlled uptake and release . 4.3. Other potential applications . . . 5. Closing remarks . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. 13 . 14 . 16 . 18 . 18 . 19 . 23 . 24 . 27 . 27 . 29 . 30 . 31 . 31 . 31

1. Introduction Aqueous microgels are an important part of water-borne polymer technologies. Cosmetics, coatings, food, and industrial processing industries employ aqueous microgels to modify rheological properties, to retain water, and for many other reasons. Commodity microgels include those based on starch, cross-linked polysodium methacrylate . w1x, and a variety of gums. However, the focus of this review is on a new class of synthetic aqueous microgels whose properties display extreme temperature sensitivity in water. Most of these systems are based on poly N-isopropylacrylamide. polyNIPAM. or related copolymers. The structure of NIPAM is shown in Fig. 1. In the absence of a universal definition of microgels, this review is restricted to dispersions with average diameters ranging between 50 nm and 5 m. The preparation and characterization of much larger gel particles has been described in the literature e.g. see Park and Hoffman w2x.. This review also includes core shell gel structures in which the shell has gel properties in water whereas the core made of water insoluble polymer such as polystyrene. The distinction between core shell gels and latex particles coated with hydrophilic polymer is not clear. For example, polystyrene particles coated with low molecular weight grafted polyethylene glycol PEG. would not, in normal circumstances, be considered a microgel. On the other hand, the properties within the PEG coating are those of a gel and the particles are temperature-sensitive w3x. In this review, no attempt is made to resolve these definitions. Instead, the core shell systems reviewed are those which seem to have been prepared specifically to display temperature-sensitive behavior in water irrespective of thickness or volume fraction of the gel layer. Thermo-responsive aqueous colloidal microgels form an interesting subset of polymer colloids the gels have properties in common with water-soluble

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 1. Monomers employed in temperature-sensitive microgel synthesis.

polymers, water-swollen macro gels, and water-insoluble latex particles. Like water-soluble polymers, the properties of microgels depend upon the subtle balance of polymerrpolymer vs. polymerrwater interactions. Like macroscopic aqueous gels, colloidal microgels are characterized by a degree of swelling, an average cross-link density and characteristic time constants for swelling and shrinking. Like hydrophobic polymer colloids, colloidal microgels can be prepared to have monodisperse particle size distributions; microgels can be characterized by standard colloidal techniques including electrophoresis, dynamic light scattering, rheology and electron microscopy. Microgels can be flocculated by salt or flocculant addition. In 1978, Philip Chibante, a high school summer student, with aspirations to become a dentist, prepared the first reported temperature-sensitive aqueous microgel under the authors supervision w4x. The microgel was a monodisperse, colloidal dispersion based on cross-linked poly N-isopropylacrylamide., herein called polyNIPAM, a polymer which has a lower critical solution temperature LCST. in water of 32 C w5,6x. All the microgel properties were sensitive functions of temperature over the range 15 50 C. A detailed description of microgel properties and characterization is given in subsequent sections. However, as a preview, the essential temperature-sensitive properties are illustrated in Fig. 2. At room temperature, the gels have a high water content, a low refractive index difference

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 2. A schematic illustration of the factors influencing microgel colloidal stability below and above the VPTT.

with water, and a few electrically charged groups on the chain ends. By contrast, at elevated temperatures the particle volume is 10-times less; the density of electrically charged groups is higher and the refractive index difference with water is greater. Since 1986 there have been many publications describing the preparation, characterization and application of temperature-sensitive microgels, most based on polyNIPAM. This activity reflects the current interest in switchable or intelligent materials w7 9x. The recent advances in temperature-sensitive microgels are reviewed. This work not only updates previous reviews w10 13x but also attempts to identify areas where more research is required. N-Isopropylacrylamide NIPAM. is the major building block for temperaturesensitive microgels. The monomer is available from specialty chemical distributors. With a structure see Fig. 1. close to acrylamide, many of the properties of NIPAM are similar to those of acrylamide. In aqueous solution it undergoes rapid free radical polymerization in water to give high molecular weight polymers at rates similar to that of acrylamide w5,14 16x. Like acrylamide, NIPAM is a suspected carcinogen and neurotoxin, however, unlike acrylamide, NIPAM has an intense odor so monomer contamination is easy to detect. Gels, either micro or macro, are temperature-sensitive if most of the polymer in the gel network displays has temperature-sensitive phase behavior in the swelling solvent. A linear polymer that displays cloud point behavior when heated can be cross-linked to give a temperature-sensitive gel network. Upon heating such a gel, one observes the gel to shrink by expelling water over a narrow temperature range, usually called the olume phase transition temperature VPTT.. PolyNIPAM-based aqueous microgels are temperature-sensitive because polyNIPAM has a lower critical solution temperature of 32 C see Schilds excellent review of the properties of linear polyNIPAM w6x. Fig. 3 shows a phase diagram for polyNIPAM in water. The cloud point temperature is not very sensitive to concentration over most of the range of water contents w6x. Indeed, there is a tendency in the literature to equate experimentally determined cloud point temperatures to the lower critical

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 3. The phase diagram for polyNIPAM in water-adapted from Heskins and Guillet, w5x.

solution temperature LCST. which is the minimum temperature for the curve in Fig. 3.

2. Microgel synthesis The following section gives an overview of the most popular methods for microgel synthesis. Homogeneous and core shell gels are considered separately. 2.1. Homogeneous gels The first published account of polyNIPAM-based microgels described a surfactant-free emulsion polymerization of aqueous NIPAM and methylene-bis-acrylamide BA see Fig. 1. w4x. This recipe was essentially the same as that used to prepare monodisperse, surfactant-free polystyrene latex w17,86x. The polymerizations were conducted at 60 70 C in order to generate free radicals by the decomposition of the persulfate initiator. However, elevated temperature was also required so that growing polyNIPAM chains phase separated to form colloidal particles some authors call this procedure precipitation polymerization. The original procedure has been reproduced by many others w18 26x. This simple polymerization procedure can produce remarkably uniform particles. Fig. 4 shows a transmission electron micrograph of typical microgel particles. The TEM sample was prepared by placing a dilute drop of aqueous particles onto the sample grid and allowing it to dry. The surface tension forces pulled the swollen particles into a closely packed array of spheres that then dehydrated in a vacuum to give disks. The original dimensions of the swollen microgels can be estimated

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 4. Transmission electron micrograph of polyNIPAM microgels w4x.

from the center-to-center spacing of the ordered disks in Fig. 4. The shape of the dehydrated spheres was confirmed to be a rather flat cap by Crowther and Vincent w27x using scanning electron microscopy an example is shown in Fig. 5. Microgel particle formation occurs by homogenous nucleation that is known sometimes to give latex dispersions with a narrow particle size distribution w28x. According to this mechanism a water-soluble sulfate radical initiates a water-soluble NIPAM monomer which then grows in solution until it reaches a critical chain length after which the growing chain collapses to become a colloidally unstable precursor particle. The precursor particles follow one of two competing processes. Either they deposit onto an existing colloidally stable polymer particle or they aggregate with other precursor particles until they form a particle sufficiently large to be colloidally stable. At the polymerization temperature 50 70 C. which is well above the LCST, the growing polyNIPAM microgel particles are colloidally stabilized by electrostatic stabilization originating from sulfate groups introduced by the persulfate initiator. In the case of surfactant-free styrene polymerization,

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 5. Scanning electron micrograph of polyNIPAM microgels taken at an angle to emphasize the shape of the dehydrated particles w27x.

the relatively hydrophilic sulfate terminated chains concentrate at the particlerwater interface so the charge density increases with particle size which in turn means the colloidal stability should increase as the particles grow. Thus, aggregating precursor particles eventually achieve a diameter where the particle is colloidally stable and a new latex particle is formed. Microgel polymerizations can yield a significant fraction of sol i.e. linear or slightly branched polymer.. The sol fraction must be removed to generate the very ordered arrays shown in Figs. 4 and 5. Centrifugation, followed by decantation and redispersion in clean water is an effective cleaning procedure. The microgel particles readily redisperse after room temperature centrifugation, showing no indication of coagulation. PolyNIPAM microgel preparation do not always yield monodisperse colloidal particles. Too much BA cross-linker or dirty monomer can give unstable suspensions. Indeed, some approaches to the synthesis of porous macroscopic temperature-sensitive gels seem to involve the aggregation of colloidal microgels w29x. Analogous to surfactant-free polystyrene latex w17,86x, cationic polyNIPAM microgels can be prepared by employing a positively charged free radical initiator such as 2,2 -azobis-2-amidiopropane. dihydrochloride w4,30 32x. The resulting microgels have covalently bonded cationic initiator fragments. The concentration of cationic groups in the microgels can be increased by copolymerization with a cationic monomer w32x, by binding sorption. of a cationic surfactant w30x or by the Mannich reaction of acrylamide containing microgels to give ionizable amine groups w33x. PolyNIPAM microgels must contain a cross-linking comonomer such as BA to prevent the gel from dissolving in water at low temperature. Furthermore, within limitations, it is possible to incorporate most other water soluble vinyl monomers. The original example is acrylamide w4x. The upper limit to the content of a third monomer is the requirement that the polymer not be soluble in water at the

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 6. The influence of SDS on the average particle size of polyNIPAM latex. Volume average diameters were determined by dynamic light scattering at 50 C w15x.

polymerization temperature. In the case of acrylamide it is difficult to incorporate more than 10 wt.%. Many other water-soluble monomers have been copolymerized with NIPAM in thermo-sensitive microgels. These include N-acryloylglycine w34x, acrylic acid w35x, 2-aminoethylmethacrylate hydrochloride w36x see Fig. 1. It is difficult to prepare small microgels by surfactant-free methods because there is simply not enough available charge to stabilize high concentrations of small particles. The surface charge density of a polyNIPAM microgel latex was found to be 3.8 eq.rg w37x which is about two orders of magnitude lower than a corresponding surfactant free polystyrene latex. Indeed, sometimes it is difficult to get good surfactant-free dispersions of any size. McPhee et al. w37x first showed that the presence of low sodium dodecyl sulfate SDS. concentrations up to 4 mmol. gave robust preparations. The average particle size in water at 25 C decreased by a factor of 10 when the SDS concentration was increased from 0.2 to 4 mmol. Fig. 6 shows the diameter of shrunken 50 C. particles as a function of SDS concentration used to prepare the microgels w15x. The fitted line on the double logarithmic wSDSxy0 .71 where D is the diameter. Although employing SDS plot predicts D extends the latitude of microgel synthesis, it does introduce difficulties. SDS binds to polyNIPAM see subsequent sections. and thus can be difficult to remove. 2.1.1. No el synthetic methods Heating a solution of linear polyNIPAM through the LCST can also form polyNIPAM particles. For example, Fig. 7 shows that the particle size of the polyNIPAM dispersions increases linearly with the cube root of the total polymer concentration w38x. The linear relationship implies that the number concentration of particles produced was independent of polyNIPAM concentration. The particles appear to be electrostatically stabilized against coagulation. The electrophoretic mobilities for the dispersions varied from y1 = 10 8 m2rVs to y4 = 10 8 m2rVs. Similarly, cationic particles form from phase separated solutions of cationic copolymers of NIPAM w39x. Upon cooling the polyNIPAM particles revert to polymer solutions. However, one could envision a process whereby polyNIPAM chains were heated to give colloidal particles that were then cross-linked, perhaps by condensation chemistry, to give microgels.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

2.2. Core shell microgels Core shell gels consist of a water-insoluble latex particle coated with a gel layer. The composite particles have properties characteristic of both the core and the shell. The core dominates light scattering turbidity. behavior whereas the colloid stability is determined by the hydrogel shell. For example, latex with temperaturesensitive colloidal stability can be achieved with a polyNIPAM shell this will be described in more detail in the section dealing with colloidal properties. The first reported preparation of polyNIPAM coated core shell particles was by Pelton w40x who described both the one-shot surfactant-free preparation of polystyrenerpolyNIPAM gels and the grafting of polyNIPAM onto existing latex particles. Unlike the homogenous microgel preparations, these polymerizations are conducted below the LCST so that polyNIPAM does not phase separate. Similar polystyrene-co-polyNIPAM surfactant-free latexes are obtained when starting with either NIPAM monomer or with polyNIPAM. This reflects the fact that NIPAM polymerizes much more quickly than does styrene in aqueous media. Thus, most of the NIPAM is converted to polyNIPAM at extremely low styrene conversions. Makino et al. w41x refined Peltons procedure with a two-stage approach. The first stage was a one-shot surfactant-free polymerization with styrene and NIPAM. In the second stage, the surface gel layer was expanded by adding more NIPAM monomer and a nonionic water-soluble free radical initiator. Zhu and Napper w42,43x employed a two-stage procedure for the preparation of polystyrene-polyNIPAM core shell particles. In the first stage styrene was added slowly to an aqueous polyNIPAM solution, presumably at room temperature, in the presence of a redox initiator. The objective of this stage was to produce a soluble polyNIPAM-co-styrene copolymer, however, the polymer from this stage was not isolated. In the second stage the remainder of the styrene was added quickly to yield a latex. Duracher et al. w44,45x investigated the preparation and properties of cationic polystyrene-polyNIPAM core shell particles. Both cationic initiator and aminoethylmethacrylate introduced positively charged groups. The cationic monomer

Fig. 7. Influence of polymer concentration on the average diameter of polyNIPAM particles formed by heating polymer solutions w37x.

10

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

gave smaller particles, however a variety of particle morphologies were obtained, depending upon the polymerization conditions. Takeuchi et al. w46x prepared polystyrene-core polyNIPAM-shell particles by an interesting macromonomer approach. The macromonomer had approximately 24 NIPAM residues with a methacrylamido end group see Fig. 1.. The core shell particles were prepared by aqueous emulsion polymerization of styrene, macromonomer and SDS at 70 C. The surfactant was necessary because at 70 C the polyNIPAM would have no capability to stabilize the latex. Presumably, the surfactant would not have been necessary if the preparations had been conducted at room temperature with a redox initiator. The resulting latex showed the temperature-sensitive colloid stability characteristics of a polyNIPAM-coated particle. The macromonomer approach is appealing because it should exclusively produce terminally attached polyNIPAM chains to give a brush structure on the surface. In contrast, the other reported methods produce more complicated polyNIPAM surface gels. More recently, Chen et al. w47x employed the macromonomer approach to prepare polystyrene-core polyNIPAM-shell particles which were used to support platinum catalysts. Many of the potential biotechnological applications of polyNIPAM microgels require the presence of reactive functional groups to act as coupling sites. Zhou et al. w26x prepared cyano functionalized gels by a two-stage polymerization. Conventional polyNIPAM microgel was prepared in the first stage and NIPAM, BA, and acrylonitrile were polymerized at 60 C in the second stage. A two-shot procedure was also used. The resulting microgels were monodisperse. Shell recipes with 50 and 75 wt.% acrylonitrile produced particles with raspberry morphologies in the SEM. Yasui et al. w48x described a particularly elegant route to core shell particles bearing reactive groups. Carboxyl terminated low molecular weight, 2500 and 11 000 Da, polyNIPAM chains were condensed onto the surface of polystyreneco-divinylbenzene-co-glycidyl methacrylate . particles by carbodimide coupling chemistry. The enzyme trypsin was coupled to the free polyNIPAM chain ends. 2.3. Non-NIPAM temperature-sensiti e microgels Most of the work discussed in this review involves microgels based on N-isopropylacrylamide. However, related monomers also yield polymers which have a lower critical solution temperature in water and a few of these monomers have been used to make temperature-sensitive hydrogels. Duracher et al. w49x reported the preparation and characterization of microgels based on N-isopropylmethacrylamide see Fig. 1.. Monodisperse microgels were prepared under the same conditions as polyNIPAM microgels, however the polymerization times were five times longer for isopropylmethacrylamide. This observation is consistent with the fact that the propagation rate constant for acrylamide is 22 times greater than that of methacrylamide w49x. A consequence of the low reactivity of isopropylmethacrylamide was that BA cross-linker was consumed much more quickly than N-isopropylmethacrylamide. This in turn could lead to a significant content of linear

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

11

poly N-isopropylmethacrylamide.. The authors identified conditions to minimize this problem. Since N-isopropylmethacrylamide contains one more methyl group than does NIPAM, one might expect the polymer to have lower cloud point temperature than polyNIPAM in water but this is not the case. Reported cloud point temperatures for poly N-isopropylmethacrylamide. range between 39 and 47 C; based on optical density measurements the value seems to be 45 C w49x. The incorporation of BA to give microgels caused the optical transmittance vs. temperature plots to broaden and shift towards lower temperature with increasing BA content. This may reflect a distribution of polymer structures arising from a batch polymerization of BA with N-isopropylmethacrylamide. Lowe et al. w50x have prepared temperature-sensitive microgels based on Nethylacrylamide. The precipitation polymerizations were conducted at 90 C, above the phase separation temperature 78 C. for the homopolymer, yielding monodisperse dispersions. The transition was less abrupt than for polyNIPAM. In contrast to the behavior of polyNIPAM gels, the DSC scans over the transition were dependent upon concentration, not reversible, and repeated measurements on the same sample were not identical. As with the homogeneous gels, only a few temperature-sensitive, core shell gels not based on polyNIPAM have been described. Most of the work has been done in Harma Kawaguchis laboratory in Japan. For example, they w51x described the surfactant-free polymerization of N-acryloylpyrrolidine copolymers with styrene. Spherical particles were obtained for particles containing more than 80 wt.%, and the resulting particles showed temperature-sensitive colloidal stability and electrophoretic mobility. Swelling data were not given, however, and in view of the high styrene content the particles are more likely to be poly N-acryloylpyrrolidine. stabilized polystyrene particles than homogeneous microgels. In a related study Hosino et al. w52x compared the properties of shells, on polystyrene cores, based on N-acryloylpyrrolidine with those based on Nacryloylpiperidine see Fig. 1.. The pyrrolidine-based particles were more hydrophilic displaying temperature sensitivity at 50 C compared with 5 C for the piperidine shell particles. Poly N-vinylisobutyramide., an isomer of polyNIPAM in which the positions of the nitrogen and carbonyl are exchanged see Fig. 1., has an LCST of 40 C. Serizawa et al. w53x, reported the preparation of monodisperse polystyrene latex stabilized with a shell of poly N-vinylisobutyramide. which was added to the polymerization as a macromonomer. The authors claim the particles are colloidally stable above the VPTC, which is difficult to explain because the particles were prepared with a nonionic initiator; thus they should have little surface charge.

3. Microgel properties 3.1. Gel structure Most of the reported microgel preparations involve batch polymerization. Fig. 8

12

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 8. Conversion curves for NIPAM and methylenebisacrylamide BA. w15x. Note that 9 mM SDS was employed in the polymerizations.

shows the consumption of NIPAM and BA cross-linker as functions of time and temperature w15x. BA is consumed more quickly than NIPAM indicating that the particles are unlikely to have a uniform composition. Indeed, it seems reasonable to speculate that there exists a zone of relatively high cross-link density in each particle. Since polyBA is more hydrophilic than polyNIPAM at elevated temperature, one might further speculate that at least part of the high cross-link density zone is situated on the particlerwater interface. A number of techniques have been applied to characterize polyNIPAM microgels, and the results of some of this work give insight into gel particle morphology. Fujimoto et al. w54x treated homogeneous microgels with a fluorescent probe and measured emission intensity and maximum wavelength as functions of temperature. Both parameters showed large changes between 30 and 35 C, and the probe indicated a more hydrophobic environment in the shrunken gels. In the same paper they compared fluorescent quenching by nitromethane in hydrogels with linear polyNIPAM at 40 C. The quenching efficiency was much lower in the hydrogel indicating that the fluorescent probe was less accessible to the nitromethane in the cross-linked network. Linear polyNIPAM at 40 C can be present either as colloidally dispersed particles or as a macroscopic coacervate phase. Although the paper did not reveal the physical state of the linear polyNIPAM during the fluorescent quenching studies, the results are interesting and this work indicates that fluorescent quenching may be a good tool for microgel characterization. A number of interesting techniques have been used to investigate macroscopic gels and may have a role in future microgel characterization. Positron annihilation lifetime spectroscopy was used by Sousa et al. w55x to measure the free volume in macroscopic polyNIPAM-co-AM gels. The free volume decreased by nearly 30% when NIPAM was replaced with AM. The explanation for the difference is not clear. Raman microscopy has been used to image the pore structure in polyNIPAM macrogels; however, the resolution of this technique seems to be too low to be of interest in microgel characterization w56x.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

13

Small angle neutron scattering shows much promise for the characterization of microgels. The only microgel work published thus far is a study of the structure of bound surfactant w57x. A series of papers describing SANS characterization of charged polyNIPAM macrogels has been published by Shibayama et al. w58,59x. Scattering originates from both spatial variations of gel structure and thermal density fluctuations 3.2. Gel swelling 3.2.1. Swelling theory Temperature-sensitive swelling is the most spectacular property of polyNIPAM gels, and this has received a lot of attention for both micro and macro gels. The driving force and the equilibrium extent of swelling should be the same for a 500-nm microgel as for a 5-cm macrogel. On the other hand, the dynamic aspects are sensitive to the size of gels. Microgels achieve steady-state swelling in less than a second when the temperature is changed whereas macrogels can take a very long time because shrinking of the exterior layer prevents water transport from the interior. Although the thermodynamic theory of gel swelling is a classical subject, there have been a number of recent theoretical articles aimed at the polyNIPAM system. Lele et al. w60x applied an extended lattice theory that accounts for hydrogen bonding, to published polyNIPAM macrogel swelling data. Their calculations indicate that the bound water content above the volume phase transition temperature is 0.4 grg. Their approach distinguishes between hydrogen-bonded bound water and other bound water. The same paper also shows that 1 H-NMR static line width measurements change dramatically around the VPTT. Prausnitzs group w61x has applied semi-empirical extended Flory Huggins theory to predict the volume phase transitions for polyNIPAM macrogels. Some of the model parameters were obtained from the experimental properties of linear polyNIPAM solutions. The predicted VPTT was about 1 C higher than the LCST of linear polyNIPAM. Choi et al. w34x employed a similar theory to predict the swelling behavior of polyNIPAM-co-N-acryloylglycine-co-BA. gels. The computed swelling vs. temperature curves was slightly broader than the experimental data. Inomata et al. w62x interpreted polyNIPAM macrogel swelling by one of Prausnitzs earlier models. The gels were swollen by equilibration with aqueous polyethylene glycol that allowed calculation of the osmotic pressure. The interpretation of polyNIPAM gel swelling behavior is aided by the observation that the BA cross-linker has little influence on the polyNIPAMrwater interaction. Thus, the driving force for swelling can be estimated from the properties of linear polyNIPAM solutions, whereas the gel elasticity opposing swelling comes mainly from the network topology, which depends upon BA concentration. Semi-empirical extensions to Flory Huggins theory are available to predict microgel swelling. However, the number of required parameters is high. For example, Choi et al. w34x requires nine for their treatment of copolymer gels, and none of the theories seem to have dealt with the non-uniformity of real gels.

14

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

3.2.2. Swelling measurements PolyNIPAM microgels can be prepared with a particle size distribution of narrow dispersity in the submicron size range. Thus, dynamic light scattering DLS. is a very convenient technique for measuring swelling, and there have been many microgel papers with DLS data. However, swollen particle diameters often approach 1000 nm that is outside the sensitive range for DLS. Furthermore, swelling changes are usually computed as volume changes, so the DLS diameters must be cubed which limits the accuracy of the volumetric data. DLS measurements give particle size, whereas the concentrations of polymer and water in the gel are usually the required quantities. Therefore, more information is required to relate particle volume to polymer or water content of the microgel. Some have avoided this problem by reporting swelling volume or diameter ratios calculated by comparing particle size to a reference particle size. However, if the average mass of polymer per gel particle is known, then the diameters are easily converted to absolute measures of swelling. A few approaches have been applied to the measurement of absolute water content in microgels. McPhee et al. w37x centrifuged polyNIPAM microgels to give an ordered, packed bed of particles, which was iridescent, indicating ordered packing. The water content of the bed was calculated from the bed mass before and after drying. Another approach is to measure the particle size at a temperature above the VPTT and assume a water content based on macrogel data, such as that given by Dong and Hoffman w63x. For example, consider a microgel with a particle diameter of 100 nm at 45 C and a corresponding diameter of 200 nm at 20 C. If we assume that the mass concentration of water in the particles at 45 C is 25%, then the corresponding concentration of water in the particles at 20 C is 0.8912. Note that this calculation is based upon the assumptions that: the density of polyNIPAM is 1269 kgrm3 w60x, the polyNIPAM density is independent of temperature over this range; and, that there is no excess volume of mixing. If the actual water content is 30% instead of 25% at 45 C i.e. an error of 20%., the corresponding water content of the particles at 20 C is 0.8996. Thus, a large 20%. error in the estimate of the water content at 45 C corresponds to only a 0.94% error in the estimated water content at 20 C. It seems therefore that this is a robust method for estimation of microgel water content in swollen gels. The water content in swollen polyNIPAM microgels has been determined by combining the results of two light scattering techniques w23x. Intensity light scattering was used to give a molecular weight of the polymer per particle and dynamic light scattering to give the corresponding hydrodynamic volume per particle. The first published DLS data for polyNIPAM microgels as functions of temperature and electrolyte concentration were by Pelton et al. w64x. Their results, shown in Fig. 9, indicate that from 10 C to approximately 30 C the diameter decreases by linearly by approximately one-quarter whereas from 30 to 35 C the diameter decreases by another quarter in monovalent electrolyte. The VPTT is usually taken as the steepest portion of the diameter vs. temperature curve. In 0.1 M CaCl 2 the low temperature swelling was less than in more dilute KCl. Furthermore, the VPTT was lower in the presence of 0.1 M CaCl 2 this is consistent with the work

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

15

Fig. 9. Average particle diameter of polyNIPAM microgel dispersions determined by dynamic light scattering at three electrolyte concentrations. Data from Pelton et al. w64x.

of Park and Hoffman w65x who have documented the influence of electrolytes on the VPTT of macrogels. Above 35 C the average diameter in 0.1 M CaCl 2 dramatically increased due to the aggregation of the shrunken polyNIPAM microgel particles. The extent to which a gel can swell is limited by the presence of cross-links usually based on BA see Fig. 1 for chemical structure .. McPhee et al. w37x published microgel diameter vs. temperature curves for a series of microgels with varying BA content an example of their results are shown in Fig. 10. As expected, the swelling ratios i.e. the diameters at 20 C divided by the diameters at 40 C. decreased with increasing BA content. Note the differences in diameters at high temperature in Fig. 10 do not reflect the influence of BA on water content above the VPTT, but instead reflect the influence of BA on the particle nucleation during latex synthesis. The lowest BA concentration gave the smallest particles at high temperature and thus gave the highest particle concentration during gel synthesis. The swelling data were interpreted by Florys gel theory. McPhee and colleagues results were recently confirmed by Kratz and Eimer w25x and Oh and Bae w66x who both neglected to mention the McPhee paper and some of the other earlier work. Wus group in Hong Kong w23,67x employed dynamic light scattering to compare the temperature sensitivity of polyNIPAM microgels to linear polyNIPAM. VPTT for the gel was slightly greater than the lower critical solution temperature for linear polyNIPAM and the transition was less sharp for the microgel. The rate of change of microgel diameter with a step change in temperature was given in the same paper. Approximately 10 ms were required for swelling or shrinking, however, the authors state that this reflects the time scale of the temperature change. Thus, microgel response seems to be very fast indeed. The same group has investigated the swelling dynamics of macroscopic poly NIPAM film formed from microgel latex w22x. The gels shrink rapidly and then the released water evaporates. In would be interesting to compare drying kinetics in

16

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 10. The influence of BA content on polyNIPAM microgel swelling w37x.

this system with film based gels such as polyacrylamide which do expel water upon heating. DSC has long been employed to characterize swelling in polyNIPAM macrogels w68x and linear polymers w69x. For example, Dong and Hoffman w68x employed DSC to estimate the bound water content as a function of temperature based on the assumption that bound water does not freeze and hence does not contribute to the heat of fusion. The Bristol University group is the first to apply DSC of poly NIPAM microgels w35x. They showed that the endotherm upon heating was identical to the exotherm upon cooling. Furthermore, the DSC results indicated a two-step process when heating polyNIPAM microgels containing 5 mol% based on NIPAM. acrylic acid. Many potential applications of polyNIPAM microgels may involve the dried gels. Agbugba et al. w70x systematically studied the properties of freeze-dried microgels. In general, freeze-drying did little damage and some types of microgels were more susceptible than others. Interestingly, they found it impossible to remove the bound water by freeze drying thus, conventional drying at high temperature is the only sure way to remove all the water. In the same paper is the first-reported moisture adsorption isotherm for freezedried microgels at both below and above the VPTT. Their results converted to absolute water contents mass waterrmass polymer. are shown in Fig. 11. The maximum water content was 1.25 grg at a relative humidity of 93% at 25 C. Fig. 11 shows that much more water was sorbed at 25 than at 50 C. Also, the sorption runs showed hysteresis typical of a vapor adsorption on a porous solid. 3.2.3. Swelling in other sol ents PolyNIPAM linear polymers and macrogels display interesting phase behavior called cononsolvency in alcohol water mixtures w71 73x. This behavior is illustrated in Fig. 12; the addition of methanol to aqueous polyNIPAM lowers the cloud point temperature of linear polymers to below 0 C for 55% by volume methanol. Higher methanol concentrations yield a dramatic increase in the cloud

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

17

Fig. 11. The moisture adsorption isotherm for freeze-dried microgel. Data calculated from results of Agbugba et al. w70x.

point temperature or the VPTT to values greater than in water. Winnik et al. w71x proposed that the phase behavior was because of the preferential adsorption of methanol on the polymer chains. McPhee et al. w37x reported polyNIPAM microgel swelling vs. temperature curves for up to 35% methanol. These results are compared with the LCST behavior of linear polyNIPAM w71x in Fig. 12. With the exception of the highest methanol concentration 35% vrv. the LCST of the linear polymer corresponded to the temperature at which the microgel vs. diameter curve had the largest negative slope. This comparison emphasized the general trend that the swelling behavior of the microgels is dominated by the interaction of the polyNIPAM segments with the solvents. Zimehl and Mielke w20,74x report isothermal swelling curves for polyNIPAM microgels as functions of the concentration of methanol, ethanol and the two isomers of propanol. Both cationic and anionic microgels were studied. They showed very interesting differences in the response of cationic and anionic microgels to propanol. However, the microgels were prepared with surfactants that were not removed. The propanol results should be repeated with clean microgels.

Fig. 12. Comparison of cloud point temperatures for polyNIPAM w71x with VPTT for polyNIPAM microgels w37x.

18

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Crowther and Vincent w27x reported polyNIPAM microgel swelling behavior as functions of temperature and the concentration of methanol, ethanol and isopropanol. Insufficient data were presented to give VPTT values as functions of alcohol content nor was any attempt made to relate the results to the significant body of macrogel swelling results and theory. In summary, polyNIPAM gels show interesting behavior in mixed alcohol water solutions. Like the water-swelling behavior, there is no evidence of any significant differences between the steady-state degree of swelling for microgels and macrogels as functions of temperature and alcohol content. Of course the microgels will approach steady-state swelling much more quickly. 3.3. Surface acti ity Linear poly N-isopropylacrylamide. is surface active aqueous solutions have a surface tension of approximately 42 mJrm2 and this value is not very temperature-dependent through the LCST of the polymer solution w6,75x. There are many examples of surfactants and water-soluble polymers that lower the surface tension of water. On the other hand, there are only a few examples of surface-active particles and these are usually very hydrophobic particles suspended on a Langmuir trough w76,77x. PolyNIPAM microgels in water share the features of both colloidal particles and soluble polymers, so we were curious to know whether microgels were surface active. Fig. 13 shows aqueous surface tensions for polyNIPAM microgel suspensions as functions of time w78x. A variety of gel morphologies was obtained by varying the amount BA cross-linker and the type polymerization batch or semibatch.. The drawings beside the curves indicate the target morphology. In all cases the microgels lowered the surface tension of water to values close to values obtained with linear polymer. The time required to achieve a steady state surface tension increased with the particle cross-linking. Direct evidence for surface activity was obtained by examining water surfaces in an environmental scanning electron microscope. Fig. 14 shows one of the images obtained for the most cross-linked microgels w79x. Clearly, the particles formed an ordered array at the airrwater interface. 3.4. Rheological properties The volume of aqueous polyNIPAM gels is a sensitive function of temperature. Therefore, one would expect the rheological properties of polyNIPAM microgel suspension to be sensitive to temperature. Kiminta et al. w21x have reported the rheological behavior of homogeneous polyNIPAM microgels as functions of temperature, shear rate and concentration. The elastic modulus decreased by an order of magnitude between 28 and 50 C reflecting the decrease in the effective volume fraction of dispersed gel phase. Lowe et al. w50x and Zimehl and Mielke w20,74x have published rheological data that show the expected decrease in viscosity with increasing temperature.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

19

Fig. 13. The influence of microgel morphology on surface tension vs. time. Data taken from Zhang and Pelton w78x.

3.5. Electrical properties The electrostatic properties of aqueous colloids influence colloidal stability and colloid interactions with dissolved polymers and surfaces. PolyNIPAM microgel dispersions have fascinating electrical properties due to the presence of covalently bonded electrically charged groups originating from the initiator. McPhee et al.

Fig. 14. Environmental Scanning Electron Micrograph of polyNIPAM microgels at the airrwater interface w79x. With permission of American Chemical Society.

20

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 15. The influence of temperature on the electrophoretic mobility of polyNIPAM microgel latex. Data taken from Pelton et al. w64x. Parameters: Eq. 1. e s 8.8 = 10y1 6 coulomb; Eq. 2. N s 4.5 mmol, f s 6 3 nm; and, Eq. 3. N s 4.5 mmol, s 0.5rnm.

w37x reported charge contents for a persulfate initiated polyNIPAM microgel to be 3.8 eq.rg of which three-quarters were carboxyl groups and one-quarter were sulfate groups. These charge contents are two orders of magnitude lower than a typical surfactant-free polystyrene latex because the NIPAM propagates to much longer chain lengths than styrene so there are many less chain ends bearing initiator fragments. Fig. 15 shows the first reported electrophoretic mobility data for polyNIPAM latex as a function of temperature and electrolyte concentration. At room temperature the mobility is nearly zero reflecting the low charge density in the particles, whereas above the VPTT the mobility increases by an order of magnitude. Since the electrical charges are covalently bonded, the total charges per particle are constant. On the other hand, the density of charges increase through the VPTT because the charges must be distributed through 1r10 the volume available below the VPTT. A few efforts have been made to model the temperature-dependent electrophoretic mobility results such as those shown in Fig. 15. A simple model, summarized by Eq. 1. w64x, relates the electrophoretic mobility of the microgels, , to particle radius, r, and the number of charges per particle, . The other parameters are: e is the charge of an electron; is the viscosity; and is the Debye Huckel parameter. This equation was based on the assumptions that: all the charges are located on the exterior surface of the microgel; the charge density is related to the potential by the Helmholtz equation s o r ; the surface potential equals the zeta potential; and, the electrophoretic mobility is related to the zeta potential by the Smoluchowski equation. If the microgel diameter is known as a function of temperature, then only , the total number of charges per particle is needed to calculate the electrophoretic mobility with Eq. 1.. The mobility vs. temperature data were fitted by Eq. 1. using the corresponding gel

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

21

diameterrtemperature data Fig. 9.. The computed curve is compared with the experimental data in Fig. 15. Note, that the noise in the computed curves arises from the noise in the gel diameter vs. temperature data used in the calculations. The fitted value of , expressed as a specific charge content based on dry polyNIPAM was 36 eq.rg which is approximately 10 times higher than McPhees w37x experimental value of 3.8 eq.rg. Therefore, Eq. 1. gives a reasonable prediction of experimental behavior, however, the predicted gel particle charge contents are too high. s y e 4 r2

1.

The applicability of Eq. 1. is questionable since there is no reason to assume that all the electrical charges are located on the exterior surface of the microgel particles, particularly in the highly swollen state at low temperature. Two more realistic models have been published and are now compared. Eq. 2. was given by Buscall et al. w80x to describe the electrophoretic mobility of a polyelectrolyte particle based on the theory of Hermans and Fujita, where f is the friction factor per repeat unit, b s fNr and N is the number of electrically charge groups per unit volume.

'
2 3

ye 3 f

q3 3

b q 2 b2 q b3 q3
2

2.

Eq. 2. was fitted to our data assuming a friction factor calculated from the Stokes equation for a sphere of 3 nm, the water content of the polyNIPAM gel particles was 20% by mass. Experimental diameter vs. temperature results shown in Fig. 9 were used to calculate the temperature-dependent volumetric charge density together with an assumed specific charge content of 4.5 eq.rg. The predicted electrophoretic mobility vs. temperature curve is compared with the experimental data in Fig. 15. The mobility values computed with Eq. 2. were a little lower than the data at low temperature and higher than the data at high temperature. Eq. 2. predicted the same mobilities as did Eq. 1. up to 32 C, the VPTT. Oshima et al. w81x presented an alternative gel layer model for electrophoresis of polyNIPAM microgels. Their model is described by the following equations. s
o r

or

.q

Don r

1r

. q 1r .

zen
2

3.

where N is the volumetric concentration of charge, is a softness parameter with units of meter, is the valency of the counter ions, n is the supporting electrolyte concentration, is the Debye parameter, and the remaining terms are given by the

22

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

following equations s kT e zN 2 n q

ln

/ 5
2 n q1

zN

1r2

2 n zN

1y

/ 5
2 n q1

zN

1r2

4.

DO N

kT e

ln

zN 2 n

/ 5
2 N q1

zN

1r2

5.

1q

/
2 n

zN

2 1r4

6.

Although this set of equations is a little more complicated than Eq. 2., this is also a two parameter model, N and . As before, N was calculated on the assumption that the specific charge content of the gel was 4.5 eq.rg whereas was considered to be an adjustable, temperature-independent parameter. The curve predicted by Eq. 3. is also shown in Fig. 15 assuming s 5 = 10 8rm. This model gave a slightly better prediction of the experimental results than did the previous models. Note, Eqs. 1. and 3. gave very similar predictions about the VPTT. Nabzar et al. w32x reported electrophoretic mobilities as functions of electrolyte concentration for polystyrene-core, polyNIPAM-co-aminoethylmethacrylate hydrochloride shell latexes. They claim that Oshimas theory was unable to fit their results for a single pair of values for N and . Their data are compared with the predictions of Eq. 3. in Fig. 16. Neither theoretical curve fits the experimental results. In summary, polyNIPAM microgels provide an extensive database of electrophoretic mobility values as functions of temperature, and electrolyte and surfac-

Fig. 16. Comparison of experimental and theoretical mobility data as functions of electrolyte. The data taken from Nabzar et al. w32x Latex DD4 Fig. 6. and the calculations are based on Eq. 3.. N s 0.52 molrl.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

23

tant concentration. It seems unlikely that electrophoresis will provide detailed microstructure information about the gels. Instead, gel data seem to be a tough test for electrokinetic models. 3.6. Gel surfactant interactions The temperature sensitivity of polyNIPAM and related materials arises from the fact that the polymer has considerable hydrophobic character. This same characteristic also evokes strong interactions with surfactants in water. Eliassaf w82x was the first to report the interaction of aqueous surfactant with polyNIPAM. Schild and Tirrell see references in Schild w6x. have investigated the influence of surfactant chain length. In general, both cationic and anionic surfactants bind to polyNIPAM whereas non-ionic polymers seem not to. The cloud point of linear polyNIPAM increases with anionic surfactant binding but is less influenced by cationic surfactants. Surfactants also bind to polyNIPAM microgels. Fig. 17 shows a binding isotherm for SDS to polyNIPAM microgel w58x. One of the advantages of studying surfactant binding to microgels compared with binding to linear polymers is that microgels are easily separated by centrifugation which, in turn, allows direct analytical determination of the amount of bound surfactant. Of course, less direct methods such as conductivity w15x, rheology w83x, and NMR w22x also give information about surfactant binding. Microgel swelling dramatically increases with bound surfactant. Fig. 18 shows microgel swelling as functions of sodium dodecyl sulfate SDS. concentration w30x and these results have been verified by others w22,23,42x. At all temperatures SDS binding gave increased swelling and also tended to raise the VPTT. Careful work with state-of-the-art DLS measurements indicates two inflection points when polyNIPAM microgel radius is plotted as a function of temperature in the presence of SDS w23x. The dramatic influence of surfactant on the VPTT may limit the utility

Fig. 17. Binding isotherm for SDS onto polyNIPAM microgels at 23 C. Data taken from Mears et al. w57x.

24

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 18. The influence of SDS on the swelling of polyNIPAM microgels w30x. With permission of the American Chemical Society.

of temperature-sensitive gels in formulated products with high surfactant concentrations. In general, discussions of the interactions of polyNIPAM interactions with SDS have drawn on the body of published work on SDS binding to polyethyleneoxide w84x. In particular, it is often assumed that polymer-bound surfactant is present as rather large micelles w22x. Mears et al. w57x reported neutron scattering studies in which the polyNIPAM gel was contrast matched with the waterrD 2 O mixture and scattering off deuterated SDS was observed. Surprisingly, the neutron results indicated that the bound surfactant was present as very small aggregates containing less than six surfactant molecules. Thus, polymer bound micelles do not appear to exist in the polyNIPAM microgel system it would be of interest to determine if this conclusion also holds for linear polyNIPAM. Finally, in very recent work we have reported isothermal calorimetric titration results w85x. In this technique, SDS is slowly added to a microgel suspension while recording the heat effect. Fig. 19 shows the heat effect with SDS addition to microgel above and below the VPTC. The binding of SDS to polyNIPAM microgel was endothermic below the VPTC and exothermic above it. Heat effects were observed at even the lowest SDS addition levels which indicates that the critical aggregation concentration CAC. for SDS binding to polyNIPAM is below 17 mmol or there is no CAC. 3.7. Colloidal stability Temperature-sensitive colloidal stability has many potential applications including coating, papermaking, selective wetting, etc. For example, latex rubber gloves are made by dipping a mold into a concentrated latex suspension. The mold is coated with coagulant that induces latex deposition and coalescence. A potential problem with this technology is that residual coagulant can remain inside the glove. On the other hand, if the rubber latex had temperature-sensitive colloidal stability,

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

25

Fig. 19. Isothermal titration of polyNIPAM with SDS below and above the VPTT. Data taken from Wang et al. w85x.

rubber gloves could be made by dipping hot glove molds into a cold latex bath. PolyNIPAM coated latexes exhibit temperature dependent colloidal stability w40x. Aqueous colloids are considered stable if they do not aggregate in the time scale of interest. Colloid stability depends upon the balance of van der Waals attraction, which causes aggregation, and steric or electrostatic forces that oppose aggregation. Below the VFTT, polyNIPAM microgels are colloidally stable at low and high electrolyte concentration, whereas above the VPTT, the microgels are only stable in low electrolyte concentrations. For example, Pelton and Chibante w4x showed that microgel suspensions in 0.1 M CaCl 2 were colloidally stable until the temperature was raised above 34 C see Fig. 9.. This behavior is now explained by standard concepts of colloid science. Below the VPTT the particles are swollen and thus mainly water see Fig. 2.. Under these conditions, the van der Waals attractive forces are relatively small. Furthermore, it seems reasonable to propose that polymer tails extend from the gel structure to act as steric stabilizers further enhancing colloid stability. Note that for surfactant free polyNIPAM microgels, the electrophoretic mobility is low see Fig. 15. w64x indicating that electrostatic repulsion does not contribute to colloidal stability below the VPTT. At elevated temperatures the water content of the gels is reduced giving a higher density and thus a greater Hamaker constant than at low temperature. The greater the Hamaker constant, the greater are the attractive van de Waals forces tending to aggregate the gels. Above the VPTT any polyNIPAM tails on the microgel surface will be collapsed upon the particle surface and thus not contributing to colloidal stability. On the other hand, at low ionic strength the electrophoretic mobility of polyNIPAM microgels is high see Fig. 15. indicating that electrostatic stabilization is operative. CaCl 2 induces coagulation above the VPTT because the electrical double layer is compressed giving low electrostatic interactions. The electrical charges in surfactant-free polyNIPAM are sulfate and carboxyl groups that originate from the ionic free radical polymerization initiator. Presumably,

26

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

microgels prepared from non-ionic initiators would be colloidally unstable above the PVTT without electrolyte addition. This remains to be proven. PolyNIPAM colloidal microgels exhibit extreme colloid stability below the VPTT, however they can be aggregated by polymer addition. Snowden and Vincent w18x showed that sodium polystyrene sulfonate gave depletion flocculation. Similarly, polyNIPAM- shell polystyrene-core latexes can be flocculated by very high molecular weight cationic polyacrylamide copolymers Cong and Pelton, unreported results .. Islam et al. w31x reported the temperature-sensitive heteroflocculation of anionic polystyrene latex with cationic polyNIPAM microgel the mixed dispersion was colloidally stable at low temperature and it aggregated above the VPTT. An unexpected observation was that flocs formed at high temperatures could be redispersed at low temperature only when the pH was 10 or greater. This could be due to a decrease in electrostatic attraction either induced by increased ionic strength or, more likely, to the alkaline hydrolysis of cationic amidine groups. There are many parallels between this work and that of Deng et al. w39x, who reported the aggregation of anionic colloidal titanium dioxide with cationic linear copolymers of polyNIPAM. Above the cloud point temperature, the copolymer was present as phase separated colloidal particles with an average diameter of 87 nm. Like the more recent Islam and Snowden results, cationic polyNIPAM particles heteroflocculated with the anionic TiO 2 . Below the cloud point, dilute soluble cationic polyNIPAM copolymer also induced aggregation of the polystyrene latex presumably by a bridging mechanism. This behavior is in contrast to the behavior of cationic microgels which did not aggregate latex below the VPTT w31x. Presumably the water swollen cationic microgel had too low a cationic content to give adsorption and thus bridging flocculation below the VPTT. Zhu and Napper w42,43x used dynamic light scattering to study the aggregation kinetics of polystyrene-core polyNIPAM-shell latex aggregation. Aggregation was initiated by either increasing temperature at constant ionic strength or by increasing ionic strength at constant temperature. The fractal dimensions of the aggregates were estimated from the slopes of log log plots of average aggregate size vs. time. The behaviors of PEG and polyNIPAM stabilized polystyrene were compared. For PEG, the fractal dimension of the aggregated latex was independent of the electrolyte concentration whereas the polyNIPAM latex showed an increase in floc fractal dimension with electrolyte concentration. This was explained in terms of the attractive interactions between polyNIPAM globules on the latex surface. PolyNIPAM often displays specific ion effects. Zhu and Napper w43x showed that the electrolyte concentration required to generate a given floc fractal dimension followed the trend NaCl - NaBr - NaNO3 - NaI - NaSCN. According to the authors, above the VPTT flocculation is induced by an attraction between poly NIPAM shells on neighboring particles driven by the release of bound water present below the VPTT. Specific ion effects arise from the ions changing the amount of bound water released with heating. The authors supported these arguments by illustrating a correlation between the salt concentration required to

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

27

give a specific fractal dimension to relative effects of the same series of salts on aqueous solution viscosity. Nabzar et al. w32x recently reported the aggregation kinetics of polystyrene-core, polyNIPAM-co-aminoethyl methacrylate-shell latexes. Stability ratios, based on turbidity measurements were presented as functions of electrolyte concentration both below and above the VPTT. In general, the results were consistent with previous work and are consistent with the expected behavior of a electro-sterically stabilized colloid below the VPTT and an electrostatically stabilized colloid at high temperature. Combined Hamaker constants were calculated from the coagulation kinetics using classical models. Modern approaches to obtaining Hamaker constants or functions. involve the application of Lifshitz theory to spectroscopic data. No such analysis has been done for polyNIPAM as a function of temperature or water content.

4. Microgel applications 4.1. Biotechnology The first reported investigation of the interaction of biological molecules with polyNIPAM microgels was by Kawaguchi et al. w19x who described the temperature-dependent sorption and desorption of human globulin. Fig. 20 shows that the maximum protein binding occurs at low pH and 40 C which is above the VPTT. Presumably binding was driven by hydrophobic interactions between the protein and isopropyl groups on the gel. When sorption occurred at 40 C, desorption could be induced by lowering the temperature to 25 C. However, the extent of desorption depended on the length of time used in the sorption step. The longer the sorption time at high temperature, the lower the amount of protein which desorbed at a low temperature. This possibly indicates that the protein slowly entered the dehydrated gel structure above the VPTT. In a subsequent article the same group compared the interactions of four proteins with polyNIPAM microgel and polystyrene latex w87x. Protein binding to polyNIPAM microgel was higher at 40 C than at 25 C. However, all four proteins showed a higher degree of adsorption onto polystyrene latex than on polyNIPAM at 40 C. The activity of peroxidase desorbed from polyNIPAM was greater than that of the enzyme desorbed from polystyrene. Thus polyNIPAM had a lower tendency than polystyrene to denature the enzyme. The studies, summarized above involved the physical sorption of proteins to polyNIPAM microgel. A number of recent reports have addressed the properties of enzymes covalently bonded to microgels. Shiroya et al. w33x grafted trypsin and peroxidase to polyNIPAM microgels. The enzyme activity went down when temperature was raised above the VPTT. This was explained by decreased substrate diffusion rates and by trapping of the enzyme in the surface layer. On the other hand, temperature-independent enzyme activities were obtained when the enzymes

28

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

Fig. 20. Influence of pH and temperature on human gamma globulin adsorption onto polyNIPAM microgels. Adapted from Kawaguchi et al. w19x.

were attached to the microgels via PEG spacers. Presumably the spacers isolate the enzymes from the temperature-sensitive polyNIPAM domains. A particularly clever approach to obtaining temperature-sensitive enzyme activity was described by Yasui et al. w48x who attached trypsin to latex surfaces with polyNIPAM linear chains MW s 2500 or 11 000.. PolyNIPAM chains not bearing enzymes surrounded the tethered enzymes. The system had two characteristic temperatures. The surface polyNIPAM chains underwent a coil-to-globule transition at ; 30 C whereas the enzyme terminated chains had a transition temperature of 38 C. The main observation was that enzyme activity increased when the temperature was raised above 30 C. The explanation is illustrated in Fig. 21. At low temperature the enzyme is surrounded by polyNIPAM which inhibits substrate transport, whereas above 30 C the enzyme-free surface polyNIPAM chains are present as compact globules leaving the enzyme terminated chains to extend into solution with unrestricted access to substrate which gave rapid kinetics. Granulocytes are cells that attack foreign bodies. Achiha et al. w88x compared the interaction of granulocytes with micro spheres latexes . including polystyrene, polystyrene-core polyNIPAM-shell, polyNIPAM microgel, and acrylamide microgel particles. The polystyrene spheres interacted strongly at both 20 and 37 C as evidenced by oxygen consumption, strong cellrlatex adhesion leading to phagocytosis engulfment of particles.. By contrast the hydrophilic acrylamide-based microgels showed little interaction at both temperatures. PolyNIPAM microgels or core shell spheres displayed temperature-sensitive interaction with the granulocytes. Below the VPTT the polyNIPAM particles were relatively inert, similar to the acrylamide based gels, whereas at 37 C the polyNIPAM spheres interacted strongly with the granulocytes. One of the most promising applications of polyNIPAM microgels is as support materials for biological testing. Pichots group has described the coupling of oligodeoxyribonucleotides ODN. to polyNIPAM microgels w89x. The microgels

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

29

Fig. 21. Schematic illustration of temperature dependent enzyme activity for polyNIPAM tethered enzymes. A key feature of the effect is that the cloud point temperature of enzyme terminated polyNIPAM is 38 C compared with 32 C for polyNIPAM chains without terminal enzyme w48x.

have an affinity for specific DNA that is determined by the ODN structure. Higher sensitivities were obtained with polyNIPAM supports than with polystyrene. Ironically this application does not directly employ the temperature sensitivity of the microgels. Instead, the key property of the particles is their hydrophilicrhydrophobic nature coupled with the fact that they can be prepared as uniform spheres. 4.2. Controlled uptake and release The temperature-sensitive uptake or release of chemicals has been one of the most intensively investigated potential applications for both microgels and macrogels. As discussed in earlier sections, microgels, unlike macrogels, exhibit very rapid swelling or shrinking in response to temperature change. Thus, temperature-triggered drug or chemical release is possible. Snowden w90x showed that fluorescein labeled dextran bound to polyNIPAM microgels below the VPTT and was subsequently released when the temperature was raised above the VPTT. Dextran has no affinity for polyethylene glycol. so it seems unlikely that it would interact with polyNIPAM. Thus, the dextran binding possibly was due to hydrophobic interactions between the fluorescein label and polyNIPAM. In a subsequent publication, Snowden and Booty w91x measured the binding of nitrate ions onto cationic microgels and there was no release of nitrate upon heating. Also investigated was the sorption of acetylsalicylic acid and

30

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

aluminum citrate onto anionic polyNIPAM microgels. Two-thirds of the material bound at room temperature was removed by heating to 40 C. No mechanism or modeling was given to explain the results. One would expect the diffusion rate of small molecules into or out of poly NIPAM microgels should be lowest at temperatures above the VPTT because of the relatively high concentration of polymer chains. This is illustrated in the results of Kato et al. w92x who studied the diffusion rate of ascorbic acid into composite particles consisting of a polyNIPAM shell around a core of cytochrome c embedded in a matrix of polystyrene-co-2-ethylhexyl methacrylate. The initial diffusion rate of ascorbic acid into the microgel spheres is shown in Fig. 22 as a function of temperature. The rates were much faster below the VPTT of the shell. Furthermore, transport was slower when the polyNIPAM shell was cross-linked. In summary, the mass transport rate of solutes through polyNIPAM microgel will be faster below the VPTT where the gel has a high water content and a low effective viscosity. On the other hand, solutes that undergo hydrophobic bonding will be released above the VPTT because soluterNIPAM interactions are replaced by NIPAMrNIPAM interactions. A note of caution-solute binding usually modifies the VPTT. 4.3. Other potential applications It seems reasonable to expect some unique applications for uniform, temperature-sensitive microgels. Some possibilities are now considered.

Fig. 22. The initial rate of ascorbic acid transport into polyNIPAM-coated cytochrome c particles as a function of temperature. Data taken from Kato et al. w92x.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

31

Snowden et al. w93x have a European patent application for the use of poly NIPAM microgels to modify the permeability of oil reservoirs. In this application the microgels aggregate to form precipitates at temperatures above the VPTT. Chen et al. w47x described the preparation of Pt catalysts supported on polystyrene-core, polyNIPAM-shell particles. The polyNIPAM chains retarded the migration and agglomeration of the 1.5-nm Pt particles. Because of this effect, catalysts were more long-lived than those supported on bare polystyrene latex. Furthermore, the catalytic activity was temperature-dependent. Reaction rates were lower above the LCST of the polyNIPAM shell. Presumably, the collapsed shell inhibited interactions between the catalyst and substrate w94x. Finally, Ashers group has shown that arrays of temperature-sensitive microgels, such as shown in Fig. 4 give temperature-sensitive Bragg diffraction which may have applications in non-linear optics w24x. 5. Closing remarks The fascinating properties of temperature-sensitive microgels are likely to continue to generate activity in the literature. Indeed, there are a few obvious opportunities for further work. There have been few attempts to characterize the morphology distributions within and among particles. Morphological characterization is likely to lead to semi-batch and other techniques for controlling gel morphology to give more exotic particle morphologies. However, in my view the biggest opportunity is the identification of new applications for temperature-sensitive microgels. Indeed, a motivation for writing this review was the belief that it might stimulate new applications. Acknowledgements My work in this area has been performed by a talented group of students and postdoctoral fellows. I am indebted to Karinne Chan, Philip Chibante, Yulin Deng, Wayne McPhee, Sara Mears, Michael Tam, Shirley Wu and Ju Zhang. References
w1x w2x w3x w4x w5x w6x w7x w8x w9x w10x M.S. Wolfe, C. Scopazzi, J. Colloid Interface Sci. 133 1989. 265 277. T.G. Park, A.S. Hoffman, J. Polym. Sci. 30 1992. 505 507. F. Hoshino, T. Fugimoto, H. Kawaguchi, Y. Ohtsuka, Polymer J. 19 1987. 241 247. R.H. Pelton, P. Chibante, Colloids Surf. 20 1986. 247 256. M. Heskins, J.E. Guillet, J. Macromol. Sci. Chem. A2 8. 1968. 1441 1455. H.G. Schild, Prog. Polym. Sci. 17 1992. 163 249. Y. Osada, S.B. Ross-Murphy, Sci. Am. May 1993. 82 87. I.Y. Galaev, Russian Chem. Rev. 64 5. 1995. 471 489. R. Dagani, C and EN, June 9 1997. 26 30. R. Pelton, X. Wu, W. McPhee, K.C. Tam, in: J.W. Goodwin, R. Buscall Eds.., Colloidal Polymer Particles, Academic Press, 1995, pp. 81 99.

32

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33

w11x H. Kawaguchi, in: M. Yalpani Ed.., Biomedical Functions and Biotechnology of Natural and Artificial Polymers, ATL Press, 1996, pp. 157 168. w12x H. Kawaguchi, in: R. Arshady Ed.., Microbeads, Microcapsules and Liposomes, STN Books, London, 1998, pp. 233 248. w13x M.J. Murray, M.J. Snowdon, The preparation, characterization, and applications of colloidal microgels, Adv. Colloid Interface Sci. 54 1995. 73 91. w14x W.C. Wooten, R.B. Blanton, H.W. Coover, Jr., J. Polym. Sci. XXV 1957. 403 412. w15x X. Wu, R.H. Pelton, A.E. Hamielec, D.R. Woods, W. McPhee, Colloid Polym. Sci. 272 1994. 467 477. w16x G. Bokias, A. Durand, D. Hourdet, Macromol. Chem. Phys. 199 1998. 1387 1392. w17x J.W. Goodwin, R.H. Ottewill, R. Pelton, G. Vianello, D. Yates, Br. Polym. J. 10 1978. 173 180. w18x M.J. Snowden, B. Vincent, J. Chem. Soc, Chem. Commun. 1992. 1103 1105. w19x H. Kawaguchi, K. Fujimoto, Y. Mizuhara, Colloid Polym. Sci. 270 1992. 53 57. w20x M. Mielke, R. Zimehl, Prog. Colloid Polym. Sci. 111 1998. 74 77. w21x D.M. Ole Kiminta, P.F. Luckham, S. Lenon, Polymer 36 25. 1995. 4827 4831. w22x C. Wu, S. Zhou, J. Polym. Sci., Part B Polym. Phys. 34 1996. 1597 1604. w23x C. Wu, S. Zhou, S.C.F. Au-Yeung, S. Jiang, Die Ang. Makro. Chemie 240 1996. 123 136. w24x J.M. Weissman, H.B. Sunkara, A.S. Tse, S.A. Asher, Science 274 1996. 950 959. w25x K. Kratz, W. Eimer, Ber. Bunsenges Phys. Chem. 102 1998. 848 854. w26x G. Zhou, A. Elaissari, Th. Delair, C. Pichot, Colloid Polym. Sci. 276 1998. 12, 1131 1139. w27x H.M. Crowther, B. Vincent, Colloid Polym. Sci. 276 1998. 46 51. w28x Gilbert, Emulsion Polymerization: A Mechanistic Approach, Academic Press, London, 1995. w29x T. Gotoh, Y. Nakatani, S. Sakahara, J. Appl. Polym. Sci. 69 1998. 895 906. w30x K.C. Tam, S. Ragaram, R.H. Pelton, Langmuir 10 1994. 418 422. w31x A.M. Islam, B.Z. Chowdhry, M.J. Snowden, J. Phys. Chem. 99 1995. 14205 14206. w32x L. Nabzar, D. Duracher, A. Elaisssari, G. Chauveteau, C. Pichot, Langmuir 14 1998. 5062 5069. w33x T. Shiroya, N. Tamura, M. Yasui, K. Fujimoto, H. Kawaguchi, Colloids Surf. B 4 1995. 267 274. w34x H.S. Choi, J.M. Kim, K. Lee, Y.C. Bae, J. Appl. Polym. Sci. 69 1998. 799 806. w35x M.J. Snowden, B.Z. Chowdhry, B. Vincent, G.E. Morris, J. Chem. Soc, Faraday Trans. 92 24. 1996. 5013 5016. w36x F. Meunier, A. Elaissari, C. Pichot, Polym. Advanced Technol. 6 1994. 489 496. w37x W. McPhee, K.C. Tam, R. Pelton, J. Colloid Interface Sci. 156 1993. 24 30. w38x K. Chan, R. Pelton, J. Zhang, Langmuir 15 1999. 4018 4020. w39x Y. Deng, H. Xiao, R. Pelton, J. Colloid Interface Sci. 179 1996. 188 193. w40x R.H. Pelton, J. Polym. Sci. 26 1988. 9 18. w41x K. Makino, S. Yamamoto, K. Fujimoto, H. Kawaguchi, H. Oshima, J. Colloid Interface Sci. 166 1994. 251 258. w42x P.W. Zhu, D.H. Napper, Phys. Rev. E. 50 2. 1994. 1360 1366. w43x P.W. Zhu, D.H. Napper, Colloids Surf. A 98 1995. 93 106. w44x D. Duracher, F. Sauzedde, A. Elaissari, A. Perrin, C. Pichot, Colloid Polym. Sci. 276 1998. 219 231. w45x D. Duracher, F. Sauzedde, A. Elaissari, C. Pichot, L. Nabzar, Colloid Polym. Sci. 276 1998. 10, 920 929. w46x S. Takeuchi, M. Oike, C. Kowitz, C. Shimasaki, K. Hasegawa, H. Kitano, Makromol. Chem. 194 1993. 551 558. w47x C.W. Chen, M.Q. Chen, T. Serizawa, M. Akashi, Chem. Commun. 1998. 831 832. w48x M. Yasui, T. Shiroya, K. Fujimoto, H. Kawaguchi, Colloids Surf. B: Biointerfaces 8 1997. 311 319. w49x D. Duracher, A. Elaissari, C. Pichot, J. Polym. Sci., Part A Polym. Chem. 27 12. 1999. 1823 1837. w50x J.S. Lowe, B.Z. Chowdhry, J. Parsonage, M.J. Snowden, Polymer 39 1998. 1207 1212. w51x H. Kawaguchi, F. Hoshino, Y. Ohtsuka, Chem. Rapid Commun. 7 1986. 109 114. w52x F. Hoshino, M. Sakai, H. Kawaguchi, Y. Ohtsuka, Polym. J. 19 1987. 383 387. w53x T. Serizawa, M.Q. Chen, M. Akashi, J. Polym. Sci. Part A: Polym. Chem. 36 1998. 2581 2587.

R. Pelton r Ad ances in Colloid and Interface Science 85 (2000) 1 33 w54x w55x w56x w57x w58x w59x w60x w61x w62x w63x w64x w65x w66x w67x w68x w69x w70x w71x w72x w73x w74x w75x w76x w77x w78x w79x w80x w81x w82x w83x w84x w85x w86x w87x w88x w89x w90x w91x w92x w93x w94x

33

K. Fujimoto, Y. Nakajima, M. Kashiwabara, H. Kawaguchi, Polym. Int. 30 1993. 237 241. R.G. Sousa, R.F.S. Freitas, W.F. Magalhaes, Polymer 39 1998. 3815 3819. R. Appel, W. Xu, T.W. Zerda, Z. Hu, Macromolecules 31 1998. 5071 5074. S.J. Mears, Y. Deng, T. Cosgrove, R. Pelton, Langmuir 13 1997. 1901 1906. M. Shibayama, K. Kawakubo, F. Ikkai, M. Imai, Macromolecules 31 1998. 2586 2592. F. Ikkai, M. Shibayama, C.C. Han, Macromolecules 31 1998. 3275 3281. A.K. Lele, M.M. Hirve, M.V. Badiger, R.A. Mashelkar, Macromolecules 30 1997. 157 159. T. Hino, J.M. Prausnitz, Polymer 39 14. 1998. 3279 3283. H. Inomata, K. Nagahama, S. Saito, Macromolecules 27 1994. 6459 6464. L. Dong, A.S. Hoffman, J. Controlled Release 30 1990. 21 31. R.H. Pelton, H.M. Pelton, A. Morphesis, R.L. Rowell, Langmuir 5 1989. 816 818. T.G. Park, A.S. Hoffman, Macromolecules 26 1993. 5045 5048. K.S. Oh, Y.C. Bae, J. Appl. Polym. Sci. 69 1998. 109 114. C. Wu, Polymer 39 1998. 4609 4619. L. Dong, A.S. Hoffman, J. Controlled Release 15 1991. 141 152. H.G. Schild, D.A. Tirrell, J. Phys. Chem. 94 1990. 4352 4356. C.B. Agbugba, B.A. Hendriksen, B.Z. Chowdhry, M.J. Snowden, Colloids Surf. A 137 1998. 155 164. F.M. Winnik, H. Ringsdorf, J. Venzmer, Macromolecules 23 1990. 2615 2616. H.G. Schild, M. Muthukumar, D.A. Tirrell, Macromolecules 24 1991. 948 952. F.M. Winnik, M.F. Ottaviani, S.H. Bossmann, M. Garcia-Garibay, N.J. Turro, Macromolecules 25 1992. 6007 6017. M. Mielke, R. Zimehl, Ber. Bunsenges. Phys. Chem. 102 1998. 1 7. J. Zhang, R. Pelton, Langmuir 12 1996. 1611 1612. A. Doroszkowski, R. Lambourne, J. Polym. Sci. Part C 34 1971. 253 264. M.J. Garvey, D. Mitchell, A.L. Smith, Colloid Polym. Sci. 257 1979. 70 74. J. Zhang , R. Pelton, Colloids Surf. 1999. to appear J. Zhang, R. Pelton, Langmuir 1999. in press. R. Buscall, T. Corner, I.J. McGowan, in: Th Tadros Ed.., Effect of Polymers on Dispersion Properties, Academic Press, New York, 1982, p. 379. H. Oshima, K. Makino, T. Kato, K. Fujimoto, T. Kondo, H. Kawaguchi, J. Colloid Interface Sci. 159 1993. 512 514. J. Eliassaf, J. Appl. Polym. Sci. 22 1978. 873 874. K.C. Tam, X.Y. Wu, R.H. Pelton, J. Polym. Sci., Part A Polym. Chem. 31 1993. 963 969. E.D. Goddard, Colloids Surf. 19 1986. 255 300. G. Wang, R. Pelton, J. Zhang, Colloids Surf. 1999. unpublished. J.W. Goodwin, R.H. Ottewill, R. Pelton, Colloid Polym. Sci. 257 1979. 61 69. K. Fujimoto, Y. Mizuhara, N. Tamura, H. Kawaguchi, J. Intelligent Mater. Syst. Struct. 4 1993. 184 189. K. Achiha, R. Ojima, Y. Kasuya, K. Fujimoto, H. Kawaguchi, Polym. Adv. Technol. 6 1995. 534 540. Th. Delair, F. Meunier, A. Elaissari, M.H. Charles, C. Pichot, Colloids Surf. A: 153 1999. 341 353. M.J. Snowden, J. Chem. Soc., Chem. Commun. 1992. 803 804. M.J. Snowden, M.T. Booty, Spec. Publ.- Roy. Chem. Soc. Encapsulation Controlled Rel. 138 1993. 141 147. T. Kato, K. Fujimoto, H. Kawaguchi, Polym. Gels Networks 2 1994. 307 313. M.J. Snowden, B. Vincent, J.C. Morgan, UK Patent Application GB 262 2 117A 1993.. C.W. Chen, M. Akashi, Langmuir. 13 1997. 6465.