- Email: [email protected]
Latex Surface and Bulk Coagulation Induced by Solvent Vapors
Journal of Colloid and Interface Science 228, 171–177 (2000) doi:10.1006/jcis.2000.6920, available online at http://www.idealibrary.com on
Latex Surface and Bulk Coagulation Induced by Solvent Vapors Melissa Braga, Maria do Carmo Vasconcelos Medeiros da Silva, Andr´e Herzog Cardoso,1 and Fernando Galembeck2 Instituto de Qu´ımica, Universidade Estadual de Campinas, P.O. Box 6154, 13083-970 Campinas SP, Brazil Received March 6, 2000; accepted April 17, 2000
erating serious problems. For this reason, we decided to continue the investigation of this effect, looking at other systems beyond those previously described. During this work, we observed a variety of interesting effects of latex exposure to the vapors of organic solvents, which appear highly specific, i.e., they depend on the specific latex and solvent considered. This specificity is a serious complication, because it adds a strong component of unpredictability to the behavior of latexes. On the other hand, since the observed phenomena are highly reproducible, knowledge on the cases already observed may help explain unexpected problems arising both in the laboratory and in the chemical plant. Some new chosen examples are described in this paper.
Latex exposure to solvent vapors leads to highly specific changes in latex stability as well as on the morphologies of the particle association products, depending on the latex and solvent used. Examples of solvent vapor-induced aggregation are given: surface films are obtained on two PS latexes; in one case, the film surface is mirrorreflective and very flat, as evidenced by AFM. Another PS latex coagulates under exposure to acetone vapors, and the morphologies of the coagula are highly sensitive to the exposure conditions. This latex yields a highly porous foam-like structure, in which particles are strongly coalesced but form percolating patches around the pores. The same latex but under other conditions produces a coagulum of large numbers of aggregated particles with a raspberry-like morphology. Density centrifugation experiments show that the effect of solvents on different latex fractions is not uniform, and some fractions show larger density changes than others, thus evidencing a variability in their swelling ability. °C 2000 Academic Press Key Words: latex film formation; PS latex stability; solvent effects on latex stability; solvent-induced coagulation.
EXPERIMENTAL PROCEDURES
Emulsion Polymerization and Latex Preparation
INTRODUCTION
There has been a great progress toward understanding colloidal stability, since the appearance of the DLVO theory (1–3), and the role of Van der Waals, electrostatic interactions, steric interactions, particle bridging, depletion, particle surface hydration layers (4), and the associated hydrodynamic interactions is now well established, in many relevant systems. In a recent work (5, 6), we have shown that the exposure of some latexes to organic solvent vapors increases the stability of the latex dispersion in the presence of salt, and later we showed that the exposure to short-chain alcohol vapors of a redox-initiated, non-ionic surfactant stabilized polystyrene latex, which has a strong stability toward salt, provokes its coagulation (7). These examples show a great susceptibility of latexes to minor contamination by solvent vapors; they are not predicted by the current theories of colloidal stability, and we did not find analogous work previously reported in the literature. However, since solvents are always found in laboratories and chemical plants, their stabilizing effect may well create unpredicted situations for latex users, which have a potential for gen1 2
Present address: Universidade Regional do Cariri, Crato CE, Brazil. To whom correspondence should be addressed.
P[SBMA], PS/PBMA, and PBMA/PS are latexes obtained by the polymerization of styrene (S) and butyl methacrylate (BMA), previously prepared and characterized in this laboratory (8). The first is a random copolymer prepared by simultaneous addition of the two monomers. The two others are core-and-shell latexes, containing separate domains of polystyrene and acrylic. Polymer chain segments are predominantly homopolymeric in these two latexes. PS-M latex was also prepared and characterized in a previous work (9), using Brij 35 and SDS as emulsifiers and potassium persulfate as the initiator. The other latexes were prepared in this laboratory, using procedures from the literature, as follows. Latexes PS-11 and PS-LEV were prepared following Ottewill and Satgurunathan (10), but Renex 300 was substituted for the Levelan stabilizer used by these authors. In the first, a mixture of ascorbic acid and H2 O2 was used as the redox initiator. The yield was 86%, and the solids content of the dispersion was 7.7%. In the second, the initiator was persulfate and NaHCO3 was used to raise the pH. The yield was 99% and solids contents was 20%. PS-THS is a polystyrene latex prepared in the absence of surfactant, using K2 S2 O8 initiator following Tamai et al. (11). The yield of the synthesis was 56% and the final solids content in the dispersion was 1.1%. PS/AAM is a polystyrenepolyacrylamide latex prepared following another paper by Tamai et al. (12), using K2 S2 O8 as the initiator, in the absence of surfactant. Synthesis yield was 76%, and solids content was 8.6%. PS/HEMA is a polystyrene-poly(2-hydroxyethyl methacrylate)
171
0021-9797/00 $35.00
C 2000 by Academic Press Copyright ° All rights of reproduction in any form reserved.
172
BRAGA ET AL.
TABLE 1 Latex Particle Effective Diameters (nm) Following Exposure to Toluene, as Determined by PCSa Time/h
PBMA/PS
PS/PBMA
P[SBMA]
PS-THS
PS-LEV
PS-11
PS-M
PS-HEMA
PS-AAM
0 24 48 96 168 336 504
67.4 ± 0.5 64.9 ± 1.5 67.3 ± 0.7 68.7 ± 0.2 68.3 ± 0.2 68.0 ± 0.1 69.5 ± 0.7
75.1 ± 0.6 74.4 ± 0.6 74.5 ± 0.9 78.3 ± 0.3 77.5 ± 0.3 77.5 ± 0.1 78.6 ± 0.9
73.6 ± 0.1 74.3 ± 0.7 75.3 ± 1.3 78.9 ± 0.6 80.8 ± 0.8 82.8 ± 0.1 77.3 ± 0.9
282 ± 1 284 ± 1 288 ± 3 288 ± 1 282 ± 1 286 ± 8 294 ± 2
432 ± 7 440 ± 6 442 ± 8 453 ± 5 462 ± 7 445 ± 3 453 ± 4
97 ± 1 101 ± 1 103 ± 1 100 ± 1 103 ± 1 102 ± 1 92.4 ± 1
99 ± 2 104 ± 1 104 ± 1 107 ± 1 108 ± 2 104 ± 1 107 ± 1
427 ± 9 453 ± 6 437 ± 3 443 ± 1 440 ± 1 439 ± 2 441 ± 1
489 ± 3 459 ± 4 474 ± 10 529 ± 1 537 ± 6 515 ± 16 486 ± 2
a
The results reported are averages of two or more measurements.
latex, prepared following Tamai et al. (12) and Kamei et al. (13). Polymerization is initiated by K2 S2 O8 , in the absence of surfactant. Yield was 94% with a 14.3% solids content. In every case, the amount of coagulated material was less than 1%. All the latexes were filtered with a 200 mesh steel sieve to remove coagulated latex; in order to remove excess emulsifiers, free monomer, and unwanted electrolyte, they were subsequently dialyzed against water with daily changes over a 1month period, until the dialyzate conductivity reached 4 µS/cm and remained unchanged for 48 h. The dialysis tubing (a Visking membrane from Sigma) was boiled in several quantities of distilled water prior to use. After dialysis, the samples were dispersed in water as required for reaching the desired concentrations. Latex Exposure to Solvent Vapor and Characterization A 1–1.5 ml amount of a latex (1 or 2% w/w solids content) was placed within a 15 ml glass vial in a closed glass container, in the presence of 3 ml of organic solvent (toluene, chloroform, bromoform, carbon tetrachloride, acetone) in a room at 25 ± 2◦ C and observed periodically over a few days. Aliquots of this latex were withdrawn (after 48 h or as given in the Results) after the desired times for the required assays. Particle size (by photon correlation spectroscopy) and zeta potential determinations were performed in a Brookhaven ZetaPlus instrument. To achieve an adequate signal intensity, 0.1 ml stock latex dispersions were diluted with 2 ml of water (for parti-
cle sizes) or 10−3 MKCl solution (for zeta potential). Palladium electrodes were used. Zonal centrifugation in density gradients was done by layering 100 µl latex zones (1% solids) on top of preformed linear sucrose density gradients, prepared by mixing water and 20% aqueous sucrose (14), followed by centrifugation at 19,000 pm. Centrifuge tube pictures were obtained with a video camera and digitalized; scattered light profiles from the tubes were obtained by line-scanning the tube pictures (15), along the tube axis. Macrographs were obtained using a SSC-C350 Sony camera, with a 55 mm Computar lens. Atomic force microscopy (AFM) was performed in a Topometrix Discoverer instrument operating in the true noncontact mode and using a silicon probe tip. Latex samples were spread on top of fresh mica surfaces, which were fastened to the sample holders with double-face adhesive tape. Scanning electron microscopy (SEM) was done in a fieldemission JSM-6340F JEOL instrument. Latex samples were deposited directly on a film of a carbon-filled conductive glue. RESULTS
Particle Diameter and Zeta Potential Aliquots of all nine latexes were exposed to an atmosphere saturated with toluene. Their effective particle diameters and zeta potentials were determined as a function of exposure time, and the results are presented in Tables 1 and 2. Particle effective
TABLE 2 Zeta Potentials (mV) Measured after Latex Exposure to Toluene Time/h
PBMA/PS
PS/PBMA
P[SBMA]
PS-THS
PS-LEV
PS-11
PS-M
PS-HEMA
PS-AAM
0a 24b 48b 96b 168b 336b 504b
−31 ± 3 −9 ± 5 1±1 −21 ± 8 −22 ± 5 −28 ± 4 −39 ± 7
−37 ± 3 −39 ± 4 −37 ± 6 −32 ± 7 −34 ± 1 −39 ± 4 −36 ± 2
−27 ± 3 −39 ± 2 −44 ± 6 −32 ± 1 −39 ± 1 −39 ± 1 −34 ± 10
−36 ± 2 −28 ± 3 −37 ± 1 −28 ± 3 −36 ± 3 −31 ± 1 −21 ± 8
−39 ± 1 −40 ± 1 −45 ± 2 −50 ± 2 −63 ± 10 −44 ± 3 −42 ± 2
−19 ± 3 −26 ± 2 −12 ± 3 −21 ± 3 −18 ± 1 −23 ± 3 −21 ± 1
−29 ± 2 −35 ± 1 −29 ± 5 −26 ± 2 −42 ± 1 −32 ± 2 −26 ± 4
−41 ± 1 −47 ± 1 −50 ± 2 −54 ± 4 −49 ± 1 −50 ± 1 −49 ± 1
−26 ± 2 −23 ± 3 −25 ± 4 −26 ± 1 −28 ± 5 −23 ± 1 −23 ± 1
a b
The results reported are averages of ten or more individual runs. Averages of duplicate measurements.
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
173
FIG. 1. Non-contact AFM images of PS-M latex films formed by surface coagulation, under chloroform vapor: (a) surface formed in contact with air; (b) surface formed in contact with liquid; (c) the same as in (b) but rinsed; (d) zoom of a field in (c) but away from the large particles.
diameters did show only minor changes, except in the case of the P[SBMA], PS-LEV, PS-M and PS-AAM latexes, which showed an increase (ca. 10%) followed by a decrease. Changes in the zeta potentials were more marked, but in most cases the zeta potential after 2 weeks exposure to toluene was within 20% of the value of the pristine latex. Large zeta potential oscillations were observed in intermediate times in the case of PBMA/PS only: its zeta potential rose from −31 to l mV and then decreased to −39 mV. Film Formation During the experiments described in the previous section, film formation on top of the PS-THS latex dispersions was observed, under exposure to toluene, as well as on PS-M under a 1-week exposure to chloroform. Allowing the PS-M film to thicken for
2 weeks under chloroform, it was strong enough to be collected over a thin nichrome wire ring; this was first immersed in the liquid, placed beneath the film, and moved upward. This film was transparent and shiny, for which reason its surface was examined in the AFM microscope (see Fig. 1). The upper film surface (formed in contact with air and chloroform vapor) is indeed very flat: over a 5 µm scanned range the maximum height difference observed is less than 7 nm, which is much less than 1/20 the visible light wavelength and justifies the observation of mirror-like light reflection in this film. As a control, Fig. 1 also presents the AFM image of the lower film surface, which was formed in contact with the liquid and was removed from it. In one picture the film was rinsed but not in the other case: individual and aggregated particles are discerned in the nonrinsed surface, and a few particles are also observed in the surface formed in contact with the liquid. Beyond these particles, this surface is
174
BRAGA ET AL.
TABLE 3 Effect of Chloroform Vapor on PS-M Latex Particle Effective Diameters (nm)
TABLE 4 Effective Diameters of PS-M Latex Exposed to Chloroform for 24 h
Time/h
Fresh latexa
Aged latexb
Zone positiona
Effective diameter/nm
0 24 48
101 ± 2 200 ± 2 220 ± 2
99 ± 2 135 ± 1 157 ± 1
Upper Middle Lower
109 ± 1 117 ± 1 228 ± 3
a b
Determined in 1997. Determined in 2000.
extremely smooth. The PS-THS film was already visible after 48 h of exposure to toluene, but this was too fragile and could not be successfully collected. Solvent Effect on Latex Sedimentation Profiles PS-M latex was chosen for more detailed examination of the effects of exposure to solvents because it was previously shown (6) to display an intriguing behavior: in the presence of chloroform vapor, its effective diameter (determined by PCS) increased significantly as described in Table 3, while in the presence of toluene, which is a good solvent for polystyrene, it showed only a modest increase. This latex displays two broad overlapping bands under isopycnic equilibrium in aqueous sucrose solutions, covering the 1.002–1.026 g cm−3 density range. After 24 h of exposure to chloroform vapor, the isopycnic profile showed bands at 1.045, 1.036, and 1.026 g cm−3 (see Fig. 2). Particle effective diameters were determined for these three isopycnic fractions, and the values thus obtained are in Table 4, showing that the particles within the denser zone have an effective diameter corresponding to about twice the original value. Consequently, the increase in the diameter of the latex particles is observed in only one of its three fractions.
Note. Particles were collected from the zones separated by centrifugation in a sucrose density gradient. a See Fig. 2.
However, under bromoform vapor a more drastic modification was observed: there was a heavy fraction reaching the bottom of the tube, thus exceeding a 1.071 g cm−3 density, together with a lighter fraction with density 1.024 g cm−3 . The effective diameter of the particles exposed to bromoform was 98 nm, agreeing with the value obtained for the pristine particles. The comparison of sedimentation and PCS data for PS-M latex under chloroform shows an apparent contradiction: effective particle diameters double, while the particle density experiences only a modest increase, ca. 3%. If particle diameter doubling was the result of latex swelling with chloroform, the particle density should increase to a weighted average between the densities of polystyrene and chloroform. Considering that doubling the diameter increase the particle volume 9-fold, the swollen particle density should be ca. 1.40 g cm−3 , much higher than the measured density. Consequently, the large increase in particle effective diameter cannot be due to overall particle swelling, but it is compatible with one of the two following hypotheses: (i) Particles undergo a limited aggregation; this is a simple assumption, but it raises another problem: why is the aggregation limited, or else, why do the aggregates not grow further? (ii) There is a release of long “hairs” from the particle, combined with intense immobilization of the solvent around the particle. Due the large change in particle diameter, we tend to favor to the first hypothesis. Coagulation of a Hydrophylic Latex
FIG. 2. Vertical line-scans of PS latex zone profiles obtained by centrifugation in sucrose gradients. (a) Pristine latex; (b) the same, but after 24 h exposure to CHCl3 vapor. The numbers 5, 10, and 30 indicate the centrifugation times.
PS-AAM latex is very stable toward salt. This latex sediments in a sucrose gradient (Fig. 3) presenting a fast-moving band and a slower band: the first reaches the isopycnic position in a very short time, and both merge at 1.043–1.045 g cm−3 (Fig. 4). The two transient bands were collected and washed with water by ultrafiltration, and particle sizes were determined. The results obtained are 436 nm for the upper band and 454 nm for the lower. In other experiments, the latex of the transient bands was separated from sucrose by dialysis, which takes much longer than ultrafiltration, and small coagula were visible in the material from the lower band. PS-AAM latex exposure to CCl4 for 48 h produces a 10% increase in the particle diameter. Centrifugation of this sample
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
175
FIG. 4. Vertical line-scans of the picture in Fig. 3, showing the latex zones within the density gradients.
with vapors of C1 –C4 alcohols (7) and acetone. On the other hand, coagulation was not observed in the presence of higher alcohols, ketones, or many other solvents. Figure 5 shows pictures of a latex placed on a watch glass, within a closed acrylic box, which also contained another a
FIG. 3. PS-AAM latex zone profiles obtained by centrifugation in density gradients. (a) Pristine latex; (b) the same, but after 48 h exposure to CCl4 vapor. The numbers 5, 10, and 20 refer to the centrifugation times.
in a density gradient yields two fractions: one sediments very fast and accumulates in the bottom of the centrifuge tube; the other makes a thin band at 1.056 g cm−3 . The denser particles reach the bottom of the tube after 2 min, but they move upward slowly at longer times thus evidencing a decrease in density, which in turn shows that these particles can de-swell in relatively short times. However, this is the only case throughout this work in which any evidence was obtained for particle de-swelling. Coagulation of PS-11 Latex, by Acetone Following PS-11 latex exposure to solvents for 48 h, coagulation was observed in the samples left in atmospheres saturated
FIG. 5. PS-11 latex (0.3 ml, 1% w/w) coagulating under acetone vapor. Exposure times in hours are (a–f) zero, 4, 12, 24, 36, and 48 h. Diameter of the watch glass: 4 cm.
176
BRAGA ET AL.
small beaker with 4 ml of acetone. A change in sample turbidity is observed after a few hours, followed by the formation of a large monolithic coagulum, immersed in aqueous acetone solution (acetone concentration increases with time, and the coagulum undergoes contraction). The supernatant was sampled after 48 h, and the particle effective diameter was determined, yielding 308 nm. Using the R = n c r (16) relationship, where R is the radius of the aggregate, n is the aggregate particle number, r is the radius of the individual particle, and c = 0.58, we obtain ca. 7 as the aggregate particle number. This shows clearly that some latex particles do not join the large coagulum, but rather they undergo only partial aggregation. Observation of coagula formed after exposure to acetone under an optical microscope showed the coexistence of at least three different types of domains (see Fig. 6): (i) smooth reflective areas surrounded by (ii) opaque solid with a rough surface and (iii) rare opalescent areas, formed after longer exposure times and covering a small fraction of the area (not shown).
FIG. 7. SEM pictures of coaguli obtained by exposure of the latex (0.3 ml) to (a) a beaker with 2 ml acetone; (b) a container with 3 ml acetone, together with another container with 2 ml water.
SEM observation (Fig. 7) of a fracture surface obtained from this coagulum shows a network of elongated polymer domains formed by coalesced particles and surrounding connected pores, ranging from one micron to many microns. This may be compared to the micrograph of a coagulum obtained in a similar way, but only adding a beaker with 3 ml water within the closed acrylic box, together with latex and acetone. In this case, many large coagula are formed, but the individual particles are clearly discerned in their pictures. These particles are not strongly coalesced as in the previous case. DISCUSSION FIG. 6. Micrographs of a coagulum formed following 48 h of exposure of PS-11 to acetone vapor: (a) overall picture; (b) rough area with bright domains; (c) smooth opaque domains.
Latex exposure to solvent vapors has as important effect, not only on latex stability but also on the morphology of the coagula obtained. In the case of PS-11 coagulation following a exposure
177
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
to acetone vapor, an interesting porous structure is formed, completely different from the structure obtained by coagulation with acetone but in the presence of a water container. Minute changes in solvent composition have been widely used in the making of polymer membranes with different pore sizes (17). The intervening mechanisms (e.g., the role of spinodal or binodal phase separation) are known in some detail. However, analogous small changes in solvent composition have not yet been used to generate solids with a desirable porosity, starting from polymer colloids, or latexes, and this may be a useful application of the present results. Solvent vapors adsorbed within a latex dispersion are partitioned between the solvent, the particles bulk and the interface. In the case of multidomain latexes (e.g., core-and-shell, sandwich type), partitioning is even more complex since many domains are available for the solvent. We consider seriously the solvent accumulation at the interface, because this is often a site of intermediate polarity between those of polymer bulk and the aqueous solvent. This accumulation may help to explain why small amounts of solvents have such profound effects on latex stability. In a recent work, it was shown that CO2 sorption in a latex exceeds by far its sorption by the bulk polymer and solvent, thus evidencing the strong and specific sorption ability of the particle–serum interface (18). We conclude that the composition and properties of latex interfaces deserve increased attention, and renewed efforts should be directed at determining their chemical composition, structure and dynamics (19, 20), and other properties, following the approaches that have been established in the literature in recent years (21, 22). Another relevant issue is the chemical heterogeneity of the latex particles and its effects on solvent sorption. This heterogeneity has been established in previous work of this group, using sedimentation and micro-chemical techniques and showing significant differences in the chemical compositions of individual particles (8, 9, 23). The subdivision and spread of the centrifuged zones shown in Figs. 2–4 show that the pristine latexes used are not homogeneous. The zone profiles obtained after exposure to solvents show that the different latex fractions do not respond identically to solvent. For instance, in Fig. 2 a two-broad-band profile is replaced by a three-band profile in the same latex, but
after exposure to chloroform, and these three bands are sharper than those in the initial material. ACKNOWLEDGMENTS F.G. acknowledges grants from Fapesp, Pronex/Finep/MCT, and CNPq. M.B. was supported by a CNPq undergraduate fellowship, and A.H.C. was on leave from the Universidade Regional do Cariri, Crato (CE), Brazil.
REFERENCES 1. Verwey, E. J. W., and Overbeek, J. Th. G., “Theory of Stability of Lyophobic Colloids.” Elsevier, Amsterdam, 1948. 2. Buscall, R., and Ottewill, R. H., in “Polymer Colloids” (R. Buscall, T. Corner, and J. F. Stageman, Eds.). Elsevier, London, 1985. 3. Hunter, R. J., “Foundations of Colloid Science,” Vol. 1. Oxford Univ. Press, Oxford, 1991. 4. Gregory, J., Crit. Rev. Environ. Control 19, 185 (1989). 5. Cardoso, A. O., and Galembeck, F., J. Colloid Interface Sci. 182, 614 (1996). 6. Cardoso, A. O., and Galembeck, F., J. Colloid Interface Sci. 204, 16. (1998). 7. Galembeck, F., Braga, M., Silva, M. C. V. M., and Cardoso, A. H., Colloids Surfaces 164, 217 (2000). 8. Cardoso, A. H., Moita-Neto, J. M, Cardoso, A. O., and Galembeck, F., Colloid Polym. Sci. 275, 244 (1997). 9. Moita-Neto, J. M., Monteiro, V. A. R., and Galembeck, F., Colloids Surfaces 108, 83 (1996). 10. Ottewill, R. H., and Satgurunathan, R., Colloid Polym. Sci. 265, 845 (1987). 11. Tamai, H., Hasegawa, M., and Suzawa, T., Colloids Surfaces 51, 271 (1990). 12. Tamai, H., Fujji, A., and Suzawa, T., J. Colloid Interface Sci. 116, 37 (1987). 13. Kamei, S., Okubo, M., and Matsumoto, T., J. Polym. Sci., Polym. Chem. Ed. 24, 3109 (1986). 14. Takayasu, M. M., and Galembeck, F., J. Colloid Interface Sci. 155, 16 (1993). 15. Cardoso, A. H., Leite, C. A. P., and Galembeck, F., Langmuir 14, 3187 (1998). 16. Takayasu, M. M., and Galembeck, F., J. Colloid Interface Sci. 202, 84 (1998). 17. Nunes, S. P., and Galembeck, F., J. Polym. Sci. Polym. Lett. Ed. 21, 49 (1983). 18. Otake, K., Webber, S. E., Munk, P., and Johnston, K. P., Langmuir 13, 3047 (1997). 19. Lyklema, J., Sringer, M. M., Shilob, V. N., and Dukhin, S. S., J. Electroanal. Chem. 198 19 (1986). 20. Chonde, Y., and Shabrang, M., J. Colloid Interface Sci. 186, 248 (1997). 21. Zhao, J., and Brown, W., J. Colloid Interface Sci. 179, 281 (1996). 22. Niu, B.-J., and Urban, M. V., J. Appl. Polym. Sci. 60, 389 (1996). 23. Galembeck, F., and Souza, E. F., “Polymer Interfaces and Emulsions” (K. Esumi, Ed.), p. 119. Marcel Dekker, Inc., New York, 1999.
Latex Surface and Bulk Coagulation Induced by Solvent Vapors Melissa Braga, Maria do Carmo Vasconcelos Medeiros da Silva, Andr´e Herzog Cardoso,1 and Fernando Galembeck2 Instituto de Qu´ımica, Universidade Estadual de Campinas, P.O. Box 6154, 13083-970 Campinas SP, Brazil Received March 6, 2000; accepted April 17, 2000
erating serious problems. For this reason, we decided to continue the investigation of this effect, looking at other systems beyond those previously described. During this work, we observed a variety of interesting effects of latex exposure to the vapors of organic solvents, which appear highly specific, i.e., they depend on the specific latex and solvent considered. This specificity is a serious complication, because it adds a strong component of unpredictability to the behavior of latexes. On the other hand, since the observed phenomena are highly reproducible, knowledge on the cases already observed may help explain unexpected problems arising both in the laboratory and in the chemical plant. Some new chosen examples are described in this paper.
Latex exposure to solvent vapors leads to highly specific changes in latex stability as well as on the morphologies of the particle association products, depending on the latex and solvent used. Examples of solvent vapor-induced aggregation are given: surface films are obtained on two PS latexes; in one case, the film surface is mirrorreflective and very flat, as evidenced by AFM. Another PS latex coagulates under exposure to acetone vapors, and the morphologies of the coagula are highly sensitive to the exposure conditions. This latex yields a highly porous foam-like structure, in which particles are strongly coalesced but form percolating patches around the pores. The same latex but under other conditions produces a coagulum of large numbers of aggregated particles with a raspberry-like morphology. Density centrifugation experiments show that the effect of solvents on different latex fractions is not uniform, and some fractions show larger density changes than others, thus evidencing a variability in their swelling ability. °C 2000 Academic Press Key Words: latex film formation; PS latex stability; solvent effects on latex stability; solvent-induced coagulation.
EXPERIMENTAL PROCEDURES
Emulsion Polymerization and Latex Preparation
INTRODUCTION
There has been a great progress toward understanding colloidal stability, since the appearance of the DLVO theory (1–3), and the role of Van der Waals, electrostatic interactions, steric interactions, particle bridging, depletion, particle surface hydration layers (4), and the associated hydrodynamic interactions is now well established, in many relevant systems. In a recent work (5, 6), we have shown that the exposure of some latexes to organic solvent vapors increases the stability of the latex dispersion in the presence of salt, and later we showed that the exposure to short-chain alcohol vapors of a redox-initiated, non-ionic surfactant stabilized polystyrene latex, which has a strong stability toward salt, provokes its coagulation (7). These examples show a great susceptibility of latexes to minor contamination by solvent vapors; they are not predicted by the current theories of colloidal stability, and we did not find analogous work previously reported in the literature. However, since solvents are always found in laboratories and chemical plants, their stabilizing effect may well create unpredicted situations for latex users, which have a potential for gen1 2
Present address: Universidade Regional do Cariri, Crato CE, Brazil. To whom correspondence should be addressed.
P[SBMA], PS/PBMA, and PBMA/PS are latexes obtained by the polymerization of styrene (S) and butyl methacrylate (BMA), previously prepared and characterized in this laboratory (8). The first is a random copolymer prepared by simultaneous addition of the two monomers. The two others are core-and-shell latexes, containing separate domains of polystyrene and acrylic. Polymer chain segments are predominantly homopolymeric in these two latexes. PS-M latex was also prepared and characterized in a previous work (9), using Brij 35 and SDS as emulsifiers and potassium persulfate as the initiator. The other latexes were prepared in this laboratory, using procedures from the literature, as follows. Latexes PS-11 and PS-LEV were prepared following Ottewill and Satgurunathan (10), but Renex 300 was substituted for the Levelan stabilizer used by these authors. In the first, a mixture of ascorbic acid and H2 O2 was used as the redox initiator. The yield was 86%, and the solids content of the dispersion was 7.7%. In the second, the initiator was persulfate and NaHCO3 was used to raise the pH. The yield was 99% and solids contents was 20%. PS-THS is a polystyrene latex prepared in the absence of surfactant, using K2 S2 O8 initiator following Tamai et al. (11). The yield of the synthesis was 56% and the final solids content in the dispersion was 1.1%. PS/AAM is a polystyrenepolyacrylamide latex prepared following another paper by Tamai et al. (12), using K2 S2 O8 as the initiator, in the absence of surfactant. Synthesis yield was 76%, and solids content was 8.6%. PS/HEMA is a polystyrene-poly(2-hydroxyethyl methacrylate)
171
0021-9797/00 $35.00
C 2000 by Academic Press Copyright ° All rights of reproduction in any form reserved.
172
BRAGA ET AL.
TABLE 1 Latex Particle Effective Diameters (nm) Following Exposure to Toluene, as Determined by PCSa Time/h
PBMA/PS
PS/PBMA
P[SBMA]
PS-THS
PS-LEV
PS-11
PS-M
PS-HEMA
PS-AAM
0 24 48 96 168 336 504
67.4 ± 0.5 64.9 ± 1.5 67.3 ± 0.7 68.7 ± 0.2 68.3 ± 0.2 68.0 ± 0.1 69.5 ± 0.7
75.1 ± 0.6 74.4 ± 0.6 74.5 ± 0.9 78.3 ± 0.3 77.5 ± 0.3 77.5 ± 0.1 78.6 ± 0.9
73.6 ± 0.1 74.3 ± 0.7 75.3 ± 1.3 78.9 ± 0.6 80.8 ± 0.8 82.8 ± 0.1 77.3 ± 0.9
282 ± 1 284 ± 1 288 ± 3 288 ± 1 282 ± 1 286 ± 8 294 ± 2
432 ± 7 440 ± 6 442 ± 8 453 ± 5 462 ± 7 445 ± 3 453 ± 4
97 ± 1 101 ± 1 103 ± 1 100 ± 1 103 ± 1 102 ± 1 92.4 ± 1
99 ± 2 104 ± 1 104 ± 1 107 ± 1 108 ± 2 104 ± 1 107 ± 1
427 ± 9 453 ± 6 437 ± 3 443 ± 1 440 ± 1 439 ± 2 441 ± 1
489 ± 3 459 ± 4 474 ± 10 529 ± 1 537 ± 6 515 ± 16 486 ± 2
a
The results reported are averages of two or more measurements.
latex, prepared following Tamai et al. (12) and Kamei et al. (13). Polymerization is initiated by K2 S2 O8 , in the absence of surfactant. Yield was 94% with a 14.3% solids content. In every case, the amount of coagulated material was less than 1%. All the latexes were filtered with a 200 mesh steel sieve to remove coagulated latex; in order to remove excess emulsifiers, free monomer, and unwanted electrolyte, they were subsequently dialyzed against water with daily changes over a 1month period, until the dialyzate conductivity reached 4 µS/cm and remained unchanged for 48 h. The dialysis tubing (a Visking membrane from Sigma) was boiled in several quantities of distilled water prior to use. After dialysis, the samples were dispersed in water as required for reaching the desired concentrations. Latex Exposure to Solvent Vapor and Characterization A 1–1.5 ml amount of a latex (1 or 2% w/w solids content) was placed within a 15 ml glass vial in a closed glass container, in the presence of 3 ml of organic solvent (toluene, chloroform, bromoform, carbon tetrachloride, acetone) in a room at 25 ± 2◦ C and observed periodically over a few days. Aliquots of this latex were withdrawn (after 48 h or as given in the Results) after the desired times for the required assays. Particle size (by photon correlation spectroscopy) and zeta potential determinations were performed in a Brookhaven ZetaPlus instrument. To achieve an adequate signal intensity, 0.1 ml stock latex dispersions were diluted with 2 ml of water (for parti-
cle sizes) or 10−3 MKCl solution (for zeta potential). Palladium electrodes were used. Zonal centrifugation in density gradients was done by layering 100 µl latex zones (1% solids) on top of preformed linear sucrose density gradients, prepared by mixing water and 20% aqueous sucrose (14), followed by centrifugation at 19,000 pm. Centrifuge tube pictures were obtained with a video camera and digitalized; scattered light profiles from the tubes were obtained by line-scanning the tube pictures (15), along the tube axis. Macrographs were obtained using a SSC-C350 Sony camera, with a 55 mm Computar lens. Atomic force microscopy (AFM) was performed in a Topometrix Discoverer instrument operating in the true noncontact mode and using a silicon probe tip. Latex samples were spread on top of fresh mica surfaces, which were fastened to the sample holders with double-face adhesive tape. Scanning electron microscopy (SEM) was done in a fieldemission JSM-6340F JEOL instrument. Latex samples were deposited directly on a film of a carbon-filled conductive glue. RESULTS
Particle Diameter and Zeta Potential Aliquots of all nine latexes were exposed to an atmosphere saturated with toluene. Their effective particle diameters and zeta potentials were determined as a function of exposure time, and the results are presented in Tables 1 and 2. Particle effective
TABLE 2 Zeta Potentials (mV) Measured after Latex Exposure to Toluene Time/h
PBMA/PS
PS/PBMA
P[SBMA]
PS-THS
PS-LEV
PS-11
PS-M
PS-HEMA
PS-AAM
0a 24b 48b 96b 168b 336b 504b
−31 ± 3 −9 ± 5 1±1 −21 ± 8 −22 ± 5 −28 ± 4 −39 ± 7
−37 ± 3 −39 ± 4 −37 ± 6 −32 ± 7 −34 ± 1 −39 ± 4 −36 ± 2
−27 ± 3 −39 ± 2 −44 ± 6 −32 ± 1 −39 ± 1 −39 ± 1 −34 ± 10
−36 ± 2 −28 ± 3 −37 ± 1 −28 ± 3 −36 ± 3 −31 ± 1 −21 ± 8
−39 ± 1 −40 ± 1 −45 ± 2 −50 ± 2 −63 ± 10 −44 ± 3 −42 ± 2
−19 ± 3 −26 ± 2 −12 ± 3 −21 ± 3 −18 ± 1 −23 ± 3 −21 ± 1
−29 ± 2 −35 ± 1 −29 ± 5 −26 ± 2 −42 ± 1 −32 ± 2 −26 ± 4
−41 ± 1 −47 ± 1 −50 ± 2 −54 ± 4 −49 ± 1 −50 ± 1 −49 ± 1
−26 ± 2 −23 ± 3 −25 ± 4 −26 ± 1 −28 ± 5 −23 ± 1 −23 ± 1
a b
The results reported are averages of ten or more individual runs. Averages of duplicate measurements.
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
173
FIG. 1. Non-contact AFM images of PS-M latex films formed by surface coagulation, under chloroform vapor: (a) surface formed in contact with air; (b) surface formed in contact with liquid; (c) the same as in (b) but rinsed; (d) zoom of a field in (c) but away from the large particles.
diameters did show only minor changes, except in the case of the P[SBMA], PS-LEV, PS-M and PS-AAM latexes, which showed an increase (ca. 10%) followed by a decrease. Changes in the zeta potentials were more marked, but in most cases the zeta potential after 2 weeks exposure to toluene was within 20% of the value of the pristine latex. Large zeta potential oscillations were observed in intermediate times in the case of PBMA/PS only: its zeta potential rose from −31 to l mV and then decreased to −39 mV. Film Formation During the experiments described in the previous section, film formation on top of the PS-THS latex dispersions was observed, under exposure to toluene, as well as on PS-M under a 1-week exposure to chloroform. Allowing the PS-M film to thicken for
2 weeks under chloroform, it was strong enough to be collected over a thin nichrome wire ring; this was first immersed in the liquid, placed beneath the film, and moved upward. This film was transparent and shiny, for which reason its surface was examined in the AFM microscope (see Fig. 1). The upper film surface (formed in contact with air and chloroform vapor) is indeed very flat: over a 5 µm scanned range the maximum height difference observed is less than 7 nm, which is much less than 1/20 the visible light wavelength and justifies the observation of mirror-like light reflection in this film. As a control, Fig. 1 also presents the AFM image of the lower film surface, which was formed in contact with the liquid and was removed from it. In one picture the film was rinsed but not in the other case: individual and aggregated particles are discerned in the nonrinsed surface, and a few particles are also observed in the surface formed in contact with the liquid. Beyond these particles, this surface is
174
BRAGA ET AL.
TABLE 3 Effect of Chloroform Vapor on PS-M Latex Particle Effective Diameters (nm)
TABLE 4 Effective Diameters of PS-M Latex Exposed to Chloroform for 24 h
Time/h
Fresh latexa
Aged latexb
Zone positiona
Effective diameter/nm
0 24 48
101 ± 2 200 ± 2 220 ± 2
99 ± 2 135 ± 1 157 ± 1
Upper Middle Lower
109 ± 1 117 ± 1 228 ± 3
a b
Determined in 1997. Determined in 2000.
extremely smooth. The PS-THS film was already visible after 48 h of exposure to toluene, but this was too fragile and could not be successfully collected. Solvent Effect on Latex Sedimentation Profiles PS-M latex was chosen for more detailed examination of the effects of exposure to solvents because it was previously shown (6) to display an intriguing behavior: in the presence of chloroform vapor, its effective diameter (determined by PCS) increased significantly as described in Table 3, while in the presence of toluene, which is a good solvent for polystyrene, it showed only a modest increase. This latex displays two broad overlapping bands under isopycnic equilibrium in aqueous sucrose solutions, covering the 1.002–1.026 g cm−3 density range. After 24 h of exposure to chloroform vapor, the isopycnic profile showed bands at 1.045, 1.036, and 1.026 g cm−3 (see Fig. 2). Particle effective diameters were determined for these three isopycnic fractions, and the values thus obtained are in Table 4, showing that the particles within the denser zone have an effective diameter corresponding to about twice the original value. Consequently, the increase in the diameter of the latex particles is observed in only one of its three fractions.
Note. Particles were collected from the zones separated by centrifugation in a sucrose density gradient. a See Fig. 2.
However, under bromoform vapor a more drastic modification was observed: there was a heavy fraction reaching the bottom of the tube, thus exceeding a 1.071 g cm−3 density, together with a lighter fraction with density 1.024 g cm−3 . The effective diameter of the particles exposed to bromoform was 98 nm, agreeing with the value obtained for the pristine particles. The comparison of sedimentation and PCS data for PS-M latex under chloroform shows an apparent contradiction: effective particle diameters double, while the particle density experiences only a modest increase, ca. 3%. If particle diameter doubling was the result of latex swelling with chloroform, the particle density should increase to a weighted average between the densities of polystyrene and chloroform. Considering that doubling the diameter increase the particle volume 9-fold, the swollen particle density should be ca. 1.40 g cm−3 , much higher than the measured density. Consequently, the large increase in particle effective diameter cannot be due to overall particle swelling, but it is compatible with one of the two following hypotheses: (i) Particles undergo a limited aggregation; this is a simple assumption, but it raises another problem: why is the aggregation limited, or else, why do the aggregates not grow further? (ii) There is a release of long “hairs” from the particle, combined with intense immobilization of the solvent around the particle. Due the large change in particle diameter, we tend to favor to the first hypothesis. Coagulation of a Hydrophylic Latex
FIG. 2. Vertical line-scans of PS latex zone profiles obtained by centrifugation in sucrose gradients. (a) Pristine latex; (b) the same, but after 24 h exposure to CHCl3 vapor. The numbers 5, 10, and 30 indicate the centrifugation times.
PS-AAM latex is very stable toward salt. This latex sediments in a sucrose gradient (Fig. 3) presenting a fast-moving band and a slower band: the first reaches the isopycnic position in a very short time, and both merge at 1.043–1.045 g cm−3 (Fig. 4). The two transient bands were collected and washed with water by ultrafiltration, and particle sizes were determined. The results obtained are 436 nm for the upper band and 454 nm for the lower. In other experiments, the latex of the transient bands was separated from sucrose by dialysis, which takes much longer than ultrafiltration, and small coagula were visible in the material from the lower band. PS-AAM latex exposure to CCl4 for 48 h produces a 10% increase in the particle diameter. Centrifugation of this sample
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
175
FIG. 4. Vertical line-scans of the picture in Fig. 3, showing the latex zones within the density gradients.
with vapors of C1 –C4 alcohols (7) and acetone. On the other hand, coagulation was not observed in the presence of higher alcohols, ketones, or many other solvents. Figure 5 shows pictures of a latex placed on a watch glass, within a closed acrylic box, which also contained another a
FIG. 3. PS-AAM latex zone profiles obtained by centrifugation in density gradients. (a) Pristine latex; (b) the same, but after 48 h exposure to CCl4 vapor. The numbers 5, 10, and 20 refer to the centrifugation times.
in a density gradient yields two fractions: one sediments very fast and accumulates in the bottom of the centrifuge tube; the other makes a thin band at 1.056 g cm−3 . The denser particles reach the bottom of the tube after 2 min, but they move upward slowly at longer times thus evidencing a decrease in density, which in turn shows that these particles can de-swell in relatively short times. However, this is the only case throughout this work in which any evidence was obtained for particle de-swelling. Coagulation of PS-11 Latex, by Acetone Following PS-11 latex exposure to solvents for 48 h, coagulation was observed in the samples left in atmospheres saturated
FIG. 5. PS-11 latex (0.3 ml, 1% w/w) coagulating under acetone vapor. Exposure times in hours are (a–f) zero, 4, 12, 24, 36, and 48 h. Diameter of the watch glass: 4 cm.
176
BRAGA ET AL.
small beaker with 4 ml of acetone. A change in sample turbidity is observed after a few hours, followed by the formation of a large monolithic coagulum, immersed in aqueous acetone solution (acetone concentration increases with time, and the coagulum undergoes contraction). The supernatant was sampled after 48 h, and the particle effective diameter was determined, yielding 308 nm. Using the R = n c r (16) relationship, where R is the radius of the aggregate, n is the aggregate particle number, r is the radius of the individual particle, and c = 0.58, we obtain ca. 7 as the aggregate particle number. This shows clearly that some latex particles do not join the large coagulum, but rather they undergo only partial aggregation. Observation of coagula formed after exposure to acetone under an optical microscope showed the coexistence of at least three different types of domains (see Fig. 6): (i) smooth reflective areas surrounded by (ii) opaque solid with a rough surface and (iii) rare opalescent areas, formed after longer exposure times and covering a small fraction of the area (not shown).
FIG. 7. SEM pictures of coaguli obtained by exposure of the latex (0.3 ml) to (a) a beaker with 2 ml acetone; (b) a container with 3 ml acetone, together with another container with 2 ml water.
SEM observation (Fig. 7) of a fracture surface obtained from this coagulum shows a network of elongated polymer domains formed by coalesced particles and surrounding connected pores, ranging from one micron to many microns. This may be compared to the micrograph of a coagulum obtained in a similar way, but only adding a beaker with 3 ml water within the closed acrylic box, together with latex and acetone. In this case, many large coagula are formed, but the individual particles are clearly discerned in their pictures. These particles are not strongly coalesced as in the previous case. DISCUSSION FIG. 6. Micrographs of a coagulum formed following 48 h of exposure of PS-11 to acetone vapor: (a) overall picture; (b) rough area with bright domains; (c) smooth opaque domains.
Latex exposure to solvent vapors has as important effect, not only on latex stability but also on the morphology of the coagula obtained. In the case of PS-11 coagulation following a exposure
177
LATEX COAGULATION INDUCED BY SOLVENT VAPOR
to acetone vapor, an interesting porous structure is formed, completely different from the structure obtained by coagulation with acetone but in the presence of a water container. Minute changes in solvent composition have been widely used in the making of polymer membranes with different pore sizes (17). The intervening mechanisms (e.g., the role of spinodal or binodal phase separation) are known in some detail. However, analogous small changes in solvent composition have not yet been used to generate solids with a desirable porosity, starting from polymer colloids, or latexes, and this may be a useful application of the present results. Solvent vapors adsorbed within a latex dispersion are partitioned between the solvent, the particles bulk and the interface. In the case of multidomain latexes (e.g., core-and-shell, sandwich type), partitioning is even more complex since many domains are available for the solvent. We consider seriously the solvent accumulation at the interface, because this is often a site of intermediate polarity between those of polymer bulk and the aqueous solvent. This accumulation may help to explain why small amounts of solvents have such profound effects on latex stability. In a recent work, it was shown that CO2 sorption in a latex exceeds by far its sorption by the bulk polymer and solvent, thus evidencing the strong and specific sorption ability of the particle–serum interface (18). We conclude that the composition and properties of latex interfaces deserve increased attention, and renewed efforts should be directed at determining their chemical composition, structure and dynamics (19, 20), and other properties, following the approaches that have been established in the literature in recent years (21, 22). Another relevant issue is the chemical heterogeneity of the latex particles and its effects on solvent sorption. This heterogeneity has been established in previous work of this group, using sedimentation and micro-chemical techniques and showing significant differences in the chemical compositions of individual particles (8, 9, 23). The subdivision and spread of the centrifuged zones shown in Figs. 2–4 show that the pristine latexes used are not homogeneous. The zone profiles obtained after exposure to solvents show that the different latex fractions do not respond identically to solvent. For instance, in Fig. 2 a two-broad-band profile is replaced by a three-band profile in the same latex, but
after exposure to chloroform, and these three bands are sharper than those in the initial material. ACKNOWLEDGMENTS F.G. acknowledges grants from Fapesp, Pronex/Finep/MCT, and CNPq. M.B. was supported by a CNPq undergraduate fellowship, and A.H.C. was on leave from the Universidade Regional do Cariri, Crato (CE), Brazil.
REFERENCES 1. Verwey, E. J. W., and Overbeek, J. Th. G., “Theory of Stability of Lyophobic Colloids.” Elsevier, Amsterdam, 1948. 2. Buscall, R., and Ottewill, R. H., in “Polymer Colloids” (R. Buscall, T. Corner, and J. F. Stageman, Eds.). Elsevier, London, 1985. 3. Hunter, R. J., “Foundations of Colloid Science,” Vol. 1. Oxford Univ. Press, Oxford, 1991. 4. Gregory, J., Crit. Rev. Environ. Control 19, 185 (1989). 5. Cardoso, A. O., and Galembeck, F., J. Colloid Interface Sci. 182, 614 (1996). 6. Cardoso, A. O., and Galembeck, F., J. Colloid Interface Sci. 204, 16. (1998). 7. Galembeck, F., Braga, M., Silva, M. C. V. M., and Cardoso, A. H., Colloids Surfaces 164, 217 (2000). 8. Cardoso, A. H., Moita-Neto, J. M, Cardoso, A. O., and Galembeck, F., Colloid Polym. Sci. 275, 244 (1997). 9. Moita-Neto, J. M., Monteiro, V. A. R., and Galembeck, F., Colloids Surfaces 108, 83 (1996). 10. Ottewill, R. H., and Satgurunathan, R., Colloid Polym. Sci. 265, 845 (1987). 11. Tamai, H., Hasegawa, M., and Suzawa, T., Colloids Surfaces 51, 271 (1990). 12. Tamai, H., Fujji, A., and Suzawa, T., J. Colloid Interface Sci. 116, 37 (1987). 13. Kamei, S., Okubo, M., and Matsumoto, T., J. Polym. Sci., Polym. Chem. Ed. 24, 3109 (1986). 14. Takayasu, M. M., and Galembeck, F., J. Colloid Interface Sci. 155, 16 (1993). 15. Cardoso, A. H., Leite, C. A. P., and Galembeck, F., Langmuir 14, 3187 (1998). 16. Takayasu, M. M., and Galembeck, F., J. Colloid Interface Sci. 202, 84 (1998). 17. Nunes, S. P., and Galembeck, F., J. Polym. Sci. Polym. Lett. Ed. 21, 49 (1983). 18. Otake, K., Webber, S. E., Munk, P., and Johnston, K. P., Langmuir 13, 3047 (1997). 19. Lyklema, J., Sringer, M. M., Shilob, V. N., and Dukhin, S. S., J. Electroanal. Chem. 198 19 (1986). 20. Chonde, Y., and Shabrang, M., J. Colloid Interface Sci. 186, 248 (1997). 21. Zhao, J., and Brown, W., J. Colloid Interface Sci. 179, 281 (1996). 22. Niu, B.-J., and Urban, M. V., J. Appl. Polym. Sci. 60, 389 (1996). 23. Galembeck, F., and Souza, E. F., “Polymer Interfaces and Emulsions” (K. Esumi, Ed.), p. 119. Marcel Dekker, Inc., New York, 1999.