Studies on the Preparation and Characterization of Monodisperse

Langmuir 2005, 21, 9911-9916

9911

Studies on the Preparation and Characterization of Monodisperse Polystyrene Latices. VI. Preparation of Zwitterionic Latices† P. S. Bolt, J. W. Goodwin, and R. H. Ottewill* School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 ITS, U.K. Received May 19, 2005. In Final Form: August 10, 2005 The preparation of zwitterionic latex particles is reported by using a mixed anionic and cationic initiator system without requiring surface-active agents. Isoelectric points were found from microelectrophoresis experiments and were in the pH range of 3.5-5. Close to the isoelectric point, the latices coagulated as expected, and good stability was achieved outside this narrow range. This range of stability was in good agreement with predictions from current theory. Redispersion after coagulation was found to be difficult as was expected for a hydrophobic colloid. The electrokinetic behavior did not result in the maximum in ζ potential at an electrolyte concentration of 1 mM unlike the situation for other hydrophobic polystyrene latex particles, and hence these systems may be even better models for other colloidal studies.

1. Introduction During the last 30 years, monodisperse polystyrene latices have become widely used as model systems for the experimental elucidation and validation of a variety of problems in colloid science. The areas of interest range from the electrical properties such as the conductivity of concentrated systems,1 dielectric spectroscopy2 as well as electrophoretic behavior,3 through the stability to aggregation,4 scattering by light and neutrons6,7 to the rheological behavior8,9 of concentrated dispersions. In addition to the spherical shape and monodispersity of the particles, the ability to build in specific chemical groups to the surface that can then be characterized with precision has led many groups to choose these systems. The surface of the particles can be predominantly hydrophobic with strong acid, weak acid, or strong base groups covering approximately 10% of the surface.10 Copolymerization of styrene with an ionic monomer will lead to a hydrophilic surface, and this has been exploited to give zwitterionic latices by Homola and James.11 One difficulty with such systems is the gellike nature of the surface that can be produced as blocks of ionic monomer †

Part of the Bob Rowell Festschrift special issue. * To whom correspondence should be addressed. E-mail: [email protected]. (1) Zukoski, C. F.; Saville, D. A. J. Colloid Interface Sci. 1987, 115, 422-436. (2) Schwann, H. P. Schwartz, G.; Maczuk, J.; Pauly, H. J. Phys. Chem. 1962, 66, 2626-2635. (3) Ottewill, R. H.; Shaw, J. N. J. Electroanal. Chem. 1972, 37, 133142. (4) Ottewill, R. H.; Shaw, J. N. Discuss. Faraday Soc. 1966, 42, 154163. (5) Cornell, R. M.; Goodwin, J. W.; Ottewill, R. H. J. Colloid Interface Sci. 1979, 71, 244-266. (6) Brown, J. C.; Pusey, P. N.; Goodwin, J. W.; Ottewill, R. H. J. Phys. A 1975, 8, 664-662. (7) Cebula, D. J.; Goodwin, J. W.; Jeffrey, G. C.; Ottewill, R. H.; Parentich, A.; Richardson, R. A. Faraday Discuss. Chem. Soc. 1983, 76, 37-52. (8) Buscall, R.; Goodwin, J. W.; Hawkins, M. W.; Ottewill, R. H. J. Chem. Soc., Faraday Trans. 1 1982, 78, 2888-2899. (9) Goodwin, J. W.; Hughes, R. W. Adv. Colloid Interface Sci. 1992, 42, 303-351. (10) Goodwin, J. W.; Ottewill, R. H.; Pelton, R.; Vianello, G.; Yates, D. E. Br. Polym. J. 1978, 10, 173-180. (11) Homola, A.; James, R. O. J. Colloid Interface Sci. 1977, 59, 123129.

and are polymerized into the polystyrene chain. In addition to producing a surface that is swellable by changes in pH or ionic strength, there is always a risk of water-soluble polyelectrolyte species that can be difficult to remove from the latex and reduce the value of the system for model studies. Particles coated with protein also suffer from the gellike nature of the surface. Oxide systems are difficult to produce in a monodisperse spherical form in quantity, but even when available, problems can arise from the solubility of the particles in aqueous media. It is therefore desirable to have a model colloidal system that is insoluble and has a typical latex type of hydrophobic surface. This article reports a method by which this may be readily achieved. 2. Experimental Section 2.1. Materials. The distilled water used was doubly distilled with the second distillation having been carried out using an all-Pyrex glass still. The sodium chloride and the sodium hydroxide used were B.D.H. analar grade, and the hydrochloric acid was B. D. H. laboratory reagent grade. The styrene was B. D. H. laboratory reagent grade material that was purified by distillation at a temperature of 40-50 °C under a nitrogen atmosphere at a pressure of ca. 5 mm of mercury. After purification, the styrene was stored at 5 °C in a nitrogen atmosphere. This material was always used within 1 week of the purification procedure. The initiators used were azobiscyanopentanoic acid (ABCPA) (purum grade from Fluorochem Ltd.) and azobisisobutyramidine dihydrochloride (ABIBA) (supplied by Fison Ltd.). Both were used without further purification. To make the ABCPA watersoluble, it was dissolved in an equimolar solution of sodium hydroxide to produce the disodium salt. 2.2.i. Preparation of Latices. The procedure used was the same as has been described previously12 with the exception that two initiators were used in this work. All of the preparations were carried out at 70 ( 1 °C under a nitrogen blanket in a 1 dm3 round-bottom flask. Distilled water (0.72 dm3) was adjusted to pH 8 by the addition of a 10-2 mol dm3 solution of sodium hydroxide and was then added to the flask. Styrene was then added (normally a charge of 36 g was used), and the mixture was stirred with a T-shaped paddle at 350 rpm. Nitrogen was passed at a slow flow rate across the surface of the reaction mixture and (12) Goodwin, J. W.; Heamrn, J.; Ho, C. C.; Ottewill, R. H. J. Polym. Sci. 1974, 252, 464-471.

10.1021/la058012r CCC: $30.25 © 2005 American Chemical Society Published on Web 09/27/2005

9912

Langmuir, Vol. 21, No. 22, 2005

Bolt et al.

Table 1. Latex Preparation Conditions, Mean Particle Diameter, and Isoelectric Points latex no. PB

monomer conc. %

total [initiator] mol dm-3 × 104

ratio -/+

separation in addition time/min

pH at iep

particle diameter /nm

12 8 15 16 10 13b 13a 14 5 8 10 11

5 5 5 10 5 5 5 5 5 5 5 5

9.2 9.6 8.12 8.12 9.6 9.6 9.6 3.06 21.2 9.6 9.6 9.6

1:1 1:1 1:1 1:1 2:1 2:1 2:1 2:1 2:1 1:1 2:1 3:1

45 60 360 360 45 200 490 45 0 60 45 45

4.4 9 3.8 4.4 4.1 4.5 4.2 3.0 3.8 9 4.1 4.3

polydisperse ∼1000 600 ∼2000 530 340 360 600 poly- disperse ∼1000 530 ∼2000

was allowed to escape through the condenser and two wash bottles, the second containing a 10% ferrous sulfate solution. A period of 30 min was allowed for the attainment of thermal equilibrium and the removal of oxygen from the system. A weighed amount of ABCPA was dissolved in 25 cm3 of distilled water containing an equimolar amount of sodium hydroxide. The appropriate amount of ABIBA was weighed and dissolved in a separate aliquot of distilled water. The initiator solutions were added to the reaction mixture and washed in with 10 cm3 in each case. In all of the preparations, the anionic initiator was added first, and this was followed by the addition of the cationic material. However, the time interval between the addition of the two initiators was varied systematically from 0 to 490 min. The other experimental conditions that were varied were the ratio of the two and the total initiator concentration. In addition, one experiment was carried out at the higher monomer concentration of 10%. A reaction time of 24 h was chosen as convenient. After this period, there was no sign of a separate monomer phase, and there was only a slight smell of styrene. The latices were then filtered through glass wool (B. D. H. lead-free material) to remove any coagulum that might have formed during the polymerization. After cooling, a 20 cm3 sample of each latex was analyzed by dry weight to give the percentage yield of polymer as latex. All latices were then transferred to well-boiled Visking dialysis tubing and dialyzed against distilled water adjusted to pH 8 using a 10-2 mol dm3 sodium hydroxide solution. The dialysis was carried out in Pyrex glass containers with a ratio of latex dialysate of 1:15. Fifteen to twenty changes of dialysate were used. The dialyzed latices were subsequently stored in polypropylene containers that had been cleaned with concentrated nitric acid and distilled water followed by extensive steaming. 2.2.ii. Particle Size Analysis. The majority of the latices were sized by transmission electron microscopy as previously described12 using a Hitachi HU 11B microscope and a Carl Zeiss TGZ3 particle size analyzer. However, two of the latices with large particle diameters (PB8, PB16) were sized by optical microscopy. These sizes are recorded as approximate values because too few particle were examined for good statistical analysis. 2.3. Determination of Electrophoretic Mobility. The mobility of the latex particles was measured as a function of pH using microelectrophoresis with a Mattson cell of the type described by Alexander and Saggers.13 Platinum black electrodes were used to minimize polarization. Dilute latex suspensions were prepared in solutions containing sodium chloride at concentrations of 1 × 10-4, 1 × 10-3, 1 × 10-2, and 5 × 10-2 mol dm3. The volume fraction of these solutions was ∼10-4. The pH of aliquots of these solutions was adjusted to the required value by the addition of dilute solutions of hydrochloric acid or sodium hydroxide. The electrophoretic mobilities were determined from the average of 20 velocity measurements with the applied potential reversed between each measurement. 2.4. Coagulation Studies. Samples of the latices were diluted to give volume fractions of 10-4 using sodium chloride solutions at concentrations of 1 × 10-4, 1 × 10-3, and 1 × 10-2 mol dm3. Dispersions (500 cm3 of each) were prepared, and 50 cm3 aliquots of each were pipetted into graduated flasks. These solutions were

coefficient of variation /%

9.9 5.0 7.5 2.6 6.2 5.0

at pH ∼7, and dilute hydrochloric acid was added to adjust the pH value of the aliquots to values between 2 < pH < 7. The latices were allowed to stand for 24 h at room temperature. After this period, the contents of each flask were gently mixed, and 15 cm3 aliquots were removed and centrifuged in a bench centrifuge for 5 min at 15 × 103 g. A sample of the supernatant was transferred to a 1 cm path length optical cell, and the optical density was measured at 436 nm in a Unicam SP600 spectrophotometer. The initial mixing process was used to ensure that no single particles were in a sedimented state, and the centrifugation time was varied to show that the turbidity of the supernatant did not vary significantly. 2.5. Redispersion of Coagulated Latex. A sample of latex PB13b was diluted with a l × 10-2 mol dm-3 sodium chloride solution to give 500 cm3 of a dispersion with a volume fraction of 2 × 10-4. Dilute hydrochloric acid was used to adjust the pH of the dispersion to a value of 4.1 (i.e., previously determined as the isoelectric point of the latex). This dilute latex was allowed to stand for 24 h before being shaken and aliquots removed. The pH of each sample was adjusted to its new value by the addition of dilute hydrochloric acid or sodium hydroxide with water to give half the starting volume fraction. These solutions were allowed to stand for another 24 h. After further gentle mixing followed by centrifugation, the percentage transmission of the supernatant dispersion was measured.

3. Results 3.1. Latex Preparation. The preparation conditions with the resulting particle diameters and coefficients of variation are given in Table 1. 3.1.i. Effect of Time of Addition and Ratios of Initiators. At a fixed ratio of initiators, the particle diameter was most sensitive to the time separation between the addition of the initiators at the early stages of the polymerization. It should be noted that a latex was formed in each preparation ca. 15 min after the first initiator addition. If the second initiator was added too early, then a polydisperse product was obtained. A reduction in particle size was achieved the later the addition of the second initiator. The particle size was also reduced as the ratio of the initiators increased in favor of the anionic initiator (which was added first). 3.1.ii. Variation of Monomer and Initiator Concentrations. Only one of the preparations (PB16) was carried out at a monomer concentration other than 5% v/v (i.e., 0.44 mol dm-3), and the resulting particle size increased from 600 nm to 2 µm, which was much greater than expected from previous polymerization studies.12 However, there is insufficient data available on this system to regard this as indicative of a significant trend. This is also true of the initiator variation except that it is interesting that a 3-fold reduction in initiator concentration between preparation PB10 and PB14 increased the particle size by only 13% from 0.53 to 0.60 µm. This is a much smaller increase than observed for single initiator systems11

Preparation of Zwitterionic Latices

Figure 1. ζ potential as a function of pH for latex PB13a at various sodium chloride concentrations.

Figure 2. ζ potential as a function of pH for latex PB10 at various sodium chloride concentrations.

Figure 3. ζ potential as a function of pH for latex PB13b at various sodium chloride concentrations.

3.2. Microelectrophoresis. The latices were found to have a point of zero mobility at a pH value within the range of 3-5. The mobility was positive at low pH values and negative at high values. Latex PB8 was the only exception to this result with a point of zero mobility at pH 9. Previous measurements of the mobility of cationic particles with amidine headgroups indicate14 that the change in mobility at similar high pH values was due to the adsorption of silicate species; therefore, it is probable that latex PB8 was predominantly cationic. The electrophoretic mobility data were converted to ζ-potentials using the tables of Ottewill and Shaw,13 and these data are plotted in Figures 1-3 for latices PB10, PB13a, and PB13b. These plots show that a clearly defined isoelectric point can be observed (i.e., the point of zero mobility), which was dependent only on the H-ion concentration and was independent of the ionic strength. The isoelectric point of all of the preparations are listed in (13) Alexander, A. E.; Saggers, L. J. Sci. Instrum. 1948, 25, 374380. (14) Goodwin, J. W.; Ottewill, R. H.; Pelton, R. Colloid Polym. Sci. 1979, 257, 61-69.

Langmuir, Vol. 21, No. 22, 2005 9913

Figure 4. Transmission of the supernatant solutions of latex PB10 at 10-3 and 10-2 M sodium chloride concentrations as a function of pH.

Figure 5. Transmission of the supernatant solutions of latex PB13a at 10-4, 10-3, and 10-2 M sodium chloride concentrations as a function of pH.

Table 1. The range of isoelectric points varied between pH 3.0-4.5. The uncertainty in these values was approximately (0.1 pH units. There was a slight tendency to produce particles with a higher isoelectric point as either the initiator ratio was reduced toward 1:1 or as the period between the addition of the initiators was increased to 200-400 min. To determine whether any contamination of the latices could have occurred during the dialysis process at pH 8,15 a sample of latex PB10 was treated with a mixed-bed ionexchange resin that had been purified by the method described by van den Hul, Vanderhoff, Tausk, and Overbeek.16 The dependence of the microelectrophoretic mobility on pH was followed at a sodium chloride concentration of 1 × 10-2 mol dm-3. Good agreement was found with the results without ion-exchange cleanup. The isoelectric point was found to be at pH 4.0 ( 0.1, and a slight increase in mobility was found at high pH values although this was within experimental error. It was therefore thought that the possibility of contamination during dialysis was not important in this case. 3.3. Coagulation Studies. The percentage transmission data obtained by measuring the supernatants from latices PB10 and PB13a at various pH values are plotted in Figures 4 and 5 for a range of salt concentrations. The marked increase in transmission that was observed close to the isoelectric point in both Figures is the lowering of the particle number concentration in the supernatants and so indicates regions of instability. In both cases, the pH range over which the particles are unstable increases with increasing electrolyte concentration. In addition, the (15) Yates, D. E.; Ottewill, R. H.; Goodwin, J. W. J. Colloid Interface Sci. 1977, 62, 356-358. (16) Vanderhoff, J. W.; van den Hul, H. J.; Tausk, R. J. M.; Overbeek, J. Th. G. In Clean Surfaces; Goldfinger, M., Ed.; Academic Press: New York, 1970.

9914

Langmuir, Vol. 21, No. 22, 2005

Figure 6. Transmission of the supernatant of latex PB13a at 10-2 M as a function of pH after attempting to redisperse material held at the isoelectric point.

boundaries of the unstable regions become most sharply defined at the lowest sodium chloride concentration (i.e., the percentage transmission changes more rapidly at the lowest electrolyte level used). 3.4. Redispersion. Figure 6 shows the spectrophotometer data obtained from the redispersion experiment using latex PB13b at 10-2 mol dm-3 sodium chloride concentration. This curve is much broader than those plotted in Figures 4 and 5, and the minimum transmission value is twice that obtained in the coagulation experiments although the initial number concentration used in this experiment was twice that used to study coagulation. These data clearly show that the latex cannot be fully redispersed after coagulation at the isoelectric point has occurred. The partial redispersion indicated by the lowering of the percentage transmission readings at pH values well away from the isoelectric point could be caused by the shear forces produced by the mechanical agitation during the readjustment of the pH breaking up some of the aggregates into smaller units that did not sediment very rapidly. 4. Discussion 4.1. Latex Preparation and Control of the Isoelectric Point. The results given in section 3.1 and Table 1 show that zwitterionic latices can be prepared by a surface-active agent free polymerization using a mixture of anionic and cationic initiators (i.e., functional monomers are not required). Monodispersity was achieved by the addition of the initiator that was required to give the lowest number of surface groups. This was the cationic material in this work. This, in combination with the ratio of the initiator concentrations, is an important feature. Control of the isoelectric point was sought through the weak acid carboxyl group rather than the relatively strong base amidinium hydrochloride group. For example, if the number density of carboxyl groups on a particle surface is twice that of the amidinium hydrochloride, then the isoelectric point would be at the pH corresponding to the pKa of the carboxyl group. This is simplistic of course because the latex particle surface should have a more or less polyelectrolyte surface because the density means that the separation between groups is normally between 1 and 3 nm. The rates of decomposition of the initiators is also a major factor. At 70 °C, the half-life of the ABCPA was 520 min, and that of the ABIBA was 100 minutes. Hence, the addition time of each species was important. The nucleation mechanism of polystyrene latex formation, in the absence of surface-active agents, has long been established as one of homogeneous initiation followed by controlled

Bolt et al.

Figure 7. Ratio of cationic to anionic free radicals in solution as a function of time from addition at a total concentration of 0.66:1 cationic/anionic.

coagulation to give nuclei that then grow. The coagulation is controlled by the increasing colloid stability of the particles as the surface charge density grows as precursor particles are added. For monodispersity to be attained, the stability must rise rapidly at low conversions so that most of the polymerization is a particle growth process. If either new nuclei are stabilized or stability becomes compromised later in the reaction, then a monodisperse product is unlikely. This latter problem is the one that is solved, in the main, by the time delay between the time of addition of the two initiators, although the ratio of concentrations is also important. As an illustration of the solution, we can calculate the ratio of the number of negatively charged radicals produced to that of positively charged radicals. If these radicals were all to initiate polymerization, then the problem of maintaining stability becomes clear if that ratio approaches unity at any time during the polymerization. Clearly a radical efficiency of 100% is unrealistic, as is the assumption that all of the polymer end groups will always be at the surface. However, in the spirit of the illustration we can assume that the radical efficiency is 100%, that at the high pH of 8 used in the polymerization all of the carboxyl groups remain as charged species on the surface, and that two-thirds of the amidinium groups also remain. Figure 7 shows how the ratio of anionic to cationic species would vary with time if both initiators were added together and if the cationic species was added with a delay of 45 min. The results clearly indicate that at zero delay time there is a change from a cationic latex to an anionic one well into the reaction. This would result in the aggregation of the product at this time, although the change to an excess negative charge would occur quite quickly. Any product in the form of a latex would therefore be polydisperse. Latex PBS falls into this category. However, in Figure 7, a delay time of 45 min for the addition of the cationic initiator is sufficient to ensure an anionic product over the whole reaction time. This is the category that PB10 falls into, for example. Another obvious feature made clear by Figure 7 is that at long times the delay time has no effect on the final ratio of charged species produced. Thus, little control of the isoelectric point can be expected by altering the addition times unless the cationic species are added very late in the polymerization. This would then tend to lower the IEP slightly. This is analogous to what happened to PB5 where many surface groups would be buried and lost in the late aggregation stage and further nucleation and polymerization would be at a much lower cationic-toanionic ratio. The main conclusion is that post addition of the minor initiator enables stability to be maintained after the initial nucleation stage so that monodispersity may be achieved but little variation in the IEP is obtained.

Preparation of Zwitterionic Latices

Figure 8. ζ potential as a function of sodium chloride concentration for latex PB13b. Upper line, pH 3.2; lower line, pH 6.2. The lines were fitted to ζ ∝ [NaCl](12.2.

4.2. Microelectrophoresis. The microelectrophoresis data shown in Figures 1-3 show that a clear isoelectric point was obtained for the latices, and these are shown in Table 1. This means that the there was no effect of the concentration of sodium chloride on the point of zero mobility; this is a similar result to that observed with oxide particles. This, however, has some implications for our perception of the latex surface. With discrete charged groups in relatively close proximity, the degree of ionization of the weak acid groups could be expected to depend on both the ionic strength and pH. If this were the case, then the point of zero mobility would drift to lower pH values with increasing sodium chloride concentration. This was not found here. Another interesting feature for these zwitterionic particles is shown in Figure 8. The ζ potential was found to decrease monotonically with salt concentration on either side of the IEP. This is in contrast to homoionic latices where a maximum in the ζ potential is found at an electrolyte concentration of -1 × 10-3 mol dm-3.17 It is similar to what is observed for oxide particles and latices prepared using charged comonomers.18 It is normally suggested that the maximum in the ζ potential for low charge density latices is due to surface conduction (e.g., Hunter19). Because the effect does not occur with these systems, it suggests that they are worthy of further electrokinetic study. 4.3. Coagulation Studies. The coagulation studies showed that there was a region of instability around the IEP of the latices and that this region became wider as the electrolyte concentration increased. A comparison of the electrophoretic data and the percentage transmission data indicates that the rise in transmission (i.e., the onset of coagulation) occurs when the ζ potential is ζ e (20 mV. Figure 9 shows some pair potential curves calculated from the DLVO theory20,21 for latex PB13a in a 1 × 10-3 mol dm-3 sodium chloride solution. The stability ratio, W, given by Fuchs22 gives the rate of coagulation relative to the most rapid rate due to Brownian encounters.23,24 The approximate form for W given by Prieve and Ruckenstein25 is

[ ( ) ]

W ) 1 + 0.25 exp

Vm -1 kT

Langmuir, Vol. 21, No. 22, 2005 9915

Figure 9. Relative pair potential as a function of interparticle separation calculated for ζ potentials of 10, 20, and 30 mV.

Figure 10. Stability ratio, W, as a function of sodium chloride concentration for latex PB10.

Figure 11. Stability ratio, W, as a function of sodium chloride concentration for latex PB13a.

The results of the calculation of W for latices PBI0 and PB1I3a under the conditions used for the coagulation experiments are shown in Figures 10 and 11. The calculated curves climb steeply on either side of the IEP, indicating a reduction in coagulation rate and an increase in colloidal stability. They are also in agreement with the experimental data in that the unstable range increases with increasing electrolyte concentration. To estimate the range of pH over which instability can be expected, the time scale available for the aggregation must be considered. In the experiment described here, the latices were allowed to undergo aggregation for a period of 24 h. Now the characteristic time for doublet formation26 can be calculated from the rate of aggregation. Hence the stability ratio required to maintain stability over the time scale of the coagulation experiment may be estimated from

where Vm is the energy mximum in the pair potential. (17) Bensley, C. N.; Hunter, R. J. J. Colloid Interface Sci. 1983, 92, 448-462. (18) Zukoski, C. F.; Saville, D. A. J. Colloid Interface Sci. 1986, 114, 45-53. (19) Hunter, R. J. Foundations of Colloid Science II; Oxford University Press: Oxford, U.K., 1989. (20) Deriaguin, B. V.; Landau, L. Acta Physicochim. URSS 1941, 14, 633-662.

(21) Verwey, E. J. W.; Overbeek, J. Th. G. “Theory of Lyophobic Colloids; Elsevier: Amsterdam, 1948. (22) Fuchs, N. Z. Phys. 1934, 89, 736-743. (23) von Smoluchowski, M. Z. Phys Chem. 1917, 92, 129-168. (24) Spielman, L. A. J. Colloid Interface Sci. 1970, 33, 562-571. (25) Prieve, D. C.; Ruckenstein, E. J. Colloid Interface Sci. 1986, 73, 539-555. (26) Russell, W. B.; Saville, D. A.; Schowalter, W. R. Colloidal Dispersions; Cambridge University Press: Cambridge, U.K., 1989.

9916

Langmuir, Vol. 21, No. 22, 2005

the experimental time:

We )

πηa3 t φkT e

Here the subscript e refers to the experiment, η is the viscosity of the electrolyte solution, a is the particle radius, and φ is the volume fraction used in the experiment. From this expression, a value of We ≈ 103 was required in these experiments. The widths of the pH ranges for instability shown in Figures 4 and 5 are in excellent agreement with the range of pH corresponding to We ≈ 103 in Figures 10 and 1 1. 5. Conclusions The experiments described in this article show how zwitterionic polystyrene latex particles can be prepared

Bolt et al.

without using ionic comonomers or surface-active agents. This is accomplished by using both an anionic and a cationic initiator, but a delay of the addition of the second initiator is required to maintain monodispersity. Clear isoelectric points are obtained, but the control of these requires further experimentation. The electrokinetic behavior of the particles was unusual, when compared to that of a homoionic low-charge latex, and suggests that these systems may be better models for the study of colloidal behavior. The coagulation behavior gives good agreement with predictions from existing theory. Complete redispersion of the coagulated dispersions was not successfully achieved, and this is also what can be expected from a lyophobic colloid and is in contrast to the behavior of lyophilic zwitterionic latices prepared with ionic comonomers. LA058012R