- Email: [email protected]
Electrophoretic Properties of Monodisperse Polystyrene Particles
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
179, 522–531 (1996)
0245
Electrophoretic Properties of Monodisperse Polystyrene Particles G. TUIN, J. H. J. E. SENDERS,
AND
H. N. STEIN 1
Laboratory of Colloid Chemistry and Thermodynamics, Department of Chemical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands Received April 3, 1995; accepted October 25, 1995
In this paper the electrophoretic properties of large monodisperse polystyrene (PS) latices, prepared by a one-step surfactantfree emulsion polymerization, are described. The latices were characterized by conductometric titration. The charged groups, arising from initiator fragments, were all strong acid groups and no carboxyl groups could be detected. The amount of surface charge groups varies with particles size. Electrophoresis measurements have shown that the electrophoretic mobility passes through a maximum as the electrolyte concentration is increased. Converting the electrophoretic mobility to z -potentials, taking into account electrophoresis and retardation effects but neglecting surface conductivity gives the same picture. The same phenomenon is observed with increasing H / concentration. Only at very high electrolyte or H / concentrations does the z -potential reach values close to zero. The z -potentials were converted to charge densities behind the electrokinetic shear plane ( sz), using the Gouy–Chapman theory. In general, PS particles behave as expected without discrepancies between experiment and theory: És0É ( s0 is the surface charge density) is larger than ÉszÉ. For one particular latex, however, a ÉszÉ larger than És0É was found. This was observed at high electrolyte concentrations, where uncertainties in the z -potential because of neglect of surface conductivity are not important. This fact was attributed to chemisorption of NO 0 3 ions. Chemisorption of NO 0 3 ions is not a common fact, but may be attributed to the pronounced hydrophobic character of the PS particles. q 1996 Academic Press, Inc. 0 Key Words: chemisorption, NO 0 3 , on polystyrene; NO 3 chemisorption, on polystyrene; zeta potential, polystyrene; polystyrene, nitrate chemisorption on; polystyrene, surface charge.
INTRODUCTION
Monodisperse polystyrene (PS) particles are widely used as model colloids for, e.g., adsorption studies. Aqueous PS latices are usually prepared by an emulsion polymerization. The emulsion polymerization can be performed either with a surfactant or without an intentionally added surfactant (‘‘soap-free’’ or emulsifier-free polymerization). The surfactant-free emulsion 1
To whom correspondence should be addressed.
522
0021-9797/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AID
JCIS 4099
/
6g0e$$$561
polymerization has the advantage that the only charged groups present on the PS latex surface are sulfate groups, arising from the initiator fragments. The preparation of PS particles by surfactant-free emulsion polymerization is well described (1–5). The particle size of these latices is usually below 1 mm, when prepared by a one-step surfactant-free emulsion polymerization. Recently, we described a method for obtaining particle sizes up to 3.2 mm in a one-step surfactant-free emulsion polymerization (6). However, it has been reported that the model PS particles have an anomalous electrokinetic behavior. Theoretically it is expected that if there is no chemisorption of ions present, an increasing electrolyte concentration at a constant pH should lead to a continuous decrease in the absolute value of the electrophoretic mobility or zeta-potential ( z ), because of compression of the electric double layer. This decrease is indeed observed, but only at rather high electrolyte concentrations ( ú0.01 M) (7). As the electrolyte concentration is increased from low to high concentrations, the electrophoretic mobility (or the zeta-potential calculated from it) first increases in absolute value and it decreases only when a certain electrolyte concentration is surpassed. The zeta-potential versus electrolyte concentration curves therefore exhibits a maximum in absolute value. This phenomenon has been observed by many authors (7–22). The observed maximum is explained by the hairy layer model, the co-ion adsorption model (7, 16–19), or by surface conductivity (22). The hairy layer model assumes that the latex particle is covered by a layer of protruding, rather flexible, polymer chains (‘‘hairs’’) having terminal end groups (the sulfate groups introduced by the initiator fragments) (7–15). As the ionic strength decreases, the repulsion between these charged groups increases and the hairy layer expands. This results in the slipping plane moving outwardly, and the net charge transported electrokinetically decreases. In addition, a hairy layer provides a convenient background for introducing electrical conductance within the shear plane. The co-ion adsorption model assumes, in the case of negatively charged particles, that anions, which are less hydrated than cations, are closer to the apolar surface (7, 16–19).
04-17-96 17:15:56
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
These anions are not believed to be chemisorbed on specific sites and the magnitude of the electrokinetic potential depends on the valency of the co-ions. In the surface conductivity model, the presence of mobile Stern-layer anions causes the mobility to decrease and the conductivity to increase, in comparison with the case where surface conductivity is absent (22). According to this model, not only does the diffuse double layer contribute to the surface conductance, but also a process of additional conductance between slipping plane and particle interface is operative. It is obvious that there seems to be a controversy about the explanation of the electrokinetic behavior of the PS latices. In this paper the electrophoretic properties of large monodisperse PS particles, of which the preparation is described elsewhere (6), are investigated. By varying the particle size of these latex particles, ka values larger than 10 are reached (where k is the Debye–Hu¨ckel parameter and a is the particle radius). For these ka values the theory for calculating z -potentials from measured mobilities (23) is reliable, as long as surface conductivity can be neglected. The PS particles are prepared by a one-step surfactant-free emulsion polymerization. In this way PS particles are obtained with a uniform surface composition. EXPERIMENTAL
Materials Water was twice distilled from an all glass apparatus. The specific conductivity (0.8 mS cm01 ) and the surface tension (72 mN/m) of this water indicated that it was free of surfaceactive impurities. Sodium nitrate, potassium nitrate, sodium hydroxide, and nitric acid (Titrisol) were obtained from Merck (pro analysi, purity ú99%) and used without purification. The ion-exchange resins DOWEX 50W-X4 (H / form) and DOWEX 1-X4 (OH 0 form) were obtained from Fluka and purified as described by Van den Hul and Vanderhoff (24). Latices The preparation of the latices used is described elsewhere (6). The particle sizes of the latices are listed in Table 1. Duplicate measurements of the particle size differed by less than 2%. Cleaning of the Latices All latices were dialyzed by the serum replacement technique as described by Vanderhoff et al. (25) and Ahmed et al. (26) in Amicon serum replacement cells. For a description of the dialysis cells see (26). The latex was first diluted to a solid content less than 5% (w/v) with twice-distilled water. Then a volume of 300 cm3 of the latex was placed in the stirred cell. This cell was
AID
JCIS 4099
/
6g0e$$$562
04-17-96 17:15:56
523
equipped with a Nucleopore membrane (Poretics Corporation, U.S.A.) with a pore size approximately 10–20% smaller than the particle size of the latex. Then a continuous stream of twice-distilled water was flown through the cell until the conductivity of the outlet water reached the same value as that of the inlet twice-distilled water (0.8 mS cm01 ). Normally this took about 24 h. First all the dissolved electrolyte was removed and then the latex was removed from the cell. This procedure was applied to all samples of latex. Determination of Solids Content The solids percentages of the different PS latices were determined by drying a known amount of the latex in an oven at 1057C, until constant weight was reached. Latex Titration The ion-exchange resins used in this study were the DOWEX 50W-X4 (cation exchange, H / form) and the DOWEX 1-X4 (anion exchange, OH 0 form). These ion-exchange resins were separately cleaned according to the same procedure as described by Van den Hul and Vanderhoff (24). After the cleaning procedure the wash water of the ionexchange resins exhibited no surface-active impurities (surface tension 72 mN/m at 207C) and the conductivity of the wash water had the same value as the inlet water used (0.8 mS cm01 ). Equal quantities of the two ion-exchange resins were then mixed together. The wash water from the mixed resin had a considerably lower conductivity (0.25 mS cm01 ) than the water used. The freshly cleaned latices (by the serum replacement technique) were diluted with ion-exchanged water to a solids percentage lower than 5% and were transferred into a Pyrex glass bottle with a large amount of mixed ion-exchange resin. The bottles were placed on a roll-bank and were rolled for 3 to 24 h. At specific times (3, 6, 9, and 24 h) samples were taken to be analyzed by conductometric titration. Latex and ion-exchange resin were, after their contact, separated by decantation followed by filtration through a glass filter. Conductometric titrations of the ion-exchanged latices were performed at 207C under a nitrogen atmosphere. The solid content was always low (1–2% w/v). The conductivity was measured using a conductivity meter (Radiometer Kopenhagen) equipped with a Schott conductivity cell. A 2.1 mM NaOH solution (nitrogen flushed) was added with a constant rate pump (LC-5000 syringe pump, ISCO). The change in conductivity was continuously monitored using a computer. The influence of the rate of adding NaOH solution on the titration was checked for several rates ( 0.1 to 40 ml /
coidas
AP: Colloid
524
TUIN, SENDERS, AND STEIN
TABLE 1 Characteristics of the PS Latices Used Latex
Dn (nm)
Ds (nm)
Dv (nm)
Pn
L-78a L-80a L-86b L-88b L-89b
736 381 1309 2165 3124
754 398 1357 2217 3253
764 407 1386 2252 3279
1.04 1.07 1.06 1.04 1.02
Note. Dn , number averaged diameter; Da , surface averaged diameter; Dv , volume averaged diameter; Pn , degree of polydispersity, Dv /Dn . a Determined with Coulter LS-130. b Determined with Coulter counter.
h ) , but no difference was found for the endpoint of the conductometric titration. Therefore a moderate rate of 15 ml / h was chosen. A typical titration curve ( specific conductivity kc versus quantity of NaOH solution added ) is shown in Fig. 1. The samples of the latices (taken after various ion-exchange times) were titrated three to four times and the average value was used. When the determined endpoints, obtained after different ion-exchange times, were the same (within experimental error), this was assumed to be the endpoint. This endpoint was always reached within 24 h of ion exchange. The endpoint (assuming complete dissociation of charged groups at the end of the conductometric titration and complete charge compensation of the charged groups by H / at the start of the titration) was then converted to the surface charge ( s0 ) and the area per negative charge (An ) assuming a quadratic arrangement of charged sites on the particle surface. This is shown in Fig. 2 for latex L-78 (as an example). z -Potential Measurements
are often called monodisperse if the value of Pn is õ1.05. This is especially the case for the larger particles. In Fig. 1 it can be seen that both the descending and the ascending legs of the plot of conductivity versus quantity of NaOH solution added are linear, with the exception of a region near the intersection point, which means that only one type of surface group is observed. This was found for all latices investigated. The charged surface groups are in origin sulfate groups ( 0SO 40 ) (27), but carboxyl ( 0COO 0 ) groups have also been reported (4, 28, 29) originating from oxidation of intermediate alkanol groups. In the polymerization procedure the so-called Kolthoff reaction can occur (30), resulting in the loss of sulfate groups and the appearance of hydroxyl ( 0OH) groups. These hydroxyl groups can be present at the PS–water interface as noncharged groups. The Kolthoff reaction can be suppressed by using KHCO3 as supporting electrolyte (31); however, this was considered not to be necessary in view of the titration data (see later). The chemical instability of the PS particles, prepared by using K2S2O8 as initiator, has also become apparent from other studies (see, e.g., Hearn et al. (31)). The PS latices may loose their sulfate groups upon storage by hydrolysis of the sulfate groups (31, 32). The rate of hydrolysis appears to vary widely. Goodall et al. (33), e.g., found rates of hydrolysis varying from 2 to 37% per month (at 257C), whereas Vanderhoff (34) reports the complete loss of sulfate groups of an ion-exchanged latex in 14 days. Most reports only mention the removal of the sulfate groups upon storage as a result of hydrolysis. Vanderhoff (34) and Goodall et al. (33) observed the appearance of weak acid groups. From this literature survey, it is obvious that the preparation and handling of the PS particles can have a large influence on the surface charge groups present. The PS particles used in this study were titrated 9 months after preparation. The fact that only one type of acid group could be detected
z -Potential measurements were performed using a Malvern Zetasizer 3 (Malvern Instruments, Malvern, England) equipped with an AZ-4 cell. This apparatus uses the laser– Doppler effect to measure the electrophoretic mobility. The latices used were the ion-exchanged PS latices in the H / form. In case of measuring the z -potential as function of pH the electrolyte used was either potassium nitrate or sodium nitrate, and nitric acid was used to set the pH to the desired value. RESULTS AND DISCUSSION
Latex Titration In Table 1 the particle sizes of the latices are listed. As can be seen from Table 1 the particle sizes of the latices are fairly monodisperse (indicated by Pn Å Dv /Dn ). Particles
AID
JCIS 4099
/
6g0e$$$562
04-17-96 17:15:56
FIG. 1. Change in specific conductivity ( kc ) versus added volume 2.1 mM NaOH solution for the conductometric titration of latex L-78.
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
525
FIG. 2. Change in surface charge density ( s0 ) and area per negative group (An ) versus time for the ion-exchange cycles of latex L-78. ‘‘Time’’ is the total duration of the ion-exchange process.
FIG. 4. Electrophoretic mobility (Em ) versus concentration of HNO3 for different latices. s, Latex L-78; /, latex L-80; h, latex L-86; n, latex L-89.
might suggest that only weak acid groups are present. It is, however, difficult to deduct which type of acid groups at the PS surface is detected by conductometric titration. The additional presence of surface hydroxyl groups is, of course, not excluded. In Fig. 2 the change in surface charge ( s0 ) and the area per surface charge group (An ) with ion-exchange time for latex L-78 is shown as an example. The increase in s0 and thus decrease in An vanishes after 9 h of ion exchange. Further ion exchange does not lead to an increase in s0 . A possible explanation for this could be that the ion-exchange resin is exhausted. The amount of ion-exchange resin was, however, large and using an excess of fresh ion-exchange resin did not result in a change in s0 . When treating tap water with the used ion-exchange resin the conductivity decreased to values as described before. The fact that the conductivity of the wash water from the used ion-exchange resin
was still as low as before the titration supports the fact that the obtained value for s0 is the correct one. In Fig. 3 the determined values of s0 and An are given as functions of the particle sizes of the latices. From Fig. 3 it can be seen that the surface charge density measured increases slightly and the area per negative charge decreases slightly with increasing particle size. The only exception in this respect is latex L-78, which was made by a different polymerization procedure (in a different reactor) (6). The surface charge of latex L-78 is almost as high as that of latex L-89, but the particle size of latex L-89 is more than four times larger than latex L-78. This means that the polymerization procedure can have a large influence on the final latex.
FIG. 3. The final determined surface charge densities ( s0 ) and area per negative group (An ) for the different particle sizes used.
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
Electrophoresis In HNO3 solutions. In Fig. 4 the dependence of the electrophoretic mobility (Em ) on the concentration of HNO3 is shown for the different latices used. In these experiments, the latices were in the H / form. As can be seen from Fig. 4 the electrophoretic mobilities of all latices show a maximum at a concentration of about 10 03 M, similar to the maxima reported by Voegtli and Zukoski for PS particles in HCl and HClO4 solutions (16), though these authors found in such solutions maxima at 10 02 M. The surface charge properties of PS particles is due to dissociation of either sulfate or carboxyl groups. The former have a pKa in the range 1 to 2 (35), while for the latter pKa values between 4 and 6 have been reported (36). Refinement of such values, by the site-binding model, appears not to be useful, since the fact that such values are dependent on pH and the electrolyte concentration raises the suspicion that they are no more than fitting parameters in the model equations (see also the criticism by Janssen and Stein (37)). Increasing absolute values of the electrophoretic mobility
coidas
AP: Colloid
526
TUIN, SENDERS, AND STEIN
FIG. 5. Zeta-potential ( z ) versus pH latex L-80 for different electrolytes. s, [NaNO3 ] Å 3.4 1 10 02 M; h, [KNO3 ] Å 3.3 1 10 02 M.
with increasing HNO3 concentration are found in a pH range in which incomplete dissociation is expected for carboxyl groups, but not for sulfate groups. Thus, the increase in electrophoretic mobility at low [HNO3 ] values indicates that in the polystyrene samples investigated here, sulfate groups are predominant and thus are the groups found by conductometric titration (see Fig. 1). The fact that at [HNO3 ] õ 0.001 M the absolute value of the electrophoretic mobility increases (instead of remaining constant) can best be attributed to increasing chemisorption of NO 30 ions with increasing [HNO3 ] (see below). An alternative explanation could be the hairy layer model: at low [HNO3 ] the hairy layer is extended which leads to a slipping plane far away from the phase boundary of the polystyrene, leading to low absolute values of the z -potential, while at higher [HNO3 ] the hairy layer is compressed, which leads to increasing absolute values of the z -potential. However, it will be shown later that the existence of a hairy layer is not easily compatible with other data. The decrease of the absolute value of Em at HNO3 concentrations higher than 0.001 M is too pronounced to be due to incomplete dissociation of sulfate groups and can best be understood by being due predominantly to the larger electrolyte concentration compressing the electrical double layer (cf. the data in NaNO3 or KNO3 solution at constant pH). Figure 5 shows values of the z -potential calculated from electrophoretic mobilities, measured in solutions in which, in addition to HNO3 , NaNO3 or KNO3 had been added ( É0.033 M). The calculation of the z -potentials was performed by using the equations of Ohshima et al. (23). These equations are an approximation of the numerical results obtained by O’Brien and White (38), which take into account electrophoretic and retardation effects, but neglect contributions to conductance by ions behind the electrokinetic slipping plane. At the electrolyte concentrations employed in
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
the experiments discussed here, however, the influence of the latter is negligible. The z -potentials appear to be independent of pH, at pH values ú4, in agreement with absence of significant quantities of carboxyl groups at the surface of the PS particles. The absolute values of the z -potentials decrease significantly only at pH values between 3 and 2, i.e., when the HNO3 concentration becomes on the same order as the NaNO3 or KNO3 concentration, in agreement with compression of the electrical double layer at large electrolyte concentrations as predominant cause of the decrease in absolute values of Em in Fig. 4. The data presented in Fig. 5 are average values based on four data points each. Application of the Wilcoxon test (42) to these data showed that the difference between the results observed in NaNO3 and KNO3 solutions is significant. This is not surprising in view of the fact that all average values found in KNO3 solutions are lower than those found in NaNO3 solutions at the same pH, while the difference between the respective values at one pH value is larger than the standard deviation of the average values concerned. The difference between z -potentials in NaNO3 and KNO3 solutions, respectively, can be tentatively understood from the different sizes of the hydrated cations concerned: a hydrated K / ion is smaller than a hydrated Na / ion; therefore, the average potential experienced by an adsorbed K / ion will be larger in absolute sense than that experienced by an adsorbed Na / ion. Then a larger part of the surface charge will be compensated by cations behind the electrokinetic slipping plane in the case of K / than in the case of Na / . In NaNO3 or KNO3 solutions at pH 5.5. At a pH of 5.5, any surface sulfate groups will be fully dissociated. In Fig. 6 the dependence of the z -potential of latex L-86 at pH 5.5 in NaNO3 and KNO3 solutions is shown as an example. Similar behavior is found for the other latices studied. A
FIG. 6. Zeta-potential ( z ) versus electrolyte concentration for latex L86 at pH 5.5. s, Electrolyte KNO3 ; h, electrolyte NaNO3 . The lines are just a guide for the eye.
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
maximum in z -potential is found at a concentration of approximately 6 1 10 03 M for the electrolyte NaNO3 and at a slightly higher concentration for the electrolyte KNO3 . A similar behavior was found for positively charged latices in different electrolytes by Hidalgo Alvarez et al. (13). The location of the maximum in the electrophoretic mobility versus electrolyte concentration curve depended on the nature of the counterion in their measurements. As can be seen from Fig. 6, at low electrolyte concentrations the ÉzÉ is larger in NaNO3 than in KNO3 solution. This was also observed in Fig. 5 and is explained by the difference in size of the hydrated cations. However, at high electrolyte concentrations the ÉzÉ are the same within experimental error. Electrophoretic mobilities were again converted into z potentials through the Oshima equations (23), and from the z -potentials the values of the surface charge density in the diffuse double layer beyond the electrokinetic slipping plane ( sz) were calculated, by means of the Gouy–Chapman equation valid at the large ka values (where k is the reciprocal value of the Debye length and a is the particle radius and valid in the cases investigated here (see, e.g., Hunter (39)) q
sz Å 0 8000 c e0er NAkT sinh
S D ze0z 2kT
,
[1]
where c is the bulk concentration of ions (mol/liter), NA is Avogadro’s number, e0 is the permittivity of vacuum, er is the solvent dielectric constant, z is the valency, e0 is the elementary charge, z is the zeta-potential, k is Boltzmann’s constant, and T is the temperature. In addition, we calculated the surface charge density in the region between the phase boundary and the slipping plane, sD , from the relation sz / sD / s0 Å 0,
[2]
where s0 is the surface charge density. Figures 7 and 8 show some typical data, referring to NaNO3 solutions, but similar results are found in KNO3 solution. The data were obtained for two polystyrene samples: latex L-78 with a rather high surface charge density ( 00.0863 C/m 2 ) and latex L-80 with a low surface charge density ( 00.0252 C/m 2 ). In discussing these data, we will start with the data at high electrolyte concentration. In these circumstances, the uncertainty of the z -potential, due to possible contributions of the electric conductivity by ions behind the slipping plane, is minimal. Also, in these circumstances minimal uncertainty to any action of a hairy layer arises. At high values of the pH, where the surface charge properties are due to dissociation of surface 0SO4H groups, És0É
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
527
is found to be larger than ÉszÉ, which makes sD ú 0 as expected for surfaces with a negative value of s0 . A similar behavior was found for all latices except latex L-80, for which data are plotted in Fig. 8. Here at large electrolyte concentrations, És0É õ ÉszÉ, which makes sD õ 0. For negative values of the surface charge properties, this indicates chemisorption of NO 30 ions. Similar cases of polystyrene particles with És0É õ ÉszÉ have been reported by Ma et al. (40). These authors even found negative z -potentials on polystyrene particles in the absence of any titratable surface charge. Abramson (41) also reported negative z -potentials for paraffin wax in KNO3 solution. Chemisorption of nitrate ions is not commonly assumed: nitrate ions are usually regarded as ‘‘inert’’ toward solid surfaces. However, much of the evidence on which this is based refers to surfaces of solids such as oxides, hydroxides, or AgI. With surfaces of a pronounced hydrophobic character such as polystyrene, the situation may be different as remarked by Voegtli and Zukoski (16). Such chemisorption of nitrate ions is expected primarily near surface regions which are far from 0SO 40 groups, since the latter are hydrophilic, while their negative charge supplies an additional repulsion toward nitrate ions. The interpretation of negative values of sD as due to chemisorption of NO 30 ions is supported by the fact that such values are found at relatively high NaNO3 concentrations. At lower electrolyte concentrations (especially õ0.001 M), the neglect of the surface conductivity, inherent in the use of the Oshima equation for calculating the z -potentials, becomes increasingly suspect. If this would be important, the absolute values of the z -potentials, and consequently also that of sz , as calculated by use of the Oshima equation, would be too low. In Figs. 7 and 8, the sD values at low electrolyte concentrations would be shifted in the negative direction; this may even lead to negative values of sD (indicating chemisorption of nitrate ions) at the lowest concentrations investigated here (0.0014 M). However, this can only be defended at such low concentrations if an extensive hairy layer is assumed to be present. An argument against the simultaneous effect of a hairy layer and neglect of ionic conductance by ions behind the electrokinetic slipping plane can be derived from Fig. 9. This figure shows sD values as a function of surface charge density for various electrolyte concentrations. These lines have a nearly identical slope, in spite of the fact that the lines refer to electrolyte concentrations varying from 0.036 to 0.0014 M. At the former concentration, a significant contribution of ions behind the electrokinetic slipping plane to electrical conductance, or of a hairy layer, is improbable, while at the latter concentration such contributions if present should be much more pronounced. The nearly equal slopes of all curves in Fig. 9 would be quite fortuitous if there would be a totally
coidas
AP: Colloid
528
TUIN, SENDERS, AND STEIN
FIG. 7. Charge density as function of [NaNO3 ] for latex L-78 at pH 5.5. Dashed line, s0 ; s, sz; n, sD .
different mechanism operative in determining the electrophoretic mobility, at low concentrations and at high concentrations. If the possibility of an important contribution to electric conductivity by ions behind the electrokinetic slipping plane and of a hairy layer can be dismissed, the maximum in the
ÉzÉ-potential versus electrolyte concentration at constant pH (Fig. 6) should be understood by the co-ion adsorption mechanism: at low electrolyte concentrations, NO 30 adsorption becomes increasingly important with increasing [NaNO3 ] or [KNO3 ], leading to an increase of the absolute value of the z -potential; at high concentrations, the compres-
FIG. 8. Charge density as function of [NaNO3 ] for latex L-80 at pH 5.5. Dashed line, s0 ; s, sz; D, sD .
AID
JCIS 4099
/
6g0e$$$564
04-17-96 17:15:56
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
529
FIG. 9. Change in sD versus surface charge density ( s0 ) at pH 5.5 at different [NaNO3 ]. l, 1.4 1 10 03 M; n, 3.2 1 10 03 M; j, 6.0 1 10 03 M; s, 1.3 1 10 02 M; /, 3.6 1 10 02 M.
sion of the double layer works in the opposite direction. It should be noted that chemisorption of nitrate ions must be assumed at electrolyte concentrations which are too large for either of the alternative mechanisms to be important. From Fig. 9, values of dsD/ds0 at constant electrolyte
FIG. 10.
AID
JCIS 4099
/
concentration can be deduced. These are shown, for NaNO3 solutions at pH 5.5, in Fig. 10. If no nitrate ions would be present behind the electrokinetic slipping plane, and if the Na / ions necessary for compensating the charges of additional surface 0SO 40 groups would be accommodated pre-
dsD/ds0 versus [NaNO3 ] at pH 5.5, calculated from the data in Fig. 9.
6g0e$$$564
04-17-96 17:15:56
coidas
AP: Colloid
530
TUIN, SENDERS, AND STEIN
dominantly behind the slipping plane, dsD/ds0 would be approximately equal to 01. It should be noted that the negative value of dsD/ds0 refers to s0 values in an algebraic, not in an absolute, sense. Such is, at high concentrations, the case in reasonable approximation (dsD/ds0 Å 00.89 at [NaNO3 ] Å 0.036 M). At lower concentrations, dsD/ds0 approaches 01. In principle, the decrease in absolute value of dsD/ds0 with increasing electrolyte concentration could be explained both by an increase of NO 30 incorporation behind the electrokinetic slipping plane at higher concentration or by the absolute values of the z -potentials being too low at low electrolyte concentrations through neglect of contributions to conductance by ions behind the electrokinetic slipping plane. However, the presence of significant quantities of nitrate ions behind the electrokinetic slipping plane at low electrolyte concentrations is probable only if an extensive hairy layer is assumed. Since arguments against the combined action of a hairy layer and underestimation of the z -potentials have been presented, the present authors regard an explanation along these lines as unconvincing, though the possibility cannot wholly be excluded. CONCLUSIONS
In this paper the electrophoretic properties of large monodisperse PS latices are described. The latices were characterized by conductometric titration. The charged groups, arising from initiator fragments, were all strong acid groups and no carboxyl groups could be detected. The amount of surface charge groups varies with particles size. Electrophoresis measurements have shown that the electrophoretic mobility passes through a maximum as the electrolyte concentration is increased. In this respect, the present work confirms conclusions by previous investigators. Converting the electrophoretic mobility to z -potentials, taking into account electrophoresis and retardation effects, but neglecting surface conductivity, gives the same picture. The same phenomenon is observed with increasing H / concentration. Only at very high electrolyte or H / concentrations does the z -potential reach values close to zero. The z -potentials were converted to charge densities behind the electrokinetic shear plane ( sz), using the Gouy– Chapman theory. In general, PS particles behave as expected without discrepancies between experiment and theory: És0É is larger than ÉszÉ. For one particular latex, however, a ÉszÉ larger than És0É was found. This was observed at high electrolyte concentrations, where uncertainties in the z -potential because of neglect of surface conductivity are not important. This fact was attributed to chemisorption of NO 30 ions. Chemisorption of NO 30 ions is not a common fact, but may be attributed to the pronounced hydrophobic character of the PS particles.
AID
JCIS 4099
/
6g0e$$$564
04-17-96 17:15:56
Such chemisorption of NO 30 ions had been supposed by previous investigators to be present on polystyrene (e.g., Ref. (40)). The present work confirms this and adds an important new argument from the dependence of sD on s0 at constant electrolyte concentrations, for excluding the possibility of explaining the results by the simultaneous actions of a hairy layer and of conductance by ions behind the electrokinetic slipping plane. Therefore, in discussing electrophoretic data on polystyrene in nitrate solution, chemisorption of nitrate ions should be taken into account. ACKNOWLEDGMENTS The financial support of the Foundation of Emulsion Polymerization (SEP) is gratefully acknowledged.
REFERENCES 1. Matsumoto, T., and Ochi, A., Kobunshi Kagaku 22, 481 (1965). 2. Kotera, A., Furasawa, K., and Takeda, Y., Kolloid-Z. Z. Polym. 239, 677 (1970). 3. Kotera, A., Furasawa, K., and Kudo, K. Kolloid-Z. Z. Polym. 240, 837 (1970). 4. Goodwin, J. W., Hearn, J., Ho, C. C., and Ottewill, R. H., Br. Polym. J. 5, 347 (1973). 5. Goodwin, J. W., Hearn, J., Ho, C. C., and Ottewill, R. H., Colloid Polym. Sci. 252, 464 (1974). 6. Tuin, G., Peters, A. C. I. A., van Diemen, A. J. G., and Stein, H. N., J. Colloid Interface Sci. 158, 508 (1993). 7. Verdegan, B. M., and Anderson, M. A., J. Colloid Interface Sci. 158, 372 (1993). 8. Midmore, B. R., Diggins, D., and Hunter, R. J., J. Colloid Interface Sci. 129, 153 (1989). 9. Chow, R. S., and Takamura, K., J. Colloid Interface Sci. 125, 226 (1988). 10. Midmore, B. R., and Hunter, R. J., J. Colloid Interface Sci. 122, 521 (1988). 11. Brouwer, W. M., and Zsom, R. L. J., Colloids Surf. 24, 195 (1987). 12. Van den Hoven, Th. J. J., and Bijsterbosch, B. H., Colloids Surf. 22, 187 (1987). 13. Hidalgo Alvarez, R., De las Nieves, F. J., Van der Linde, A. J., and Bijsterbosch, B. H., Colloids Surf. 21, 259 (1986). 14. Goff, J. R., and Luner, P., J. Colloid Interface Sci. 99, 468 (1984). 15. Van der Put, A. G., and Bijsterbosch, B. H., J. Colloid Interface Sci. 92, 499 (1983). 16. Voegtli, L. P., and Zukoski, C. F., J. Colloid Interface Sci. 141, 92 (1991). 17. Elimelech, M., and O’Melia, C., Colloids Surf. 44, 165 (1990). 18. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 114, 32 (1986). 19. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 114, 45 (1986). 20. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 107, 322 (1985). 21. Bensley, C. N., and Hunter, R. J., J. Colloid Interface Sci. 92, 448 (1983). 22. Van der Linde, A. J., and Bijsterbosch, B. H., Croat. Chem. Acta 63(3), 455 (1990). 23. Ohshima, H., Healy, T. W., and White, L. R., J. Chem. Soc. Faraday Trans. 2 79, 1613 (1983).
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES 24. Van den Hul, H. J., and Vanderhoff, J. W., J. Electroanal. Chem. 37, 161 (1972). 25. Vanderhoff, J. W., Van den Hul, H. J., Tausk, R. J. M., and Overbeek, J. Th. G., in ‘‘Clean Surfaces: Their Preparation and Characterization for Interfacial Studies’’ (G. Goldfinger, Ed.), p. 15. Dekker, New York, 1970. 26. Ahmed, S. M., El-Aasser, M. S., Pauli, G. H., Poehlein, G. W., and Vanderhoff, J. W., J. Colloid Interface Sci. 73(2), 388 (1980). 27. Van den Hul, H. J., and Vanderhoff, J. W., Br. Polym. J. 2, 121 (1970). 28. Stone-Masui, J., and Watillon, A., J. Colloid Interface Sci. 52, 479 (1975). 29. Laaksonen, J., Lebell, J. C., and Stenius, P., J. Electroanal. Chem. 64, 207 (1975). 30. Kolthoff, I. M., and Miller, I. K., J. Am. Chem. Soc. 73, 3055 (1955). 31. Hearn, J., Wilkinson, M. C., and Goodall, A. R., Adv. Colloid Interface Sci. 14, 173 (1981). 32. Furasawa, K., Bull. Chem. Soc. Jpn. 55, 48 (1982). 33. Goodall, A. R., Hearn, J., and Wilkinson, M. C., J. Polym. Sci. Polym. Chem. Ed. 17, 1019 (1979).
AID
JCIS 4099
/
6g0e$$$565
04-17-96 17:15:56
531
34. Vanderhoff, J. W., Pure Appl. Chem. 52, 1263 (1980). 35. James, R. O., in ‘‘Polymer Colloids’’ (R. Buscal, T. Corner, and J. F. Stageman, Eds.), p. 69. Elsevier, Amsterdam, 1985. 36. Ottewill, R. H., and Shaw, J. N., Kolloid-Z. Z. Polym. 218, 34 (1967). 37. Janssen, M. G. J., and Stein, H. N., J. Colloid Interface Sci. 111, 112 (1986). 38. O’Brien, R. W., and White, L. R., J. Chem. Soc. Faraday Trans. 2 74, 1607 (1978). 39. Hunter, R. J., ‘‘Foundations of Colloid Science,’’ Vol. 1, p. 335. Clarendon, Oxford, 1991. 40. Ma, C. M., Micalle, F. J., El-Aasser, M. S., and Vanderhoff, J. W., in ‘‘Emulsion Polymers and Emulsion Polymerization’’ (D. R. Basset and A. E. Hamielec, Eds.), ACS Symposium Series, Vol. 165, p. 251. Am. Chem. Soc., Washington, DC, 1981. 41. Abramson, M. D., ‘‘Electrokinetic Phenomena and their Application to Biology and Medicine.’’ ACS Monograph Series 66, Chem. Catalog Co., New York, 1934. 42. M. L. Wijvekate, ‘‘Verklarende statistiek,’’ p. 155. Aula boeken, Utrecht, 6th printing 1964.
coidas
AP: Colloid
179, 522–531 (1996)
0245
Electrophoretic Properties of Monodisperse Polystyrene Particles G. TUIN, J. H. J. E. SENDERS,
AND
H. N. STEIN 1
Laboratory of Colloid Chemistry and Thermodynamics, Department of Chemical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands Received April 3, 1995; accepted October 25, 1995
In this paper the electrophoretic properties of large monodisperse polystyrene (PS) latices, prepared by a one-step surfactantfree emulsion polymerization, are described. The latices were characterized by conductometric titration. The charged groups, arising from initiator fragments, were all strong acid groups and no carboxyl groups could be detected. The amount of surface charge groups varies with particles size. Electrophoresis measurements have shown that the electrophoretic mobility passes through a maximum as the electrolyte concentration is increased. Converting the electrophoretic mobility to z -potentials, taking into account electrophoresis and retardation effects but neglecting surface conductivity gives the same picture. The same phenomenon is observed with increasing H / concentration. Only at very high electrolyte or H / concentrations does the z -potential reach values close to zero. The z -potentials were converted to charge densities behind the electrokinetic shear plane ( sz), using the Gouy–Chapman theory. In general, PS particles behave as expected without discrepancies between experiment and theory: És0É ( s0 is the surface charge density) is larger than ÉszÉ. For one particular latex, however, a ÉszÉ larger than És0É was found. This was observed at high electrolyte concentrations, where uncertainties in the z -potential because of neglect of surface conductivity are not important. This fact was attributed to chemisorption of NO 0 3 ions. Chemisorption of NO 0 3 ions is not a common fact, but may be attributed to the pronounced hydrophobic character of the PS particles. q 1996 Academic Press, Inc. 0 Key Words: chemisorption, NO 0 3 , on polystyrene; NO 3 chemisorption, on polystyrene; zeta potential, polystyrene; polystyrene, nitrate chemisorption on; polystyrene, surface charge.
INTRODUCTION
Monodisperse polystyrene (PS) particles are widely used as model colloids for, e.g., adsorption studies. Aqueous PS latices are usually prepared by an emulsion polymerization. The emulsion polymerization can be performed either with a surfactant or without an intentionally added surfactant (‘‘soap-free’’ or emulsifier-free polymerization). The surfactant-free emulsion 1
To whom correspondence should be addressed.
522
0021-9797/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AID
JCIS 4099
/
6g0e$$$561
polymerization has the advantage that the only charged groups present on the PS latex surface are sulfate groups, arising from the initiator fragments. The preparation of PS particles by surfactant-free emulsion polymerization is well described (1–5). The particle size of these latices is usually below 1 mm, when prepared by a one-step surfactant-free emulsion polymerization. Recently, we described a method for obtaining particle sizes up to 3.2 mm in a one-step surfactant-free emulsion polymerization (6). However, it has been reported that the model PS particles have an anomalous electrokinetic behavior. Theoretically it is expected that if there is no chemisorption of ions present, an increasing electrolyte concentration at a constant pH should lead to a continuous decrease in the absolute value of the electrophoretic mobility or zeta-potential ( z ), because of compression of the electric double layer. This decrease is indeed observed, but only at rather high electrolyte concentrations ( ú0.01 M) (7). As the electrolyte concentration is increased from low to high concentrations, the electrophoretic mobility (or the zeta-potential calculated from it) first increases in absolute value and it decreases only when a certain electrolyte concentration is surpassed. The zeta-potential versus electrolyte concentration curves therefore exhibits a maximum in absolute value. This phenomenon has been observed by many authors (7–22). The observed maximum is explained by the hairy layer model, the co-ion adsorption model (7, 16–19), or by surface conductivity (22). The hairy layer model assumes that the latex particle is covered by a layer of protruding, rather flexible, polymer chains (‘‘hairs’’) having terminal end groups (the sulfate groups introduced by the initiator fragments) (7–15). As the ionic strength decreases, the repulsion between these charged groups increases and the hairy layer expands. This results in the slipping plane moving outwardly, and the net charge transported electrokinetically decreases. In addition, a hairy layer provides a convenient background for introducing electrical conductance within the shear plane. The co-ion adsorption model assumes, in the case of negatively charged particles, that anions, which are less hydrated than cations, are closer to the apolar surface (7, 16–19).
04-17-96 17:15:56
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
These anions are not believed to be chemisorbed on specific sites and the magnitude of the electrokinetic potential depends on the valency of the co-ions. In the surface conductivity model, the presence of mobile Stern-layer anions causes the mobility to decrease and the conductivity to increase, in comparison with the case where surface conductivity is absent (22). According to this model, not only does the diffuse double layer contribute to the surface conductance, but also a process of additional conductance between slipping plane and particle interface is operative. It is obvious that there seems to be a controversy about the explanation of the electrokinetic behavior of the PS latices. In this paper the electrophoretic properties of large monodisperse PS particles, of which the preparation is described elsewhere (6), are investigated. By varying the particle size of these latex particles, ka values larger than 10 are reached (where k is the Debye–Hu¨ckel parameter and a is the particle radius). For these ka values the theory for calculating z -potentials from measured mobilities (23) is reliable, as long as surface conductivity can be neglected. The PS particles are prepared by a one-step surfactant-free emulsion polymerization. In this way PS particles are obtained with a uniform surface composition. EXPERIMENTAL
Materials Water was twice distilled from an all glass apparatus. The specific conductivity (0.8 mS cm01 ) and the surface tension (72 mN/m) of this water indicated that it was free of surfaceactive impurities. Sodium nitrate, potassium nitrate, sodium hydroxide, and nitric acid (Titrisol) were obtained from Merck (pro analysi, purity ú99%) and used without purification. The ion-exchange resins DOWEX 50W-X4 (H / form) and DOWEX 1-X4 (OH 0 form) were obtained from Fluka and purified as described by Van den Hul and Vanderhoff (24). Latices The preparation of the latices used is described elsewhere (6). The particle sizes of the latices are listed in Table 1. Duplicate measurements of the particle size differed by less than 2%. Cleaning of the Latices All latices were dialyzed by the serum replacement technique as described by Vanderhoff et al. (25) and Ahmed et al. (26) in Amicon serum replacement cells. For a description of the dialysis cells see (26). The latex was first diluted to a solid content less than 5% (w/v) with twice-distilled water. Then a volume of 300 cm3 of the latex was placed in the stirred cell. This cell was
AID
JCIS 4099
/
6g0e$$$562
04-17-96 17:15:56
523
equipped with a Nucleopore membrane (Poretics Corporation, U.S.A.) with a pore size approximately 10–20% smaller than the particle size of the latex. Then a continuous stream of twice-distilled water was flown through the cell until the conductivity of the outlet water reached the same value as that of the inlet twice-distilled water (0.8 mS cm01 ). Normally this took about 24 h. First all the dissolved electrolyte was removed and then the latex was removed from the cell. This procedure was applied to all samples of latex. Determination of Solids Content The solids percentages of the different PS latices were determined by drying a known amount of the latex in an oven at 1057C, until constant weight was reached. Latex Titration The ion-exchange resins used in this study were the DOWEX 50W-X4 (cation exchange, H / form) and the DOWEX 1-X4 (anion exchange, OH 0 form). These ion-exchange resins were separately cleaned according to the same procedure as described by Van den Hul and Vanderhoff (24). After the cleaning procedure the wash water of the ionexchange resins exhibited no surface-active impurities (surface tension 72 mN/m at 207C) and the conductivity of the wash water had the same value as the inlet water used (0.8 mS cm01 ). Equal quantities of the two ion-exchange resins were then mixed together. The wash water from the mixed resin had a considerably lower conductivity (0.25 mS cm01 ) than the water used. The freshly cleaned latices (by the serum replacement technique) were diluted with ion-exchanged water to a solids percentage lower than 5% and were transferred into a Pyrex glass bottle with a large amount of mixed ion-exchange resin. The bottles were placed on a roll-bank and were rolled for 3 to 24 h. At specific times (3, 6, 9, and 24 h) samples were taken to be analyzed by conductometric titration. Latex and ion-exchange resin were, after their contact, separated by decantation followed by filtration through a glass filter. Conductometric titrations of the ion-exchanged latices were performed at 207C under a nitrogen atmosphere. The solid content was always low (1–2% w/v). The conductivity was measured using a conductivity meter (Radiometer Kopenhagen) equipped with a Schott conductivity cell. A 2.1 mM NaOH solution (nitrogen flushed) was added with a constant rate pump (LC-5000 syringe pump, ISCO). The change in conductivity was continuously monitored using a computer. The influence of the rate of adding NaOH solution on the titration was checked for several rates ( 0.1 to 40 ml /
coidas
AP: Colloid
524
TUIN, SENDERS, AND STEIN
TABLE 1 Characteristics of the PS Latices Used Latex
Dn (nm)
Ds (nm)
Dv (nm)
Pn
L-78a L-80a L-86b L-88b L-89b
736 381 1309 2165 3124
754 398 1357 2217 3253
764 407 1386 2252 3279
1.04 1.07 1.06 1.04 1.02
Note. Dn , number averaged diameter; Da , surface averaged diameter; Dv , volume averaged diameter; Pn , degree of polydispersity, Dv /Dn . a Determined with Coulter LS-130. b Determined with Coulter counter.
h ) , but no difference was found for the endpoint of the conductometric titration. Therefore a moderate rate of 15 ml / h was chosen. A typical titration curve ( specific conductivity kc versus quantity of NaOH solution added ) is shown in Fig. 1. The samples of the latices (taken after various ion-exchange times) were titrated three to four times and the average value was used. When the determined endpoints, obtained after different ion-exchange times, were the same (within experimental error), this was assumed to be the endpoint. This endpoint was always reached within 24 h of ion exchange. The endpoint (assuming complete dissociation of charged groups at the end of the conductometric titration and complete charge compensation of the charged groups by H / at the start of the titration) was then converted to the surface charge ( s0 ) and the area per negative charge (An ) assuming a quadratic arrangement of charged sites on the particle surface. This is shown in Fig. 2 for latex L-78 (as an example). z -Potential Measurements
are often called monodisperse if the value of Pn is õ1.05. This is especially the case for the larger particles. In Fig. 1 it can be seen that both the descending and the ascending legs of the plot of conductivity versus quantity of NaOH solution added are linear, with the exception of a region near the intersection point, which means that only one type of surface group is observed. This was found for all latices investigated. The charged surface groups are in origin sulfate groups ( 0SO 40 ) (27), but carboxyl ( 0COO 0 ) groups have also been reported (4, 28, 29) originating from oxidation of intermediate alkanol groups. In the polymerization procedure the so-called Kolthoff reaction can occur (30), resulting in the loss of sulfate groups and the appearance of hydroxyl ( 0OH) groups. These hydroxyl groups can be present at the PS–water interface as noncharged groups. The Kolthoff reaction can be suppressed by using KHCO3 as supporting electrolyte (31); however, this was considered not to be necessary in view of the titration data (see later). The chemical instability of the PS particles, prepared by using K2S2O8 as initiator, has also become apparent from other studies (see, e.g., Hearn et al. (31)). The PS latices may loose their sulfate groups upon storage by hydrolysis of the sulfate groups (31, 32). The rate of hydrolysis appears to vary widely. Goodall et al. (33), e.g., found rates of hydrolysis varying from 2 to 37% per month (at 257C), whereas Vanderhoff (34) reports the complete loss of sulfate groups of an ion-exchanged latex in 14 days. Most reports only mention the removal of the sulfate groups upon storage as a result of hydrolysis. Vanderhoff (34) and Goodall et al. (33) observed the appearance of weak acid groups. From this literature survey, it is obvious that the preparation and handling of the PS particles can have a large influence on the surface charge groups present. The PS particles used in this study were titrated 9 months after preparation. The fact that only one type of acid group could be detected
z -Potential measurements were performed using a Malvern Zetasizer 3 (Malvern Instruments, Malvern, England) equipped with an AZ-4 cell. This apparatus uses the laser– Doppler effect to measure the electrophoretic mobility. The latices used were the ion-exchanged PS latices in the H / form. In case of measuring the z -potential as function of pH the electrolyte used was either potassium nitrate or sodium nitrate, and nitric acid was used to set the pH to the desired value. RESULTS AND DISCUSSION
Latex Titration In Table 1 the particle sizes of the latices are listed. As can be seen from Table 1 the particle sizes of the latices are fairly monodisperse (indicated by Pn Å Dv /Dn ). Particles
AID
JCIS 4099
/
6g0e$$$562
04-17-96 17:15:56
FIG. 1. Change in specific conductivity ( kc ) versus added volume 2.1 mM NaOH solution for the conductometric titration of latex L-78.
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
525
FIG. 2. Change in surface charge density ( s0 ) and area per negative group (An ) versus time for the ion-exchange cycles of latex L-78. ‘‘Time’’ is the total duration of the ion-exchange process.
FIG. 4. Electrophoretic mobility (Em ) versus concentration of HNO3 for different latices. s, Latex L-78; /, latex L-80; h, latex L-86; n, latex L-89.
might suggest that only weak acid groups are present. It is, however, difficult to deduct which type of acid groups at the PS surface is detected by conductometric titration. The additional presence of surface hydroxyl groups is, of course, not excluded. In Fig. 2 the change in surface charge ( s0 ) and the area per surface charge group (An ) with ion-exchange time for latex L-78 is shown as an example. The increase in s0 and thus decrease in An vanishes after 9 h of ion exchange. Further ion exchange does not lead to an increase in s0 . A possible explanation for this could be that the ion-exchange resin is exhausted. The amount of ion-exchange resin was, however, large and using an excess of fresh ion-exchange resin did not result in a change in s0 . When treating tap water with the used ion-exchange resin the conductivity decreased to values as described before. The fact that the conductivity of the wash water from the used ion-exchange resin
was still as low as before the titration supports the fact that the obtained value for s0 is the correct one. In Fig. 3 the determined values of s0 and An are given as functions of the particle sizes of the latices. From Fig. 3 it can be seen that the surface charge density measured increases slightly and the area per negative charge decreases slightly with increasing particle size. The only exception in this respect is latex L-78, which was made by a different polymerization procedure (in a different reactor) (6). The surface charge of latex L-78 is almost as high as that of latex L-89, but the particle size of latex L-89 is more than four times larger than latex L-78. This means that the polymerization procedure can have a large influence on the final latex.
FIG. 3. The final determined surface charge densities ( s0 ) and area per negative group (An ) for the different particle sizes used.
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
Electrophoresis In HNO3 solutions. In Fig. 4 the dependence of the electrophoretic mobility (Em ) on the concentration of HNO3 is shown for the different latices used. In these experiments, the latices were in the H / form. As can be seen from Fig. 4 the electrophoretic mobilities of all latices show a maximum at a concentration of about 10 03 M, similar to the maxima reported by Voegtli and Zukoski for PS particles in HCl and HClO4 solutions (16), though these authors found in such solutions maxima at 10 02 M. The surface charge properties of PS particles is due to dissociation of either sulfate or carboxyl groups. The former have a pKa in the range 1 to 2 (35), while for the latter pKa values between 4 and 6 have been reported (36). Refinement of such values, by the site-binding model, appears not to be useful, since the fact that such values are dependent on pH and the electrolyte concentration raises the suspicion that they are no more than fitting parameters in the model equations (see also the criticism by Janssen and Stein (37)). Increasing absolute values of the electrophoretic mobility
coidas
AP: Colloid
526
TUIN, SENDERS, AND STEIN
FIG. 5. Zeta-potential ( z ) versus pH latex L-80 for different electrolytes. s, [NaNO3 ] Å 3.4 1 10 02 M; h, [KNO3 ] Å 3.3 1 10 02 M.
with increasing HNO3 concentration are found in a pH range in which incomplete dissociation is expected for carboxyl groups, but not for sulfate groups. Thus, the increase in electrophoretic mobility at low [HNO3 ] values indicates that in the polystyrene samples investigated here, sulfate groups are predominant and thus are the groups found by conductometric titration (see Fig. 1). The fact that at [HNO3 ] õ 0.001 M the absolute value of the electrophoretic mobility increases (instead of remaining constant) can best be attributed to increasing chemisorption of NO 30 ions with increasing [HNO3 ] (see below). An alternative explanation could be the hairy layer model: at low [HNO3 ] the hairy layer is extended which leads to a slipping plane far away from the phase boundary of the polystyrene, leading to low absolute values of the z -potential, while at higher [HNO3 ] the hairy layer is compressed, which leads to increasing absolute values of the z -potential. However, it will be shown later that the existence of a hairy layer is not easily compatible with other data. The decrease of the absolute value of Em at HNO3 concentrations higher than 0.001 M is too pronounced to be due to incomplete dissociation of sulfate groups and can best be understood by being due predominantly to the larger electrolyte concentration compressing the electrical double layer (cf. the data in NaNO3 or KNO3 solution at constant pH). Figure 5 shows values of the z -potential calculated from electrophoretic mobilities, measured in solutions in which, in addition to HNO3 , NaNO3 or KNO3 had been added ( É0.033 M). The calculation of the z -potentials was performed by using the equations of Ohshima et al. (23). These equations are an approximation of the numerical results obtained by O’Brien and White (38), which take into account electrophoretic and retardation effects, but neglect contributions to conductance by ions behind the electrokinetic slipping plane. At the electrolyte concentrations employed in
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
the experiments discussed here, however, the influence of the latter is negligible. The z -potentials appear to be independent of pH, at pH values ú4, in agreement with absence of significant quantities of carboxyl groups at the surface of the PS particles. The absolute values of the z -potentials decrease significantly only at pH values between 3 and 2, i.e., when the HNO3 concentration becomes on the same order as the NaNO3 or KNO3 concentration, in agreement with compression of the electrical double layer at large electrolyte concentrations as predominant cause of the decrease in absolute values of Em in Fig. 4. The data presented in Fig. 5 are average values based on four data points each. Application of the Wilcoxon test (42) to these data showed that the difference between the results observed in NaNO3 and KNO3 solutions is significant. This is not surprising in view of the fact that all average values found in KNO3 solutions are lower than those found in NaNO3 solutions at the same pH, while the difference between the respective values at one pH value is larger than the standard deviation of the average values concerned. The difference between z -potentials in NaNO3 and KNO3 solutions, respectively, can be tentatively understood from the different sizes of the hydrated cations concerned: a hydrated K / ion is smaller than a hydrated Na / ion; therefore, the average potential experienced by an adsorbed K / ion will be larger in absolute sense than that experienced by an adsorbed Na / ion. Then a larger part of the surface charge will be compensated by cations behind the electrokinetic slipping plane in the case of K / than in the case of Na / . In NaNO3 or KNO3 solutions at pH 5.5. At a pH of 5.5, any surface sulfate groups will be fully dissociated. In Fig. 6 the dependence of the z -potential of latex L-86 at pH 5.5 in NaNO3 and KNO3 solutions is shown as an example. Similar behavior is found for the other latices studied. A
FIG. 6. Zeta-potential ( z ) versus electrolyte concentration for latex L86 at pH 5.5. s, Electrolyte KNO3 ; h, electrolyte NaNO3 . The lines are just a guide for the eye.
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
maximum in z -potential is found at a concentration of approximately 6 1 10 03 M for the electrolyte NaNO3 and at a slightly higher concentration for the electrolyte KNO3 . A similar behavior was found for positively charged latices in different electrolytes by Hidalgo Alvarez et al. (13). The location of the maximum in the electrophoretic mobility versus electrolyte concentration curve depended on the nature of the counterion in their measurements. As can be seen from Fig. 6, at low electrolyte concentrations the ÉzÉ is larger in NaNO3 than in KNO3 solution. This was also observed in Fig. 5 and is explained by the difference in size of the hydrated cations. However, at high electrolyte concentrations the ÉzÉ are the same within experimental error. Electrophoretic mobilities were again converted into z potentials through the Oshima equations (23), and from the z -potentials the values of the surface charge density in the diffuse double layer beyond the electrokinetic slipping plane ( sz) were calculated, by means of the Gouy–Chapman equation valid at the large ka values (where k is the reciprocal value of the Debye length and a is the particle radius and valid in the cases investigated here (see, e.g., Hunter (39)) q
sz Å 0 8000 c e0er NAkT sinh
S D ze0z 2kT
,
[1]
where c is the bulk concentration of ions (mol/liter), NA is Avogadro’s number, e0 is the permittivity of vacuum, er is the solvent dielectric constant, z is the valency, e0 is the elementary charge, z is the zeta-potential, k is Boltzmann’s constant, and T is the temperature. In addition, we calculated the surface charge density in the region between the phase boundary and the slipping plane, sD , from the relation sz / sD / s0 Å 0,
[2]
where s0 is the surface charge density. Figures 7 and 8 show some typical data, referring to NaNO3 solutions, but similar results are found in KNO3 solution. The data were obtained for two polystyrene samples: latex L-78 with a rather high surface charge density ( 00.0863 C/m 2 ) and latex L-80 with a low surface charge density ( 00.0252 C/m 2 ). In discussing these data, we will start with the data at high electrolyte concentration. In these circumstances, the uncertainty of the z -potential, due to possible contributions of the electric conductivity by ions behind the slipping plane, is minimal. Also, in these circumstances minimal uncertainty to any action of a hairy layer arises. At high values of the pH, where the surface charge properties are due to dissociation of surface 0SO4H groups, És0É
AID
JCIS 4099
/
6g0e$$$563
04-17-96 17:15:56
527
is found to be larger than ÉszÉ, which makes sD ú 0 as expected for surfaces with a negative value of s0 . A similar behavior was found for all latices except latex L-80, for which data are plotted in Fig. 8. Here at large electrolyte concentrations, És0É õ ÉszÉ, which makes sD õ 0. For negative values of the surface charge properties, this indicates chemisorption of NO 30 ions. Similar cases of polystyrene particles with És0É õ ÉszÉ have been reported by Ma et al. (40). These authors even found negative z -potentials on polystyrene particles in the absence of any titratable surface charge. Abramson (41) also reported negative z -potentials for paraffin wax in KNO3 solution. Chemisorption of nitrate ions is not commonly assumed: nitrate ions are usually regarded as ‘‘inert’’ toward solid surfaces. However, much of the evidence on which this is based refers to surfaces of solids such as oxides, hydroxides, or AgI. With surfaces of a pronounced hydrophobic character such as polystyrene, the situation may be different as remarked by Voegtli and Zukoski (16). Such chemisorption of nitrate ions is expected primarily near surface regions which are far from 0SO 40 groups, since the latter are hydrophilic, while their negative charge supplies an additional repulsion toward nitrate ions. The interpretation of negative values of sD as due to chemisorption of NO 30 ions is supported by the fact that such values are found at relatively high NaNO3 concentrations. At lower electrolyte concentrations (especially õ0.001 M), the neglect of the surface conductivity, inherent in the use of the Oshima equation for calculating the z -potentials, becomes increasingly suspect. If this would be important, the absolute values of the z -potentials, and consequently also that of sz , as calculated by use of the Oshima equation, would be too low. In Figs. 7 and 8, the sD values at low electrolyte concentrations would be shifted in the negative direction; this may even lead to negative values of sD (indicating chemisorption of nitrate ions) at the lowest concentrations investigated here (0.0014 M). However, this can only be defended at such low concentrations if an extensive hairy layer is assumed to be present. An argument against the simultaneous effect of a hairy layer and neglect of ionic conductance by ions behind the electrokinetic slipping plane can be derived from Fig. 9. This figure shows sD values as a function of surface charge density for various electrolyte concentrations. These lines have a nearly identical slope, in spite of the fact that the lines refer to electrolyte concentrations varying from 0.036 to 0.0014 M. At the former concentration, a significant contribution of ions behind the electrokinetic slipping plane to electrical conductance, or of a hairy layer, is improbable, while at the latter concentration such contributions if present should be much more pronounced. The nearly equal slopes of all curves in Fig. 9 would be quite fortuitous if there would be a totally
coidas
AP: Colloid
528
TUIN, SENDERS, AND STEIN
FIG. 7. Charge density as function of [NaNO3 ] for latex L-78 at pH 5.5. Dashed line, s0 ; s, sz; n, sD .
different mechanism operative in determining the electrophoretic mobility, at low concentrations and at high concentrations. If the possibility of an important contribution to electric conductivity by ions behind the electrokinetic slipping plane and of a hairy layer can be dismissed, the maximum in the
ÉzÉ-potential versus electrolyte concentration at constant pH (Fig. 6) should be understood by the co-ion adsorption mechanism: at low electrolyte concentrations, NO 30 adsorption becomes increasingly important with increasing [NaNO3 ] or [KNO3 ], leading to an increase of the absolute value of the z -potential; at high concentrations, the compres-
FIG. 8. Charge density as function of [NaNO3 ] for latex L-80 at pH 5.5. Dashed line, s0 ; s, sz; D, sD .
AID
JCIS 4099
/
6g0e$$$564
04-17-96 17:15:56
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES
529
FIG. 9. Change in sD versus surface charge density ( s0 ) at pH 5.5 at different [NaNO3 ]. l, 1.4 1 10 03 M; n, 3.2 1 10 03 M; j, 6.0 1 10 03 M; s, 1.3 1 10 02 M; /, 3.6 1 10 02 M.
sion of the double layer works in the opposite direction. It should be noted that chemisorption of nitrate ions must be assumed at electrolyte concentrations which are too large for either of the alternative mechanisms to be important. From Fig. 9, values of dsD/ds0 at constant electrolyte
FIG. 10.
AID
JCIS 4099
/
concentration can be deduced. These are shown, for NaNO3 solutions at pH 5.5, in Fig. 10. If no nitrate ions would be present behind the electrokinetic slipping plane, and if the Na / ions necessary for compensating the charges of additional surface 0SO 40 groups would be accommodated pre-
dsD/ds0 versus [NaNO3 ] at pH 5.5, calculated from the data in Fig. 9.
6g0e$$$564
04-17-96 17:15:56
coidas
AP: Colloid
530
TUIN, SENDERS, AND STEIN
dominantly behind the slipping plane, dsD/ds0 would be approximately equal to 01. It should be noted that the negative value of dsD/ds0 refers to s0 values in an algebraic, not in an absolute, sense. Such is, at high concentrations, the case in reasonable approximation (dsD/ds0 Å 00.89 at [NaNO3 ] Å 0.036 M). At lower concentrations, dsD/ds0 approaches 01. In principle, the decrease in absolute value of dsD/ds0 with increasing electrolyte concentration could be explained both by an increase of NO 30 incorporation behind the electrokinetic slipping plane at higher concentration or by the absolute values of the z -potentials being too low at low electrolyte concentrations through neglect of contributions to conductance by ions behind the electrokinetic slipping plane. However, the presence of significant quantities of nitrate ions behind the electrokinetic slipping plane at low electrolyte concentrations is probable only if an extensive hairy layer is assumed. Since arguments against the combined action of a hairy layer and underestimation of the z -potentials have been presented, the present authors regard an explanation along these lines as unconvincing, though the possibility cannot wholly be excluded. CONCLUSIONS
In this paper the electrophoretic properties of large monodisperse PS latices are described. The latices were characterized by conductometric titration. The charged groups, arising from initiator fragments, were all strong acid groups and no carboxyl groups could be detected. The amount of surface charge groups varies with particles size. Electrophoresis measurements have shown that the electrophoretic mobility passes through a maximum as the electrolyte concentration is increased. In this respect, the present work confirms conclusions by previous investigators. Converting the electrophoretic mobility to z -potentials, taking into account electrophoresis and retardation effects, but neglecting surface conductivity, gives the same picture. The same phenomenon is observed with increasing H / concentration. Only at very high electrolyte or H / concentrations does the z -potential reach values close to zero. The z -potentials were converted to charge densities behind the electrokinetic shear plane ( sz), using the Gouy– Chapman theory. In general, PS particles behave as expected without discrepancies between experiment and theory: És0É is larger than ÉszÉ. For one particular latex, however, a ÉszÉ larger than És0É was found. This was observed at high electrolyte concentrations, where uncertainties in the z -potential because of neglect of surface conductivity are not important. This fact was attributed to chemisorption of NO 30 ions. Chemisorption of NO 30 ions is not a common fact, but may be attributed to the pronounced hydrophobic character of the PS particles.
AID
JCIS 4099
/
6g0e$$$564
04-17-96 17:15:56
Such chemisorption of NO 30 ions had been supposed by previous investigators to be present on polystyrene (e.g., Ref. (40)). The present work confirms this and adds an important new argument from the dependence of sD on s0 at constant electrolyte concentrations, for excluding the possibility of explaining the results by the simultaneous actions of a hairy layer and of conductance by ions behind the electrokinetic slipping plane. Therefore, in discussing electrophoretic data on polystyrene in nitrate solution, chemisorption of nitrate ions should be taken into account. ACKNOWLEDGMENTS The financial support of the Foundation of Emulsion Polymerization (SEP) is gratefully acknowledged.
REFERENCES 1. Matsumoto, T., and Ochi, A., Kobunshi Kagaku 22, 481 (1965). 2. Kotera, A., Furasawa, K., and Takeda, Y., Kolloid-Z. Z. Polym. 239, 677 (1970). 3. Kotera, A., Furasawa, K., and Kudo, K. Kolloid-Z. Z. Polym. 240, 837 (1970). 4. Goodwin, J. W., Hearn, J., Ho, C. C., and Ottewill, R. H., Br. Polym. J. 5, 347 (1973). 5. Goodwin, J. W., Hearn, J., Ho, C. C., and Ottewill, R. H., Colloid Polym. Sci. 252, 464 (1974). 6. Tuin, G., Peters, A. C. I. A., van Diemen, A. J. G., and Stein, H. N., J. Colloid Interface Sci. 158, 508 (1993). 7. Verdegan, B. M., and Anderson, M. A., J. Colloid Interface Sci. 158, 372 (1993). 8. Midmore, B. R., Diggins, D., and Hunter, R. J., J. Colloid Interface Sci. 129, 153 (1989). 9. Chow, R. S., and Takamura, K., J. Colloid Interface Sci. 125, 226 (1988). 10. Midmore, B. R., and Hunter, R. J., J. Colloid Interface Sci. 122, 521 (1988). 11. Brouwer, W. M., and Zsom, R. L. J., Colloids Surf. 24, 195 (1987). 12. Van den Hoven, Th. J. J., and Bijsterbosch, B. H., Colloids Surf. 22, 187 (1987). 13. Hidalgo Alvarez, R., De las Nieves, F. J., Van der Linde, A. J., and Bijsterbosch, B. H., Colloids Surf. 21, 259 (1986). 14. Goff, J. R., and Luner, P., J. Colloid Interface Sci. 99, 468 (1984). 15. Van der Put, A. G., and Bijsterbosch, B. H., J. Colloid Interface Sci. 92, 499 (1983). 16. Voegtli, L. P., and Zukoski, C. F., J. Colloid Interface Sci. 141, 92 (1991). 17. Elimelech, M., and O’Melia, C., Colloids Surf. 44, 165 (1990). 18. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 114, 32 (1986). 19. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 114, 45 (1986). 20. Zukoski, C. F., and Saville, D. A., J. Colloid Interface Sci. 107, 322 (1985). 21. Bensley, C. N., and Hunter, R. J., J. Colloid Interface Sci. 92, 448 (1983). 22. Van der Linde, A. J., and Bijsterbosch, B. H., Croat. Chem. Acta 63(3), 455 (1990). 23. Ohshima, H., Healy, T. W., and White, L. R., J. Chem. Soc. Faraday Trans. 2 79, 1613 (1983).
coidas
AP: Colloid
ELECTROPHORETIC PROPERTIES OF MONODISPERSE PS LATICES 24. Van den Hul, H. J., and Vanderhoff, J. W., J. Electroanal. Chem. 37, 161 (1972). 25. Vanderhoff, J. W., Van den Hul, H. J., Tausk, R. J. M., and Overbeek, J. Th. G., in ‘‘Clean Surfaces: Their Preparation and Characterization for Interfacial Studies’’ (G. Goldfinger, Ed.), p. 15. Dekker, New York, 1970. 26. Ahmed, S. M., El-Aasser, M. S., Pauli, G. H., Poehlein, G. W., and Vanderhoff, J. W., J. Colloid Interface Sci. 73(2), 388 (1980). 27. Van den Hul, H. J., and Vanderhoff, J. W., Br. Polym. J. 2, 121 (1970). 28. Stone-Masui, J., and Watillon, A., J. Colloid Interface Sci. 52, 479 (1975). 29. Laaksonen, J., Lebell, J. C., and Stenius, P., J. Electroanal. Chem. 64, 207 (1975). 30. Kolthoff, I. M., and Miller, I. K., J. Am. Chem. Soc. 73, 3055 (1955). 31. Hearn, J., Wilkinson, M. C., and Goodall, A. R., Adv. Colloid Interface Sci. 14, 173 (1981). 32. Furasawa, K., Bull. Chem. Soc. Jpn. 55, 48 (1982). 33. Goodall, A. R., Hearn, J., and Wilkinson, M. C., J. Polym. Sci. Polym. Chem. Ed. 17, 1019 (1979).
AID
JCIS 4099
/
6g0e$$$565
04-17-96 17:15:56
531
34. Vanderhoff, J. W., Pure Appl. Chem. 52, 1263 (1980). 35. James, R. O., in ‘‘Polymer Colloids’’ (R. Buscal, T. Corner, and J. F. Stageman, Eds.), p. 69. Elsevier, Amsterdam, 1985. 36. Ottewill, R. H., and Shaw, J. N., Kolloid-Z. Z. Polym. 218, 34 (1967). 37. Janssen, M. G. J., and Stein, H. N., J. Colloid Interface Sci. 111, 112 (1986). 38. O’Brien, R. W., and White, L. R., J. Chem. Soc. Faraday Trans. 2 74, 1607 (1978). 39. Hunter, R. J., ‘‘Foundations of Colloid Science,’’ Vol. 1, p. 335. Clarendon, Oxford, 1991. 40. Ma, C. M., Micalle, F. J., El-Aasser, M. S., and Vanderhoff, J. W., in ‘‘Emulsion Polymers and Emulsion Polymerization’’ (D. R. Basset and A. E. Hamielec, Eds.), ACS Symposium Series, Vol. 165, p. 251. Am. Chem. Soc., Washington, DC, 1981. 41. Abramson, M. D., ‘‘Electrokinetic Phenomena and their Application to Biology and Medicine.’’ ACS Monograph Series 66, Chem. Catalog Co., New York, 1934. 42. M. L. Wijvekate, ‘‘Verklarende statistiek,’’ p. 155. Aula boeken, Utrecht, 6th printing 1964.
coidas
AP: Colloid