Influence of the peroxodisulfate counterion on the dodecyl sulfate adsorption onto polystyrene latex particles

Influence of the peroxodisulfate counterion on the dodecyl sulfate adsorption onto polystyrene latex particles

Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16 Contents lists available at ScienceDirect Colloids and Surfaces A: Physicochemica...

1MB Sizes 1 Downloads 15 Views

Recommend Documents

Adsorption of amphiphilic polysaccharides onto polystyrene latex particles
Adsorption properties of hydrophobically modified carboxymethylpullulans (HMCMP) with different degree of modification o

The aggregation of large polystyrene latex particles
The rate of aggregation of monodisperse polystyrene latices in various electrolyte concentrations has been studied. A ch

Adsorption of Polyampholytes on Polystyrene Latex: Effect on Colloid Stability
The interfacial characteristics of low-charge-density polyelectrolytes and polyampholytes synthesized by a microemulsion

polystyrene latex particles
Emulsion polymerization of styrene was carried out using dendrimer DAB-dendr-(NH2)64 as seed. The size and size distribu

Adsorption of Sodium Dodecyl Sulfate and Butanol onto Acidic and Basic Alumina
The adsorption of sodium dodecylsulfate (SDS) and the co-adsorption of SDS and 1-butanol from aqueous electrolyte soluti

Photoreactions on latex surfaces: Excimer formation of 1-pyrenemethanol on polystyrene latex particles
Excimer formation from 1-pyrenemethanol (PyM) and several other pyrene derivatives is investigated in aqueous dispersion

Dynamics of polyelectrolyte adsorption and colloidal flocculation upon mixing studied using mono-dispersed polystyrene latex particles
•Dynamic aspects of polyelectrolyte adsorption and subsequent flocculation are analyzed in terms of collision rate.•Fold

Acoustic study on the kinetics for the dissociation-recombination reaction between micelle and counterion in sodium dodecyl sulfate solution
The ultrasonic absorption of sodium dodecyl sulfate (SDS) in aqueous solution was measured at 30°C. in the range of freq

Adsorption of sodium dodecyl sulfate and sodium dodecyl benzenesulfonate on poly(vinyl chloride) latexes
The adsorption of sodium dodecyl sulfate (SDS) and sodium dodecyl benzenesulfonate (SDBS) on poly(vinyl chloride) (PVC)

Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

Influence of the peroxodisulfate counterion on the dodecyl sulfate adsorption onto polystyrene latex particles Klaus Tauer ∗ , Hatice Kaplan Can 1 , Dingding He 2 Max Planck Institute of Colloids and Interfaces, Am M¨ uhlenberg, D-14476 Golm, Germany

a r t i c l e

i n f o

Article history: Received 21 December 2007 Received in revised form 9 April 2008 Accepted 10 April 2008 Available online 18 April 2008 Keywords: Dodecyl sulfate adsorption Emulsion polymerization Peroxodisulfate counterion Polystyrene particles

a b s t r a c t Results of an experimental study are reported on the influence of the nature of the peroxodisulfate counterion (NH4 + , K+ , and Na+ ) on the properties of polystyrene latexes. The data show that the adsorption of sodium dodecyl sulfate onto polystyrene particles during the ab initio batch emulsion polymerization depends strongly on both the peroxodisulfate concentration and the nature of the counterion. For sodium and ammonium counterions the adsorbed amount increases and decreases with increasing initiator concentration, respectively. © 2008 Elsevier B.V. All rights reserved.

1. Introduction

CI is the molar initiator concentration:

An ab initio batch emulsion polymerization is self-optimizing regarding the interfacial area between the organic phase and the water that can be stabilized at given recipe and polymerization conditions. The stabilizing ability of a given recipe and procedure, expressed by the value of the interfacial area, is determined by the concentration of hydrophilic groups, which, in general, mainly stem from initiator primary radicals, or hydrophilic comonomers, or adsorbed surfactant molecules. For recipes comprising peroxodisulfate as initiator, sodium dodecyl sulfate (SDS) as emulsifier, and only styrene as monomer the optimized interfacial area is governed by the initiator end group concentration and the amount of adsorbed SDS. The initiator end group concentration is determined by the rate of initiation (ri , equation (1)), which is directly proportional to the initiator concentration. In equation (1) ki is the initiation rate constant in l mol−1 s−1 , R is the molar concentration of primary free (initiator) radicals, Mw is the molar monomer concentration at the locus of initiation (here the aqueous phase), f is the efficiency factor of the reaction of a primary radical with a monomer molecule, kd is the initiator decomposition rate constant in s−1 , and

ri = ki R Mw = 2f kd CI

∗ Corresponding author. Tel.: +49 331 567 9511; fax: +49 331 567 9512. E-mail address: [email protected] (K. Tauer). 1 Permanent address: Faculty of Science, Department of Chemistry, Hacettepe University, Beytepe, Ankara TR-06532, Turkey. 2 Present address: Institute of Microsystems Engineering, University of Freiburg, ¨ Georges-Kohler-Allee 102, D-79110 Freiburg, Germany. 0927-7757/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2008.04.022

(1)

The SDS properties such as critical micelle concentration (CMC) and the surface area per molecule in a saturated adsorption layer depend on the ionic strength and hence, during the emulsion polymerization mainly on the initiator concentration. As both SDS adsorption and initiator end group concentration depend on the initiator concentration also the optimized interfacial area depends on it. In turn, at given peroxodisulfate concentration and temperature the optimized interfacial area depends on the SDS concentration at given solids content. So do also the average particle number and the rate of polymerization. However, the rate of polymerization is not considered here. Three peroxodisulfate salts are commercially available: potassium peroxodisulfate (PPS), sodium peroxodisulfate (SPS), and ammonium peroxodisulfate (APS), which differ regarding their water solubility considerably. At 20 ◦ C the solubilities in water are 2.375–3.550, 2.297, and 0.175 M for APS [1,2], SPS [3], and PPS [3], respectively. A few experimental hints are available that the nature of the counterion also has an influence on the thermal decomposition of the peroxodisulfate anion [4] and on the polymerization kinetics [5]. The aim of this study is to investigate whether there is an influence of the nature of the peroxodisulfate counterion on the properties of polystyrene latexes and the SDS adsorption. The polymerization procedure was an ab initio batch emulsion polymerization at 80 ◦ C with constant water to monomer volume

8

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

ratio. The peroxodisulfate concentration, the nature of the counterion (XPS, X: A, P, S), and the SDS concentration were varied. The latexes were characterized regarding solids content, average particles size, and surface tension liquid–vapor (i.e. latex to air). Moreover, the surfactant properties were investigated by surface tension measurements at room temperature in dependence on the ionic strength that is, in the presence of the XPS type and concentration used also for the polymerization. The surface tension concentration plots were evaluated regarding the CMC, the occupied surface area in saturation layers, and the relation between the interfacial tension and the concentration of the dodecyl sulfate (DS) ions in water below the CMC. It turned out that these investigations were not possible with PPS as precipitates were formed at any SDS concentration at room temperature. This observation is in accordance with former results showing that the solubility of PPS is at temperatures below 38 ◦ C quite low [6] and that precipitation occurred if the molar ratio K+ to C12 OSO3 − exceeded 0.5 [7]. Another strong motivation to carry out this study is the following. In non-polymerizing systems, that is at given latex particle surface, SDS adsorption can be nicely described by a Langmuir-type adsorption isotherm, which has been proven with post-added SDS [8,9]. An interesting question is regarding the situation at the end of an ab initio batch emulsion polymerization where the interfacial area is the result of self-optimization under the particular polymerization conditions. Is under these conditions the relation between the adsorbed amount of surfactant (Sads ) and the surfactant concentration in water (SW ) also of Langmuir-type? As one can expect different Sads - and SW -values by changing both the XPS and the SDS concentration for the polymerizations, the experimental determination of SW by measuring the surface tension of the final latexes should allow the determination of a “polymerization-influenced” adsorption isotherm. 2. Experimental part 2.1. Materials Styrene (STY, 99% purity) from Sigma–Aldrich was distilled under reduced pressure before use. The peroxodisulfates (purity ≥ 98%) all from Sigma–Aldrich and SDS ultra pure from Roth were used as received. The water for all experiments was taken from a Seral purification system (PURELAB PlusTM ) with a conductivity of 0.06 ␮S cm−1 and degassed prior to use for the polymerization. 2.2. Polymerizations The polymerizations were carried out in glass vials placed in ¨ a rotation thermostat VLM20 (VLM GmbH, Leopoldshohe, Germany) at 80 ◦ C for 15 h. The agitation of the reaction mixture is due to an end-over-end rotation of the glass vials. About 20 revolutions per minute provide sufficient agitation. The following recipe was used: 10 ml of water, 1 ml of styrene monomer, and variable amounts of XPS and SDS as given in the polymerization matrix in Table 1. All ingredients were charged in the glass vials with volumeprecalibrated micropipettes. The added volumes correspond to 10.983 g of water and 0.902 g of monomer as it was checked gravimetrically. The use of the rotation thermostat allows polymerizing simultaneously 24 samples. Each polymerization was repeated at least four times. After the polymerization and before any characterizations the latexes were passed through glass frits in order to remove coagulum. Note, the highest PPS concentrations are above the solubility at room temperature.

Table 1 XPS and SDS concentration range employed in the study; the crosses indicate XPS–SDS concentration pairs investigated, the letter code given in parenthesis for each initiator and surfactant concentration allows an unambiguous identification of the polymerization condition for each sample CXPS (mM)

3.7 (I1) 7.4 (I2) 11.1 (I3) 14.8 (I4) 18.5 (I5) 22.2 (I6)

CSDS (mM) 3.33 (S1)

3.04 (S2)

2.50 (S3)

2.13 (S4)

1.85 (S5)

1.53 (S6)

×

×

×

× ×

×

×

×

×

× × × × × ×

×

×

× × ×

× ×

×

×

2.3. Analytics The surface tension measurements were carried out with a tensiometer TD1 (Lauda) according to the Du Nouy ring method at room temperature. The latexes were characterized regarding the solids content (SC) with a HR73 Halogen Moisture Analyzer (Mettler Toledo, Gießen, Germany) and average particle size (Di , intensity weighted average particle size) with dynamic light scattering (NICOMP particle sizer model 370, particle sizing systems, Santa Barbara, CA, USA). Additionally some latex samples (APS-I1/S6, PPS-I6/S1, SPS-I4/S3) were investigated regarding the particle size distribution by means of transmission electron microscopy (TEM) with a Zeiss EM 912 Omega microscope operating at 100 kV after suspension preparation and an Analytical Ultracentrifuge (AUC) (Beckman Coulter Optima XLI, Palo Alto, CA) at 2000, 10,000 and 20,000 rpm increasing with decreasing particle size at 25 ◦ C. 2.4. Data evaluation and calculations Data evaluation is based entirely on the experimental values of solids content (FG in %), average particle size by dynamic light scattering (intensity weighted average diameter Di in nm), and latex surface tension  in mN/m. Moreover, experimentally determined relations between the surface tension and the free SDS concentration in water were used. Such relations were measured for each type of peroxodisulfate at three concentrations and additionally with NaCl at the same ionic strengths. Solids content of latex (theoretical), FGth , where I, S, M, and W are the mass of initiator, surfactant, monomer, and water, respectively, in gram per polymerization run: FGth =

I+S+M × 100 I+S+W+M

Auxiliary content, HG: HG =

I+S × 100 I+S+M+W

Polymer content (theoretical), PGth : PGth = FGth − HG Polymer content (experimental), PG: PG = FG − HG Particle number (per ml of water), N, where D is the intensity average particle size determined by dynamic light scattering and p is the polymer density (1.05 g cm−3 ): N=

PG (100 − FG)(/6)Di3 p × 10−21

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

9

Total particle surface area (cm2 per ml of water), AP : Ap = N Di2 × 10−14 Free surfactant concentration in water of the final latex, SW in M, where z1 , z2 are constants, which were determined from surface tension—SDS concentration plots at the corresponding ionic strength and have the unit of measurement mN/m: SW = 10(−+z2 )/z1

(2)

The formula for calculating SW is based on the experimentally determined and linearised relation between the surface tension () of a surfactant solution and the surfactant concentration below the critical micelle concentration:  = z1 log SW + z2 both z1 and z2 are negative (cf. Fig. 3). Adsorbed amount of surfactant at the particle water interface, Sads in M from surfactant mass balance:

Fig. 1. Surface tension–ln CSDS plots in the presence of various types and concentrations of peroxodisulfate.

Sads = Stot − SW  Adsorbed amount of surfactant per surface area Sads in mol/cm2 :  Sads

=

Sads

(3)

Ap × 103

Adsorption isotherm according to [8] for SDS adsorption onto differ in mol/cm2 : ent surfaces including PS-latex particles; Sads  Sads =

2.5 × 106 SW

(4)

1 + (8 × 103 ) SW

2.5. Experimental errors and reproducibility For the primary polymerization data solids content, average particle size, and surface tension the overall experimental error was determined by calculating the average value and the standard deviation from the individual repeats of each polymerization (at least 4 repeats). Considering all experiments of this study the relative errors of the solids content, the average particle size, and the interfacial tension are less than 0.06, 0.09, and 0.01, respectively. From these data the relative error of the overall particle surface and the surfactant adsorption, which scales with the particle surface, can be estimated to be less 0.15. The latter value was used to estimate the standard deviations as plotted in the corresponding graphs. 3. Results and discussions 3.1. SDS properties in dependence on added ionic strength (add-IS) The properties of SDS solutions, as for any other surfactant solution, depend on the add-IS [10]. For these investigations we

used XPS at the same concentrations as employed in the polymerizations. Fig. 1 shows the surface tension concentration plots (–ln CSDS plots) for APS and SPS at three different XPS concentrations. Note, the add-IS is by a factor of 3 higher. As the SDS was used as received these curves show shallow minima, which are caused by the impurities (mainly dodecanol). The surfactant was not cleaned as hydrolysis of the alkylsulfate to the corresponding alcohol practically cannot be avoided during application in aqueous heterophase polymerization and also storage. Despite the impurity the results clearly show the expected influence of the add-IS as the CMC is shifted to lower values with increasing ionic strength. Moreover, already these data reveal a pretty strong influence of the nature of the counterion. All data extracted from these measurements are summarized in Table 2. For the calculation of the area occupied per DS ion in a saturated adsorption layer (as ) at add-IS = 0 (for pure SDS solutions) the factor 1/2 was considered in the Gibbs adsorption isotherm. If the factor 1/2 is not considered the as -value for pure SDS is with only 23 A˚ 2 much too low. At high add-IS this factor has not to be considered but in a transition range, which is, based on these data not possible to specify, it increases to 1. Compared to the value for pure SDS the data in Table 2 and Fig. 2 show a distinct decrease in as in the presence of peroxodisulfate, which makes sense as the Debye-length decreases and the surfactant molecules can pack more closely. Moreover, this is in concordance with data for tritiated SDS determined by radiotracer measurements and published in [11,12]. Not only the sign of the slope of the dependence on add-IS but also the range of the numerical as -values agrees with the published data [12]. A closer look reveals that the as -values slightly increase with increasing add-IS but remain even at the highest ionic strength clearly below the value for pure SDS. Also this

Table 2 Summary of the SDS surface tension data (the values in the brackets are interpolated values used for the calculations) CXPS (mM)

CSDS (mM)

+

3.7 7.4 11.1 14.8 18.5 22.2 0

3.33 3.04 2.50 2.13 1.85 1.53 –

aS (A´˚ 2 )

CMC (mM) +

Na

NH4

4.78 – 2.86 – – 2.27

4.26 – 2.57 – – 1.96 6.94

z1 (mN/m)

+

+

Na

NH4

32.5

36.9

39.5

35.2

40.5

38.1 45.8 *

+

z2 (mN/m) +

Na

NH4

−27.55 (−25.00) −22.35 (−25.50) (−25.20) −24.35 −40.67

−19.16 (−24.25) −29.63 (−30.75) (−31.50) −32.46

Na+

NH4 +

−29.31 (−26.70) −24.02 (−23.90) (−23.70) −23.40 −37.01

−24.10 (−24.90) −25.71 (−25.50) (−25.20) −24.92

10

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

Fig. 2. Dependence of the area per DS chain in a saturated adsorption layer (as ) on the added ionic (add-IS) strength realized by APS, SPS, and NaCl; the solid line is just for guiding the eye in the range of add-IS > 0.

result seems reasonable as an increased counter or co-ion condensation at higher ionic strength reduces the effective charge of the DS chains. The linear parts of the surface tension–log concentration plots below the CMC, which are used to determine the concentration of the DS ions in the continuous phase of the final latexes are summarized in Fig. 3. Again, a clear influence of both the XPS concentration and the nature of the counterion are observed. The z1 - and z2 -values determined by linear regression from the data of Fig. 3 are plotted in Fig. 4a in dependence on the peroxodisulfate concentration. It is interestingly to note that both the z1 - and z2 -values show a contradictory dependence on the peroxodisulfate concentration for ammonium and sodium counterions. Especially, the z1 -values for both initiators behave in a quite different way. Comparison measurements with the 1:1 electrolyte sodium chloride reveal that the z1 - and z2 -values depend not only on the ionic strength but also strongly on the nature of the electrolyte (cf. Fig. 4b). The influence of the nature of the electrolyte is especially obvious for APS as for this electrolyte both the z1 and the z2 dependence has a different curvature compared with the other salts. Moreover, the data in Fig. 4b suggest that the z1 -value is mainly governed by the cation (NH4 + vs. Na+ ) whereas the z2 -value is determined by the anion (S2 O8 2− vs. Cl− ).

Fig. 3. Surface tension–log CSDS plots below the CMC for various APS and SPS concentrations; the points are the experimental data and the lines are the linear regressions.

The CMC data obtained in the presence of NaCl are in a similar range like the values for XPS. At add-IS of 11.1, 33.3, and 66.6 mM NaCl (which correspond to I1, I2, and I3) the CMC values were determined as 4.5, 3.5, and 1.8 mM. Moreover, these values agree well with the data published in [12]. 3.2. Polymerization data In this part the primary polymerization data are evaluated that is, the dependence of the solids content, the average particle size, and the interfacial tension on the initiator type and concentration as well as on the SDS concentration. It is necessary to point that the experimental conditions regarding CI and Stot cover quite a broad field and thus, the particle size distributions as well as the average particle sizes might be expected from very broad to almost monodisperse and from below 100 nm to above 500 nm, respectively. In order to check this assumption selected samples were investigated regarding the particle size distribution (Fig. 5). Expectedly, the particle size distribution depends strongly on the particular initiator and SDS concentration. Very broad distributions as obtained with low initiator and low SDS concentration can cause problems in the further data evaluation based on average values. In general, the data presented in Fig. 5 show a quite good agreement regarding the width of the particle size distribu-

Fig. 4. (a) Dependence of z1 , z2 on the peroxodisulfate concentration for APS and SPS; smaller symbol are interpolated values, the numerical values of z1 , z2 were determined from the data of Fig. 3 are valid for the SDS concentration in the aqueous phase in M. (b) Dependence of z1 , z2 on the added ionic strength realized by APS, SPS, and NaCl (add-IS); the numerical values of z1 , z2 were determined from the data of Fig. 3 are valid for the SDS concentration in the aqueous phase in M.

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

11

Fig. 5. Illustration of the range of particle size distributions of the latexes obtained with lowest (A), highest (B), and intermediate (C) initiator and surfactant concentration; TEM images (upper row) and comparison of the particle size distributions (lower row) as determined by enumerating TEM pictures (filled symbols) and by AUC (open symbols); (A) APS-I1S6 (minimum initiator and SDS concentration), (B) PPS-I6S1 (highest initiator and SDS concentration), and (C) NPS-I4S3 (intermediate initiator and SDS concentration).

tion, especially if one considers that in the case of TEM only less than 500 particles have been enumerated. The intensity weighted average diameters put together in Table 3, which scale with D7 /D6 , also show good agreement between the various sizing techniques. Only for the sample with the broadest particle size distribution the average value obtained by DLS deviates strongly. This seems to be a specific problem of the DLS equipment due to the extremely broad particle size distribution of that particular sample. For the further data evaluation, which relies on the average values of DLS, the problems with the very broad particle size distributions do not apply because the samples obtained with low initiator concentrations are not considers as discussed below. Fig. 6 shows the dependence of the solids content on the initiator concentration for the SDS concentration S3 and reveals that the lowest initiator concentration does not lead at 80 ◦ C within a polymerization time of 15 h to complete conversion (also the typical styrene smell was observed at the end of the polymerization).

Thus, the data obtained with I1 were treated with caution in the further evaluation. Moreover, the data summarized in Fig. 6 shows, if any, only negligible differences between the three types of peroxodisulfate. The average particle size generally decreases with both increasing XPS concentration at given SDS concentration and increasing

Table 3 Comparison of the average particle size (Di in nm) obtained with different sizing techniques Sample

TEM

AUC

DLS

APS-I1S6 NPS-I4S3 PPS-I6S1

898 248 74

862 213 70

(559) 243 89

Fig. 6. Dependence of the solids content of the final latexes on the peroxodisulfate concentration; emulsifier concentration S3 (cf. Table 1); the solid line corresponds to complete monomer conversion.

12

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

Fig. 7. Change of the average particles size with increasing total SDS concentration (Stot ) for APS at different concentrations; line is just for guiding the eye through the data points for I3.

SDS concentration at given peroxodisulfate concentration as shown in Fig. 7 for APS. However, the data for I3 show a somewhat different trend than for the higher and lower XPS concentrations. At SDS concentrations below 2.2 mM (Stot < 2.2 mM) the average particle size for this initiator concentration (I3) increases with increasing amount of SDS. The data in Fig. 8 prove that this trend is obviously real as a similar behavior is observed for all three peroxodisulfates. For this particular initiator concentration (I3) the dependence of D on Stot changes clearly at a SDS concentration between 2.1 and 2.4 mM for all three peroxodisulfate salts. An explanation of this experimental fact cannot be given at the moment. However, it is clear that this has apparently nothing to do with the CMC as the range where the change occurs is below the CMC at this add-IS (cf. Table 2). This should also hold for the conditions during the polymerization as the CMC of anionic surfactants increases with increasing temperature [13] and so also for SDS [14] but the adsorbed amount of surfactant decreases [10]. In general, the surface tension of the final latexes increases with increasing SDS concentration (at least at larger XPS concentrations) and increases also with increasing XPS concentrations (at least at larger SDS concentrations) as it is shown in Fig. 9. Note, the values of the surface tension at the lower XPS concentrations might be influenced by the lower conversion although the data in Fig. 9 give no proof that a free styrene film is on the liquid vapor interface as in that case one would expect interfacial tension values in order of 30 mN/m.

Fig. 8. Dependence of the average particle size on the total SDS concentration for APS, SPS, and PPS at peroxodisulfate concentration I3.

Fig. 9. Surface tension liquid vapor () of the final latexes in dependence on the total SDS concentration (Stot ) for APS.

Increasing surface tension values with increasing total amount of SDS are obtained for all peroxodisulfate salts as it is shown in Fig. 10 for I6. In the range of lower SDS concentrations the data in Fig. 10 suggest an influence of the nature of the peroxodisulfate counterion in a way that ammonium leads to the lowest values. Note, the surface tension liquid–vapor of the final latexes are in any case clearly higher than the corresponding values at the CMC’s (cf. Fig. 1). Moreover, the highest -values at the highest total surfactant concentration are about 70 mN/m and indicate that in these cases almost all SDS molecules are adsorbed at the particle water interface. 3.3. Interfacial area optimization during emulsion polymerization The overall particle–water interface develops during the ab initio batch emulsion polymerization according to thermodynamic and kinetic requirements. Thermodynamics always tries to minimize the excess interfacial energy and competes with kinetics, which freezes the system in at higher conversion due to viscosity increase (additional high Tg of polystyrene). In addition, the experimental data show that both the XPS and the SDS concentrations influence the optimum interfacial area. The following is noteworthy to mention in this context, after more than 2 years in the shelf at room temperature the latexes are still stable, which is proven by DLS measurements. The particular values are 88.6 ± 6 nm vs.

Fig. 10. Dependence of the surface tension liquid–vapor of the final latexes on the total SDS concentration for different peroxodisulfate salts at I6.

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

13

decreases with increasing SDS concentration, which is unusual but at least might be partly caused by the lower conversion. However, the data as observed for I3 shows that the lower conversion is not the sole reason as the dependence changes its sign at a SDS concentration between 2.1 and 2.4 mM. In this context it should be mentioned that also for the average particle size—Stot dependence in the same SDS concentration range a significant jump for all types of peroxodisulfates is observed (cf. Fig. 8). 3.4. Evaluation of the surfactant adsorption

Fig. 11. Change of the overall particle surface in the final latexes with the XPS concentration for SDS concentration S3.

91.4 nm, 243 ± 12 nm vs. 234 nm, and 570 ± 70 nm vs. 558 nm for the samples PPS-I6S1, NPS-I4S3, and APS-I1S6, respectively. For all peroxodisulfate salts a very similar dependence of the overall particle area (AP ) on the initiator concentration (CI ) is observed at given SDS concentration (cf. Fig. 11). Despite four outliers the overall particle water interfacial area clearly increases with increasing initiator concentration (or addIS) at given surfactant concentration. This is a rather surprising result as one might expect also the reverse dependence as the increasing ionic strength should reduce the stabilizing ability of SDS. However, the data of the surface area in saturated SDS layers obtained in dependence on add-IS (which was adjusted by addition of the particular amount of XPS) as depicted in Fig. 2 support the observed dependence between AP and CI . The data in Fig. 2 show that the as -values for add-IS > 0 increase with increasing add-IS, which means that also the stabilizing ability of the SDS molecules increases with increasing ionic strength. In this sense both data sets in Figs. 2 and 11 confirm each other. One would also expect a clear increase in AP with increasing SDS concentration as at given other conditions a larger interfacial area can be stabilized with increasing number of surfactant molecules. However, the experimental data show surprisingly that this relation strongly depends on the range of both the SDS and the XPS concentration as exemplary shown for SPS in Fig. 12. The data in Fig. 12 clearly disclose that only for the highest peroxodisulfate concentration AP increases over the whole range with increasing Stot . At the lowest initiator concentration AP even

Fig. 12. Change of the particle water interfacial area with increasing SDS concentration for different SPS concentrations; the lines are just for guiding the eye.

The evaluation of the experimental data by means of equations (2)–(4) allows a decision whether the surfactant adsorption is influenced by the peroxodisulfate concentration and/or the nature of the counterion. One might expect differences in the surfactant adsorption if the hydrophilicity of the particle water interface is different due to a changing degree of covalently bound sulfate endgroups with varying peroxodisulfate concentration. The counterion influence might be caused by a difference in the degree of dissociation and hence by different properties of the electrical double layer. Unfortunately the following evaluation is only possible for SPS and APS due to the above mentioned solubility problems of the PPS–SDS combination. Moreover, the evaluation is restricted to CI > I1 in order to avoid problems caused by not complete conversion during the polymerizations and the broad particle size distributions.  With SW from equation (2) and Sads from equation (3) it is possible to get a “polymerization-influenced” or “polymerizationcontrolled” adsorption isotherm. This isotherm can be compared with the Langmuir-isotherm from equation (4), which was obtained for SDS adsorption studies onto various substrates including polystyrenes and polymethacrylate latex particles with postaddition of the surfactant [8]. There are obviously several possibilities to evaluate the experimental data obtained in this study regarding the SDS adsorption behavior. (1) Neglecting the influence of the add-IS by the initiator, which means the use of the z1 - and z2 -values obtained for the pure SDS solution from the –CSDS plots as depicted in Fig. 3 and summarized in Table 2. With these z1 - and z2 -values SW is obtained  is calculated with equation (4). from equation (2) and then Sads This Langmuir-type adsorption isotherm is plotted in graph A of Fig. 13. The data in graph A show no conspicuousness and  even the order of magnitude of Sads agrees with the data given in [8]. (2) Neglecting the influence of add-IS as in case 1 but calculating  Sads according to equation (3) via the total surfactant mass balance. The corresponding data are plotted in graph B of Fig. 13, which shows a complete different dependence (in fact the opposite behavior) than in graph A. This discrepancy already gives a hint that the particular polymerization recipe (the XPS nature and concentration as well as the SDS concentration) influences surfactant adsorption. Astonishingly, most of the  data points show negative adsorption that is, Sads -values less than 0. These are unrealistic results very likely due to the use of wrong z1 , z2 values for the calculation of SW . (3) Considering the influence of add-IS and using the z1 , z2 values obtained in the presence of the corresponding type and  via the Langmuir-type adsorpamount of XPS to calculate Sads tion isotherm given by equation (4) gives the dependence as depicted in graph C of Fig. 13. Compared with graph A both the range and the shape have changed. (4) Using again the SDS mass balance as in case 2 but now in combination with the specific z1 , z2 values for each XPS (regarding the nature of the counterion and the concentration) results in

14

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

Fig. 13. Evaluation of SDS adsorption by means of different models applied to the data obtained after polymerization (experimental data except for I1); SW determined by    via equation (4); (B) z1 , z2 values for pure SDS, Sads via equation (3); (C) z1 , z2 values for the particular add-IS, Sads means of equation (2); (A) z1 , z2 values for pure SDS, Sads  via equation (4); (D) z1 , z2 values for the particular add-IS, Sads via equation (3).  –S dependence as depicted in graph D of Fig. 13. Only in a Sads W this case a really surprising result is obtained showing that SPS and APS behave differently. As the experimental data points disclose there is not only a gradual or quantitative difference between both types of peroxodisulfate but a really fundamental or qualitative difference. For SPS (sodium peroxodisulfate)  –S the Sads W relation has a positive slope whereas the slope is negative for ammonium peroxodisulfate (APS). In contrast to APS, SPS shows the expected behavior that is, the adsorbed amount of SDS increases with increasing SDS concentration in the continuous phase. Moreover a few data points obtained  from the polymerizations with APS show negative Sads values. This is the case for all initiator–surfactant combinations CI > I3 and CSDS < S4, that is for the polymerizations concentrated in the lower right corner of the polymerization matrix (Table 1) which were carried out with higher APS and lower SDS concentrations.

The results summarized in Fig. 13 allow the following conclusions: (a) SDS adsorption during emulsion polymerization (“polymerization-controlled” adsorption) does not follow a Langmuir-type adsorption isotherm pattern, (b) the use of specific SDS adsorption parameters (here z1 - and z2 -values), which were determined in the presence of the real electrolyte (here SPS and APS) is obviously a necessary prereq –S uisite to get reliable Sads W relations, and (c) neither the parameters for pure SDS nor for an indifferent or inert electrolyte such as sodium chloride can be used to mimic

the influence of the type of peroxodisulfate on SDS adsorption during emulsion polymerization. The comparison of the data in graphs C and D of Fig. 13 shows   values or in other words that the Sads -values are lower than the Sads the Langmuir adsorption isotherm (equation (4)) calculates greater adsorbed amounts than the mass balance (equation (3)). Obviously the parameters in the Langmuir adsorption isotherm overestimate the SDS adsorption. The observed difference might be due to a high hydrophilic initiator end group concentration on the particle surface, which on the one hand contributes to latex stability and on the other hand occupies potential SDS adsorption sides. Consequently some excess surfactant is present at the end of the polymerization, which means there is more surfactant than necessary for the interfacial area stabilization. Note, hydrophilic end groups are generated due to initiator decomposition during the whole duration of the polymerization whereas the increase in the interfacial area declines steadily with increasing conversion. Moreover, the amount of adsorbed surfactant decreases with increasing conversion due to droplet vanishing and particle shrinking.  –S  A negative slope of the Sads W relation and negative Sads values mean that the measured interfacial tension, which is used for the calculation is too low and SW is calculated too high. Obviously other molecules than the dodecyl sulfate ions contribute to . These might be residual monomer, oligomers, or counterions. Residual monomer can be excluded as the data for the lowest initiator concentration have not been evaluated. Surface active oligomers might be an issue but not very likely as the observed effect is counterion specific. Indeed, the observed fundamental difference in the

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

15

Fig. 14. Summary of primary experimental polymerization and SDS adsorption data for SPS and APS; change of solids content (A), average particle size (B), interfacial tension of the final latexes (C) in dependence of total SDS concentration for I6, and change of z1 , z2 -values describing the change of the surface tension of SDS solutions below the CMC in dependence on the peroxodisulfate concentration.

action of ammonium and sodium counterions of peroxodisulfate on the SDS adsorption behavior as shown in graph D of Fig. 13 clearly points to a specific influence of ammonium ions. The argumentation, based on polymerization kinetics, does not account for the experimentally observed influence of the nature of the counterion. So the question arises what is special with ammonium compared with sodium cations? The primary experimental data put together in Fig. 14 clearly prove that the effect as depicted in graph D of Fig. 13 is caused by the influence of the counteri-

ons of peroxodisulfate on the SDS adsorption behavior. The only fundamental difference in the primary experimental data between sodium and ammonium peroxodisulfate is regarding the dependence of the z1 - and z2 -values on the peroxodisulfate concentration (graph D in Fig. 14). This causes also a diverse dependence of SW on the peroxodisulfate concentration for both kinds of counterions as displayed in Fig. 15 for the final latexes.  vs. The data points for SPS in Fig. 15 as well as in Fig. 16 (Sads CI ) show the expected behavior, that is a decreasing SDS concen-

Fig. 15. Change of the SDS concentration in the aqueous phase in the final latexes (SW ) in dependence on the peroxodisulfate concentration for SPS and APS at S3.

 Fig. 16. Change of the adsorbed amount of SDS in the final latexes (Sads ) in dependence on the peroxodisulfate concentration for SPS and APS at S3.

16

K. Tauer et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 325 (2008) 7–16

tration in the aqueous phase and an increasing amount of adsorbed emulsifier with increasing ionic strength [10]. In addition, the experimental data points in both figures scatter for the ammonium counterions much more than for the sodium ions. 4. Conclusions In overall conclusion, the data of this study clearly show that the SDS adsorption behavior is governed by the nature of the peroxodisulfate counterion. But the reason for this rather unexpected effect is still obscure. Trying to find a satisfying explanation needs further investigations which should take into account the following considerations. Firstly, in the case of APS the electrical double layer around the particles contains chemically different co- and counterions (sodium stemming from the emulsifier and ammonium from the initiator). For the combination SPS/SDS co- and counterions are identical. Each of the cations can have a specific influence on the water structure close to the particle surface. But both are not that far away positioned in the Hofmeister series of cations regarding salting-out of proteins [15]. The sulfates of ammonium and sodium are effective salting-out agents for instance for deoxygenated sickle hemoglobin whereas sodium sulfate is more effective than ammonium sulfate [16]. Acknowledgements The authors gratefully acknowledge financial support from the Max Planck Society and technical assistance from Mrs. Ursula Lubahn, Mrs. Sylvia Pirok, Mrs. Rona Pitschke (TEM), and Mrs. Antje ¨ Volkel (AUC). H.K.C. is thankful for fellowships provided by the German Research Foundation (DFG)/The Scientific and Technological ¨ Research Council of Turkey (TUBITAK) and the German Academic Exchange Service (DAAD).

References [1] J.F. Gall, G.L. Church, R.L. Brown, The solubility of (NH4 )2 S2 O8 in water and in solutions of H2 SO4 and (NH4 )2 SO4 , J. Phys. Chem. 47 (1943) 645–649. [2] J. Balej, A. Regner, Correction of the solubility of ammonium persulfate in water, Coll. Czech. Chem. Commun. 31 (1966) 361–363. [3] A. Apelblat, E. Korin, E. Manzurola, Solubilities and vapour pressures of saturated aqueous solutions of sodium peroxydisulfate and potassium peroxydisulfate, J. Chem. Thermodyn. 33 (2001) 61–69. [4] M.G. Levi, E. Migliorini, Sopra la scomposizione del persolfati, Gaz. Chim. Ital. 36B (1906) 599–619. ¨ [5] K. Tauer, H. Muller, On the role of initiator in emulsion polymerization, Colloid Polym. Sci. 281 (2003) 52–65. [6] R.L. Neman, Physical properties of potassium dodecylsulfate, Tex. J. Sci. 20 (1968) 199–205. [7] B. Wu, S.D. Christian, J.F. Scamehorn, Recovery of surfactant from micellarenhanced ultrafiltration using a precipitation process, Prog. Colloid Polym. Sci. 109 (1998) 60–73. [8] J.G. Brodnyan, E.L. Kelley, Adsorption onto synthetic latexes, J. Polym. Sci., Polym. Symp. No. 27 (1969) 263–273. [9] S.-Y. Lin, C. Dong, T.-J. Hsu, C.-T. Hsu, Determination of adsorption of an ionic surfactant on latex from surface tension measurements, Colloid Surf. A: Physicochem. Eng. Aspects 196 (2002) 189–198. [10] S. Paria, K.C. Khilar, A review on experimental studies of surfactant adsorption at the hydrophilic solid–water interface, Adv. Colloid Interface Sci. 110 (2004) 75–95. [11] K. Tajima, Radiotracer studies on adsorption of surface-active substance at aqueous surface. II. Effect of excess salt on the adsorption of sodium dodecyl sulfate, Bull. Chem. Soc. J. 43 (1970) 3063–3066. [12] K. Tajima, Radiotracer studies on adsorption of surface-active substance at aqueous surface. III. Effects of salt on the adsorption of sodium dodecyl sulfate, Bull. Chem. Soc. J. 44 (1971) 1767–1771. [13] Y. Moroi, MIcelles: Theoretical and Applied Effects, Plenum Press, New York, 1992. [14] P. Mukerjee, K.J. Mysels, Critical Micelle Concentrations of Aqueous Surfactant Systems (NSRDS-NBS 36), 1971. [15] R.A. Curtis, J. Ulrich, A. Montaser, J.M. Prausnitz, H.W. Blanch, Protein–protein interactions in concentrated electrolyte solutions: Hofmeister-series effects, Biotechnol. Bioeng. 79 (2002) 367–380. [16] W.N. Poillon, J.F. Bertles, Deoxygenated sickle hemoglobin. Effects of lyotropic salts on its solubility, J. Biolog. Chem. 254 (1979) 3462–3467.