Investigation of Ouzo Effect Colloid Formation Via Organosilica Nanoparticles

Investigation of Ouzo Effect Colloid Formation via Organosilica Nanoparticles

James H. Whittemore IV*, Andrew T. Jones, Sharathkumar K. Mendon, James W. Rawlins

Corresponding Author

* Address: School of Polymers and Higher Performance Materials, The University of Southern Mississippi, 118 College Drive #5217, Hattiesburg, MS 39406. Email: , Tel: +1-662-574-9500, Fax: +1-601-266-5504.

Colloid and Polymer Science

1.Organosilica synthesis

Sil1_10

4 mL of NH4OH was added to 50 mL of stirring 190 proof ethanol in a 100 mL round bottom flask. While stirring magnetically, 4 mL of TEOS was quickly injected into the above solution. The solution was stirred for two hours, 1 mL of diethoxydimethyl silane was added and the solution was allowed to stir overnight. The reaction mixture was washed via centrifugation thrice with ethanol and twice with DI water.

Sil2_10

6 mL of NH4OH was added to 50 mL of stirring 190 proof ethanol in a 100 mL round bottom flask. While magnetically stirring, 4 mL of TEOS was quickly injected into the above solution. The solution was stirred for two hours, 1 mL of diethoxydimethyl silane was added and the solution was allowed to stir overnight. The reaction mixture was washed via centrifugation thrice with ethanol and twice with DI water.

Sil3_10

Three solutions were prepared: (I) 3.75 mL of TEOS and 22.5 mL of ethanol, (II) 0.625 mL of TEOS, 1 g of n-octyltriethoxy silane and 7.5 mL of ethanol (III) 9.5 mL NH4OH in 50 mL of ethanol. Solution III was stirred while solution I was fed into it at a rate of 0.167 mL/min. After complete addition of solution I, solution II was fed into the mixture at a rate of 0.167 mL/min. The solution was allowed to stir overnight and then washed via centrifugation thrice with ethanol and twice with DI water.

2.organoSilica particle size control

Organosilica particles varying in size and surface chemistry were synthesized via modified Stöber methods. The surface chemistry of organosilica particles was controlled by addition of an organosilane at the end of the reaction, as reported in literature.[21,23] Particle size distribution (PSD) of the silica particles was controlled by altering the ratio of reactants or the feed rate of reactants (Figure 1S)

Figure 1S. PSD of organosilica particles.

Sil1_10 and sil2_10 used similar formulation, but a higher concentration of NH4OH was added in the synthesis of sil2_10. The increased NH4OH concentration resulted in a higher rate of tetraethyl orthosilicate (TEOS) hydrolysis and an enhanced monomer condensation rate (Scheme 1S). Increasing the hydrolysis and condensation rates resulted in increased polymerization and growth, and decreased nucleation, thus, by increasing NH4OH concentration in sil2_10 synthesis an increase in particle size was observed.[24] Particle size was controlled by means of drop wise monomer addition and sil3_10 contained the largest organosilica particles. Controlled monomer feeding starved the sol gel, emphasizing growth of already formed nuclei.

a. Si(OR)4 + H2O  (OR)3Si(OH) + ROH

b. (RO)3Si(OH) +H2O  SiO2 + 3ROH

Scheme 1S. Shows the (a) hydrolysis of TEOS to a singly-hydrolyzed TEOS monomer intermediate, and then the (b) condensation of the monomer intermediate to form silica.

3.Dominance of Number particle size distribution in DLS

There are benefits and drawbacks to using one angle DLS to determine the PSD of heterogeneous colloidal solutions. In one angle DLS, many subtleties in the PSD are “rounded-out”, resulting in certain trends in PSD being highlighted at the expense of others. The most numerous set of particles in solution will scatter most of the light back to a detector, effectively drowning out the detection of the least numerous particles.

As an example, sil1_10 (~0.16 m in diameter) and sil3_10 (~0.60 m in diameter) were dispersed into an aqueous surfactant solution, at a ratio of 1:1, sil1_10:sil3_10, by weight, and their PSD was obtained via DLS (Figure 2S). A very strong peak was noted for sil1_10 at approximately 0.16 m, while sil3_10 was barely detectable. Even when sil1_10 and sil3_10 particles were analyzed at a 1:1 weight ratio, the smaller particles dominated detection. If it is assumed that these two particles are perfectly monodisperse and have the same density, then sil1_10 particles are 10 times as numerous as sil3_10 particles in solution. When the weight loading of sil3_10 was increased by a magnitude, making the weight ratio of sil1_10 to sil3_10 1:10 and the number ratio 1:1, a bimodal PSD was observed, with the sil3_10 peak as a shoulder on the sil1_10 peak. When the weight loading of sil3_10 was raised by another factor of 10 to a sil1_10 to sil3_10 weight ratio of 1:100 and a number ratio 1:10, sil1_10 was not detected as it was completely masked by the increased number of sil3_10 particles.

Figure 2S. PSD for mixtures of sil1_10 and sil3_10 at ratios of 1:1, 1:10, and 1:100.

Detection of a bimodal colloid system using DLS is difficult when the two colloids have similar particle sizes, because peak blurring is the result. The PSD of a sil1_10 (~0.16 m diameter) and sil2_10 (~0.30 m diameter) colloid solution, at a 1:1 weight ratio showed only one particle with a size in-between sil1_10 and sil2_10 (Figure 3S). The same behavior was observed for a sample made of sil2_10 and sil3_10 mixed at a 1:1 weight ratio (Figure 3S).

Figure 3S. PSD of mixtures of organosilica solutions include sil1_10:Sil2_10 1:1 and sil2_10:Sil3_10 1:1.

In this investigation, the weakness of the instrumentation was used as a strength. As long as the particle number is a magnitude higher than any other colloid, a discreet particle type in a bimodal or trimodal colloidal solution can be exclusively monitored without signal interference from other colloids. If a colloid with a known particle size is introduced into another colloidal solution of known particle size, and a unimodal peak in between the two PSDs is observed, then it is known that signals from the two colloids blurred. Additionally, particles become more difficult to detect when particle size decreases below 200 nm because the scattering coefficient decreases logarithmically with particle size (Figure 4S).[25]

Figure 4S. Scattering coefficients at 780 nm of polystyrene particles dispersed in water versus the polystyrene particles size.[25]

4.Ouzo Effect Particle Size Control

The particle size of colloids formed through nanoprecipitation has generally been controlled via two variables, degree of nanoprecipitation and solute concentration.[1, 4, 7] As the degree of supersaturation increases, particle size decreases. An increase in solute concentration results in larger particles.[1, 4, 7] The same response was observed in Figure 6S. Three colloids of varying particle size were formed by adjusting the degree of supersaturation and solute concentration. A colloid with a ~ 300 nm particle size was formed by using a low degree of supersaturation (50% poor solvent), and a low solute concentration (0.05 wt% PMMA). A colloid with a ~80 nm particle size was formed by increasing the degree of supersaturation to 90% poor solvent at the same solute concentration of 0.05 wt% PMMA. A colloid with a particle size intermediate between the two previous colloids, at ~160 nm, was produced by using a higher degree of supersaturation (90% poor solvent), and a high solute concentration of 0.5 wt% PMMA.

Figure 5S. The effect of supersaturation and solute concentration on the PSD of PMMA colloids formed via nanoprecipitation.

5.TGA Determination of PMMA Shell Formation

During processing a surfactant, IGEPAL CO-887, was used to maintain the colloidal stability. Because residual surfactant might interfere with TGA analysis of PMMA deposition on ANEs, the thermal decomposition under air flow for both IGEPAL CO-887 and PMMA were run independently via TGA (Figure 6S). IGEPAL CO-887 degraded quickly between 200° C and 250 °C, while PMMA degraded quickly between 250 °C and 300 °C. As the two moieties of interest degrade in two different temperature zones differentiating the two materials is more straightforward.

Figure 6S. Thermal decomposition of PMMA and IGEPAL CO-887 in air.

Figure 7S shows the thermal decomposition of select sil1_10 solvent shifting samples. The following formulations where analyzed; sil1_10, sil1_10:PMMA 2:1 90% poor solvent 0.5 wt% PMMA, sil1_10:PMMA 2:1 90% poor solvent 0.05 wt% PMMA, sil1_10:PMMA 5:1 50% poor solvent 0.05 wt% PMMA and sil1_10:PMMA 2:1 50% poor solvent 0.05 wt% PMMA. All of the samples followed the same thermal degradation trends as the native sil1_10, however, the nanoprecipitated samples showed an earlier onset of degradation, which was expected because of the presence of IGEPAL CO-887.

Figure 7S. Thermal stability of sil1_10 and sil1_10 nanoprecipitated with PMMA at various ratio with either 50% or 90% poor solvent, and at a PMMA concentration of 0.05 wt% or 0.5 wt%.

Weight loss at three different temperature ranges, 200-250 °C, 250-300 °C and 300-600 °C, for each of the samples mentioned above was charted in a bar graph (Figure 8S). All sil1_10 samples nanoprecipitated in the presence of PMMA showed similar weight losses in the expected thermal degradation ranges for IGEPAL CO-887 and PMMA. Although data indicated that IGEPAL CO-887 and PMMA were on the sil1_10 surface, PMMA weight loss did not surpass 5% for any sample. The expected weight loss for samples formed via the nucleation and growth Ouzo effect mechanism was 30%. This strongly indicated that PMMA was not gathered by ANEs in a nucleation and growth mechanism.

Figure 8S. Bar graph showing the percent weight loss of sil1_10 samples through three pertinent heating ranges.

Figure 9S shows the thermal decomposition of select samples from sil2_10 nanoprecipitated in the presence of PMMA. The following formulations where analyzed; sil2_10, sil2_10:PMMA 2:1 90% poor solvent 0.5 wt% PMMA, sil2_10:PMMA 2:1 90% poor solvent 0.05 wt% PMMA, sil2_10:PMMA 5:1 50% poor solvent 0.05 wt% PMMA and sil2_10:PMMA 2:1 50% poor solvent 0.05 wt% PMMA. Samples followed the same thermal degradation trends as native sil2_10, except for the sil2_10:PMMA 5:1 50% poor solvent 0.05 wt% PMMA sample. As compared to the other samples, the onset of thermal degradation began early and thermal degradation continued to be greater through all temperature ranges, indicating that there was an excess of IGEPAL CO-887 that must have remained through experimental error during washing.

Figure 9S. Thermal stability of sil2_10 and sil2_10 nanoprecipitated with PMMA at various ratio with either 50% or 90% poor solvent, and at a PMMA concentration of 0.05 wt% or 0.5 wt%.

Thermal degradation weight loss at three different temperature ranges, 200-250 °C, 250-300 °C and 300-600 °C, for each of the samples mentioned above, was charted in a bar graph (Figure 10S). Weight loss associated with PMMA thermal decomposition barely exceeded 5%, and only for sample, sil2_10:PMMA 50% poor solvent 0.05 wt%, while the rest of the samples had PMMA weight loss near 3%. If the particles were ANEs in a nucleation and growth mechanism then the PMMA weight loss should be close to 30%, indicating that PMMA was not gathered by ANEs in a nucleation and growth mechanism.

Figure 10S. Bar graph showing the percent weight loss of sil2_10 samples through three different heating ranges.

Figure 11S. Percent weight loss of sil3_10 samples through three different heating ranges.