- Email: [email protected]
Surfactant Molar Ratio and Ammonia Concentration
Journal of Colloid and Interface Science 211, 210 –220 (1999) Article ID jcis.1998.5985, available online at http://www.idealibrary.com on
Synthesis of Nanosize Silica in a Nonionic Water-in-Oil Microemulsion: Effects of the Water/Surfactant Molar Ratio and Ammonia Concentration F. J. Arriagada1 and K. Osseo-Asare2 Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802 Received December 15, 1997; accepted November 18, 1998
The effect of ammonia concentration on the region of existence of single-phase water-in-oil microemulsions has been investigated for the system polyoxyethylene (5) nonylphenyl ether (NP-5)/ cyclohexane/ammonium hydroxide. The presence of ammonia decreases the size of the microemulsion region. A minimum concentration of surfactant (estimated at about 1.1 wt%) is required for solubilization of the aqueous phase; this value is not significantly affected by ammonia concentration. As indicated by fluorescence spectral data, the transition between bound and free water occurs when the water-to-surfactant molar ratio is about 1 and the presence of ammonium hydroxide does not appear to have a significant effect on this. Ultrafine (30 –70 nm diameter), monodisperse silica particles produced by hydrolysis of tetraethoxysilane (TEOS) in the microemulsion show a complex dependence of the particle size on the water-to-surfactant molar ratio (R) and on the concentration of ammonium hydroxide. At relatively low ammonia concentration in the aqueous pseudophase (1.6 wt% NH3) the particle size decreases monotonically with increase in R. However, for higher ammonia concentrations (6.3–29.6 wt% NH3) a minimum in particle size occurs as R is increased. These trends are rationalized in terms of (a) the effects of the concentration, structure, and dynamics of the NP-5 reverse micelles on the hydrolysis and condensation reactions of TEOS, and (b) the effects of ammonia concentration on the stability of the microemulsion phase, the hydrolysis/condensation reactions of TEOS, and the depolymerization of siloxane bonds. © 1999 Academic Press Key Words: silica; w/o microemulsion; nanoparticles; tetraethoxysilane; TEOS; polyoxyethylene nonylphenyl ether; phase equilibria; hydrolysis/condensation.
INTRODUCTION
The growing scientific and technological interest in nanosize particles is fueling a quest for new synthesis methods that are capable of molecular-level control of particle characteristics. In this connection, the microheterogeneous nature of microemulsion media offers attractive possibilities, as already demonstrated for a variety of materials, including metals, chalco1 Present address: Idesol Ingenieros S.A., Av. San Martin 255 Oficina 53, Iquique, Chile. 2 To whom correspondence should be addressed.
0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
genides, metal halides, carbonates, oxides, and organic polymers (1– 6). In our laboratory we are investigating the relationship between microemulsion properties and the characteristics of the ensuing nanoparticles (7–13). In an earlier paper (7) we reported on the preparation of nanosize monodisperse silica particles by controlled hydrolysis of tetraethoxysilane (TEOS) in a nonionic surfactant/cyclohexane/ammonium hydroxide water-in-oil microemulsion system. Particles in the range of 50 to 70 nm were produced with standard deviations below 8.5% around the mean diameter. As the water-to-surfactant molar ratio (R) increased from 0.7 to 2.3, the particle size decreased and the size distribution became narrower. Further work has now shown that the dependence of particle size on the waterto-surfactant molar ratio becomes much more complicated as a wider range of R values is utilized. In the previous work (7) the ammonium hydroxide concentration in the aqueous pseudophase was maintained at the same constant value (29.6 wt%) for all the synthesis experiments. We have now discovered that variations in ammonium hydroxide concentration can lead to dramatic changes in particle characteristics (12). This paper presents and discusses these new findings. The work reported here includes the determination of the one-phase microemulsion phase boundaries (with emphasis on the effects of temperature and ammonium hydroxide concentration), and the investigation of the effects of the water-to-surfactant molar ratio and ammonium hydroxide concentration on the size and size distributions of the resulting particles. The nonionic surfactant selected for this study was a polyoxyethylene nonylphenyl ether with an average of 5 oxyethylene groups per molecule (NP-5). It has an HLB number of 10 (14), and at room temperature is readily soluble in saturated hydrocarbons like cyclohexane. Nonylphenyl and octylphenyl polyoxyethylene ether compounds have been used in studies on reverse micellar catalysis (15, 16), solubilization and phase behavior of reverse microemulsions (17–21), aggregation phenomena in nonaqueous systems (19, 22–25), and microemulsion-based particle synthesis (26 –30). Thus, the literature contains much relevant data for these surfactants on the effects of temperature, electrolyte content and composition on solubili-
210
SILICA IN WATER-IN-OIL MICROEMULSION
zation, aggregation numbers, and the state of solubilized water. In the case of silica synthesis via the alkoxide sol-gel process, an extensive literature exists for homogeneous solution systems (31– 49). EXPERIMENTAL PROCEDURES
Materials The polyoxyethylene (5) nonylphenyl ether surfactant (NP-5) used was the commercial product Igepal CO-520 from GAF Chemicals Corporation. The as-received material was a viscous liquid and contained ca. 0.2 wt% of water, as determined by Karl Fisher titration. Water and low-boiling impurities (such as free alcohols) were removed by treatment at 60 –70°C under vacuum for 26 h (21). The molecular weight of NP-5 is 440.6 g mol21. HPLC grade cyclohexane (99.9%, Aldrich) was used after storage under molecular sieves (8 3 12 mesh, 0.4 nm, Union Carbide). Absolute ethanol (anhydrous, Quantum Chemical) and reagent grade ammonium hydroxide (29.6 wt% NH3, Baker) were used as received. Water was purified by reverse osmosis (Millipore-RO4) and deionization (Millipore Milli-Q). Tetraethoxysilane (TEOS, 991 wt%, Alfa) and fluorescent probe tris(2,29-bipyridyl)ruthenium(II) chloride hexahydrate (Ru(Bpy), Aldrich) were used as received. Microemulsion Characterization Phase equilibria and solubilization. The solubility domains of water and ammonium hydroxide solutions in the NP-5 surfactant system were determined by two methods. In one set of experiments, the surfactant/oil molar ratio was fixed, and the phase behavior of a number of samples with different contents of aqueous phase was observed as a function of temperature. A solution of NP-5 in cyclohexane and the required amounts of water (or ammonium hydroxide solutions) were weighed and sealed in glass ampoules (5 mL, Thomas). After thorough hand-mixing, the ampoules were placed in a thermostated water chamber. While subjected to slow temperature changes (at a rate of 1°C/h), the number of phases present at a given temperature in each sample was recorded. In a separate set of experiments, the phase behavior was determined at 22°C for samples of variable surfactant and aqueous phase content. In this case, the solubilization limits were determined by inspection of the mixture upon small, sequential additions of aqueous phase to a given surfactant/oil solution. Once permanent turbidity was observed (indicating the onset of phase separation), a few samples were backtitrated with surfactant/oil solution to double-check the transition points. Fluorescence spectroscopy. The nature of solubilized water (i.e., bound to the oxyethylene groups or free) within the surfactant aggregates was studied through the fluorescence emission spectra of the hydrophilic probe Ru(Bpy). Fluores-
211
cence spectra were measured at 23°C on a Shimadzu RF-5000 spectrofluorimeter. Excitation wavelength was 460 nm with a 1.5-nm bandwidth for both excitation and emission. Stock solutions of the probe were freshly prepared and stored in the dark before use. Probe concentration (with respect to the total microemulsion volume) was 9 3 1026 M. In the previous study (7), the water-to-surfactant molar ratio (R) was varied by the addition of different amounts of water (or ammonium hydroxide solutions) to a surfactant solution of NP-5 at a fixed concentration (0.3 M) in cyclohexane. In order to ascertain whether in this surfactant system the transition from bound to free water molecules is only a function of the water-to-surfactant molar ratio (R), additional experiments were conducted in which the parameter R was varied by changing the surfactant concentration (in the range of 0.05 to 0.29 M), while keeping constant the amount of water added to the samples. Ammonium hydroxide solutions with ammonia concentrations in the range of 1.6 to 29.6 wt% NH3 (0.9 to 15 M) were investigated. Particle Preparation A set of standard conditions was used in most of the synthesis experiments. Typically TEOS concentrations in the range of 0.02 to 0.04 M, water-to-TEOS molar ratios (h) of above 6, and concentrated ammonium hydroxide (29.6 wt% NH3 in the aqueous solution) were used. Hereafter, the concentrations quoted refer to the total microemulsion volume unless stated otherwise. The water content was in excess of that required for complete conversion of TEOS to silica. Given the low TEOS concentrations used, the resulting solid content in the samples was limited to below 0.2 wt%. In order to maintain all relevant parameters ([TEOS], [H2O], [NH4OH]) constant while exploring the effect of the water-tosurfactant molar ratio (R), samples with different surfactant concentrations were used. In each sample, the value of R was calculated by considering only the surfactant associated with the reverse micelles according to the expression, R 5 @H2O#/~@NP-5# 2 @NP-5# o! 5 @H2O#/@NP-5# m,
[1]
where [NP-5] is the total surfactant concentration and [NP-5]o is the surfactant concentration in the bulk oil phase, i.e., not adsorbed at the oil/water interface. The quantity [NP-5]m is, thus, the concentration of surfactant associated with the reverse micelles. As shown below, the value of [NP-5]o is ca. 0.02 M of NP-5 in cyclohexane at 22°C. A microemulsion plus second reactant synthesis route (3, 7) was used; i.e., TEOS was added to a previously prepared NP-5/cyclohexane/ammonium hydroxide microemulsion to initiate the synthesis. The experiments were conducted in 8-mL glass bottles, sealed by Teflon-lined screw caps (Supelco). All synthesis experiments were conducted at 22°C. A fixed total microemulsion volume (ca. 5 mL) with NP-5 concentrations in
212
ARRIAGADA AND OSSEO-ASARE
dispersed in the continuous cyclohexane phase. It can be seen that replacement of water by ammonium hydroxide solutions has a strong effect on the phase stability in this system; the solubilization curve is shifted to lower temperatures. This effect is more pronounced with concentrated ammonium hydroxide. For both concentrated (29.6 wt% NH3) and dilute (14.6 wt% NH3) ammonium hydroxide solutions, only the solubilization curve (i.e., upper temperature limit) was observed for a 5.83 wt% NP-5 solution. These samples remained clear upon cooling to about 5°C, where cyclohexane freezes. The effect of ammonium hydroxide on the phase equilibria may be interpreted as the result of the competition between hydroxyl ions and the oxyethylene groups of the surfactant for interaction with water molecules; i.e., ammonium hydroxide is acting as a lyotropic salt reducing the mutual solubility between water and surfactant (50). Similar results were obtained by Kon-no and Kitahara (21) who studied the effects of sodium hydroxide and salt additions in NP-8/tetrachloroethylene microemulsions. FIG. 1. Solubilization diagram for H2O and NH4OH solutions in the NP-5/cyclohexane system; 5.83 wt% NP-5.
the range of 0.05 to 0.3 M was normally used, to which constant amounts of aqueous phase and then of TEOS (typically 30 and 25 mL, respectively) were added with a microsyringe (Eppendorf, 100 mL). In most cases, all components were weighed to determine precisely the R values and the final sample composition. Characterization of Particles A Phillips 420 transmission electron microscope was used to characterize particle size and size distribution. Samples for TEM observation were prepared by depositing a drop of the dispersion on Formvar-covered carbon-coated copper grids (300#, 3 mm, Structure Probe) and drying on filter paper at room temperature. Prior to extraction of samples, the bottles were sonicated for 30 s to ensure the extraction of representative dispersion volumes. Number-average particle diameters and size distributions were determined on enlarged TEM micrographs with a ZIDAS image analysis system (Zeiss). Several hundred particles were measured for each sample. RESULTS AND DISCUSSION
Phase Equilibria and Solubilization Effect of temperature on phase equilibria. The solubilization diagram of water and ammonium hydroxide solutions as a function of temperature in cyclohexane containing 5.83 wt% NP-5 (ca. 0.1 M) is shown in Fig. 1. The region between the solubilization and solubility curves is a homogeneous transparent phase where water-swollen NP-5 reverse micelles are
Effects of NP-5 and ammonia concentrations on phase equilibria at 22°C. It was of interest to evaluate microemulsion stability as a function of surfactant concentration at a fixed temperature of 22°C, which is the temperature of the particle synthesis experiments reported in this work. A section of the ternary phase diagram for the NP-5/cyclohexane/aqueous solution system is shown in Fig. 2, illustrating the microemulsion regions for water and concentrated ammonium hydroxide solutions. As can be seen in Fig. 2, water solubilization increases significantly for NP-5 concentrations above about 6 wt%. At these relatively high surfactant concentrations, the one-phase microemulsion region is bound by the solubilization curve, beyond which phase separation involves the equilibrium between water-in-oil microemulsion and excess water. The range of microemulsion compositions used in the synthesis experiments is also illustrated in Fig. 2. As described above, most experiments were conducted with a constant concentration of aqueous phase, i.e., in a range of microemulsion compositions parallel to the NP-5/cyclohexane axis. Determination of unaggregated surfactant concentration. Following Johnson and Shah (51), the data presented in Fig. 2 are plotted in Fig. 3 as the maximum water and ammonium hydroxide uptake vs the initial NP-5 concentration. Extrapolation of the solubilized amount to zero content of aqueous phase indicates the “critical microemulsion concentration” (cmc), i.e., the concentration of surfactant which is not adsorbed at the surfactant/water interface (51). Based on Fig. 3, a cmc value of about 1.1 wt% is obtained by extrapolation of the water solubilization data, which furthermore coincides very closely with the onset of solubilization of ammonium hydroxide. This value agrees qualitatively with cmc values estimated from solubilization data available in the literature for similar polyoxyethylene compounds in cyclohexane (0.6 wt% of NP-7.5 solubilizing 0.1 M NaOH solutions (52) and 1 wt% of NP-6 solubilizing
SILICA IN WATER-IN-OIL MICROEMULSION
213
FIG. 2. One-phase w/o microemulsion regions in the ternary phase diagram for the NP-5/cyclohexane/water and NP-5/cyclohexane/NH4OH (29.6 wt% NH3) systems, T 5 228C.
1.4 M FeCl2 solutions (53)). It is worth mentioning that the cmc values in nonaqueous solvents are, in general, considerably higher than the corresponding critical micelle concentration (cmc) values typically found for nonionic surfactants in aqueous solutions (;1023 wt%, (54)). Fluorescence spectroscopy and the nature of solubilized water. Figure 4 presents the fluorescence spectra of Ru(Bpy) in NP-5/cyclohexane/water microemulsions. It must be noted that, in these experiments, R was varied (in the range of 0.50 to 4.80) by changing the surfactant concentration (in the range of 0.29 to 0.05 M). This is in contrast to the previous work (7) in which the surfactant concentration was maintained constant. It can be seen that at the low water contents (R 5 0.5 and 0.73), the emission spectra exhibit a maximum intensity at about 570 nm with a red shoulder at about 625 nm. With further increases in the water content, the intensity at 625 nm increases at the expense of the 570 nm peak (see R 5 0.93). It is interesting to note that in the case of the spectra for R 5 1.24 and R 5 4.80, only the 625-nm peak remains. The red shift in fluorescence emission is attributable to the fact that, as the water content increases, there is a corresponding increase in the polarity of the environment surrounding the probe molecule. Since the maximum emission intensity at the higher R values occurs at the same wavelength as that recorded for
Ru(Bpy) in bulk water (620 – 630 nm) (55), it may be concluded that R 5 1 marks the transition from bound to “free” water molecules. The maximum emission intensity shown in Fig. 4 follows a trend which is similar to that observed in the case of the previous experiments (7) that were based on a fixed surfactant concentration and with ammonium hydroxide as the solubilizate. Thus, within the surfactant concentration range investigated, the transition from bound to free water molecules is only a function of the water-to-surfactant molar ratio, R. Effects of Synthesis Variables on Particle Characteristics TEOS hydrolysis and condensation. As shown by the early work of Stober et al. (31), under basic conditions TEOS undergoes hydrolysis and polycondensation reactions which result in the formation of monodisperse spherical particles of amorphous silica. The hydrolysis reaction producing silanol groups may be represented as Si(OR)4 1 xH2O 5 Si(OR)42x~OH! x 1 xROH,
[2]
where R 5 C2H5. The base-catalyzed hydrolysis occurs by a nucleophilic substitution mechanism, where the hydroxyl anion attacks the silicon atom (32–35), as shown in Eq. [3].
214
ARRIAGADA AND OSSEO-ASARE
FIG. 3. Maximum H2O and H2O-NH4OH (29.6 wt% NH3) uptake as a function of the initial concentration of NP-5 in cyclohexane; T 5 228C.
RO RO HO– + Si©OR RO
RO
OR
d– HO
d– Si
OR
OR OR HO©Si + OR– .
OR
[3]
OR
The rate of TEOS hydrolysis is affected by the nature of the solvent, and this has been attributed to both steric effects and hydrogen bonding (32, 36 –38). Byers and Harris (36) reported that the hydrolysis rate is faster in higher alcohols (such as butanol) than in short-chain alcohols such as 1-propanol. According to these authors, the strong hydrogen bonding existing between the water molecules and the lower alcohols reduces the effective availability of water molecules and thus the hydrolysis rate, probably by reducing the mobility of OH2 ions (see Eq. [3]). Interaction between hydrolyzed species results in condensation (formation of silicon-oxygen-silicon bonds), which involves the attack of a nucleophilic deprotonated silanol on a protonated silanol (Eq. [4]) or on an ethoxysilane group (Eq. [5]), producing respectively water or ethanol (32, 35): [Si-OH 1 HO-Si[ 5 [Si-O-Si[ 1 H2O
[4]
[Si-OH 1 RO-Si[ 5 [Si-O-Si[ 1 ROH
[5]
Effect of the ammonia concentration. Particles sampled after 930 h and which were produced at selected R values with different ammonia concentrations are shown in the TEM micrographs of Fig. 5. For both R 5 4.5 and R 5 1.7, the particle morphology becomes more irregular as the ammonia concentration decreases; however, the effect appears to be more significant at high R values (Figs. 5A to 5D), as seen in the sample with the lowest ammonia concentration (Fig. 5D).
High ammonia concentrations (i.e., relatively strong alkaline conditions) are thus needed to obtain morphological control in the NP-5 reverse micellar system. In contrast, ammonia concentrations of 0.9 M (the lowest concentrations used in this work) promote the formation of quite spherical particles in homogeneous media (31, 41, 43). Effect of the water-to-surfactant molar ratio (R). A series of samples of different water-to-surfactant molar ratios (R in the range of 0.4 to 6.7) and various constant ammonium hydroxide concentrations in the solubilized aqueous phase was prepared. The number-average particle diameter (d n) for 930-h samples (930 h aging time) is shown in Fig. 6. Referring to the data for concentrated ammonium hydroxide (29.6 wt% NH3) as the solubilizate, it can be seen that as R increases, the particle size decreases sharply, reaches a minimum at an R value of about 1.8 and then increases again. A similar but less marked trend is observed for more dilute ammonia solutions, although the minimum in particle size is shifted to higher R values (to about 2.9) for both intermediate ammonia concentrations (i.e., 6.3 and 14.6% NH3) as the ammonia concentration increases. In the most dilute ammonia solution utilized in this work (1.6% NH3), no minimum is observed within the R range investigated, and the particle size decreases continuously. Effect of TEOS concentration. The particle size and size distribution were determined for silica particles obtained in experiments with variable TEOS concentrations. The spread of the particle size distributions was measured by the normalized
FIG. 4. Fluorescence spectra of Ru(Bpy) solubilized in the NP-5/cyclohexane/water reverse micellar system, [Ru(Bpy)] 5 9 3 1026 M, T 5 238C.
SILICA IN WATER-IN-OIL MICROEMULSION
215
FIG. 5. Effect of ammonia concentration (wt% NH3) on the morphology of SiO2 particles; [TEOS] 5 0.026 M, T 5 228C; (A–D) R ca 4.5, (A) 29.6 wt%, (B) 14.6 wt%, (C) 6.3 wt%, (D) 1.6 wt%; (E–H) R ca 1.7, (E) 29.6 wt%, (F) 14.6 wt%, (G) 6.3 wt%, (H) 1.6 wt%.
standard deviation, which corresponds to the standard deviation over the mean (number-average) particle diameter in percentage. Figure 7 shows the number-average particle diameters observed after 21 h of reaction and the final diameters from samples aged for 930 h. Similar to the reported effects of TEOS concentration in alcoholic media (41), a high TEOS concentration in the reverse micellar system promotes the formation of smaller particles (see 930 h data). However, as can be appreciated from Fig. 7, particles produced at high TEOS concentration grow faster, even though the water-to-
TEOS ratio is small. After 21 h of reaction, particles produced with a water/TEOS ratio (h) of 3.2 have already grown to 80% of their final size, while for the largest water/TEOS ratio used (ca. 20) the equivalent figure is only about 70%. As seen further in Fig. 7, the size distribution is narrow (normalized standard deviation of the order of 5%) and relatively insensitive to TEOS concentration. In contrast, in the conventional synthesis of silica in alcoholic media, size distribution tends to increase as the TEOS concentration increases (41).
216
ARRIAGADA AND OSSEO-ASARE
actions is justified, in part, by the fact that in most of the synthesis experiments the water content in the system was well in excess of that required for complete TEOS conversion. Therefore, the water-to-surfactant molar ratio R (which governs the surfactant aggregation phenomena) is basically constant.
FIG. 6. Effect of the water-to-surfactant molar ratio (R) on the mean diameter of SiO2 particles at different ammonia concentrations; [TEOS] 5 0.026 M, T 5 228C.
Amphiphilic nature of partially hydrolyzed TEOS. It is known that more highly hydrolyzed TEOS species become more hydrophilic (56), a behavior which is due to the presence of polar silanol groups in the molecule. Interfacial tension measurements reported elsewhere (57) suggest that the partially hydrolyzed species become surface active, and therefore remain associated with the reverse micellar aggregates once they are formed. The issue of surface activity of partially hydrolyzed TEOS molecules has also been considered in the literature (58 – 60), in particular in relation to particle synthesis by spontaneous emulsification processes (60). It is suggested, therefore, that all further reactions (hydrolysis and condensation) are restricted to the locale of the reverse micellar aggregates, and thus the overall mechanism of particle nucleation and growth must be analyzed in terms of both intra- and intermicellar events. It is likely that hydrolysis may occur in each reverse micelle (basically an intramicellar process). On the other hand, condensation of silanol groups
Conceptual Model of the Nucleation Process in NP-5 Microemulsions It has been shown above that stable dispersions of ultrafine monodisperse silica particles are produced in the NP-5/cyclohexane reverse micellar system. The narrow size distributions obtained (Fig. 7) under different experimental conditions suggest, in principle, that in the microemulsion system the events of nucleation and growth of silica particles are very well separated in time. A qualitative attempt is made here to reconcile this view of the particle formation process with the observed minimum in particle size at intermediate R values and with the effects of the ammonia concentration on particle size (Fig. 6). For this purpose, the effects of the concentration, structure, and dynamics of the reverse micellar aggregates on the TEOS hydrolysis and condensation reactions of interest are considered. To initiate this analysis, it is assumed, as a first approximation, that the aggregation numbers (hence the concentration) of the reverse micelles present in the solution are not significantly affected by the addition of TEOS or by the subsequent reactions, at least at the early stages of the process where nucleation is occurring. Under these conditions, the number of reverse micellar aggregates in the system is only a function of the surfactant concentration and of the water-to-surfactant molar ratio R. The assumption that the reverse micellar system remains essentially unperturbed by the ongoing chemical re-
FIG. 7. Effect of the water/TEOS ratio on the particle size and size distribution of SiO2 particles; R 5 2.32, T 5 228C.
SILICA IN WATER-IN-OIL MICROEMULSION
TABLE 1 Aggregate Statistics in the NP-5/Cyclohexane Systema [NP-5]m(M)
R
Nb
N m 3 10218c
Nw
Nt
0.14 0.09 0.07 0.05 0.04 0.03
1.2 1.8 2.3 3.0 4.2 6.1
47 71 89 113 160 227
9.2 3.9 2.5 1.5 0.7 0.4
54 128 203 334 677 1370
8.6 20.3 32.3 53.1 107.4 217.8
Data for samples with concentrated NH4OH and h 5 6.3. Aggregation numbers from Kitahara’s data (22). c Referred to the total microemulsion volume (ca. 5 mL). a b
(nucleation and particle growth) may occur within a given reverse micelle or by intermicellar contacts. Upon interaction between TEOS molecules and a given surfactant aggregate, the rate and extent of the hydrolysis, initial condensation (nucleation), and depolymerization processes are likely to depend on the following factors: (a) the state of the solubilized water molecules in the reverse micelle, i.e., the water “reactivity,” (b) the structure of the reverse micelle, i.e., the existence or absence of a well-defined aqueous environment, (c) the number of TEOS (or partially hydrolyzed TEOS) molecules associated with a given reverse micelle, and (d) the concentration of hydroxyl ions in the aqueous domain. By considering one isolated average reverse micelle, the conditions favoring hydrolysis, condensation, and depolymerization can be envisaged. Aggregate statistics. Quantification of aggregate statistics was first undertaken in order to evaluate reverse micellar populations and to estimate the average number of TEOS molecules which, in theory, would interact with each reverse micelle at the beginning of the reaction. Such calculations assume, in principle, that all TEOS molecules are associated with the reverse micelles. The aggregation number (N) at different R values and the concentrations of NP-5 ([NP-5]m) in each sample were combined to calculate the mean number of reverse micellar aggregates (N m) in the microemulsion volume (ca. 5 mL), the number of water molecules per aggregate (N w), and the number of (initial) TEOS molecules per aggregate (N t) as functions of R. Table 1 shows the results obtained for selected R values. As can be seen from Table 1, the number of reverse micellar aggregates (N m) decreases about 20-fold when R increases from 1.2 to 6.1. Thus, as R increases, fewer but larger reverse micellar aggregates are present in the system. Accordingly, both the number of water molecules per aggregate (N w) and the number of TEOS molecules which (in theory) interact with each aggregate (N t) increase significantly as R increases. In a given aggregate, however, the water-to-TEOS molar ratio (h) would be constant under these conditions (h 5 6.3 for these samples).
217
Effects of R on particle size. Consideration of the effect of R on the proportion of free (“reactive”) water in the reverse micellar aggregates and on the number of TEOS molecules per reverse micelle (N t) indicates that, at low R values, hydrolysis and nucleation are not favored in a given reverse micelle. Under low R conditions, water is mostly bound to the surfactant molecules (Fig. 4), and the mobility of the OH2 ions is reduced. Furthermore, there are on the average few TEOS monomers per aggregate (N t, Table 1), so that intramicellar nucleation is less probable. In addition, the interpenetrated surfactant structure of the reverse micelle at low R (7) is such that water molecules are effectively shielded by the oxyethylene chains, which may restrict the access of TEOS to the polar domain. At low R values hydrolysis may also be inhibited due to steric effects. As discussed above (Eq. [3]), maximum charge separation is achieved when the attacking hydroxyl ion and the leaving ethoxy group are on opposite sides of the silicon atom, and inversion of the silicon tetrahedron occurs when the ethoxy group is released (34, 35). Thus, at low R the formation of highly hydrolyzed monomers may be inhibited by steric interactions with the surfactant tails, which may restrict the mobility of a TEOS molecule and thus reduce its ability to release all of its ethoxy groups. At high R values, on the other hand, both hydrolysis and condensation (nucleation) are, in principle, favored in a given reverse micelle. Under high R conditions, water molecules are mostly “free” (Fig. 4), and there is a well-defined aqueous environment. Thus, the production of more highly hydrolyzed TEOS species in the reverse micelle is favored. It is also probable that, due to the existence of a water pool and the enhanced mobility of hydrolyzed TEOS species, the catalytic role of the hydroxyl ions becomes more effective. In addition, according to Table 1, under relatively high R conditions, a larger number of TEOS monomers (N t) will be present in each reverse micelle; therefore, the probability of interaction between neighboring silanol groups to form Si-O-Si bonds (nucleation) is higher. At high R values, then, intramicellar nucleation would be favored. The essential features of the nucleation process according to the above discussion can be summarized as schematized in Fig. 8. The hydrolysis process is represented in Fig. 8a, which shows that few monomers are present per reverse micelle at low R. Conversely, a larger number of hydrolyzed TEOS species is expected to be present in a reverse micelle at high R values. As shown in Fig. 8b, nucleation in the aggregates containing few monomers may depend heavily on intermicellar collisions, while in contrast each reverse micelle would be able to produce a nucleus at high R values. According to the above ideas, the formation of a higher number of nuclei is favored at high R values. Therefore, if nucleation occurs in the reverse micellar system over a limited period of time as suggested above, then the final particle size should decrease continuously as R increases. As shown in Fig.
218
ARRIAGADA AND OSSEO-ASARE
FIG. 8. Formation of silica nuclei by hydrolysis of TEOS in reverse micellar systems. (a) Partial hydrolysis of TEOS and association to reverse micelles, (b) intra- and intermicellar nucleation.
6, however, this trend conforms to the experimental data only for the samples prepared with the lowest ammonia concentration (1.6 wt% NH3). At higher ammonia concentrations, a minimum in particle size is observed as R is increased. The minima in particle size observed in Fig. 6 can be rationalized by considering the effects of intermicellar interactions on the particle formation process. As noted under Experimental Procedures, R was varied (in the range of 0.50 to 4.80) by changing the surfactant concentration. It is well established that the rate of intermicellar matter exchange (k ex) increases with the water/surfactant molar ratio (R) and that the closer the solubilization curve is approached, the higher the exchange rate (61– 63). Referring to Fig. 2 and the straight line depicting the composition range of the microemulsion compositions utilized, increasing R amounts to moving toward the cyclohexane apex. Thus, increase in R pushes the microemulsion composition toward the solubilization curve, and therefore to increased intermicellar interaction. Increases in intermicellar interaction will promote the aggregation of the silica particles contained in the communicating reverse micelles. Thus, it is suggested here that the rise in particle size at high R values in Fig. 6 is attributable to particle growth by aggregation. It is interesting
to note here that aggregative growth has also been proposed by Bogush et al. and Bogush and Zukoski (47, 48) for silica formation via TEOS hydrolysis in ethanolic solutions. This proposed micellar dynamics effect is consistent with the observed effect of ammonia concentration on the R value (R min) corresponding to the minimum in particle size. As shown in Fig. 6, R min decreases with increase in ammonia concentration. Referring to Fig. 2, it can be seen that the highest R used (i.e., the lowest NP-5 concentration used) approaches the solubilization curve more closely when concentrated ammonia (29.6 wt% NH3) instead of water is the microemulsion aqueous phase. This means that significant intermicellar interaction (and correspondingly, particle aggregation) will occur at a lower R value when the ammonia concentration is relatively high. Effects of ammonia concentration on particle size. Examination of Fig. 6 reveals that at relatively low R (e.g., R 5 0.5), the particle size goes through a minimum as ammonia concentration is increased. In contrast, at relatively high R (e.g., R 5 5), the particle size increases monotonically with ammonia concentration. At this time, no definitive explana-
SILICA IN WATER-IN-OIL MICROEMULSION
tions can be offered for these trends. However, it is not unreasonable to expect that, as in the case of silica formation in homogeneous alcohol-alkoxide systems (31, 41– 45), the observed particle size dependence on ammonia concentration may be related to the ability of the OH2 ion (generated by the hydrolysis of ammonia) to catalyze both hydrolysis and condensation reactions and the tendency for siloxane bonds to break in highly alkaline solutions (32, 35, 39). As noted above, at low R alkoxide hydrolysis and silica nucleation are inhibited. Raising the ammonia concentration then increases the OH2 concentration which, in turn, increases hydrolysis and nucleation rates. The subsequent increase in the number of nuclei leads to smaller particles. The increase in particle size at relatively high ammonia concentrations may be due to the base-catalyzed depolymerization (32, 35, 39) which decreases the number of stable nuclei. In the case of relatively large R, it was argued above that nucleation should be favorable. Thus, it is likely that the observed increase of particle size with increase in ammonia concentration is associated with growth-related processes. As already noted above, both high R and high ammonia concentration should favor intermicellar interaction. Thus, it is proposed here that at high R, increase in ammonia concentration increases intermicellar exchange rates and therefore particle aggregation. SUMMARY AND CONCLUSIONS
The work presented and discussed here reveals that, in the preparation of nanosize silica particles by controlled hydrolysis of TEOS in the NP-5/cyclohexane/ammonium hydroxide w/o microemulsion system, variations in the water-to-surfactant molar ratio (R) and ammonia concentration can have a significant effect on particle size. Specifically, the particle size goes through a minimum as R is increased. Further, the R value (R min) corresponding to the minimum in particle size becomes smaller as ammonia concentration is increased. In order to relate these variations in particle size to the physicochemical characteristics of the corresponding microemulsion phase, complementary experiments have been conducted to examine the effects of the water-to-surfactant molar ratio and/or ammonia concentration on (a) the stability domain of the single phase microemulsion, (b) the minimum surfactant concentration needed to solubilize the aqueous phase, and (c) the transition from bound to free water in the microemulsion polar region. Phase equilibria investigations show that introduction of ammonia into the NP-5/cyclohexane/water microemulsion system results in a decrease in size of the water-in-oil microemulsion stability domain. On the other hand, ammonia has no significant effect on the transition from bound to free water, which occurs at a water-to-surfactant molar ratio of about 1. The minimum surfactant concentration needed to solubilize the aqueous phase is of the order of 1.1 wt% and this is not significantly affected by ammonia concentration.
219
Given the fact that water is needed for the alkoxide hydrolysis reaction (Eq. [2]), it is reasonable to consider that the microemulsion water pools constitute the locale of the particle formation process. At low R values, the number of water molecules per surfactant aggregate (N w) is low, the number of TEOS molecules per surfactant aggregate (N t) is low, and there is little free water. Accordingly, hydrolysis and silica nucleation are inhibited. The result is that only a few nuclei are produced and these then grow to become large particles. In contrast, when R is relatively high, N w, N t, and the number of free water molecules are also relatively high. Thus, both hydrolysis and nucleation are favored. In principle, this should lead to the production of a large number of nuclei and, eventually, a large number of small particles. Contrary to expectation, however, the particle size does not decrease monotonically with increase in R. It is argued here that the observed increase in particle size with R at relatively high large values is attributable to particle aggregation associated with intermicellar exchange. It is well established that the rate of intermicellar exchange increases with increases in the water-to-surfactant molar ratio (61– 63). Further, the exchange rate increases as the microemulsion composition approaches the solubilization curve (61– 63). Accordingly, it is further suggested that increases in ammonia concentration, by resulting in shrinkage of the microemulsion stability region, should lead to an increase in the intermicellar exchange rate. This should promote particle aggregation and give rise to larger particles. This is exactly what is found experimentally, as demonstrated by the high R region of Fig. 6. It is apparent, however, that the nucleation of silica in the NP-5 reverse micellar system involves a number of factors which have not been considered in the qualitative model discussed above. Further elucidation of the factors which determine the number of particles produced as a function of R at a given ammonia concentration requires a more quantitative approach. Rather than a single reverse micelle, the whole of the micellar population must be taken into account since intermicellar events are occurring. For example, nucleation at low R values is considered difficult in each reverse micelle (see above), but intermicellar nucleation may be important since a large concentration of aggregates is present at low R. Similarly, each reverse micelle is assumed to be capable of producing a nucleus at high R values, but then there are fewer reverse micelles under these conditions. Also, it is possible that only a fraction of the TEOS molecules in the system is associated with the reverse micellar pseudophase; this would reduce significantly the average number of TEOS molecules per micelle as compared to the N t figures reported in Table 1. Work is in progress to develop a statistical model for the nucleation process. The roles of intermicellar events and of the dynamics of the reverse micellar system are considered, as well as the issue of TEOS partition between the reverse micellar pseudophase and the bulk oil phase.
220
ARRIAGADA AND OSSEO-ASARE
REFERENCES 1. Fendler, J. H., Chem. Rev. 87, 877 (1987). 2. Ward, A. J. I., and Friberg, S. E., Mater. Res. Bull. 14, 41 (1989). 3. Osseo-Asare, K., and Arriagada, F. J., in “Ceramic Powder Science III” (G. L. Messing, S. Hirano, and H. Hausner, Eds.), pp. 3–16. Am. Ceramic Soc., Westerville, OH, 1990. 4. Khan-Lodhi, A., Robinson, B. H., Towey, T., Hermann, C., Knoche, W., and Thesing, U., in “The Structure, Dynamics and Equilibrium Properties of Colloidal Systems” (D. M. Bloor and E. Wyn-Jones, Eds.), pp. 373– 383. Kluwer Academic, Boston, MA, 1990. 5. Pileni, M. P., J. Phys. Chem. 97, 6961 (1993). 6. Candau, F., Macromol. Symp. 92, 169 (1995). 7. Osseo-Asare, K., and Arriagada, F. J., Colloids Surf. 50, 321 (1990). 8. Arriagada, F. J., and Osseo-Asare, K., in “Refractory Metals: Extraction, Processing and Applications” (K. Liddell, D. R. Sadoway, and R. G. Bautista, Eds.), pp. 259 –269. TMS-AIME, Warrendale, PA, 1991. 9. Arriagada, F. J., and Osseo-Asare, K., Colloids Surf. 69, 105 (1992). 10. Arriagada, F. J., and Osseo-Asare, K., J. Dispersion Sci. Technol. 15, 59 (1994). 11. Boakye, E., Radovic, L. R., and Osseo-Asare, K., J. Colloid Interface Sci. 163, 120 (1994). 12. Arriagada, F. J., and Osseo-Asare, K., in “Advances in Chemistry Series, No. 234” (H. Bergna, Ed.), pp. 113–128. Am. Chem. Soc., Washington, DC, 1994. 13. Arriagada, F. J., and Osseo-Asare, K., J. Colloid Interface Sci. 170, 8 (1995). 14. GAF Chemicals Corporation, Technical bulletin, Igepal CO nonionic surfactants. 15. Hinze, W., and Fendler, J. H., J. Chem. Soc. Dalton, 238 (1975). 16. Fuestel, D., Runge, F., and Ilgenfritz, G., in “Reactions in Compartmentalized Liquids” (W. Knoche and R. Shoemacker, Eds.), p. 21. SpringerVerlag, Berlin, 1989. 17. Kizling, J., and Stenius, P., J. Colloid Interface Sci. 118, 482 (1987). 18. Aveyard, R., Binks, B. P., and Fletcher, P. D. I., Langmuir 5, 1210 (1989). 19. Ravey, J. C., Buzier, M., and Picot, C., J. Colloid Interface Sci. 97, 9 (1984). 20. Shinoda, K., and Saito, H., J. Colloid Interface Sci. 26, 70 (1968). 21. Kon-no, K., and Kitahara, A., J. Colloid Interface Sci. 34, 221 (1970). 22. Kitahara, A., J. Phys. Chem. 69, 2788 (1965). 23. Sheih, P. S., and Fendler, J. H., J. Chem. Soc. Faraday Trans. I, 73, 1480 (1977). 24. Tanaka, R., J. Colloid Interface Sci. 122, 220 (1988). 25. Christenson, H., Friberg, S. E., and Larsen, D. W., J. Phys. Chem. 84, 3633 (1980). 26. Wilcoxon, J. P., and Williamson, R. L., in “Macromolecular Liquids” (C. R. Safinya, S. A. Safran, and P. A. Pincus, Eds.), p. 269. Mater. Res. Soc., Pittsburgh, 1990. 27. Nagy, J., Colloids Surf. 35, 201 (1989). 28. Kandori, K., Shizuka, N., Kon-no, K., and Kitahara, A., J. Dispersion Sci. Technol. 8, 477 (1987). 29. Guizard, C., et al. in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), Vol. 121, p. 115. Mater. Res. Soc. Symp. Proc., Mater. Res. Soc., Pittsburgh, 1988. 30. Espiard, P., Mark, J. E., and Guyot, A., Polymer Bull. 24, 173 (1990). 31. Stober, W., Fink, A., and Bohn, E., J. Colloid Interface Sci. 26, 62 (1968). 32. Brinker, C. J., and Scherer, G. W., “Sol-Gel Science.” Academic Press, San Diego, 1990.
33. Aelion, R., Loebel, A., and Eirich, F., J. Am. Chem. Soc. 72, 5705 (1950). 34. Keefeer, K. D., in “Better Ceramics through Chemistry” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), p. 15. Elsevier, New York, 1984. 35. Iler, R. K., “The Chemistry of Silica.” Wiley, New York, 1979. 36. Byers, C. H., and Harris, M. T., in “Ultrastructure Processing of Advanced Ceramics” (J. D. Mackenzie and D. R. Ulrich, Eds.), p. 843. Wiley, New York, 1988. 37. Byers, C. H., Harris, M. T., and Williams, D. F., Ind. Eng. Chem. Res. 26, 1916 (1987). 38. Artaki, I., Bradley, M., Zerda, T. W., and Jonas, J., J. Phys. Chem. 89, 4399 (1985). 39. Greenberg, S. A., J. Phys. Chem. 61, 960 (1957). 40. Baes, C. F., and Mesmer, R. E., “The Hydrolysis of Cations.” Wiley, New York, 1976. 41. Van Helden, A. K., Jansen, J. W., and Vrij, A., J. Colloid Interface Sci. 81, 354 (1981). 42. Matsoukas, T., and Gulari, E., J. Colloid Interface Sci. 132, 13 (1989). 43. Bogush, G. H., Tracy, M. A., and Zukoski, C. F., J. Non-Cryst. Solids 104, 95 (1988). 44. Adams, J., Baird, T., Braterman, P. S., Cairns, J. A., and Segal, D. L., in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark and D. R. Ulrich, Eds.), p. 361. Mater. Res. Soc., Pittsburgh, 1988. 45. Badley, R. D., Ford, W. T., McEnroe, F. J., and Assink, R. A., Langmuir 6, 792 (1990). 46. Bagnall, C. M., Howarth, L. G., and James, P. F., J. Non-Cryst. Solids 121, 56 (1990). 47. Bogush, G. H., Dickstein, G. L., Lee, K. C. P., and Zukoski, C. F., in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), p. 57. Mater. Res. Soc., Pittsburgh, 1988. 48. Bogush, G. H., and Zukoski, C. F., J. Colloid Interface Sci. 142, 1 (1991). 49. Brinker, C. J., in “The Colloid Chemistry of Silica, Advances in Chemistry Series” (H. Bergna, Ed.), pp. 361– 402. Am. Chem. Soc., Washington, DC, 1994. 50. Firman, P., Haase, D., Jen, J., Kahlweit, M., and Strey, R., Langmuir 1, 718 (1985). 51. Johnson, K. A., and Shah, D. O., in “Surfactants in Solution” (K. L. Mittal and P. Bothorel, Eds.). Plenum, New York, 1986. 52. Kitahara, A., and Kon-no, K., J. Phys. Chem. 70, 3394 (1966). 53. Gobe, M., Kon-no, K., Kandori, K., and Kitahara, A., J. Colloid Interface Sci. 93, 293 (1983). 54. Becher, P., in “Nonionic Surfactants” (M. J. Schick, Ed.), Chap. 15. Dekker, New York, 1966. 55. Meisel, D., Matheson, M. S., and Rabani, J., J. Am. Chem. Soc. 100, 117 (1978). 56. Kamiya, K., and Yoko, T., J. Mater. Sci. 21, 842 (1986). 57. Arriagada, F. J., and Osseo-Asare, K., in preparation. 58. Kaufman, V. R., and Avnir, D., Langmuir 2, 717 (1986). 59. Matsui, K., and Nakazawa, T., Bull. Chem. Soc. Jpn. 63, 11 (1990). 60. Minehan, W. T., Messing, G. L., and Pantano, C. G., J. Non-Cryst. Solids 108, 163 (1989). 61. Clark, S., Fletcher, P. D. I., and Ye, X., Langmuir 6, 1301 (1990). 62. Almgren, M., Stam, J. V., Swarup, S., and Lofroth, J. E., Langmuir 2, 432 (1986). 63. Bommarius, A. S., Holzwarth, J. F., Wang, D. I. C., and Hatton, T. A., J. Phys. Chem. 94, 7232 (1990).
Synthesis of Nanosize Silica in a Nonionic Water-in-Oil Microemulsion: Effects of the Water/Surfactant Molar Ratio and Ammonia Concentration F. J. Arriagada1 and K. Osseo-Asare2 Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802 Received December 15, 1997; accepted November 18, 1998
The effect of ammonia concentration on the region of existence of single-phase water-in-oil microemulsions has been investigated for the system polyoxyethylene (5) nonylphenyl ether (NP-5)/ cyclohexane/ammonium hydroxide. The presence of ammonia decreases the size of the microemulsion region. A minimum concentration of surfactant (estimated at about 1.1 wt%) is required for solubilization of the aqueous phase; this value is not significantly affected by ammonia concentration. As indicated by fluorescence spectral data, the transition between bound and free water occurs when the water-to-surfactant molar ratio is about 1 and the presence of ammonium hydroxide does not appear to have a significant effect on this. Ultrafine (30 –70 nm diameter), monodisperse silica particles produced by hydrolysis of tetraethoxysilane (TEOS) in the microemulsion show a complex dependence of the particle size on the water-to-surfactant molar ratio (R) and on the concentration of ammonium hydroxide. At relatively low ammonia concentration in the aqueous pseudophase (1.6 wt% NH3) the particle size decreases monotonically with increase in R. However, for higher ammonia concentrations (6.3–29.6 wt% NH3) a minimum in particle size occurs as R is increased. These trends are rationalized in terms of (a) the effects of the concentration, structure, and dynamics of the NP-5 reverse micelles on the hydrolysis and condensation reactions of TEOS, and (b) the effects of ammonia concentration on the stability of the microemulsion phase, the hydrolysis/condensation reactions of TEOS, and the depolymerization of siloxane bonds. © 1999 Academic Press Key Words: silica; w/o microemulsion; nanoparticles; tetraethoxysilane; TEOS; polyoxyethylene nonylphenyl ether; phase equilibria; hydrolysis/condensation.
INTRODUCTION
The growing scientific and technological interest in nanosize particles is fueling a quest for new synthesis methods that are capable of molecular-level control of particle characteristics. In this connection, the microheterogeneous nature of microemulsion media offers attractive possibilities, as already demonstrated for a variety of materials, including metals, chalco1 Present address: Idesol Ingenieros S.A., Av. San Martin 255 Oficina 53, Iquique, Chile. 2 To whom correspondence should be addressed.
0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
genides, metal halides, carbonates, oxides, and organic polymers (1– 6). In our laboratory we are investigating the relationship between microemulsion properties and the characteristics of the ensuing nanoparticles (7–13). In an earlier paper (7) we reported on the preparation of nanosize monodisperse silica particles by controlled hydrolysis of tetraethoxysilane (TEOS) in a nonionic surfactant/cyclohexane/ammonium hydroxide water-in-oil microemulsion system. Particles in the range of 50 to 70 nm were produced with standard deviations below 8.5% around the mean diameter. As the water-to-surfactant molar ratio (R) increased from 0.7 to 2.3, the particle size decreased and the size distribution became narrower. Further work has now shown that the dependence of particle size on the waterto-surfactant molar ratio becomes much more complicated as a wider range of R values is utilized. In the previous work (7) the ammonium hydroxide concentration in the aqueous pseudophase was maintained at the same constant value (29.6 wt%) for all the synthesis experiments. We have now discovered that variations in ammonium hydroxide concentration can lead to dramatic changes in particle characteristics (12). This paper presents and discusses these new findings. The work reported here includes the determination of the one-phase microemulsion phase boundaries (with emphasis on the effects of temperature and ammonium hydroxide concentration), and the investigation of the effects of the water-to-surfactant molar ratio and ammonium hydroxide concentration on the size and size distributions of the resulting particles. The nonionic surfactant selected for this study was a polyoxyethylene nonylphenyl ether with an average of 5 oxyethylene groups per molecule (NP-5). It has an HLB number of 10 (14), and at room temperature is readily soluble in saturated hydrocarbons like cyclohexane. Nonylphenyl and octylphenyl polyoxyethylene ether compounds have been used in studies on reverse micellar catalysis (15, 16), solubilization and phase behavior of reverse microemulsions (17–21), aggregation phenomena in nonaqueous systems (19, 22–25), and microemulsion-based particle synthesis (26 –30). Thus, the literature contains much relevant data for these surfactants on the effects of temperature, electrolyte content and composition on solubili-
210
SILICA IN WATER-IN-OIL MICROEMULSION
zation, aggregation numbers, and the state of solubilized water. In the case of silica synthesis via the alkoxide sol-gel process, an extensive literature exists for homogeneous solution systems (31– 49). EXPERIMENTAL PROCEDURES
Materials The polyoxyethylene (5) nonylphenyl ether surfactant (NP-5) used was the commercial product Igepal CO-520 from GAF Chemicals Corporation. The as-received material was a viscous liquid and contained ca. 0.2 wt% of water, as determined by Karl Fisher titration. Water and low-boiling impurities (such as free alcohols) were removed by treatment at 60 –70°C under vacuum for 26 h (21). The molecular weight of NP-5 is 440.6 g mol21. HPLC grade cyclohexane (99.9%, Aldrich) was used after storage under molecular sieves (8 3 12 mesh, 0.4 nm, Union Carbide). Absolute ethanol (anhydrous, Quantum Chemical) and reagent grade ammonium hydroxide (29.6 wt% NH3, Baker) were used as received. Water was purified by reverse osmosis (Millipore-RO4) and deionization (Millipore Milli-Q). Tetraethoxysilane (TEOS, 991 wt%, Alfa) and fluorescent probe tris(2,29-bipyridyl)ruthenium(II) chloride hexahydrate (Ru(Bpy), Aldrich) were used as received. Microemulsion Characterization Phase equilibria and solubilization. The solubility domains of water and ammonium hydroxide solutions in the NP-5 surfactant system were determined by two methods. In one set of experiments, the surfactant/oil molar ratio was fixed, and the phase behavior of a number of samples with different contents of aqueous phase was observed as a function of temperature. A solution of NP-5 in cyclohexane and the required amounts of water (or ammonium hydroxide solutions) were weighed and sealed in glass ampoules (5 mL, Thomas). After thorough hand-mixing, the ampoules were placed in a thermostated water chamber. While subjected to slow temperature changes (at a rate of 1°C/h), the number of phases present at a given temperature in each sample was recorded. In a separate set of experiments, the phase behavior was determined at 22°C for samples of variable surfactant and aqueous phase content. In this case, the solubilization limits were determined by inspection of the mixture upon small, sequential additions of aqueous phase to a given surfactant/oil solution. Once permanent turbidity was observed (indicating the onset of phase separation), a few samples were backtitrated with surfactant/oil solution to double-check the transition points. Fluorescence spectroscopy. The nature of solubilized water (i.e., bound to the oxyethylene groups or free) within the surfactant aggregates was studied through the fluorescence emission spectra of the hydrophilic probe Ru(Bpy). Fluores-
211
cence spectra were measured at 23°C on a Shimadzu RF-5000 spectrofluorimeter. Excitation wavelength was 460 nm with a 1.5-nm bandwidth for both excitation and emission. Stock solutions of the probe were freshly prepared and stored in the dark before use. Probe concentration (with respect to the total microemulsion volume) was 9 3 1026 M. In the previous study (7), the water-to-surfactant molar ratio (R) was varied by the addition of different amounts of water (or ammonium hydroxide solutions) to a surfactant solution of NP-5 at a fixed concentration (0.3 M) in cyclohexane. In order to ascertain whether in this surfactant system the transition from bound to free water molecules is only a function of the water-to-surfactant molar ratio (R), additional experiments were conducted in which the parameter R was varied by changing the surfactant concentration (in the range of 0.05 to 0.29 M), while keeping constant the amount of water added to the samples. Ammonium hydroxide solutions with ammonia concentrations in the range of 1.6 to 29.6 wt% NH3 (0.9 to 15 M) were investigated. Particle Preparation A set of standard conditions was used in most of the synthesis experiments. Typically TEOS concentrations in the range of 0.02 to 0.04 M, water-to-TEOS molar ratios (h) of above 6, and concentrated ammonium hydroxide (29.6 wt% NH3 in the aqueous solution) were used. Hereafter, the concentrations quoted refer to the total microemulsion volume unless stated otherwise. The water content was in excess of that required for complete conversion of TEOS to silica. Given the low TEOS concentrations used, the resulting solid content in the samples was limited to below 0.2 wt%. In order to maintain all relevant parameters ([TEOS], [H2O], [NH4OH]) constant while exploring the effect of the water-tosurfactant molar ratio (R), samples with different surfactant concentrations were used. In each sample, the value of R was calculated by considering only the surfactant associated with the reverse micelles according to the expression, R 5 @H2O#/~@NP-5# 2 @NP-5# o! 5 @H2O#/@NP-5# m,
[1]
where [NP-5] is the total surfactant concentration and [NP-5]o is the surfactant concentration in the bulk oil phase, i.e., not adsorbed at the oil/water interface. The quantity [NP-5]m is, thus, the concentration of surfactant associated with the reverse micelles. As shown below, the value of [NP-5]o is ca. 0.02 M of NP-5 in cyclohexane at 22°C. A microemulsion plus second reactant synthesis route (3, 7) was used; i.e., TEOS was added to a previously prepared NP-5/cyclohexane/ammonium hydroxide microemulsion to initiate the synthesis. The experiments were conducted in 8-mL glass bottles, sealed by Teflon-lined screw caps (Supelco). All synthesis experiments were conducted at 22°C. A fixed total microemulsion volume (ca. 5 mL) with NP-5 concentrations in
212
ARRIAGADA AND OSSEO-ASARE
dispersed in the continuous cyclohexane phase. It can be seen that replacement of water by ammonium hydroxide solutions has a strong effect on the phase stability in this system; the solubilization curve is shifted to lower temperatures. This effect is more pronounced with concentrated ammonium hydroxide. For both concentrated (29.6 wt% NH3) and dilute (14.6 wt% NH3) ammonium hydroxide solutions, only the solubilization curve (i.e., upper temperature limit) was observed for a 5.83 wt% NP-5 solution. These samples remained clear upon cooling to about 5°C, where cyclohexane freezes. The effect of ammonium hydroxide on the phase equilibria may be interpreted as the result of the competition between hydroxyl ions and the oxyethylene groups of the surfactant for interaction with water molecules; i.e., ammonium hydroxide is acting as a lyotropic salt reducing the mutual solubility between water and surfactant (50). Similar results were obtained by Kon-no and Kitahara (21) who studied the effects of sodium hydroxide and salt additions in NP-8/tetrachloroethylene microemulsions. FIG. 1. Solubilization diagram for H2O and NH4OH solutions in the NP-5/cyclohexane system; 5.83 wt% NP-5.
the range of 0.05 to 0.3 M was normally used, to which constant amounts of aqueous phase and then of TEOS (typically 30 and 25 mL, respectively) were added with a microsyringe (Eppendorf, 100 mL). In most cases, all components were weighed to determine precisely the R values and the final sample composition. Characterization of Particles A Phillips 420 transmission electron microscope was used to characterize particle size and size distribution. Samples for TEM observation were prepared by depositing a drop of the dispersion on Formvar-covered carbon-coated copper grids (300#, 3 mm, Structure Probe) and drying on filter paper at room temperature. Prior to extraction of samples, the bottles were sonicated for 30 s to ensure the extraction of representative dispersion volumes. Number-average particle diameters and size distributions were determined on enlarged TEM micrographs with a ZIDAS image analysis system (Zeiss). Several hundred particles were measured for each sample. RESULTS AND DISCUSSION
Phase Equilibria and Solubilization Effect of temperature on phase equilibria. The solubilization diagram of water and ammonium hydroxide solutions as a function of temperature in cyclohexane containing 5.83 wt% NP-5 (ca. 0.1 M) is shown in Fig. 1. The region between the solubilization and solubility curves is a homogeneous transparent phase where water-swollen NP-5 reverse micelles are
Effects of NP-5 and ammonia concentrations on phase equilibria at 22°C. It was of interest to evaluate microemulsion stability as a function of surfactant concentration at a fixed temperature of 22°C, which is the temperature of the particle synthesis experiments reported in this work. A section of the ternary phase diagram for the NP-5/cyclohexane/aqueous solution system is shown in Fig. 2, illustrating the microemulsion regions for water and concentrated ammonium hydroxide solutions. As can be seen in Fig. 2, water solubilization increases significantly for NP-5 concentrations above about 6 wt%. At these relatively high surfactant concentrations, the one-phase microemulsion region is bound by the solubilization curve, beyond which phase separation involves the equilibrium between water-in-oil microemulsion and excess water. The range of microemulsion compositions used in the synthesis experiments is also illustrated in Fig. 2. As described above, most experiments were conducted with a constant concentration of aqueous phase, i.e., in a range of microemulsion compositions parallel to the NP-5/cyclohexane axis. Determination of unaggregated surfactant concentration. Following Johnson and Shah (51), the data presented in Fig. 2 are plotted in Fig. 3 as the maximum water and ammonium hydroxide uptake vs the initial NP-5 concentration. Extrapolation of the solubilized amount to zero content of aqueous phase indicates the “critical microemulsion concentration” (cmc), i.e., the concentration of surfactant which is not adsorbed at the surfactant/water interface (51). Based on Fig. 3, a cmc value of about 1.1 wt% is obtained by extrapolation of the water solubilization data, which furthermore coincides very closely with the onset of solubilization of ammonium hydroxide. This value agrees qualitatively with cmc values estimated from solubilization data available in the literature for similar polyoxyethylene compounds in cyclohexane (0.6 wt% of NP-7.5 solubilizing 0.1 M NaOH solutions (52) and 1 wt% of NP-6 solubilizing
SILICA IN WATER-IN-OIL MICROEMULSION
213
FIG. 2. One-phase w/o microemulsion regions in the ternary phase diagram for the NP-5/cyclohexane/water and NP-5/cyclohexane/NH4OH (29.6 wt% NH3) systems, T 5 228C.
1.4 M FeCl2 solutions (53)). It is worth mentioning that the cmc values in nonaqueous solvents are, in general, considerably higher than the corresponding critical micelle concentration (cmc) values typically found for nonionic surfactants in aqueous solutions (;1023 wt%, (54)). Fluorescence spectroscopy and the nature of solubilized water. Figure 4 presents the fluorescence spectra of Ru(Bpy) in NP-5/cyclohexane/water microemulsions. It must be noted that, in these experiments, R was varied (in the range of 0.50 to 4.80) by changing the surfactant concentration (in the range of 0.29 to 0.05 M). This is in contrast to the previous work (7) in which the surfactant concentration was maintained constant. It can be seen that at the low water contents (R 5 0.5 and 0.73), the emission spectra exhibit a maximum intensity at about 570 nm with a red shoulder at about 625 nm. With further increases in the water content, the intensity at 625 nm increases at the expense of the 570 nm peak (see R 5 0.93). It is interesting to note that in the case of the spectra for R 5 1.24 and R 5 4.80, only the 625-nm peak remains. The red shift in fluorescence emission is attributable to the fact that, as the water content increases, there is a corresponding increase in the polarity of the environment surrounding the probe molecule. Since the maximum emission intensity at the higher R values occurs at the same wavelength as that recorded for
Ru(Bpy) in bulk water (620 – 630 nm) (55), it may be concluded that R 5 1 marks the transition from bound to “free” water molecules. The maximum emission intensity shown in Fig. 4 follows a trend which is similar to that observed in the case of the previous experiments (7) that were based on a fixed surfactant concentration and with ammonium hydroxide as the solubilizate. Thus, within the surfactant concentration range investigated, the transition from bound to free water molecules is only a function of the water-to-surfactant molar ratio, R. Effects of Synthesis Variables on Particle Characteristics TEOS hydrolysis and condensation. As shown by the early work of Stober et al. (31), under basic conditions TEOS undergoes hydrolysis and polycondensation reactions which result in the formation of monodisperse spherical particles of amorphous silica. The hydrolysis reaction producing silanol groups may be represented as Si(OR)4 1 xH2O 5 Si(OR)42x~OH! x 1 xROH,
[2]
where R 5 C2H5. The base-catalyzed hydrolysis occurs by a nucleophilic substitution mechanism, where the hydroxyl anion attacks the silicon atom (32–35), as shown in Eq. [3].
214
ARRIAGADA AND OSSEO-ASARE
FIG. 3. Maximum H2O and H2O-NH4OH (29.6 wt% NH3) uptake as a function of the initial concentration of NP-5 in cyclohexane; T 5 228C.
RO RO HO– + Si©OR RO
RO
OR
d– HO
d– Si
OR
OR OR HO©Si + OR– .
OR
[3]
OR
The rate of TEOS hydrolysis is affected by the nature of the solvent, and this has been attributed to both steric effects and hydrogen bonding (32, 36 –38). Byers and Harris (36) reported that the hydrolysis rate is faster in higher alcohols (such as butanol) than in short-chain alcohols such as 1-propanol. According to these authors, the strong hydrogen bonding existing between the water molecules and the lower alcohols reduces the effective availability of water molecules and thus the hydrolysis rate, probably by reducing the mobility of OH2 ions (see Eq. [3]). Interaction between hydrolyzed species results in condensation (formation of silicon-oxygen-silicon bonds), which involves the attack of a nucleophilic deprotonated silanol on a protonated silanol (Eq. [4]) or on an ethoxysilane group (Eq. [5]), producing respectively water or ethanol (32, 35): [Si-OH 1 HO-Si[ 5 [Si-O-Si[ 1 H2O
[4]
[Si-OH 1 RO-Si[ 5 [Si-O-Si[ 1 ROH
[5]
Effect of the ammonia concentration. Particles sampled after 930 h and which were produced at selected R values with different ammonia concentrations are shown in the TEM micrographs of Fig. 5. For both R 5 4.5 and R 5 1.7, the particle morphology becomes more irregular as the ammonia concentration decreases; however, the effect appears to be more significant at high R values (Figs. 5A to 5D), as seen in the sample with the lowest ammonia concentration (Fig. 5D).
High ammonia concentrations (i.e., relatively strong alkaline conditions) are thus needed to obtain morphological control in the NP-5 reverse micellar system. In contrast, ammonia concentrations of 0.9 M (the lowest concentrations used in this work) promote the formation of quite spherical particles in homogeneous media (31, 41, 43). Effect of the water-to-surfactant molar ratio (R). A series of samples of different water-to-surfactant molar ratios (R in the range of 0.4 to 6.7) and various constant ammonium hydroxide concentrations in the solubilized aqueous phase was prepared. The number-average particle diameter (d n) for 930-h samples (930 h aging time) is shown in Fig. 6. Referring to the data for concentrated ammonium hydroxide (29.6 wt% NH3) as the solubilizate, it can be seen that as R increases, the particle size decreases sharply, reaches a minimum at an R value of about 1.8 and then increases again. A similar but less marked trend is observed for more dilute ammonia solutions, although the minimum in particle size is shifted to higher R values (to about 2.9) for both intermediate ammonia concentrations (i.e., 6.3 and 14.6% NH3) as the ammonia concentration increases. In the most dilute ammonia solution utilized in this work (1.6% NH3), no minimum is observed within the R range investigated, and the particle size decreases continuously. Effect of TEOS concentration. The particle size and size distribution were determined for silica particles obtained in experiments with variable TEOS concentrations. The spread of the particle size distributions was measured by the normalized
FIG. 4. Fluorescence spectra of Ru(Bpy) solubilized in the NP-5/cyclohexane/water reverse micellar system, [Ru(Bpy)] 5 9 3 1026 M, T 5 238C.
SILICA IN WATER-IN-OIL MICROEMULSION
215
FIG. 5. Effect of ammonia concentration (wt% NH3) on the morphology of SiO2 particles; [TEOS] 5 0.026 M, T 5 228C; (A–D) R ca 4.5, (A) 29.6 wt%, (B) 14.6 wt%, (C) 6.3 wt%, (D) 1.6 wt%; (E–H) R ca 1.7, (E) 29.6 wt%, (F) 14.6 wt%, (G) 6.3 wt%, (H) 1.6 wt%.
standard deviation, which corresponds to the standard deviation over the mean (number-average) particle diameter in percentage. Figure 7 shows the number-average particle diameters observed after 21 h of reaction and the final diameters from samples aged for 930 h. Similar to the reported effects of TEOS concentration in alcoholic media (41), a high TEOS concentration in the reverse micellar system promotes the formation of smaller particles (see 930 h data). However, as can be appreciated from Fig. 7, particles produced at high TEOS concentration grow faster, even though the water-to-
TEOS ratio is small. After 21 h of reaction, particles produced with a water/TEOS ratio (h) of 3.2 have already grown to 80% of their final size, while for the largest water/TEOS ratio used (ca. 20) the equivalent figure is only about 70%. As seen further in Fig. 7, the size distribution is narrow (normalized standard deviation of the order of 5%) and relatively insensitive to TEOS concentration. In contrast, in the conventional synthesis of silica in alcoholic media, size distribution tends to increase as the TEOS concentration increases (41).
216
ARRIAGADA AND OSSEO-ASARE
actions is justified, in part, by the fact that in most of the synthesis experiments the water content in the system was well in excess of that required for complete TEOS conversion. Therefore, the water-to-surfactant molar ratio R (which governs the surfactant aggregation phenomena) is basically constant.
FIG. 6. Effect of the water-to-surfactant molar ratio (R) on the mean diameter of SiO2 particles at different ammonia concentrations; [TEOS] 5 0.026 M, T 5 228C.
Amphiphilic nature of partially hydrolyzed TEOS. It is known that more highly hydrolyzed TEOS species become more hydrophilic (56), a behavior which is due to the presence of polar silanol groups in the molecule. Interfacial tension measurements reported elsewhere (57) suggest that the partially hydrolyzed species become surface active, and therefore remain associated with the reverse micellar aggregates once they are formed. The issue of surface activity of partially hydrolyzed TEOS molecules has also been considered in the literature (58 – 60), in particular in relation to particle synthesis by spontaneous emulsification processes (60). It is suggested, therefore, that all further reactions (hydrolysis and condensation) are restricted to the locale of the reverse micellar aggregates, and thus the overall mechanism of particle nucleation and growth must be analyzed in terms of both intra- and intermicellar events. It is likely that hydrolysis may occur in each reverse micelle (basically an intramicellar process). On the other hand, condensation of silanol groups
Conceptual Model of the Nucleation Process in NP-5 Microemulsions It has been shown above that stable dispersions of ultrafine monodisperse silica particles are produced in the NP-5/cyclohexane reverse micellar system. The narrow size distributions obtained (Fig. 7) under different experimental conditions suggest, in principle, that in the microemulsion system the events of nucleation and growth of silica particles are very well separated in time. A qualitative attempt is made here to reconcile this view of the particle formation process with the observed minimum in particle size at intermediate R values and with the effects of the ammonia concentration on particle size (Fig. 6). For this purpose, the effects of the concentration, structure, and dynamics of the reverse micellar aggregates on the TEOS hydrolysis and condensation reactions of interest are considered. To initiate this analysis, it is assumed, as a first approximation, that the aggregation numbers (hence the concentration) of the reverse micelles present in the solution are not significantly affected by the addition of TEOS or by the subsequent reactions, at least at the early stages of the process where nucleation is occurring. Under these conditions, the number of reverse micellar aggregates in the system is only a function of the surfactant concentration and of the water-to-surfactant molar ratio R. The assumption that the reverse micellar system remains essentially unperturbed by the ongoing chemical re-
FIG. 7. Effect of the water/TEOS ratio on the particle size and size distribution of SiO2 particles; R 5 2.32, T 5 228C.
SILICA IN WATER-IN-OIL MICROEMULSION
TABLE 1 Aggregate Statistics in the NP-5/Cyclohexane Systema [NP-5]m(M)
R
Nb
N m 3 10218c
Nw
Nt
0.14 0.09 0.07 0.05 0.04 0.03
1.2 1.8 2.3 3.0 4.2 6.1
47 71 89 113 160 227
9.2 3.9 2.5 1.5 0.7 0.4
54 128 203 334 677 1370
8.6 20.3 32.3 53.1 107.4 217.8
Data for samples with concentrated NH4OH and h 5 6.3. Aggregation numbers from Kitahara’s data (22). c Referred to the total microemulsion volume (ca. 5 mL). a b
(nucleation and particle growth) may occur within a given reverse micelle or by intermicellar contacts. Upon interaction between TEOS molecules and a given surfactant aggregate, the rate and extent of the hydrolysis, initial condensation (nucleation), and depolymerization processes are likely to depend on the following factors: (a) the state of the solubilized water molecules in the reverse micelle, i.e., the water “reactivity,” (b) the structure of the reverse micelle, i.e., the existence or absence of a well-defined aqueous environment, (c) the number of TEOS (or partially hydrolyzed TEOS) molecules associated with a given reverse micelle, and (d) the concentration of hydroxyl ions in the aqueous domain. By considering one isolated average reverse micelle, the conditions favoring hydrolysis, condensation, and depolymerization can be envisaged. Aggregate statistics. Quantification of aggregate statistics was first undertaken in order to evaluate reverse micellar populations and to estimate the average number of TEOS molecules which, in theory, would interact with each reverse micelle at the beginning of the reaction. Such calculations assume, in principle, that all TEOS molecules are associated with the reverse micelles. The aggregation number (N) at different R values and the concentrations of NP-5 ([NP-5]m) in each sample were combined to calculate the mean number of reverse micellar aggregates (N m) in the microemulsion volume (ca. 5 mL), the number of water molecules per aggregate (N w), and the number of (initial) TEOS molecules per aggregate (N t) as functions of R. Table 1 shows the results obtained for selected R values. As can be seen from Table 1, the number of reverse micellar aggregates (N m) decreases about 20-fold when R increases from 1.2 to 6.1. Thus, as R increases, fewer but larger reverse micellar aggregates are present in the system. Accordingly, both the number of water molecules per aggregate (N w) and the number of TEOS molecules which (in theory) interact with each aggregate (N t) increase significantly as R increases. In a given aggregate, however, the water-to-TEOS molar ratio (h) would be constant under these conditions (h 5 6.3 for these samples).
217
Effects of R on particle size. Consideration of the effect of R on the proportion of free (“reactive”) water in the reverse micellar aggregates and on the number of TEOS molecules per reverse micelle (N t) indicates that, at low R values, hydrolysis and nucleation are not favored in a given reverse micelle. Under low R conditions, water is mostly bound to the surfactant molecules (Fig. 4), and the mobility of the OH2 ions is reduced. Furthermore, there are on the average few TEOS monomers per aggregate (N t, Table 1), so that intramicellar nucleation is less probable. In addition, the interpenetrated surfactant structure of the reverse micelle at low R (7) is such that water molecules are effectively shielded by the oxyethylene chains, which may restrict the access of TEOS to the polar domain. At low R values hydrolysis may also be inhibited due to steric effects. As discussed above (Eq. [3]), maximum charge separation is achieved when the attacking hydroxyl ion and the leaving ethoxy group are on opposite sides of the silicon atom, and inversion of the silicon tetrahedron occurs when the ethoxy group is released (34, 35). Thus, at low R the formation of highly hydrolyzed monomers may be inhibited by steric interactions with the surfactant tails, which may restrict the mobility of a TEOS molecule and thus reduce its ability to release all of its ethoxy groups. At high R values, on the other hand, both hydrolysis and condensation (nucleation) are, in principle, favored in a given reverse micelle. Under high R conditions, water molecules are mostly “free” (Fig. 4), and there is a well-defined aqueous environment. Thus, the production of more highly hydrolyzed TEOS species in the reverse micelle is favored. It is also probable that, due to the existence of a water pool and the enhanced mobility of hydrolyzed TEOS species, the catalytic role of the hydroxyl ions becomes more effective. In addition, according to Table 1, under relatively high R conditions, a larger number of TEOS monomers (N t) will be present in each reverse micelle; therefore, the probability of interaction between neighboring silanol groups to form Si-O-Si bonds (nucleation) is higher. At high R values, then, intramicellar nucleation would be favored. The essential features of the nucleation process according to the above discussion can be summarized as schematized in Fig. 8. The hydrolysis process is represented in Fig. 8a, which shows that few monomers are present per reverse micelle at low R. Conversely, a larger number of hydrolyzed TEOS species is expected to be present in a reverse micelle at high R values. As shown in Fig. 8b, nucleation in the aggregates containing few monomers may depend heavily on intermicellar collisions, while in contrast each reverse micelle would be able to produce a nucleus at high R values. According to the above ideas, the formation of a higher number of nuclei is favored at high R values. Therefore, if nucleation occurs in the reverse micellar system over a limited period of time as suggested above, then the final particle size should decrease continuously as R increases. As shown in Fig.
218
ARRIAGADA AND OSSEO-ASARE
FIG. 8. Formation of silica nuclei by hydrolysis of TEOS in reverse micellar systems. (a) Partial hydrolysis of TEOS and association to reverse micelles, (b) intra- and intermicellar nucleation.
6, however, this trend conforms to the experimental data only for the samples prepared with the lowest ammonia concentration (1.6 wt% NH3). At higher ammonia concentrations, a minimum in particle size is observed as R is increased. The minima in particle size observed in Fig. 6 can be rationalized by considering the effects of intermicellar interactions on the particle formation process. As noted under Experimental Procedures, R was varied (in the range of 0.50 to 4.80) by changing the surfactant concentration. It is well established that the rate of intermicellar matter exchange (k ex) increases with the water/surfactant molar ratio (R) and that the closer the solubilization curve is approached, the higher the exchange rate (61– 63). Referring to Fig. 2 and the straight line depicting the composition range of the microemulsion compositions utilized, increasing R amounts to moving toward the cyclohexane apex. Thus, increase in R pushes the microemulsion composition toward the solubilization curve, and therefore to increased intermicellar interaction. Increases in intermicellar interaction will promote the aggregation of the silica particles contained in the communicating reverse micelles. Thus, it is suggested here that the rise in particle size at high R values in Fig. 6 is attributable to particle growth by aggregation. It is interesting
to note here that aggregative growth has also been proposed by Bogush et al. and Bogush and Zukoski (47, 48) for silica formation via TEOS hydrolysis in ethanolic solutions. This proposed micellar dynamics effect is consistent with the observed effect of ammonia concentration on the R value (R min) corresponding to the minimum in particle size. As shown in Fig. 6, R min decreases with increase in ammonia concentration. Referring to Fig. 2, it can be seen that the highest R used (i.e., the lowest NP-5 concentration used) approaches the solubilization curve more closely when concentrated ammonia (29.6 wt% NH3) instead of water is the microemulsion aqueous phase. This means that significant intermicellar interaction (and correspondingly, particle aggregation) will occur at a lower R value when the ammonia concentration is relatively high. Effects of ammonia concentration on particle size. Examination of Fig. 6 reveals that at relatively low R (e.g., R 5 0.5), the particle size goes through a minimum as ammonia concentration is increased. In contrast, at relatively high R (e.g., R 5 5), the particle size increases monotonically with ammonia concentration. At this time, no definitive explana-
SILICA IN WATER-IN-OIL MICROEMULSION
tions can be offered for these trends. However, it is not unreasonable to expect that, as in the case of silica formation in homogeneous alcohol-alkoxide systems (31, 41– 45), the observed particle size dependence on ammonia concentration may be related to the ability of the OH2 ion (generated by the hydrolysis of ammonia) to catalyze both hydrolysis and condensation reactions and the tendency for siloxane bonds to break in highly alkaline solutions (32, 35, 39). As noted above, at low R alkoxide hydrolysis and silica nucleation are inhibited. Raising the ammonia concentration then increases the OH2 concentration which, in turn, increases hydrolysis and nucleation rates. The subsequent increase in the number of nuclei leads to smaller particles. The increase in particle size at relatively high ammonia concentrations may be due to the base-catalyzed depolymerization (32, 35, 39) which decreases the number of stable nuclei. In the case of relatively large R, it was argued above that nucleation should be favorable. Thus, it is likely that the observed increase of particle size with increase in ammonia concentration is associated with growth-related processes. As already noted above, both high R and high ammonia concentration should favor intermicellar interaction. Thus, it is proposed here that at high R, increase in ammonia concentration increases intermicellar exchange rates and therefore particle aggregation. SUMMARY AND CONCLUSIONS
The work presented and discussed here reveals that, in the preparation of nanosize silica particles by controlled hydrolysis of TEOS in the NP-5/cyclohexane/ammonium hydroxide w/o microemulsion system, variations in the water-to-surfactant molar ratio (R) and ammonia concentration can have a significant effect on particle size. Specifically, the particle size goes through a minimum as R is increased. Further, the R value (R min) corresponding to the minimum in particle size becomes smaller as ammonia concentration is increased. In order to relate these variations in particle size to the physicochemical characteristics of the corresponding microemulsion phase, complementary experiments have been conducted to examine the effects of the water-to-surfactant molar ratio and/or ammonia concentration on (a) the stability domain of the single phase microemulsion, (b) the minimum surfactant concentration needed to solubilize the aqueous phase, and (c) the transition from bound to free water in the microemulsion polar region. Phase equilibria investigations show that introduction of ammonia into the NP-5/cyclohexane/water microemulsion system results in a decrease in size of the water-in-oil microemulsion stability domain. On the other hand, ammonia has no significant effect on the transition from bound to free water, which occurs at a water-to-surfactant molar ratio of about 1. The minimum surfactant concentration needed to solubilize the aqueous phase is of the order of 1.1 wt% and this is not significantly affected by ammonia concentration.
219
Given the fact that water is needed for the alkoxide hydrolysis reaction (Eq. [2]), it is reasonable to consider that the microemulsion water pools constitute the locale of the particle formation process. At low R values, the number of water molecules per surfactant aggregate (N w) is low, the number of TEOS molecules per surfactant aggregate (N t) is low, and there is little free water. Accordingly, hydrolysis and silica nucleation are inhibited. The result is that only a few nuclei are produced and these then grow to become large particles. In contrast, when R is relatively high, N w, N t, and the number of free water molecules are also relatively high. Thus, both hydrolysis and nucleation are favored. In principle, this should lead to the production of a large number of nuclei and, eventually, a large number of small particles. Contrary to expectation, however, the particle size does not decrease monotonically with increase in R. It is argued here that the observed increase in particle size with R at relatively high large values is attributable to particle aggregation associated with intermicellar exchange. It is well established that the rate of intermicellar exchange increases with increases in the water-to-surfactant molar ratio (61– 63). Further, the exchange rate increases as the microemulsion composition approaches the solubilization curve (61– 63). Accordingly, it is further suggested that increases in ammonia concentration, by resulting in shrinkage of the microemulsion stability region, should lead to an increase in the intermicellar exchange rate. This should promote particle aggregation and give rise to larger particles. This is exactly what is found experimentally, as demonstrated by the high R region of Fig. 6. It is apparent, however, that the nucleation of silica in the NP-5 reverse micellar system involves a number of factors which have not been considered in the qualitative model discussed above. Further elucidation of the factors which determine the number of particles produced as a function of R at a given ammonia concentration requires a more quantitative approach. Rather than a single reverse micelle, the whole of the micellar population must be taken into account since intermicellar events are occurring. For example, nucleation at low R values is considered difficult in each reverse micelle (see above), but intermicellar nucleation may be important since a large concentration of aggregates is present at low R. Similarly, each reverse micelle is assumed to be capable of producing a nucleus at high R values, but then there are fewer reverse micelles under these conditions. Also, it is possible that only a fraction of the TEOS molecules in the system is associated with the reverse micellar pseudophase; this would reduce significantly the average number of TEOS molecules per micelle as compared to the N t figures reported in Table 1. Work is in progress to develop a statistical model for the nucleation process. The roles of intermicellar events and of the dynamics of the reverse micellar system are considered, as well as the issue of TEOS partition between the reverse micellar pseudophase and the bulk oil phase.
220
ARRIAGADA AND OSSEO-ASARE
REFERENCES 1. Fendler, J. H., Chem. Rev. 87, 877 (1987). 2. Ward, A. J. I., and Friberg, S. E., Mater. Res. Bull. 14, 41 (1989). 3. Osseo-Asare, K., and Arriagada, F. J., in “Ceramic Powder Science III” (G. L. Messing, S. Hirano, and H. Hausner, Eds.), pp. 3–16. Am. Ceramic Soc., Westerville, OH, 1990. 4. Khan-Lodhi, A., Robinson, B. H., Towey, T., Hermann, C., Knoche, W., and Thesing, U., in “The Structure, Dynamics and Equilibrium Properties of Colloidal Systems” (D. M. Bloor and E. Wyn-Jones, Eds.), pp. 373– 383. Kluwer Academic, Boston, MA, 1990. 5. Pileni, M. P., J. Phys. Chem. 97, 6961 (1993). 6. Candau, F., Macromol. Symp. 92, 169 (1995). 7. Osseo-Asare, K., and Arriagada, F. J., Colloids Surf. 50, 321 (1990). 8. Arriagada, F. J., and Osseo-Asare, K., in “Refractory Metals: Extraction, Processing and Applications” (K. Liddell, D. R. Sadoway, and R. G. Bautista, Eds.), pp. 259 –269. TMS-AIME, Warrendale, PA, 1991. 9. Arriagada, F. J., and Osseo-Asare, K., Colloids Surf. 69, 105 (1992). 10. Arriagada, F. J., and Osseo-Asare, K., J. Dispersion Sci. Technol. 15, 59 (1994). 11. Boakye, E., Radovic, L. R., and Osseo-Asare, K., J. Colloid Interface Sci. 163, 120 (1994). 12. Arriagada, F. J., and Osseo-Asare, K., in “Advances in Chemistry Series, No. 234” (H. Bergna, Ed.), pp. 113–128. Am. Chem. Soc., Washington, DC, 1994. 13. Arriagada, F. J., and Osseo-Asare, K., J. Colloid Interface Sci. 170, 8 (1995). 14. GAF Chemicals Corporation, Technical bulletin, Igepal CO nonionic surfactants. 15. Hinze, W., and Fendler, J. H., J. Chem. Soc. Dalton, 238 (1975). 16. Fuestel, D., Runge, F., and Ilgenfritz, G., in “Reactions in Compartmentalized Liquids” (W. Knoche and R. Shoemacker, Eds.), p. 21. SpringerVerlag, Berlin, 1989. 17. Kizling, J., and Stenius, P., J. Colloid Interface Sci. 118, 482 (1987). 18. Aveyard, R., Binks, B. P., and Fletcher, P. D. I., Langmuir 5, 1210 (1989). 19. Ravey, J. C., Buzier, M., and Picot, C., J. Colloid Interface Sci. 97, 9 (1984). 20. Shinoda, K., and Saito, H., J. Colloid Interface Sci. 26, 70 (1968). 21. Kon-no, K., and Kitahara, A., J. Colloid Interface Sci. 34, 221 (1970). 22. Kitahara, A., J. Phys. Chem. 69, 2788 (1965). 23. Sheih, P. S., and Fendler, J. H., J. Chem. Soc. Faraday Trans. I, 73, 1480 (1977). 24. Tanaka, R., J. Colloid Interface Sci. 122, 220 (1988). 25. Christenson, H., Friberg, S. E., and Larsen, D. W., J. Phys. Chem. 84, 3633 (1980). 26. Wilcoxon, J. P., and Williamson, R. L., in “Macromolecular Liquids” (C. R. Safinya, S. A. Safran, and P. A. Pincus, Eds.), p. 269. Mater. Res. Soc., Pittsburgh, 1990. 27. Nagy, J., Colloids Surf. 35, 201 (1989). 28. Kandori, K., Shizuka, N., Kon-no, K., and Kitahara, A., J. Dispersion Sci. Technol. 8, 477 (1987). 29. Guizard, C., et al. in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), Vol. 121, p. 115. Mater. Res. Soc. Symp. Proc., Mater. Res. Soc., Pittsburgh, 1988. 30. Espiard, P., Mark, J. E., and Guyot, A., Polymer Bull. 24, 173 (1990). 31. Stober, W., Fink, A., and Bohn, E., J. Colloid Interface Sci. 26, 62 (1968). 32. Brinker, C. J., and Scherer, G. W., “Sol-Gel Science.” Academic Press, San Diego, 1990.
33. Aelion, R., Loebel, A., and Eirich, F., J. Am. Chem. Soc. 72, 5705 (1950). 34. Keefeer, K. D., in “Better Ceramics through Chemistry” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), p. 15. Elsevier, New York, 1984. 35. Iler, R. K., “The Chemistry of Silica.” Wiley, New York, 1979. 36. Byers, C. H., and Harris, M. T., in “Ultrastructure Processing of Advanced Ceramics” (J. D. Mackenzie and D. R. Ulrich, Eds.), p. 843. Wiley, New York, 1988. 37. Byers, C. H., Harris, M. T., and Williams, D. F., Ind. Eng. Chem. Res. 26, 1916 (1987). 38. Artaki, I., Bradley, M., Zerda, T. W., and Jonas, J., J. Phys. Chem. 89, 4399 (1985). 39. Greenberg, S. A., J. Phys. Chem. 61, 960 (1957). 40. Baes, C. F., and Mesmer, R. E., “The Hydrolysis of Cations.” Wiley, New York, 1976. 41. Van Helden, A. K., Jansen, J. W., and Vrij, A., J. Colloid Interface Sci. 81, 354 (1981). 42. Matsoukas, T., and Gulari, E., J. Colloid Interface Sci. 132, 13 (1989). 43. Bogush, G. H., Tracy, M. A., and Zukoski, C. F., J. Non-Cryst. Solids 104, 95 (1988). 44. Adams, J., Baird, T., Braterman, P. S., Cairns, J. A., and Segal, D. L., in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark and D. R. Ulrich, Eds.), p. 361. Mater. Res. Soc., Pittsburgh, 1988. 45. Badley, R. D., Ford, W. T., McEnroe, F. J., and Assink, R. A., Langmuir 6, 792 (1990). 46. Bagnall, C. M., Howarth, L. G., and James, P. F., J. Non-Cryst. Solids 121, 56 (1990). 47. Bogush, G. H., Dickstein, G. L., Lee, K. C. P., and Zukoski, C. F., in “Better Ceramics through Chemistry III” (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.), p. 57. Mater. Res. Soc., Pittsburgh, 1988. 48. Bogush, G. H., and Zukoski, C. F., J. Colloid Interface Sci. 142, 1 (1991). 49. Brinker, C. J., in “The Colloid Chemistry of Silica, Advances in Chemistry Series” (H. Bergna, Ed.), pp. 361– 402. Am. Chem. Soc., Washington, DC, 1994. 50. Firman, P., Haase, D., Jen, J., Kahlweit, M., and Strey, R., Langmuir 1, 718 (1985). 51. Johnson, K. A., and Shah, D. O., in “Surfactants in Solution” (K. L. Mittal and P. Bothorel, Eds.). Plenum, New York, 1986. 52. Kitahara, A., and Kon-no, K., J. Phys. Chem. 70, 3394 (1966). 53. Gobe, M., Kon-no, K., Kandori, K., and Kitahara, A., J. Colloid Interface Sci. 93, 293 (1983). 54. Becher, P., in “Nonionic Surfactants” (M. J. Schick, Ed.), Chap. 15. Dekker, New York, 1966. 55. Meisel, D., Matheson, M. S., and Rabani, J., J. Am. Chem. Soc. 100, 117 (1978). 56. Kamiya, K., and Yoko, T., J. Mater. Sci. 21, 842 (1986). 57. Arriagada, F. J., and Osseo-Asare, K., in preparation. 58. Kaufman, V. R., and Avnir, D., Langmuir 2, 717 (1986). 59. Matsui, K., and Nakazawa, T., Bull. Chem. Soc. Jpn. 63, 11 (1990). 60. Minehan, W. T., Messing, G. L., and Pantano, C. G., J. Non-Cryst. Solids 108, 163 (1989). 61. Clark, S., Fletcher, P. D. I., and Ye, X., Langmuir 6, 1301 (1990). 62. Almgren, M., Stam, J. V., Swarup, S., and Lofroth, J. E., Langmuir 2, 432 (1986). 63. Bommarius, A. S., Holzwarth, J. F., Wang, D. I. C., and Hatton, T. A., J. Phys. Chem. 94, 7232 (1990).