Synthesis, Polymerization, and Assembly of Nanosilica Particles below the Isoelectric Point.

The particle growth of silica below the isoelectric point plays a key role in oil well cements, production of silica gels and production of nanosilica by dissolving silicates. In this article, we study the particle growth of silica below the isoelectric point using olivine, a silicate mineral, and sodium silicate solutions as silica sources in acid, where the particle size, soluble silica concentration, specific surface area and gelling time were measured. The size of the primary particles detected by laser light scattering is 3 nm in the experiments with sodium silicate solutions. These particles grow then by aggregation forming linear chains which in time will start to branch. The particle growth follows a quadratic polynomial function and particles as large as 100 and 500 nm are detected in the final stages of experiments using sodium silica solutions and olivine, respectively. Based on these findings, a comprehensive model describing the silica particle development below the isoelectric point is proposed. This model gives fundamental information about the growth mechanism and the properties of silica (e.g., particle size of the primary particles, size of the aggregates) at the different growth stages.

Introduction

Colloidal silica and soluble silica are among of the most important materials in nature as well as in industry with a projected market of $8.8 billion in 2020.[1] Silica is soluble in small quantities, and it is present in small amounts in natural waters playing an important role in the life of plants, animals and bacteria. The most important properties of amorphous silica are its particle size (primary and aggregate size), pore size and state (e.g., as gel or as a sol) since they influence the final properties of the material and, therefore, its potential applications. These three parameters depend on the condensation and polymerization reactions. The polymerization of silica above the isoelectric point (pH = 2) has been extensively covered[2−9] and the assembly mechanism is well-known for all its stages. According to Iler,[2] the polymerization of silica above the isoelectric point (IEP) consists of three stages: (1) nucleation of particles forming ring structures; (2) growth of the nuclei forming larger particles; and (3) linking of particles into branched chains and networks. The polymerization of silica proposed by Iler between pH 2 and 10 is summarized in Figure . This figure shows the reaction path to form the primary particles as well as the two possible growth routes depending on the pH and salt concentration, which result either in silica gel or a stable colloidal dispersion.

Figure 1

Polymerization behavior of silica between pH 2 and 10 (from Iler[2]).

The polymerization of silica occurs most of the times in nature and in the industrial processes above the IEP; however, there are other processes where the polymerization takes place below the IEP such as the matrix acidization in oil well cements,[10] the production of silica gels,[11] and the synthesis of nanosilica by the dissolution of silicate minerals.[12−14] The matrix acidization in oil well cements causes pore blockage dramatically reducing the permeability and decreasing the oil production.[15,16] Silica gels below pH 0 are used for the production of batteries, showing the following advantages over traditional batteries: simplified handling and mixing, no liquid separation, improved gel strength, and low cost.[11] The synthesis route of nanosilica using silicate minerals in acidic media has been proved to be more sustainable and with lower production costs than current commercial production processes due to the low reaction temperature and to the exothermic reaction of the silicates dissolution.[12,17] Since the polymerization of silica below the IEP plays an essential role on these processes, having a better understanding is crucial to increase oil production, to obtain improved batteries and to produced sustainable nanosilicas. In spite of the significance of polymerization below the IEP, there are only few studies below the IEP[10,18−22] due to the fast polymerization kinetics of silica, which makes it difficult to study and control the reaction.

Zerda[20] and Kelts[21] studied the initial stages of silica polymerization below the IEP and found that silica grows forming linear oligomers instead of cyclic oligomers as pointed out by Iler.[2] Okkerse[18] synthesized silica-gels neutralizing sodium silicate solutions (SSS) below the IEP and found that the size of the primary particles was influenced by the rate and duration of the condensation reaction. Gorrepati focused on silica polymerization at the intermediate and final stages and with 135 mM of SSS finding that particles grow into clusters of 3 μm in 3 days in 4 M hydrochloric acid. Despite the interesting findings of these authors, further investigation is necessary to obtain a comprehensive understanding of silica particle growth during all the stages.

The aim of this study is to clarify the growth mechanism of silica below the IEP and to propose a comprehensive model to understand and predict the influence of polymerization on the properties of silica during the whole growth process. This paper is structured in three main sections: monitoring particle size at different polymerization stages using sodium silicate and olivine as silica sources; quantification of the soluble silica content; and identifying the gelling point. Based on the results presented here, a silica growth model is proposed.

Materials and Methods

Four types of experiments are reported in this article: (1) particle growth using olivine as silica source, (2) particle growth using SSS as silica source, (3) concentration of soluble silica using olivine and SSS, and (4) gelling time during the dissolution of olivine. The particle growth was determined using either a Malvern ZS (single angle) or an ALV light scattering device (multiple angle). The difference between these two laser devices is that the ALV provides all the information and calculation of the diffusivity coefficients at different measurements angles, while the Malvern ZS directly calculates the particle size measured at 173°. The results with the Malvern ZS are less accurate, especially at low size ranges, but can be used at higher particle concentration because multiple scattering is reduced.[23] The ALV equipment used was an ALV/CGS-3 MD-4 compact goniometer system equipped with a multiple tau digital real time correlator (ALV-7004). Measurements were performed over an angular range between 60 and 120°, performing 3 × 30 s acquisitions at each angle.

Particle Growth Experiments Using Olivine

Silica particle growth during the olivine dissolution was investigated using Norwegian dunite (a rock rich in olivine mineral). The dunite rock was supplied by Sibelco containing 89% olivine and was grounded for 30 min with a planetary ball mill to increase the specific surface area of the mineral, and, therefore, the dissolution kinetics. The particle size distribution of this grounded material (see Figure S1 in the Supporting Information) was in the range 0.7 to 300 μm with an average diameter (d50) of 116 μm and a geometric surface area of 0.09 m2/g. Once this granular material is brought together with the acid, olivine is dissolved releasing silica into the environment.

The dissolution experiments were carried out at 20 °C in a vigorously stirred reactor of two liters with sulfuric acid. The experimental variables were the concentration of acid and the olivine mass (see Table ). The concentration of hydrogen ions was determined by titration with a 0.05 M Na2B4O7 solution, from which the amount of silica released was calculated. The silica particle size was measured with the Malvern ZS after passing the samples through a filter of 1 μm to avoid the presence of large olivine particles.

Table 1

Conditions of the Silica Polymerization Experiments Using Olivinea

Exp.	mol (g/L)	[Si]eq (mM)	[H+] (M)	T (°C)
PO-1	50	307	2	20
PO-2	50	307	4	20
PO-3	100	613	4	20
PO-4	70	429	4	20

PO refers to polymerization experiments, mol to olivine mass concentration, and [Si]eq to the equivalent concentration of silica in the olivine.

Particle Growth Experiments Using Sodium Silicate Solutions

A sodium silicate solution (SSS) was used instead of olivine because it is supposed to be particle free, making simpler and more reliable the particle growth study by laser light scattering, especially in the low size range (below 20 nm). A 0.36 M solution was prepared using SSS supplied by Sigma-Aldrich (reagent grade) with a SiO2 and Na2O content of 26.5% and 10.6% (m/v), respectively, resulting in a solution with a pH of 11.5. Additional NaOH was added up to a pH of 12.4 resulting in a SiO2/Na2O ratio of 1.9. Surprisingly, particles were detected in this solution, which was unexpected since SSS is supposed to be completely soluble at this pH.[2] The solution was then centrifuged at 4700 rpm for 45 min and passed through a 200 nm filter. This procedure had to be carried out twice. After performing this cleaning procedure, almost no particles were found, and if so, the intensity was extremely low and the particle size detected was below 1 nm. The syringe pump was adjusted to 0.05 mL/min, and the reactor was filled with 0.5 L of 3 M sulfuric acid at 20 °C. This experiment (PO-5) was duplicated, showing the good reproducibility when the SSS does not have (or has only a few small) silica particles.

Reactive Molybdate Silica

The concentration of soluble silica in solution was measured photometrically, following the molybdosilicate method.[24,25] The equipment used was a USB2000+ UV–vis spectrophotometer (Ocean Optics) and a light source of wavelength range 395–750 nm (BluLoop Multi-LED, Ocean Optics). Ammonium molybdate at pH around 1.2 reacts with monomeric silica to produce silicomolybdic acids. However, the amount of monomeric silica can vary with time due to polymerization/depolymerization reaction so that silicic acid may still be consumed/formed. At this pH, monomeric silica as well as oligomers that are small enough to be fully depolymerized during the analysis time react with the molybdate.[25] The resulting silicomolybdic acid is a yellow compound that is titrated spectrophotometrically by monitoring the optical density at λ = 400. A more detailed study of the silica subspecies (i.e., monomers, dimers, oligomers, etc.), as was proposed by Belton,[26] even if it could provide valuable information (specially before the primary particles are formed) was not conducted here because the focus of this paper is to understand how particles physically grow to form large agglomerates.

The concentration of soluble silica was determined for experiments using olivine (experiments PO-3 and PO-4) and using SSS (experiment PO-5). In addition, the polymeric silica also known as non molybdate reactive silica (%) and the total silica (mM) were also determined. The total silica is calculated using the kinetic model of the olivine dissolution described in[27] or a silica mass balance in the case of SSS; the polymeric silica is defined as the difference between the total and the soluble silica.

Gelling Experiments

The gelling experiments were performed using a Coesfeld Visco-Mix device coupled with a reactor stirrer to register the voltage, which is related to the torque of the slurry during the dissolution of olivine and in turn to its viscosity. Table lists the experimental conditions of the gelling experiments and the gelling time, concentration of the hydrogen ion at the gelling point (GP) and the estimated amount of silica released from the olivine at the GP. The amount of silica was estimated using our kinetic model about the dissolution of olivine.[27] The ζ potential is also an influencing parameter on the silica gelling, which in turns depends on the pH and ionic strength. The ζ potential of the slurries cannot be measured due to the extremely high conductivity at such acidic conditions. However, it is known that the ζ potential is positive below the IEP, and the more distant from the IEP, the higher the ζ potential. Thus, the influence of ζ potential on the GP was investigated in TO-7, where experiment TO-1 was repeated and 0.8 mol of concentrated sulfuric acid 5 min before the GP was added.

Table 2

Experimental Conditions of the Gelling Experimentsa

name	mol (g/L)	[H]i (M)	T (°C)	time (min)	[H]GP (M)	[Si]GP (mM)
TO-1	120	6.0	90	20	3.7	599
TO-2	120	8.3	90	15	5.5	709
TO-3	120	6.0	71	53	4.0	424
TO-4	120	6.0	51	175	4.8	304
TO-5	120	3.8	90	52	1.6	556
TO-6	80	3.8	90	66	1.8	491
TO-7*	120	6.0	90	19	5.0	673

mol refers to the mass of olivine, [H]i and [H]GP the hydrogen ions concentration at the initial time and at the GP, and [Si]GP the concentration of silica at the GP. *The initial concentration of H2SO4 in TO-7 was 3 M and 5 min before the GP 0.8 mol of concentrated sulfuric acid was added.

Results and Discussion

Particle Growth Using Olivine as Silica Source

The particle size development of silica synthesized from the dissolution of olivine was analyzed with a laser light scattering technique. Olivine is dissolved in acid releasing silica to the medium according toresulting in a mixture of magnesium/iron salts, amorphous silica, unreacted olivine, and inert minerals. The total concentration of silica produced in experiments PO-1, PO-2, and PO-3 was stoichiometrically calculated from the amount of hydrogen ions neutralized (see Figure a). The higher the acid concentration, the faster the release of silica, as can be observed from experiments PO-1 and PO-2 with the same olivine amount but different [H+]. The higher the olivine content the faster the release of silica as can be observed from experiments PO-2 and PO-3 with the same [H+]. These results are in agreement with the kinetics of the dissolution of olivine in acidic conditions.[27] PO-4 experiment shows kinetics between PO-1 (the slowest experiment) and PO-3 (the fastest).

Figure 2

(A) Silica released from the dissolution of olivine. (B) Silica particle growth. Points were fitted to a quadratic (a) and exponential (b) equation for the sake of visibility.

The particle size distribution of the resulting mixture measured by laser light scattering techniques was considerably polydisperse showing either several independent peaks or very broad peaks, even after passing the sample through a 1 μm filter. In addition, there is no possibility of clearly distinguishing if certain peaks were related to nanosilica or to olivine grains. We excluded the particles larger than 200 nm at the initial stages of the experiment, since these particles could not be silica. The error related to the average particle sizes were in the range 10 to 30%, showing that these results can only be considered as an indication to observe the general trend of the particle growth of silica. The average particle size of silica for experiments PO-1, PO-2, and PO-3 is shown in Figure b. The silica particles grow exponentially with time, even though the rate of silica release from the olivine decreases with time. The smallest particles detected at the beginning of the experiments had a size of 20 nm and the largest one at the final stages of the experiments had a size around 500 nm. The growth of the silica particles presents a trend similar to the one found in Gorrepati’s experiments using SSS,[10] but with a slower particle growth rate. The reason for that is because, in the case of Gorrepati, all the silica was available from the beginning of the experiment, while in our experiments the silica was released slowly over time as olivine dissolves.

The concentration of soluble silica at early stages of experiments with 100 g (PO-3) and 70 g (PO-4) of olivine was measured using the molybdate method (Figure a). The highest concentration of soluble silica detected was 4.7 mM after 15 min of reaction, decreasing until a constant value around 1 mM afterward. This constant value corresponds to the solubility of silica.[28] Around the time when solubility values were reached, silica particles of 55 nm (at 1260 min) and 19 nm (at 1345 min) were detected for experiments with 100 and 70 g of olivine, respectively. At 15 min (time when the maximum soluble silica is detected) the amount of nonsoluble silica (i.e., polymeric silica) is 43%, and it increases above 99% after 24 h (see Figure b).

Figure 3

Solubility of silica. (A) Concentration of soluble silica in the solution at early reaction times. (B) Soluble and polymeric silica and total silica concentration for PO-3. Lines are drawn for visual reasons.

The silica primary particles observed with transmission electron microscopy (TEM; see Figure ) have an average diameter of 4 ± 1 nm. Figure b shows a cluster of aggregates of the same sample, the size of the aggregates being 21 ± 4 nm. These aggregates likewise form larger clusters in the range 0.1 to 20 μm.[29] It is noteworthy that, in spite of the fact that the silica of this experiment grew forming large clusters, the identity of primary particles can still be observed. Larger TEM pictures and particle size distributions of the primary particles are presented in Figures S2–S7 in the Supporting Information.

Figure 4

TEM pictures of nanosilica produced below the IEP. (A) Shows the primary particles and (B) shows the aggregates of olivine nanosilica (exp. NS-2[13]).

Particle Growth Using Sodium Silicate Solutions As Silica Source

The size of silica synthesized using sodium silicate was investigated by pumping SSS into sulfuric acid, where the monomeric silica starts to polymerize. The silica particle size of PO-5 is presented in Figure a. Figure b illustrates the particle size and the concentration of total silica of this experiment during the first 24 h, when nucleation takes place. The silica particle size follows a quadratic polynomial function with R2 of 99%, showing that the kinetics speeds up with time. The first colloidal particles observed by dynamic light scattering (DLS) are around 3 nm, which are comparable to the values of 4 nm measured by TEM, which in turn are also in line with the values obtained by others researchers.[2,3,10,18,30−35] The measured particles by TEM are slightly larger because the samples were taken at the end of the experiment, having more time for the particles to age.

Figure 5

Silica particle growth of experiment PO-5. (A) Particle size during the first 60 h. (B) Particle size and total silica concentration during the first 24 h. Dh refers to the hydrodynamic diameter measured with the light scattering device.

According to the literature, the nucleation should take place at silica concentrations between 3 and 16 mM.[28,36,37] These concentrations are reached between the second and eighth hour of the experiment; however, 15 h passed until 3 nm particles were measured at a silica concentration of 30 mM. This time difference could be because a certain amount of time is required between the nucleation of the first silica oligomers (which cannot be detected by DLS) and the formation of detectable 3 nm colloids. As already described in a previous investigation by Carcouët,[38] primary particles of 2 nm remain stable due to electrostatic forces until a critical aggregation concentration is reached and the repulsive forces between the particles are overcome. At this point, the primary particles collapse, becoming denser, and form larger aggregates of 5 nm. The findings of Carcouët[38] can explain the long time until the first particles are detected in our experiments.

Table presents the numbers of peaks detected and the intensity of the light scattering measurements for certain samples of experiment PO-5. The light intensity curve of this experiment is showed in Figure , where only the first 24 h samples are plotted because, after that time, the solutions were diluted. In the first 15 h the light intensity was so low that we assumed that no particles were present in the solution; then the intensity increased and the first particles were found to show a polydisperse distribution; After 22 h, the system became monomodal and the light intensity increased. The initial polydispersity can be explained based on the random collision of particles, forming larger aggregates of different sizes. The reason for the monodispersity after 22 h is not clear. Random collision still is likely to take place, but another phenomenon must also be occurring. There could be a rearrangement of the aggregates in order to obtain higher dense particles and a more stable structure, similar to the Ostwald ripening process, which makes silica colloids monomodal.

Table 3

Reaction Time, Number of Peaks and Light Intensity of PO-5 Samplesa

samples	t (h)	nPeaks	light intensity (kHz)
0	0	-	1.5–3
1	14.2	-	2–3.5
4	17.3	2–4	3–8
8	23.5	1	15–25
10	39.6	1	90–150
14	46.5	1	150–270

Samples 10 and 14 were diluted by a factor of 10 before measuring.

Figure 6

Light intensity curve of experiment PO-5.

Figure a shows the solubility of silica of experiment PO-5 determined by the molybdate method. Figure b presents the soluble silica (%), the polymeric silica (%), and the total silica (mM) for this experiment. The maximum solubility value detected in PO-5 is 3.8, which is similar to the values obtained with olivine (see Figure a). After the maximum, the concentration of soluble silica decreases, even though silica is still released into the reactor. Therefore, the silica release rate must be slower than the incorporation of silica monomers to the particle aggregates. The solubility value after 2 days is around 2.3 mM, being higher than the values obtained with olivine (∼1 mM). Even considering the high errors associated with the molybdate method, the difference between the solubilities obtained using olivine and SSS is significant. This difference could be due to the diverse way the silica has been added to the system. In the experiment with SSS, the silica addition is constant throughout the experiment, while in the case of olivine, it decreases exponentially according to its dissolution kinetics.

Figure 7

(A) Silica solubility. (B) Soluble and polymeric silica and concentration of total silica for PO-5.

When the total silica concentration in the reactor was 3.8 mM (see experiment PO-5 at 247 min), 55% of the silica was already in the polymeric form, which means that condensation of silica has already taken place. Possible explanations for this are (1) the supersaturation point for this silica at these conditions is 3 mM; (2) the molybdate method has a significant error, especially at lower concentrations, and underestimates the amount of soluble silica; and (3) some small colloids (e.g., 0.5–2 nm) are initially present in the SSS; thus, condensation could be favored at lower concentrations

Gelling during the Dissolution of Olivine

The gelling point of silica slurries produced from the olivine dissolution was identified monitoring the voltage of the stirrer. Gelling takes place when small silica colloids link together by van der Waals forces into branched chains to form 3D networks. The gel structure can be broken by stirring, allowing the silica particles to polymerize and grow. Once the particles are larger than 100 nm, they do not usually gel anymore.[2] The purposes of the gelling experiments are to know if and when gelling takes place and to observe the influence of the main variables on the GP.

The GP depends on pH, particle size, concentration of silica colloids, ionic strength, and temperature.[2,39] A typical voltage consumption curve is given in Figure a. In this curve, different sections can be observed: (1) slow increase in voltage, (2) fast increase; (3) stable voltage; and (4) slow decrease. The GP is determined here as the point at the end of section 2, which is related to an in increase in the viscosity of the slurry due to the van der Waals forces. The disappearing of gelling takes place when the viscosity decreases (section 4), which is related to the growth of silica into colloids probably as large as 100 nm. Figure b shows the silica concentration released by the dissolution of olivine versus the gelling time. The gelling time is reduced with the silica concentration, being in the range from 300 to 700 mM. Figures a and b present the influence of the sulfuric acid concentration and temperature on the gelling time, respectively. From this figure, it can be observed that the gelling time decreases with the sulfuric acid concentration and with the reaction temperature, which is in agreement with the theory of Iler.[2] In addition, the acid concentration and temperature affect positively the silica released from the olivine, which in turn accelerates the gelling. By comparing TO-1 and TO-7, it is clear that an additional amount of sulfuric acid 5 min before the gelling (resulting in a higher ζ potential in TO-7) delayed the gelling time only shortly. Therefore, it can be concluded that the increase in the ζ potential due to this pH change plays a minor role compared to the silica concentration.

Figure 8

(A) Voltage consumed of the stirrer over time. (B) Silica concentration versus gelling time.

Figure 9

Gelling point. (A) At 90 °C and different sulfuric acid concentrations. (B) With 3M sulfuric acid and at different temperatures. *The initial concentration of H2SO4 in TO-7 was 3 M and 5 min before the GP 0.8 mol of concentrated sulfuric acid was added.

Particle Growth Model

A comprehensive growth model for silica particles below the IEP is proposed here, which makes it possible to explain the phenomena involved during all the stages of the particle growth and to predict the morphological properties of these particles.

At time zero, the solution is free of silica. Once we start the experiment, the concentration of silica increases in our system (in both cases the dissolution of olivine and the SSS). After a certain time, the solution becomes supersaturated in silica and nucleation occurs. The concentration of nuclei increases until a critical aggregation concentration (CAC) is reached, at which time the first 3 nm aggregates are formed. In our experiments, that happened at silica concentrations of 30 mM.

Once these first detectable primary particles are formed, the growth theory can be classified based on two main mechanisms: growth by condensation of soluble silica or by aggregation of colloidal particles. Gorrepati[10] stated that, under the experimental conditions of negative pH and a SSS concentration of 135 mM, aggregation is the ruling mechanism. The condensation was discarded as the growth mechanism because the soluble silica decreased from 135 mM to close to solubility values in the first minutes of the experiments, but no particle formation was observed at these times. Therefore, Gorrepati[10] concluded that the presence of the first detectable particles cannot be related to the condensation of soluble silica, but to the aggregation of already formed silica particles. The case studied here is different since silica is continuously brought into the reactor. Thus, there is always soluble silica available in the reactor, and it should condense onto already present nuclei, as this is more thermodynamically favorable than creating new nuclei. The condensation should take place at the lowest solubility point,[2] which is in the junction between particles (see Figure ).

Figure 10

Schematic representation of the condensation in the junction between particles.

To better understand the growth mechanism, estimations of the particle size can be performed based on the following assumptions. Considering that at 30 mM of sodium silicate, 3 nm particles are formed, it can be calculated that the total number of primary particles is 5.76 × 1019 (see Table S1 in the Supporting Information). If these particles grew only by condensation, the number of particles would be constant, while their size would increase from 3 to 4.7 nm for a concentration of 100 mM, which is the concentration at the end of the experiments with SSS (see Table S2 in the Supporting Information). Since the particles grow up to 120 nm at the end of the experiment, condensation can be ruled out as the main growth mechanism. In addition, if fully condensed spherical particles were formed of the sizes determined by the DLS (see Figure a), then the specific surface area would decrease from 905 to 12 m2/g over the experiment (see Table S3 in the Supporting Information). Furthermore, we know from our previous studies[29] that the BET specific surface area of silica at the end of the olivine dissolution is normally in the range 300–400 m2/g and that the cluster size is in the range 0.1 to 20 μm. Summarizing, the following facts support aggregation as the main growth mechanism: particle development accelerates with time; particles observed by TEM are not spherical (typical for condensation); and still recognizable primary particles are present in the aggregates observed by TEM. Therefore, we conclude that aggregation is the dominant mechanism and condensation plays a minor role below the IEP.

Once the primary particles of 3 nm are shaped, linear chains should be formed, which is supported by the fact that the aggregation is the main growth mechanism as well as that the lowest repulsion force between chain particles is at the sides[2,20] producing longer chains. Once the chains are long enough, so the repulsion forces can be overcome, branching starts to happen.[20] The larger the particles grow, the more compact they become having a grape-cluster morphology as can be seen from the TEM pictures (see Figure ) and from the decrease of the BJH pore volume and the SSABET with the reaction conversion (see Table S4 in the Supporting Information). This is due to a better packing of the particles, which is in agreement with the findings of Okkerse,[18] who stated that a better packing was obtained for silicas gels after longer gelling times. Even if the pore volume is substantially reduced with the reaction conversion, nanosilicas synthesized under these conditions at the end of the reaction are still highly porous, showing SSABET above 300 m2/g.

Considering all these data, the silica growth below the IEP could be summarized as presented schematically in Figure . Initially, soluble silica nucleates and primary particles around 3 nm are formed after the CAC is reached. The silica particles grow mainly via an aggregation route forming linear chains. At a certain point, and depending on the silica concentration and the stirring action among other factors, gelling could take place. As the growth continues, the silica chains keep increasing in size, and at a certain moment, branching starts to occur forming 3D network aggregates. In addition, gelling disappears once the particles grow in size (around 100 nm)[2]. Over time, the aggregates become larger and slightly more compact, but they still maintain most of its porosity, its 3D network structure, and recognizable primary particles. The final particles can reach particle sizes above 1 μm.

Figure 11

Particle growth model of silica below the IEP.

Conclusion

We have proposed a comprehensive model for the silica particle growth below the isoelectric point and with a continuous release of silica. Particle growth experiments were performed using olivine mineral and sodium silicate solutions as silica sources and the particle sizes were determined using laser light scattering techniques. The concentration of soluble silica in these experiments was also analyzed using the molybdate method. In addition, the gelling time was determined by measuring the voltage of the mechanical stirrer.

We have found that the first primary silica particles observed by DLS have a diameter of 3 nm, which appeared at concentrations of 30 mM in the experiments with SSS. The considerably high concentration until the primary particles were detected is probably because a critical aggregation concentration must be reached. Individual primary particles of 4 nm in the final aggregates produced from the olivine dissolution can still be observed in the TEM pictures. We have also demonstrated that the particle growth of silica is mainly ruled by aggregation and that can be fitted to a quadratic polynomial function with a R2 of 99%. Particles as large as 100 and 500 nm were detected at the end of the experiments using SSS and olivine, respectively. Furthermore, clusters in the microrange can be developed when larger quantities of silica are released into the medium.

Based on these findings, we propose a comprehensive particle growth model encompassing from the initial stages (i.e., condensation of monomeric silica) to the final stages (i.e., formation of agglomerates as large as 1 μm). This model gives essential information about the growth mechanism and the properties of silica (e.g., particle size of the primary particles, size of the aggregates) at the different growth stages. The model is based on our results using olivine and SSS, but it could also be used under any other circumstances where silica is released into the medium below the IEP.