Synthesis of Monodisperse Silica Particles by Controlled Regrowth

Abstract

The development of a simple and reproducible method for the synthesis of monodisperse silica particles is of considerable interest from the point of view of their numerous applications in photonics, biosensing, and biomedicine. When using the well-known Stober method, there is a continuous formation and growth of seeds, which leads to the synthesis of polydisperse colloids. In this work, we used the method of successive growth of silica particles obtained by hydrolytic condensation of tetraethylorthosilicate in an alcoholic-aqueous medium using an alkaline catalyst. It is shown that this technique makes it possible to obtain colloids with a particle size from 50 nm to 3 μm and a standard deviation of less than 5%. An additional advantage of the developed method of stepwise growth is the possibility to include fluorophores and SERS tags into the silica matrix.

INTRODUCTION

Silica (SiO₂) nanoparticles are of considerable interest for many applications, including photonics [1], biomedicine [2], biosensing [3], catalysis [4] and many others [5, 6], due to the ease of synthesis, the ability to obtain particles in a wide range of sizes, low toxicity and high chemical stability. Accurate control over sample size and polydispersity is essential for various promising applications. For example, to obtain close-packed two-dimensional colloidal crystals that are used as components of gas detectors [7] or photonic devices [8], only colloids with a polydispersity of less than 5% are suitable, and the position of the maximum light reflection in such a crystal is determined by the particle size according to the Bragg law [9]. The polydispersity of particles is a significant obstacle to obtaining high-quality synthetic opals [10] and gold nanoshells on silica cores [11].

In 1968, Stober et al. published a method for the preparation of silicate nanoparticles by hydrolytic condensation of silicate precursors in alcohol medium using ammonia as a catalyst [12]. The method consisted of simply mixing the reactants at room temperature and reacting for 2–3 h with constant stirring. The particle size in the original work ranged from tens to hundreds of nanometers and was controlled by changing the concentration of ammonia, the type of alcohol (methanol, ethanol, isopropanol), as well as the type of silicate precursor. The simplicity of the Stober method led to its widespread adoption by numerous scientific groups over the course of fifty years. Many researchers successfully utilized modified versions of the method to synthesize SiO₂ nanoparticles in their laboratories, resulting in over twelve thousand citations for the original article. To date, there is a consensus that the most simple and reproducible is the modification of the method using ethanol as a solvent and tetraethylorthosilicate as a precursor [13]. The size control here is carried out by varying the concentration of tetraethylorthosilicate, ammonia and water. The two-stage hydrolytic condensation reaction is as follows:

$${\text{Si}}{{\left( {{{{\text{C}}}_{{\text{2}}}}{{{\text{H}}}_{{\text{5}}}}{\text{O}}} \right)}_{{\text{4}}}}{\text{ + 4}}{{{\text{H}}}_{{\text{2}}}}{\text{O}} \to {\text{Si}}{{\left( {{\text{OH}}} \right)}_{{\text{4}}}}{\text{ + 4}}{{{\text{C}}}_{{\text{2}}}}{{{\text{H}}}_{{\text{5}}}}{\text{OH}},$$

(1)

$${\text{Si(OH}}{{)}_{4}} \to {\text{Si}}{{{\text{O}}}_{2}} \downarrow + \,\,{\text{2}}{{{\text{H}}}_{{\text{2}}}}{\text{O}}.$$

(2)

Despite the simplicity and reproducibility of the reaction, the Stober method has a number of limitations. First, it is impossible to obtain monodisperse colloid with a particle diameter of less than 50 nm (some authors indicate a lower threshold of 70 nm [14]). It should be noted that other methods are more often used for the synthesis of such small particles, such as synthesis in micelles and microemulsions [15, 16] or hydrolysis in the presence of amino acids in an aqueous medium [17, 18], the synthesis of porous silicate nanoparticles with the addition of surfactants is also common [19]. However, these approaches are inferior in simplicity and reproducibility to the Stober method. Secondly, relatively monodisperse colloids (<5%) are obtained only for particle sizes of 120–500 nm [20, 21], while for smaller and larger particles, polydispersity reaches 20% [20]. This is due to the continuous formation and growth of seed particles during synthesis, and the addition of too little or too much tetraethylorthosilicate to the reaction mixture upsets the balance between the rates of hydrolysis and condensation. Over the past 50 years, many attempts have been made to build reaction kinetic curves and obtain optimal ratios of components to obtain monodisperse colloids.

Many different parameters have been investigated, including the concentrations of the reagents, the rate and order of their addition, the reaction temperature, the stirring rate, and many others [5]. Particularly noteworthy is the work of Bogush and Zukowski [22], who proposed a process of multistage growth of particles, which consists in dividing the total required amount of TEOS into portions and adding them step by step to the reaction medium after the completion of the previous growth stage. Subsequently, the method of multistage growth of particles to obtain particles of a given size was widely used by many researchers [20, 23]. Despite this, there is still no single universal approach that allows one to obtain monodisperse colloids in a wide range of sizes.

One of the possible solutions to the problem of polydispersity can be the use of a multi-stage protocol, in which the particles obtained at a certain stage are used as condensation centers (“seeds”) for the growth of larger particles. The main tasks in the implementation of this approach are the production of small (less than 50 nm) monodisperse seeds for the first stage of growing and further control of the absence of formation of secondary seeds. This methodology was previously used to obtain silica nanoparticles with diameters from 20 to 200 nm by hydrolysis of tetraethylorthosilicate in an aqueous medium in the presence of l-arginine [21]. Previously, the authors of [21] and we in [24] showed that these particles can be used as seeds for growing into larger particles by the modified Stober method. In the present work, we demonstrate that the multistage growth technique can be used to obtain silica particles with a controlled size in the range of 50–3000 nm and a polydispersity of less than 5%. Additional advantages of the developed method of stepwise growth are the absence of the need to use absolute ethanol and the possibility of including functional molecules and nanoparticles, for example, fluorophores and SERS tags, into the silicate matrix.

EXPERIMENTAL

Materials

The following reagents were used. Ammonia (30% aqueous solution, CAS number: 1336-21-6), fluorescein isothiocyanate (FITC, >97%), 4-nitrothiophenol (97%), sodium citrate (>99%), chloroauric acid (99.99%), polyacrylic acid (mol. wt. 1800, CAS number: 9003-01-4), polyvinylpyrrolidone (mol. wt. 10000, CAS number: 9003-39-8) were obtained from Sigma-Aldrich. Tetraethyl orthosilicate (TEOS, 98%) was obtained from Maclin. We used 96% ethanol (Kirov BioKhim, Russia) and deionized water purified using an UVOI-MF unit (Medianа filter, Russia).

Synthesis of 30–50 nm Silica “Seeds”

To obtain 50 nm silica particles, 2.5 mg of polyacrylic acid was dissolved in 30 mL of ethanol. Then, 1.5 mL of ammonia solution was added to this solution (hereinafter, a 30% solution without dilution was used). Then, using a syringe pump, 0.75 mL of tetraethylorthosilicate was added dropwise over 4 h at room temperature with constant stirring (700 rpm). After completion of the addition of TEOS, the reaction was continued with stirring for another 2 h. The obtained particles were used further without further purification. For the preparation of silicate particles with an average size of 30 nm, the protocol was exactly the same, except that the amount of added polyacrylic acid was 5 mg.

Synthesis of Silica Particles with a Diameter from 80 to 3000 nm

The developed procedure for the synthesis of silica nanoparticles is schematically shown in Fig. 1. Each step of the synthesis was as follows. In a flask with a volume of 50 mL, 24 (or 18 for growing particles larger than 1 µm) mL of ethanol, 2 mL of water, 1.5 mL of ammonia solution, 6 (or 12 for growing particles larger than 1 µm) mL of seeds were poured. The colloid was placed on a magnetic stirrer at 700 rpm. Then, 0.75 mL of tetraethylorthosilicate was added dropwise over 4 h using a syringe pump. After completion of the addition of TEOS, the reaction was continued with stirring for another 2 h until complete condensation of the precursor. This process can be observed visually by stopping the turbidity of the colloid. The obtained particles were centrifuged at 10 000–200 g depending on the particle size and resuspended in 30 mL of ethanol using ultrasound. The centrifugation/resuspension procedure was repeated 6 times to completely wash out the reaction products and small secondary particles. The colloid obtained at each stage was used as “seeds” in the next step of growing. The synthesis was considered complete when the required particle size was reached.

Fig. 1.

Schematic representation of the process of synthesis of silica nanoparticles by the method of multistage gro-wing.

Silica Nanoparticles with Embedded Fluorescent Molecules

Fluorescein isothiocyanate was used as a model fluorophore. At the first stage, a silanized dye precursor was obtained by mixing 1 mg of FITC with 5 mg of 3-aminopropyltrimethoxysilane in 2 mL of ethanol. The solution was incubated for 24 h. At the second stage, 24 mL of ethanol, 2 mL of water, 1.5 mL of an ammonia solution, and 6 mL of a colloid of “germ” particles with a diameter of 302 ± 12 nm and a silica concentration of 8 mg/mL were poured into a 50 mL flask. The colloid was placed on a magnetic stirrer at 700 rpm. Then, 30 µL of TEOS and 5 µL of a silanized fluorophore solution were added every 12 min; 10 additions were made. The obtained particles were centrifuged at 5000 g and resuspended in 30 mL of water using ultrasound.

Silica Nanoparticles with Embedded SERS Tags

At the first stage, 13 nm gold nanoparticles were obtained by the Frens method [25]. To this end, 100 mL of a 0.01% aqueous solution of chloroauric acid was heated in a flask with a water reflux condenser to 100°С with constant stirring on a magnetic stirrer (700 rpm). Next, 3 mL of 1% sodium citrate solution was added to the resulting boiling solution. The gold reduction reaction can be visually observed by the color change of the colloid from pale yellow to bright red. At the second stage, 3-aminopropyltrimethoxysilane (final concentration 1%) was added to the colloid of silica particles. Nanoparticles were incubated for 1 h at room temperature. The obtained aminated particles were centrifuged at 5000 g for 10 min and resuspended in 30 mL of water using ultrasound. In the third step, 6 mL of silica colloid was added to 100 mL of 13 nm gold nanoparticles. Due to electrostatic interaction, negatively charged gold particles were adsorbed onto the surface of positively charged aminated silicate particles. It is important to use at least a 10-fold excess of gold nanoparticles to prevent colloid aggregation. The mixture was incubated for 1 h with stirring at 700 rpm. For additional stabilization, polyvinylpyrrolidone was added to a final concentration of 1 mg/mL. Next, the particles were centrifuged with 4000 g for 10 min and resuspended in 6 mL of ethanol. 20 μL of 2 mM 4‑nitrothiophenol alcohol solution was added and colloid was incubated for 1 h. Due to the formation of Au-S, the reporter molecules bind to the surface of the gold nanoparticle. The resulting colloid, consisting of silica cores with gold nanoparticles and 4-thionitrophenol molecules on the surface, was used as “seeds” for further growth cycles as described previously.

Measurements

Transmission electron microscopy (TEM) was used as the main method for determining the size and polydispersity of nanoparticles. Among the advantages of this technique, it should be noted the possibility of simultaneously determining the size and polydispersity, as well as, in many cases, visualization of the particle structure (for example, for objects of the core/shell type). The well-known disadvantages of TEM, associated with laboriousness, a limited sample of analyzed particles, and aggregation during drying, are not so relevant for our type of samples.

Silica nanoparticles were examined using a Libra 120 electron microscope (Carl Zeiss, Germany) with an accelerating voltage of 120 kV at the Symbiosis Center for Collective Use at the Institute of Biochemistry and Physiology of Plants and Microorganisms, Saratov Scientific Center, Russian Academy of Sciences. Copper grids coated with a formvar film were used as substrates. 3 µL of an ethanol dispersion of nanoparticles was applied to the substrate and kept for 30 min.

The concentration of nanoparticles was determined gravimetrically by weighing the dry residue after colloid lyophilization on an analytical balance. The total weight of particles in 1 mL of colloid, together with TEM data, makes it possible to unambiguously determine the number concentration of particles using the formula:

$$N = \frac{{{\text{3}}m}}{{{\text{4}}\pi \rho {{R}^{{\text{3}}}}}},$$

(3)

where m is the mass of silica, expressed in mg, $\rho$ is the density of silica, expressed in mg/cm³, R is the particle radius, expressed in cm.

The extinction spectra of the samples were measured using a Specord S300 spectrophotometer (Analytik Jena, Germany) in the wavelength range of 350–800 nm using cuvettes with an optical path length of 1 cm. The initial colloids were resuspended in a tenfold volume of ethanol.

The hydrodynamic size of nanoparticles was measured using a Zetasizer ZS setup (Malvern). The measurements were carried out in four-sided plastic cuvettes, the particles were dispersed in ethanol, the optical density of the colloid at a wavelength of 633 nm was about 0.1. We used the standard procedure for accumulating and processing the correlation function proposed by the setup manufacturer.

The fluorescence spectra of nanoparticle samples with incorporated fluorophores were measured using a Cary Eclipse spectrofluorimeter (USA) in 1 cm quadrilateral quartz cuvettes under excitation at a wavelength of 488 nm. To reduce the inner filter effect [26], the initial colloids were diluted 40 times.

The SERS spectra of nanoparticle samples with incorporated SERS tags were measured using a PeakSeeker Pro setup (Ocean optics, USA) in 1 cm four-sided quartz cuvettes under excitation at a wavelength of 785 nm. The laser radiation power was 30 mW, the signal accumulation time was 10 s. To reduce the inner filter effect, the laser beam was focused near the cell wall.

RESULTS AND DISCUSSION

For the synthesis of silicate “seed” particles with sizes in the range of 30–50 nm, we used a modified Stober method with the addition of polyacrylic acid (PAA) to the reaction mixture. In the presence of ammonia, PAA molecules agglomerate to form an ethanol-insoluble PAA–NH₃ complex [27]. This complex serves as a seed for further silica growth. This, together with the dropwise addition of tetraethylorthosilicate over 4 h, results in the absence of secondary seeds and the synthesis of a monodisperse colloid. The importance of the gradual addition of TEOS to obtain monodisperse colloids was previously reported by Bogush et al. [22]. A detailed study of the effect of the rate of addition of precursors was later also carried out in [23]. The simultaneous introduction of a large amount of TEOS leads to a significant excess of the critical concentration of hydrolysis products and, as a result, to the multimodality of particle size distribution and their aggregation.

Figure 2a shows an electron microscopic image of silicate nanoparticles obtained using addition of 2.5 mg of PAA to the reaction mixture. The particles have a quasi-spherical shape. The diameter distribution histogram is shown in the inset in Fig. 2a. The average particle size is 48 nm, and the standard deviation of the diameters is 2 nm. Here and further in the article, the electron microscopic determination of the average size and standard deviation was performed when counting 200 nanoparticles in images. The fundamentally important role of PAA in the formation of a relatively monodisperse colloid with a particle diameter of less than 70 nm should be noted. For example, under all synthesis conditions except for the addition of PAA, a polydisperse colloid is formed with an average particle diameter of 170 nm (see Fig. 7 in theAppendix). An increase in PAA concentration leads to an increase in condensation centers and, consequently, to a decrease in the average size of silicate particles. For example, Fig. 2b shows an electron microscope image of particles obtained by adding 5 mg of PAA to the reaction mixture. The quasi-spherical shape of silica is preserved, the average diameter decreases to 29 nm. A further increase in the amount of added PAA to 10 mg leads to the formation of small particles of various shapes with an average size of less than 20 nm. Presumably, these structures can be characterized as a complex of condensed silica and polyacrylic acid. Another important point on the way to obtaining monodisperse particles is the low rate of TEOS entering the reaction mixture. We observed that changing the reaction protocol from a four-hour injection to a single addition of the entire precursor resulted in the formation of an aggregated colloid (see Fig. 8 in the Appendix). Finally, we found the importance of using low molecular weight PAA (1.8 kDa). When using PAA with molecular weights of 35 and 100 kDa, turbidity of the growth solution and precipitation are observed even before the addition of tetraethylorthosilicate. Obviously, the interaction of ammonia with high molecular weight polyacrylic acid results in the formation of polydisperse micro and nanoparticles, which cannot serve as “seeds” for the further synthesis of monodisperse silica particles.

Fig. 2.

Electron microscopic images of silicate nanoparticles obtained by hydrolytic condensation of tetraethyl orthosilicate with the addition of 2.5 (a) and 5 (b) mg of polyacrylic acid to the reaction mixture. The insets show histograms of particle size distribution. The scale bar on TEM images is 200 nm.

The next step in the work was the implementation of the technique for growing the “seed” particles to the required size. As primary “seeds” we used the particles shown in Fig. 2a. Figure 3 shows typical electron microscopic images of silicate particles obtained at 1−9 growth steps. All particles are spherical, relatively isodisperse, and not aggregated. The size of the obtained particles increases sequentially with each growth stage from 50 nm for the initial particles to 3 μm at the last step. The average particle diameters measured according to electron microscopy data together with their standard deviations (based on the results of counting at least 200 particles) are shown in Table 1. Note, that in all cases the standard deviation did not exceed 5%. At each growing step from the first to the sixth, the particle size increases by an average of 70%, which is in good agreement with the mass balance data. Indeed, the ratio of TEOS added to hydrolyzed during the formation of nuclei is 5 to 1. Therefore, the particle size should increase to the cubic root of 5 times, that is, about 1.7 times. The mass balance indirectly indicates that no new nuclei are formed; the surface area of the seeds is sufficient to “utilize” all forms of the resulting silica under the given specific conditions of the process. The reverse situation is observed for later growing steps. We noticed that when using 6 mL of “seed” particles in the reaction, there is only a slight increase in their size, and the colloid contains a huge amount of additional small-sized silicate particles (see Fig. 9 in the Appendix). We assume that this is due to a significant decrease in the total surface area of the “seed” particles at the later stages of growing. Indeed, if the measured mass concentration of silica in all colloids did not differ much (varied from 8 to 11 mg/mL), then the number of particles differs by a factor of 10⁴ for the first and sixth growing steps according to formula 3. Thus, the total surface area drops by a factor of 20. In the case of obtaining designed silica particles with a diameter of 1 μm or more, washing off secondary particles (whose sizes are less than 200 nm according to TEM) is not difficult. In our work, we used a sixfold sedimentation/resuspension, and the acceleration during centrifugation was chosen in such a way as to minimize the sedimentation of secondary nanoparticles. Additionally, we have doubled the amount of buds added in steps 6–10 from six milliliters to twelve milliliters. This made it possible to reduce the losses of TEOS due to the formation of side particles, however, it reduced the maximum possible step of increasing the size at each stage of growing from 70 to 30. The particle size can be controlled not only by the number of growing steps, but also by the amount of TEOS added at each stage. For example, reducing the amount of TEOS from 0.75 to 0.3 mL in the second step reduces the size of the obtained particles from 176 to 128 nm (see Table 1). Similarly, a decrease in the amount of TEOS at the fifth step to 0.25 and 0.5 mL makes it possible to obtain particles with sizes of 700 and 860 nm.

Fig. 3.

Typical electron microscopic images of silicate particles obtained at steps 1 (a)–9 (k) of growing “seed” particles. The scale bars on TEM images are: 200 (a, b), 500 nm (c, d), and 1 µm (e−j).

Table 1.   Particle sizes obtained at various steps of the multistage hydrolytic condensation reaction, as well as reaction parameters

Let discuss the optical properties of the resulting particles. Figure 4 shows the extinction spectra of silica nanoparticles obtained at the 1–10th growing step, as well as the spectrum of the initial 50 nm “seeds.” All curves are normalized to a silica concentration of 4 mg/mL. The spectra are typical of nonabsorbing spherical particles. In particular, decreasing curves without extrema are observed for particles smaller than 2 µm, which is explained by more efficient light scattering in the short-wavelength spectral range. For larger particles, there is a scattering maximum in the visible region due to interference effects. Under the conditions of single scattering [28], the optical density of the colloid is the sum of the energy loss of the light beam on individual particles and can be represented as:

$$\tau = N\pi {{R}^{2}}Q(R,n),$$

(4)

where N is the number of scattering particles per unit volume, R is their radius, n is the refractive index of the particles, $Q(R,n)$ is the scattering coefficient. The smooth nature of the dependence of extinction or optical density over a wavelength interval allows us to approximate the turbidity spectrum by the Angström relation [29]:

$$\tau \sim {{\lambda }^{{ - w}}}.$$

(5)

Fig. 4.

Normalized optical extinction spectra of silicate seeds (curve 0) and silica particles obtained at 1–10 growing steps. The spectra are normalized to an optical path length of 1 cm and a silica concentration of 4 mg/mL.

The wavelength exponent w is numerically equal to the tangent of the slope of the turbidity spectrum in double logarithmic coordinates, taken with a minus sign.

By substituting (4) into (5), one can obtain the dependence of the wavelength exponent on the particle radius and their refractive index. Previously, we obtained calibration dependences for determining the average size of monodisperse silicate particles from the wavelength exponent [30].

Another common method for determining particle sizes is dynamic light scattering [31]. We decided to compare the results of particle sizing by three methods: TEM, spectroturbidimetry (STT), and dynamic light scattering (DLS). The results of comparative measurements are shown in Table 2. We can see good agreement between the data for electron microscopy and spectroturbidimetry, which is expected, given the monodispersity and sphericity of the colloids. It should also be noted that the results of the DLS method are overestimated compared to other methods, which is often explained by the presence of a near-surface layer that affects the hydrodynamic properties of particles.

Table 2.   Average diameters of particles obtained at different steps of the multistage hydrolytic condensation reaction, determined using TEM, STT, and DLS methods

In recent years, much attention has been paid to the synthesis and application of functional nanoparticles and nanomaterials [32]. From this point of view, the developed multistage growth procedure seems to be promising due to the possibility of including functional molecules and/or nanoparticles layer by layer into the volume or onto the particle surface. To demonstrate this possibility, we synthesized nanoparticles with incorporated fluorescent molecules. The synthesis procedure corresponded to stage 4 of growing silica particles; however, at the time of growing, an aminosilane-conjugated dye (fluorescein isothiocyanate, FITC) was added to the reaction mixture. The technique makes it possible to obtain particles with any dye capable of reacting with the amino group. For example, we obtained similar results with activated esters of borondipyrromethene.

Figure 5a shows electron microscopic images of synthesized silica particles with FITC molecules incorporated into the matrix. The average particle size obtained was 388 nm. In the enlarged image of the particle in the inset in Fig. 5a, a surface fluorescent layer about 40 nm thick can be seen. Note that the incorporation of fluorescent molecules did not affect the colloidal properties, i.e., did not lead to an increase in the of polydispersity, particle aggregation, or their ability to act as seeds for further growth. Figure 5b shows the optical density and fluorescence spectra of the obtained composite particles. The fluorescence spectrum was measured upon excitation with light at a wavelength of 488 nm. The extinction of composite particles is a superposition of the extinction spectrum of a silicate colloid and a molecular dye. This makes it possible to estimate the number of dye molecules in one particle. The mass concentration of silica in the colloid determined by weighing the dry residue is 4 mg/mL. Taking into account the silica density of 1.7 g/cm³ and the particle diameter of about 400 nm, the numerical concentration of the particles is 7.4 × 10¹⁰ mL^–1. The FITC concentration determined by spectrophotometric calibration [33] is 0.27 μg/mL. Taking into account the molar mass of FITC equal to 389 g/mol, the numerical concentration of the dye molecules is 4.16 × 10¹⁴ mL^–1. Thus, there are about 5600 dye molecules per particle. If we assume a uniform distribution of dye molecules in a 40 nm shell, then the average intermolecular distance will be about 14 nm. At such distances, the effect of concentration quenching of fluorescence is unlikely. This is confirmed by the observed fluorescence spectrum of the nanoparticles (Fig. 5b), which exactly matches the spectrum of the molecular dye, both in position and intensity. The brightness of a single label (particles with incorporated dye molecules) is sufficient to visualize single particles. As an example, Fig. Fig. 10 in the Appendix shows a confocal fluorescent microscopic image of a HeLa cell incubated with fluorescent nanoparticles. A significant obstacle to the use of organic dyes in fluorescence microscopy is photobleaching. It is currently believed that the main mechanism of photobleaching is a photochemical reaction with singlet oxygen [34]. From this point of view, the inclusion of fluorescent dyes in a silica matrix should increase their photostability. To test this assumption, we measured the fluorescence intensity from a 1 mM FITC solution and dye-incorporated silica particles under constant irradiation for 300 s. Irradiation was performed using a Leica DMI 3000-B fluorescent microscope using a 40× objective and a blue light filter (average wavelength 480 nm). Fluorescence photographs were taken every 15 s with constant exposure and aperture. Next, the average intensity of the pixels in the image was digitized using the ImageJ program and normalized to the maximum intensity at the initial time. In Fig. 5c, the top row shows a series of fluorescence images of FITC and silica nanoparticles (and their aggregates) obtained during 75 s of irradiation. Almost complete photobleaching of the dye in the dissolved state is obvious, while when FITC is included in the silica matrix, the fluorescence intensity decreases by no more than 15% in 75 s of irradiation. Figure 5c also shows the evolution of photobleaching during irradiation for 300 s. Triangles show data for FITC solution, circles for nanoparticles and their aggregates. Defined as the tangent of the slope at the zero point, the photobleaching rate was reduced by a factor of five when the dye was incorporated into the silica matrix.

Fig. 5.

(a) Typical electron microscopic images of silica particles with incorporated FITC molecules. The inset shows an enlarged image of a single particle. Scale bars are 200 nm. (b) Normalized optical density and fluorescence spectra of the indicated particles. The fluorescence spectrum was measured under excitation with light at a wavelength of 488 nm. (c) Series of fluorescence images of FITC and silica nanoparticles obtained during 75 s of irradiation. Photobleaching kinetics of FITC and silica nanoparticles during 300 s of irradiation.

Another example of the efficient incorporation of functional objects into a silicate matrix is the synthesis of nanoparticles with incorporated SERS tags. SERS tags consist of metal nanoparticles with adsorbed molecules with a high Raman cross section. In our work, we used 13 nm gold nanoparticles with adsorbed nitrobenzenethiol molecules. These SERS tags were adsorbed on silicate particles before the 4th growing step. Figure 6 shows TEM images of 300 nm silicate spheres used as cores (panel (a)), the same spheres after adsorption of SERS tags (panel (b)) and after growing an additional silicate shell (panel (c)). The particles retain monodispersity and colloidal stability at all stages of obtaining a composite sample. As a result SERS tags are incorporated inside the silicate particle, and the thickness of the secondary silicate shell is about 100 nm. Due to the large thickness of the shell, the gold particles inside the silica are not visible in a conventional TEM image. However, when using the scattered electron mode and the filter for the scattering energy corresponding to silicon, SERS tags can be easily visualized (see inset in panel (c)). Figure 6d shows the SERS spectra of the obtained composite particle colloid. In accordance with our previous measurements [35], the SERS spectra for all samples are dominated by nitrobenzene peaks: π (CC) at 723 cm^–1, π (CH) at 854 cm^–1, ν (CS) at 1081 cm^–1, δ (CH) at 1110 cm^–1, ν (NO₂) at 1343 cm^–1, and ν (CC) at 1569 cm^–1, with the most intense peak at 1343 cm^–1 associated with the vibration of the nitro group.

Fig. 6.

Typical electron microscopic images of 300 nm silica particles used as cores for the synthesis of composite particles (a), silica nanoparticles with adsorbed SERS tags (b), and composite particles coated with an additional layer of silica (c). The inset shows an image in the electron scattering mode, which allows visualization of the SERS tags inside the silica matrix. The scale bar in all TEM images is 200 nm. (d) SERS spectrum from a colloid of composite nanoparticles.

CONCLUSION

In this work, we used the method of successive growth of silicate particles obtained by hydrolytic condensation of tetraethylorthosilicate in an alcohol medium using an alkaline catalyst to obtain particles of large diameter (up to 3 μm). The key point is the creation of conditions for directed hydrolytic condensation of tetraethylorthosilicate on the surface of seeds by varying the concentration of particles, reducing the rate of addition of the precursor, and removing secondary seeds by centrifugation/resuspension of the sample. This technique makes it possible to obtain particles with a size in the range from 50 nm to 3 μm. The size of the resulting particles is controlled by the number of growing steps and the amount of added tetraethylorthosilicate. The samples obtained were characterized by transmission electron microscopy, spectroturbidimetry, and dynamic light scattering.

The synthesis of monodisperse silicate particles in a wide range of sizes is not the only advantage of the multistage growing technique. It is shown that the developed approach makes it possible to obtain functional nanoparticles by including fluorescent molecules and SERS tags without changing the colloidal stability and monodispersity of the sample. We believe that the obtained nanoparticles can find applications in various fields, including photonics and biomedicine.

Appendix

Fig. 7.

Electron microscopic image of silica nanoparticles obtained by the Stober method. Composition of the reaction mixture 30 mL of ethanol, 1.5 mL of ammonia, 0.75 mL of TEOS.

Fig. 8.

Electron microscopic image of aggregated silica “seeds” obtained by simultaneous addition of TEOS to the reaction solution.

Fig. 9.

Electron microscopic image of silica particles obtained at the 8th step of overgrowth under non-optimal conditions and without washing from secondary particles .

Fig. 10.

Fluorescent, phase contrast image and their overlay obtained for HeLa cell after incubation with fluorescent silicate nanoparticles. The arrows show the image of single particles .