SiBP alone as catalyst

Reaction kinetics were studied via OD measurements of the particles and GC analysis of tetraethyl orthosilicate (TEOS). OD profiles of the particle formation are shown in Figure 1a, and representative GC spectra of the control group are shown in Figure 1b. No particle formation or TEOS hydrolysis was observed in the negative control (no catalyst) group within the measurement timeframe. As seen from the OD profiles (Figure 1a) and TEOS conversion rates (Figure 1c), distinct profiles were observed when SiBP was added alone or in combination with NH3.

Figure 1: (a) OD profiles of particle formation (dashed line: no SiBP, empty markers: SiBP alone, solid markers: SiBP + 0.45 M NH3). (b) GC spectra of the control (NH3 only) group at different time points. (c) TEOS hydrolysis rates (dashed line: no SiBP, empty markers: SiBP alone, solid markers: SiBP + 0.45 M NH3). (d) Effect of peptide concentration on reaction rate and yield (circles: TEOS conversion at 120 min, triangles: rate constant, empty markers: SiBP alone, solid markers: SiBP + 0.45 M NH3).

SiBP was able to hydrolyze TEOS and produce SiO2 particles when used alone. The increase in peptide concentration resulted in a dose-dependent increase in the reaction rate and yield (Figure 1d). Despite the same amount of precursors added to all reactions, the yields of the reactions with SiBP alone at plateau were lower compared to reactions containing NH3. Possible reasons of this observation will be discussed below.

No significant differences were observed regarding particle size or morphology depending on the SiBP concentration when SiBP was used alone as catalyst (Figure 2). At all concentrations, 17–20 nm particles were obtained with polydispersity indices (PDI) of 0.027, 0.050, and 0.010 for 0.04, 0.40, and 1 mM SiBP, respectively, indicating highly monodisperse (PDI < 0.080) particles. At all concentrations, the particles were organized into a branched fibrillar network, which is characteristic to the gel state of colloidal SiO2 (Figure 2) [26]. Colloidal gels are formed when colloidally suspended particles form a branched fibrillar network of particle strands through interparticle attractions. Under alkaline conditions, electrostatic repulsion between the SiO2 particles prevents formation of interconnecting particle strands. However, the presence of a cationic emulsifier allows for stable interparticle interactions and coagulation of the particles into interconnected particle strands [26]. Our findings indicate that, when used alone, the positively charged SiBP can also act as a cationic emulsifier resulting in the branched fibrillar networks observed by SEM.

Figure 2: SEM images of the SiO2 particles formed by SiBP alone at concentrations of (a) 0.04 mM, (b) 0.4 mM, and (c) 1 mM. (d) Particle size frequency distributions obtained from DLS measurements.

In this aspect, when used alone, the SiBP mimicked the in vitro behavior of biosilicification-related proteins (BSRPs), such as silicateins and silaffins. BSRPs facilitate the formation of inorganic silica structures in marine organisms [27,28]. It has been reported that isolated silicateins and silaffins or certain repeating motifs of these proteins can facilitate silica precipitation in vitro [29-32]. The catalytic activity of these proteins is thought to be similar to the serine–histidine–aspartic acid (SHD) catalytic triad [33,34]. In this model, a hydrogen bond between serine and histidine increases the nucleophilicity of serine. Aspartic acid stabilizes the favorable orientation of histidine. Then a nucleophilic attack by serine on the Si–O bond of the precursor molecule results in a Ser–O–Si(OR)3 transitory complex. The hydrolysis is completed by the addition of water, separating the protein and the hydrolyzed precursor molecule, and the release of ethanol.

Although SiBP contains an N-terminal serine and two arginine residues, it does not contain histidine, aspartic acid, or another residue that can act as a H bond acceptor for serine. However, serine residues can form hydrogen bonds among themselves. Therefore, one can speculate that a hydrogen bond formed between the serine residues of two peptide molecules can increase the nucleophilicity. If this is the case, the nucleophilic attack of serine can facilitate hydrolysis of TEOS. However, a second and more likely speculation is that the SiBP mediates the hydrolysis through arginine residues. Arginine is a very strong proton acceptor with a side chain pKa of 13.80 [35]. Therefore, locally increased concentrations of OH− by arginine could facilitate the hydrolysis of silicon alkoxides, since OH− is also a potent nucleophile. Future studies where serine and arginine residues of the SiBP are substituted could help to elucidate the nature of the catalytic activity of SiBP.

As mentioned above, the yield of the reactions with SiBP alone were lower compared to reactions including NH3 despite the same amount of initial precursor molecules (Figure 1c). A probable reason of this observation is the high affinity of the peptide to SiO2[25]. As colloidally stable silica particles start to form, the peptide starts to adhere to the surface of the particles, effectively being removed from the solution before all the precursor molecules available are hydrolyzed.

SiBP + NH3 as catalyst

When NH3 was added, the SiBP had a negligible effect on the reaction rate and the yield (Figure 1c), indicating that SiBP contributed very little to the catalytic process. This can also be the result of high affinity of SiBP to SiO2. Since particle formation occurs relatively fast when NH3 is added, it is possible that SiBP binding to the particles results in the negligible effect on reaction rate and yield.

All groups containing NH3 yielded spherical submicrometer particles characteristic to the Stöber method (Figure 3a–d). However, a decrease in average particle size and size distribution was observed with increasing SiBP concentrations (Figure 3e,f). The PDI for the NH3 + 1 mM SiBP was lower compared to other groups containing NH3, indicating a narrower size distribution. However, the other groups also yielded monodisperse distributions with PDIs below 0.080 (Figure 3a–d). The second hypothesis of the study is that because of its high affinity to SiO2, SiBP can form a dense layer on the particle surface that improves the stability of nanoparticles [36]. These observations support this hypothesis, as smaller and more narrowly distributed particles were obtained with higher SiBP concentrations. Capping agents are not only used to regulate growth and size of colloidal nanoparticles, but also to control the physicochemical or biological characteristics. As will be demonstrated below, in addition to the size of the particles, the SiBP also changes the surface charge of the particles, resulting in improved self-assembly.

Figure 3: SEM images and size distributions of the particles formed with (a) NH3 only and NH3 with (b) 0.04 mM, (c) 0.4 mM, and (d) 1 mM SiBP (arrows: coalescing particles). (e) Particle size distributions obtained from DLS measurements. (f) Effect of SiBP on the particle size when NH3 was added.

An interesting and distinct OD profile was observed in the NH3 + 1 mM SiBP reaction, which started with a fast increase reaching to a peak higher than the plateau of the other groups after approximately 40 min. This was followed by a steep decrease falling back to the plateau of the other groups after approximately 20 min (Figure 1a). This was at first thought to be the result of an error in experimental methods or measurement. The reactions were repeated with freshly prepared solutions and different brands of 96-well plates, but the distinct profile was observed in each repetition. To further investigate this interesting profile, samples were collected from the NH3 + 1 mM SiBP reaction and the NH3 only reaction in the middle of the steep increase (20 min), at the peak point (45 min), and in the middle of the steep decrease (55 min) (Figure 4a). SEM analysis showed that very different particle formation regimes occurred in the presence and absence of the SiBP. After 20 min, NH3 + 1 mM SiBP yielded a high amount of monodisperse particles of approx. 10 nm (Figure 4b), while NH3 alone yielded polydisperse particles of 10–40 nm (Figure 4c). After 45 min, NH3 + 1 mM SiBP yielded a mixture of irregularly shaped particles (Figure 4d; arrows), clusters of small 10 nm particles (Figure 4d; asterisk), and larger particles between 30 and 50 nm, while NH3 alone yielded spherical particles of 20 to 50 nm (Figure 4e). After 55 min, NH3 + 1 mM SiBP yielded particles of 60–100 nm. Small particle clusters or irregular particles were not observed (Figure 4f), while NH3 alone yielded spherical particles of 100 to 130 nm (Figure 4g).

Figure 4: (a) UV–vis spectra of the reactions with NH3 alone and NH3 + 1 mM SiBP. SEM images of the SiO2 particles collected after (b, c) 20 min, (d, e) 45 min, and (f, g) 55 min (arrows: coalescing particles, asterisk: clusters of approx. 10 nm particles).

Based on the OD and SEM observations, SiBP seems to drastically change the particle formation/growth regime above a threshold concentration. In the reactions of NH3 alone, NH3 + 0.04 mM SiBP, and NH3 + 0.4 mM SiBP, the growth regime followed the classical aggregative growth and monomer addition model [37]. According to this model, at the early stages of the reaction, the dominant regime is homogeneous nucleation of SiO2 particles. The growth continues by coalescence and Ostwald ripening of the particles. Growth by coalescence and Ostwald ripening is a fundamental process that plays a dominant role in nanoparticle formation. In coalescence, two or more particles combine to form a larger particle, whereas in Ostwald ripening, small particles dissolve in a solution and redeposit to form large masses. The process is mainly driven by the differences in chemical potential and surface energy between particles with different size and shape. In SiO2 synthesis, smaller particles with high surface energy dissolve via cleavage of siloxane bonds on the surface. The released silicic acid is then deposited onto particles with larger radius. Evidence of coalescence was observed in our study as well (shown by the arrows in Figure 3 and Figure 4). At later stages of the reaction, when the precursor concentration drops below the nucleation threshold, the dominant regime becomes growth by monomer addition to the surface of the particles.

At 1 mM concentration, however, the SiBP alters this profile and delays the aggregation of early approx. 10 nm particles into larger particles (Figure 4b). The most likely mechanism for this is the SiBP binding on the surface of the particles and, in turn, stabilizing the early particles. The resulting high number of small particles with very large surface area results in the high OD at 45 min, which is higher than the plateau of the positive control group (Figure 4a). When the fraction of the peptide to primary particles exceeds a critical value, the primary particles rapidly aggregate into larger particles, which results in the rapid drop in the OD (Figure 4a,d). Further studies are underway to take advantage of this interesting effect in synthesizing colloidally stable approx. 10 nm SiO2 particles at high volume fractions.

Self-assembly of the particles

The effect of the SiBP on the self-assembly of the as-grown particles was investigated via SEM and UV–vis spectroscopy. Single-layer and multilayer assemblies were investigated by using different dilutions of the as-synthesized particles.

SEM imaging showed that the particles from the NH3 + 1 mM SiBP reaction assembled into ordered single-layer (Figure 5a) or multilayer (Figure 5c) opal-like structures. Opal is a naturally occurring mineraloid with silica as the principal chemical constituent. The optical behavior of iridescent opal is a result of the regularly stratified structure of silica particles, in which the alternate layers differ in refractive index. The periodic difference in the refractive index creates photonic band gaps, in which certain wavelengths of the light cannot propagate, depending on the size of the periodic structures and the differences in the refractive indices. These structures are referred to as photonic crystals and can be manufactured synthetically for various optical applications. Similar optical behavior of the opal-like structures formed in this study will be demonstrated and discussed below.

The fast Fourier transform of the SEM images revealed a hexagonal close-packed structure (insets in Figure 5a,c). In fact, the tendency of the particles formed with NH3 + 1 mM SiBP to assemble into ordered structures was visible on samples not prepared by vertical deposition but simply by dripping on a surface and vacuum drying (Figure 3d). The particles formed in the reaction with NH3 alone assembled into less ordered single-layer (Figure 5b) or multilayer (Figure 5d) structures.

Figure 5: SEM micrographs of (a, c) single-layer and (b, d) multilayer self-assembled SiO2 particles (insets: corresponding fast Fourier transform diffraction pattern of the images). (e) UV–vis spectrogram of the self-assembled particles on a quartz surface. (f) Zeta potential of the particles in deionized water in the presence and absence of the SiBP.

In UV–vis spectra of the self-assembled particles, a broad peak was observed around 611 nm with NH3 only, while with NH3 + 1 mM SiBP, a narrower and stronger peak was observed around 457 nm (Figure 5e). The position and the width of the absorbance peaks in the UV–vis spectroscopy are determined by the Bragg diffraction of the light in photonic crystal structures and depend on the particle size, extent of the periodicity (i.e., quality of the assembly), and the angle of the incident light. The position of the absorbance peak can be calculated using Bragg’s law (Equation 1):

(1)

where λ is the wavelength of the peak absorbance, d is the diameter of the particles, f is the packing factor (0.74 for hcp structures), n is the refractive index (1.46 for SiO2[38] and 1 for air), and θ is the angle of the incident light. Using this formula and the average particle sizes obtained from the particle size analysis, the absorbance peak was calculated to be 593 nm for the NH3-only reaction and 453 nm for the NH3 + 1 mM SiBP reaction. The slight difference between the calculated and measured Bragg maxima is likely due to the hydrodynamic radius measured in particle size analysis being slightly larger than the actual size of the particles. Considering this, Figure 5e shows that the calculated and measured values agree well. The broader and weaker peak observed with the NH3-only group is the result of the less ordered assembly and broader particle size distribution. This is better observed when the measured λ values are fitted to the theoretical Bragg reflection maximum (see Supporting Information File 1, Figure S4).

The improvement in the self-assembly by the SiBP is likely achieved by reducing the negative surface charge of the SiO2 particles. The surface of the SiO2 particles is negatively charged above pH 2–3 [39,40]. Therefore, the as-synthesized particles exhibit little cohesion because of electrostatic repulsion. Usually, a post-assembly sintering process or long waiting periods [41] are required to improve the assembly. When the positively charged SiBP binds to the particle surface, it reduces the negative charge of the particle surface (Figure 5f). The particles synthesized with NH3 only had a ζ potential of −44.90 mV, while the particles synthesized with NH3 + 1 mM SiBP had a ζ potential of −30.87 mV. At the meniscus where the solvent dries out, the decreased surface charge reduces the distance between the particles, enabling them to assemble into more ordered and close-packed structures.

Angular dependence of the Bragg reflections and the uniformity of the assembled particles from the NH3 + 1 mM SiBP group was qualitatively investigated by assembling the particles on the inner surface of a cylindrical beaker by letting the reaction solution evaporate under vacuum. Following the curvature of the surface, different colors were observed depending on the angle of the incident light (Figure 6a). When the beaker was rotated under the same angle of incident light, the observed colors did not change, indicating a long-range uniform assembly. Side-by-side visual comparison of the uniformity of the self-assembly from the NH3 alone and NH3 + 1 mM SiBP reactions are demonstrated in Figure 6b.

Figure 6: (a) Qualitative demonstration of the long-range homogeneity of self-assembly and angular dependence of Bragg reflection of as-synthesized SiO2 particles formed with NH3 + 1 mM SiBP. (b) Qualitative comparison of the long-range homogeneity of self-assembly of as-synthesized SiO2 particles formed with NH3 alone and NH3 + 1 mM SiBP.