Organic Base-Free Synthesis of Tunable Mesoporous Silica Nanoparticles Triggered by CO₂‑Derived Cyclic Carbonates

Abstract

Mesoporous silica nanoparticles (MSNs) with controlled size (d ∼ 30–70 nm) and monodisperse size distribution are attracting significant interest for multiple applications, especially drug delivery. Particles with such structural features could be obtained in some cases through aqueous protocols employing TEOS (tetraethyl orthosilicate) as the silicon source and CTAB (cetyltrimethylammonium bromide) as the surfactant, but in the presence of toxic, volatile, and flammable triethylamine (TEA). There is a lack of safer protocols able to produce MSNs with the above-mentioned desirable characteristics by using nonvolatile inorganic bases. In this work, we explore the formation of MSNs using CTAB, TEOS, and NaOH as the base in the presence of cosolvents of different polarities in well-defined molar ratios. When using highly polar, CO₂-based cyclic carbonates as cosolvents above a specific CTAB/cyclic carbonate ratio, formation of monodisperse MSNs with a tunable size below 100 nm (and specifically in the 40–60 nm size range for propylene carbonate) and a well-defined starburst morphology was observed. Based on experimental evidence, a growth mechanism for carbonate-based MSNs is derived where strong coordination of micelle and silica condensation nuclei by the cyclic carbonates drives particle growth, yield, and morphology.

Introduction

Silica is an extremely abundant, green, and safe material that is part of crucial applications of daily life and research. Available in a variety of different particle sizes and morphologies, , silica-based materials are used in catalysis, − biomedicine, , electronics, energy storage, coatings, , and many others. Remarkably, with the emergence of nanotechnology, most of these applications are currently carried out by using silica nanomaterials. , In a further leap forward, mesoporous silica nanoparticles (herein MSNs) were developed. , The presence of mesopores enormously amplifies the surface area of silica nanoparticles with clear benefits under several aspects, for instance, by facilitating the diffusion of reagents or active site accessibility , in catalysis or by increasing the loading of bioactive payloads in biomedical applications. Additionally, it is worth noting that the use of monodisperse silica nanoparticles affords higher reproducibility and degree of control over the final particle properties and performance.

While silica particles with a size close to 1 μm have also attracted interest, , increasing focus has been placed on monodisperse MSNs. In particular, MSNs with (tunable) sizes below 100 nm find remarkable applications. For instance, MSNs with diameters larger than 10 nm and lower than 100 nm are important for drug delivery. , MSNs with a size in the 50–100 nm range were found to perform as ideal nanocarriers for anticancer treatment. In particular, cellular uptake of drug-functionalized MSNs was found to strongly vary with particle size in the 30–280 nm range and reached an optimum for a size of 50 nm. MSNs with a size smaller than 100 nm have also found application in the preparation of dental nanocomposites (45 nm) in the preparation of antireflective coatings and as stimulants of immune response in plants. Therefore, there has been significant interest in the preparation of monodisperse MSNs with tunable size in the 40–70 nm range using TEOS as the silicon precursor and CTAB as a surfactant and pore template. Semeykina and Zharov prepared MSNs in high yields with size in the 30–40 nm range by a protocol using water as the solvent and TEA (triethylamine) as the base. Further size modulation could be introduced by adding ethanol (EtOH) to the reaction mixture (Table S1, entries 1, 2). Similar results were reported in the past by using triethanolamine in place of TEA. ,, MSNs with size in the 45–70 nm range could also be obtained in an EtOH/water mixture using aqueous ammonia as the base. Nevertheless, given the toxicity, volatility, flammability, and environmental issues related to the use of organic amines, , the high environmental toxicity of ammonia, and the classification of TEA as a regulated volatile organic compound, methodologies using simple, inexpensive inorganic bases and environmentally benign reaction components would be preferable. However, the use of nonvolatile inorganic bases such as NaOH in aqueous systems for the synthesis of MSNs generally leads to the formation of relatively large particles. Candela-Noguera et al. have shown that when using water as the solvent and NaOH as the base, MSNs with MCM-41 morphology and high surface area were obtained with a size in the 90–100 nm range (Table S1, entry 3). Including EtOH in the solvent mixture, as reported by Luo et al., led to the formation of particles with a 200–300 nm size, high surface area, and MCM-48 morphology (Table S1, entry 4). The latter result is in line with the observation that EtOH decreased the rate of TEOS hydrolysis, leading to the formation of fewer nuclei for MSN formation and resulting in larger MSNs. Kim et al. used methanol as the solvent and NaOH as the base, leading to the preparation of monodispersed MSNs with a 15–56 nm diameter and high surface area, but the reaction yields were relatively low, and TEOS had to be added through several hours at room temperature using a syringe pump (Table S1, entry 5), with the subsequent reaction lasting an additional 24 h. Other methodologies have been reported for the preparation of MSNs with tunable size (70–380 nm) but in the presence of large amounts (11–22 mol % with respect to the amount of water used) of additives such as aldehydes, , significantly affecting the sustainability and cost of the process. Similar limitations apply for other literature-reported procedures replacing CTAB with more sophisticated templating agents such as oligomeric fluorinated alkyl-chain surfactants, which are not readily available and were found to suppress porosity, or other combinations of compounds. Overall, it is challenging to obtain monodisperse MSNs with (tunable) size below 100 nm through practical strategies without employing volatile compounds, toxic bases, and by using media that contain exclusively water and only moderate amounts of environmentally benign cosolvents or additives for tuning particle size. Interestingly, most efforts to tune MSN size and morphology by the addition of cosolvents have been focused on the use of additives with lower polarity than water such as alcohols. However, these compounds have partial solubility in the micelle, leading to its swelling and an increase in size. , Conversely, highly polar cosolvents could interact with the charged micelle interface, avoiding its swelling. Indeed, it has been observed that the polarity of the solvent has an effect on particle size. Wang et al. have shown that silica nanoparticle size decreased remarkably by changing the reaction solvent from isopropanol to ethanol and methanol. The decrease in size was attributed to the increase of the dielectric constant of the medium, hindering aggregation at the nucleation sites. However, alcohols have significantly lower polarity than water (for instance, the dielectric constant of water is 78 versus 25 of EtOH). Therefore, it is expected that highly polar compounds with a dielectric constant similar to or even higher than that of water could provide a more efficient solvation of the nucleation sites, hindering their excessive growth. In this context, cyclic carbonates are an increasingly important class of CO₂-derived highly polar compounds that can be prepared under mild nonreductive conditions − by using structurally simple and even biobased catalytic systems. − Thus-produced cyclic carbonates serve for multiple applications such as solvents, , chemical intermediates, , and monomers for sustainable polymers, , enabling CO₂ upcycling to chemicals , for everyday life applications. Therefore, expanding the field of application of cyclic carbonates to new areas is of utmost importance, especially when their use commands unique structural features and functionalities. While propylene carbonate has been used in the past as a medium for the generation of nanometric metal oxide particles (Cu₂O), to the best of our knowledge, cyclic carbonates have not been used as media to produce and control the size of silica nanoparticles. In this work, we show that cyclic carbonates such as propylene carbonate (PC) and glycerol carbonate (GC) can be used as cosolvents to control the synthesis of spherical, monodispersed MSNs in the absence of organic and/or volatile bases, leading to nanoparticles with starburst morphology and size generally in the 40–100 nm range (Table S1) that could be optimized by varying the amount of cyclic carbonates in the reaction medium (Figure ). Within a rare , comparative analysis of multiple cosolvents including also ethanol (EtOH), cyclopentanol (CyP), and glycerol (Gly), we show that such control on size and morphology could be achieved only by using cyclic carbonates. We also show that cyclic carbonate-based MSNs generally afforded tunable, monodispersed, and smaller sizes (Figure ) than for the other cosolvents used in this work, without strong variations in other morphological properties such as surface area and pore size.

1.

Overall synthesis of MSNs using different cosolvents with their resulting particle size (d _Part.), pore size (r _Pore), and pore volume (V _Pore).

Experimental Section

General information on commercially available materials and analytical instrumentation used in this study is provided in the Supporting Information.

Synthesis of MSNs

The synthesis of MSNs was adapted from Y. Y. Pue et al.; in a round-bottom flask, 0.56 mmol of CTAB was dissolved in 100 mL of Milli-Q water. Then, the desired molar amount of the cosolvent (GC, EtOH, PC, CyP, or Gly) with respect to the concentration of CTAB was added, and the solution was stirred to reach the target CTAB/cosolvent molar ratio (Table ). The solution was stirred for 30 min at 80 °C (1000 rpm). Afterward, 0.7 mL of 2 M NaOH followed by 1 mL of TEOS was added dropwise into the solution. After allowing the reaction to continue for 2 h, the MSNs were collected via filtration or centrifugation at 15,000 rpm for 15 min. The obtained MSNs were washed via centrifugation using an EtOH/HCl mixture with a ratio of 40 mL of EtOH to 5 mL of HCl several times. After drying, the MSNs underwent calcination at 550 °C for 6 h to remove all organics present. The obtained samples were denoted as GC, EtOH, PC, CyP, or Gly, followed by the ratio between CTAB and the listed cosolvent, CTAB/cosolvent (e.g., GC 1:1). The sample without any cosolvent was denoted CTAB 1:0. To note, attempts were made to prepare MSNs without the use of CTAB, but no silica precipitation was observed within 2 h.

1. Cosolvent Conditions Tested in the Synthesis of MSNs.

CTAB (mmol)	cosolvent (mmol)	CTAB/cosolvent ratio
0.56	0	1:0
	0.28	1:0.5
	0.56	1:1
	1.12	1:2
	1.68	1:3
	2.24	1:4
	2.80	1:5
	3.36	1:6

Results and Discussion

In order to explore the role of different carbonate-based cosolvents as alternatives to traditional cosolvents such as primary alcohols, we selected glycerol carbonate (GC) as a CO₂-based and highly polar α-functionalized alcohol bearing a cyclic carbonate moiety. GC has recently emerged as a highly attractive chemical for multiple applications. − Moreover, we selected propylene carbonate (PC) as a traditional nonprotic carbonate to explore the effect of the hydroxy substituent compared to GC. Moreover, we selected cyclopentanol (CyP) as an alicyclic alcohol without a carbonate functionality and glycerol (Gly) as a polyol which serves as the GC precursor. The selected solvents have different polarities as indicated by the varying relative permittivity values (ε_r, Figure ), ranging from very high values for cyclic carbonates, due to the presence of the polarized carbonyl group, to moderate or low values for the other alcohols. The selected set of compounds was investigated for the preparation of MSNs using NaOH as the base and CTAB as the surfactant. We fixed the CTAB amount and selected various CTAB/cosolvent ratios where the cosolvent amount relative to CTAB increased progressively (Table ) from 1:0.5 to 1:6. Additionally, we investigated the process in the absence of cosolvents.

2.

Cosolvents studied in the preparation of MSNs with their relative permittivity (ε_r) values.

The formed MSN samples were characterized through various techniques. In the first place, the absence of organic materials in the silica samples was demonstrated by FTIR investigation of representative CTAB/cosolvent ratios (i.e., 1:1, 1:3, and 1:6, Figure S1). The spectra generally showed the absence of aliphatic C–H stretching (2700–3000 cm^–1) and only silica-related peaks such as the broad absorption in the 3000–3700 cm^–1 region relative to hydrogen-bonded nonisolated silanol groups (ν_O–H), along with some adsorbed water (bending vibration at 1638 cm^–1). The strong adsorption centered at about 1060 cm^–1, and the remaining signals at 960, 794, and 450 cm^–1 all correspond to vibrational modes of the Si–O–Si network. , The MSNs were also investigated by X-ray diffraction (XRD). The XRD patterns of the samples obtained for a CTAB: cosolvent ratio of 1:6 are reported in Figure ; those for other CTAB: cosolvent ratios are provided in the Supporting Information (Figure S2). The provided XRD diagrams with the broad diffraction signal centered between 20 and 25 2θ values confirm the amorphous nature of the silica samples in all cases.

3.

XRD patterns of obtained MSNs using different cosolvents.

To correlate the trends between the added amount of cosolvent and MSN size (diameter) in relation to the obtained MSN yields (Figure ), we collected SEM images (Figures –) of all samples.

4.

Average particle size (black) and yield (colored) for all MSNs made: (a) GC, (b) PC, (c) EtOH, (d) CyP, and (e) Gly.

5.

SEM images of MSNs made (a) without GC and made with CTAB/GC ratios of (b) 1:0.5, (c) 1:1, (d) 1:2, (e) 1:3, (f) 1:4, (g) 1:5, and (h) 1:6.

9.

SEM images of MSNs made (a) without Gly and made with CTAB/Gly ratios of (b) 1:0.5, (c) 1:1, (d) 1:2, (e) 1:3, (f) 1:4, (g) 1:5, and (h) 1:6.

8.

SEM images of MSNs made (a)­without CyP and made with CTAB/CyP ratios of (b) 1:0.5, (c) 1:1, (d) 1:2, (e) 1:3, (f) 1:4, (g) 1:5, and (h) 1:6.

The MSNs prepared without the addition of cosolvents (CTAB/cosolvent 1:0, shown as reference figure “a” for each cosolvent SEM micrograph) yielded a mixture of spherical and elliptical particles, reflecting low control on particle morphology, with a relatively large average particle size and broad standard deviation (107 ± 27 nm). MSNs prepared using low GC amounts (Figure ) showed irregular spherical or granular shapes (Figure b,c). However, the GC-based MSNs evolved into highly uniform spheres at higher GC loading, i.e., for CTAB/GC 1:2 and above (Figure d–h).

The particle size and yield trend presented in Figure a showed an initial decrease in particle size by increasing the amount of GC up to CTAB: GC 1:2 from about 110 nm (no GC) to a significantly smaller size of 40–50 nm, with the particles being obtained in high yields. For these MSNs, the particle size histogram displayed in Figure S3 showed relatively broad size distributions (SD > 15%). A further increase in the amount of added GC led to a slight increase in particle size (70–80 nm range) and to a strong decrease in MSN yield. Nevertheless, more uniform and monodisperse (SD ∼ 10%) spherical particles were obtained. Similarly, when using PC as the cosolvent, irregularly shaped spherical particles or grains were obtained for low added volumes of PC (Figure b,c), which progressively evolved into a monodispersed spherical shape (Figure d–h) at higher CTAB: PC ratios. As in the GC case, the particle size initially decreased to a minimum upon increasing the PC amount up to CTAB/PC 1:2 (Figure b), leading to small particles (d ∼ 35 nm) for CTAB/PC 1:2, but with relatively broad size distribution (Figure S4). An increase in the added amount of PC led to a marked drop in MSN yields (Figure b) but, unlike the GC case, only to a relatively small increase in particle size. The obtained particles for CTAB/PC 1:4 to 1:6 showed a size in the 47–57 nm range with a good monodispersity (SD ∼ 10% or lower). For the other cosolvents (Figures –), the particle morphology was generally less regular (mixture of spherical and elliptical particles) than for cyclic carbonates, and, in some cases, large aggregates were observed. For EtOH and CyP, significantly broader size distributions and larger sizes compared to cyclic carbonates were obtained (Figures S5 and S6, respectively). For these cosolvents, the particle size always exceeded 100 nm and was always larger than in the absence of a cosolvent (Figure c,d).

6.

SEM images of MSNs made (a) without PC and made with CTAB/PC ratios of (b) 1:0.5, (c) 1:1, (d) 1:2, (e) 1:3, (f) 1:4, (g) 1:5, and (h) 1:6.

7.

SEM images of MSNs made (a) without EtOH and made with CTAB/EtOH ratios of (b) 1:0.5, (c) 1:1, (d) 1:2, (e) 1:3, (f) 1:4, (g) 1:5, and (h) 1:6.

In the case of Gly, some similarity with the carbonate-based MSNs was observed, with formation of slightly smaller and more spherical particles (Figure ) than EtOH and CyP and with particle size progressively declining by increasing the Gly volume up to a minimum for intermediate CTAB/Gly ratios, followed by a sharp increase in size for the highest cosolvent loadings (Figure e). However, different from cyclic carbonates, the lowest particle size observed for Gly was in the 70–80 nm range, with a broad size distribution (min. 20% SD, Figure S7), especially for the largest particle sizes. The MSN yields also followed a distinct trend for EtOH, CyP, and Gly compared to the carbonate-based MSNs (Figure ). For the former cosolvents, instead of a sudden drop in yield upon an increase in the cosolvent volume over a certain threshold, the MSN yield tended to a maximum for intermediate CTAB/cosolvent ratios, which also corresponded to the condition in which the lowest particle size was observed. This observation likely indicates the formation of numerous nucleation sites at intermediate CTAB/cosolvent ratios for EtOH, CyP, and Gly, leading to higher yields and lower particle sizes. Conversely, for cyclic carbonates, this situation was observed at low cosolvent loadings (CTAB/cyclic carbonate 1:2 or lower). For larger added amounts of cyclic carbonates, fewer nucleation sites were likely formed, resulting in lower yields and slightly larger particle size, but without reaching the sizes observed for EtOH, CyP, and Gly (except for CTAB/GC 1:6). These differences, and the fact that cyclic carbonates led to smaller, more monodisperse, and shape-defined MSNs than the other cosolvents, suggest that cyclic carbonates, due to their high polarity, provided a stronger interaction with the silica nucleation sites via electrostatic interactions hindering their excessive growth; a similar role was attributed in the past to lysine during the growth of fluorescent silica nanoparticles. At the same time, high concentrations of cyclic carbonates, due to their high solvent power, may interfere with micelle formation by hindering the formation of nucleation sites, leading to reduced MSN yields. Overall, this initial morphological investigation and the explored correlation between particle sizes and MSN yields demonstrate a strong difference in behavior between cyclic carbonate cosolvents and the other cosolvents used in MSN synthesis.

BET analysis was performed to investigate the surface area, pore size, and pore volume of all samples (see Figure for an overview). The N₂-adsorption isotherms of all samples are given in the Supporting Information (Figure S8). As a general trend, all samples showed type IV isotherms with H1-type hysteresis loops typical of mesoporous silica materials. , Pore size analysis of all samples given in Figure S9 shows the coexistence of both micropores and mesopores in the majority of samples as typically observed for other mesoporous silica materials such as MCM-41. The occurrence of larger pores (macropores) was observed in some cases for cyclic carbonates at lower cosolvent concentrations when particles with a relatively low surface area were obtained. With respect to surface area, pore size, and volume, each cosolvent displayed different trends by varying the CTAB/cosolvent ratio, reflecting different mechanistic behavior during particle formation and arising, in turn, from the different polarity and structural features.

10.

Textural properties of MSNs prepared using (a) GC, (b) PC, (c) Gly, (d) EtOH, and (e) CyP.

Nevertheless, cyclic carbonates PC and GC shared some significant similarities. For instance, as shown in Figure S9b,c, both cyclic carbonates showed similar trends of pore size distribution by increasing the CTAB: cyclic carbonate ratio. These trends show that the addition of cyclic carbonates even in small amounts had a dramatic effect on porosity. Indeed, compared to the MSNs prepared in the absence of cosolvents (Figure S9a), the presence of even very low amounts of cyclic carbonates (CTAB/cyclic carbonate 1:0.5) led to a loss of porosity due to the nearly complete disappearance of micropores with only some residual mesoporosity. A slight increase in cyclic carbonate concentration (CTAB/cyclic carbonate 1:1) led to the appearance of macropores with no significant formation of mesopores and micropores. A further increase in the added amount of cyclic carbonates (CTAB/cyclic carbonate 1:≥3 for GC, 1:≥2 for PC) led to the abundant formation of mesopores and micropores with the progressive disappearance of macropores. Compared to the carbonate-based MSNs, the use of Gly led to a totally different trend in pore size distribution (Figure S9f) that resembled that obtained in the absence of cosolvents (Figure S9a), apart from a general increase in mesoporosity compared to microporosity, and did not strongly change by varying the amount of Gly. Other cosolvents (EtOH, Figure S9d; CyP, Figure S9e) provided less predictable size distribution profiles. In general, the latter profiles resembled those obtained without cosolvents, but with a lower micropore to mesopore ratio; in some cases, samples with low porosity were observed. Figure a,b shows that for cyclic carbonate cosolvents, the BET surface area initially decreased from the value observed in their absence (i.e., CTAB/cosolvent 1:0) to a minimum for the CTAB/cosolvent 1:1 ratio. Such a change in surface area was somewhat limited for GC (from 600 m²/g to a minimum of 502 m²/g) and more evident for PC (from 600 m²/g to a minimum of about 300 m²/g). This trend is justified by the TEM images in Figure b,e, showing that at low carbonate concentrations (CTAB/cyclic carbonate, 1:1), a significant change in morphology with respect to the cosolvent-free MSNs occurred (Figure a).

11.

TEM of MSNs prepared using (a) CTAB/cosolvent 1:0 and CTAB/cosolvent ratios as listed: (b) 1:1 (GC), (c) 1:3 (GC), (d) GC 1:6 (GC), (e) 1:1 (PC), (f) PC 1:3 (PC), (g) 1:6 (PC), (h) 1:1 (EtOH), (i) 1:3 (EtOH), (j) 1:6 (EtOH), (k) Gly 1:1 (Gly), (l) 1:3 (Gly), and (m) 1:6 (Gly).

Accordingly, the carbonate-based particles obtained at this CTAB/cyclic carbonate ratio presented a less regular shape, with a smaller average size, in agreement with the above-mentioned discussion of SEM microscopy and particle size distribution. Despite their smaller size, the lower surface area of the carbonate-based particles GC 1:1 is due to their much lower porosity compared to the carbonate-free particles featuring ordered cylindrical pores. Indeed, the carbonate-based particles GC 1:1 generally showed the presence of macropores with some mesopores, as supported by the trend in pore size that reached a maximum of 14.2 nm for GC and 17.7 nm for PC for a CTAB/cosolvent ratio 1:1 (Figure a,b), i.e., the lowest surface area values. Nevertheless, by increasing the amount of cyclic carbonate, the surface area increased over the value observed for carbonate-free particles and reached a maximum value (832 m²/g for GC and 873 m²/g for PC) for a CTAB/cyclic carbonate ratio between 1:2 and 1:4 depending on the cyclic carbonate. This increase in surface area corresponded to the appearance of mesopores and a decrease or disappearance of macropores that was observed for GC 1:3 and PC 1:2 (Figure S9b,c). This trend was also confirmed by the decrease in pore size in Figure a,b. In the case of GC, pore volume followed the same trend as pore size, while in the case of PC, pore size and pore volume trends were slightly shifted, with the pore volume reaching a maximum for the particles with the highest surface area. This trend of surface area increase and mesopore formation was confirmed by the evident morphological change of the particles as observed by TEM for the CTAB/cyclic carbonate 1:3 ratio (Figure c,f), showing extremely regular and monodispersed starburst morphologies likely containing a large number of mesoporous cavities between the silica fibers projecting from the solid core. A comparable starburst morphology was observed in the past by using EtOH as the cosolvent, although by using a much larger CTAB/EtOH ratio of 1:1260, leading to low monodispersity and larger particle size. Additionally, the obtained morphology is reminiscent of the fibrous silica nanoparticles reported by Polshettiwar et al., although those particles were generally larger in size (d∼200–500 nm) and coreless. The results in Figure show that a comparable starburst morphology could not be observed with any of the other cosolvents, which generally led to formation of larger particles with irregular oval shapes. Further increase in the concentration of cyclic carbonates above the optimal value, i.e., from GC 1:5 and PC 1:3, resulted in some decrease in the MSN surface area (Figure a,b) but without significantly altering pore volume and size. This effect is likely due to the slight increase in particle size (especially for the case of GC, Figure d, and to a lower extent for PC, Figure g) while still maintaining the starburst morphology. In the case of Gly, a distinct surface area trend was observed (Figure c), with the surface area rising suddenly by the addition of small amounts of Gly (Gly 1:0.5) and nearly doubling compared to the value obtained in its absence (from 600 to 1119 m²/g), finally reaching a very high surface area of 1314 m²/g for Gly 1:1. Gly was used in the past as a templating agent of the synthesis of MSNs, leading to a remarkable increase of surface area compared to the material produced in its absence. Due to the presence of multiple H-bonding hydroxy moieties, Gly is likely to drive aggregation of silica condensation nuclei, leading to the formation of narrow wormhole-like channels with a 3D interconnected morphology upon its removal. Therefore, the strong increase of surface area by addition of Gly is in agreement with the pore size distribution in Figure S9f, showing the appearance of abundant sub-5 nm mesopores and micropores, with the pore size and volume reaching a minimum for the sample with the highest surface area (Gly 1:1, Figure c). A further increase in the concentration of Gly had the effect of reducing particle size (Figure S7) but led to a decrease in the surface area (Figure c) due to a general decrease in microporosity. For Gly content above the Gly 1:5 ratio, the particle size rapidly increased along with the amount of micropores, causing a new increase of the surface area to about 900 m²/g. As in the case of cyclic carbonates, especially GC, the pore size and pore volume for the Gly-based particles reached a maximum for the MSNs with the lowest surface area, while the minimum pore size and pore volume were observed for the particles with the highest surface area. However, these situations occurred at different CTAB/cosolvent ratios for cyclic carbonates and Gly. Moreover, the TEM micrographs in Figure k–m show that Gly generally provided larger, irregularly shaped particles compared to PC and GC without starburst morphology. For EtOH and CyP, the surface area and pore geometry trends followed less regular profiles (Figure d,e). The surface area tended to increase with the addition of small concentrations of alcohol cosolvents, reaching a maximum around 1000 m²/g for a 1:0.5 CTAB/cosolvent ratio, corresponding to a minimum in pore size in analogy to what was observed with Gly. Above this optimal ratio, the surface area tended to decrease by following an irregular trend due to the increase in particle size (Figure cd) with variable porosity (Figure S10d,e). The morphology of the EtOH-based MSNs (Figure h–j) generally displayed a relatively large elliptical geometry and low monodispersity.

Based on the observations made in this work and on previous literature reports, a plausible mechanism for the formation of MSNs in carbonate-based media is reported in Figure . MSN formation in alcoholic aqueous media (Figure a) typically starts from CTAB micelles, which serve as initial sites of silica nucleation by interacting with the negatively charged silicate molecules produced by the hydrolysis of TEOS by the base. Such a process leads to the formation of condensation nuclei that eventually lead to the porous particle with cylindrical channels. Because alcohol provides weaker H-bonding than water, the micelles and condensation nuclei are likely coordinated by water, which participates in the silica condensation process (TEOS hydrolysis). To note, among the selected cosolvents, only cyclic carbonates have a polarity comparable to water (Figure , ε_r of water is ∼ 80 at r.t.) and can, therefore, compete with water for binding at positively charged sites even when used in relatively low concentrations. Moreover, different from cyclic carbonates, less polar cosolvents such as EtOH and CyP can be incorporated in the micelle, leading to its swelling, thus increasing the size of the final particles. At low cyclic carbonate concentrations, we observed the formation of MSNs with no micro- or mesoporosity, with tiny size and high yields with respect to higher cyclic carbonate concentrations. These observations indicate the presence of abundant nucleation sites, but a different aggregation mechanism compared to traditional MSNs prepared in the presence of alcohols. Indeed, in carbonate-based media, the positively charged portion of the micelle, which is exposed to the polar solvent, strongly interacts with the highly polar cyclic carbonate through its partially negatively charged carbonyl oxygen (Figure b). The produced condensation nuclei are also expected to coordinate cyclic carbonate molecules via H-bonding between the abundant surface silanols and the cyclic carbonate carbonyl. The aggregation of the latter nuclei is hindered by cyclic carbonate coordination that allows only the growth of dense particles by inserting individual silicate molecules until their nearly complete consumption. As observed, increasing the concentration of cyclic carbonate led to a strong drop in MSN yield, to a slight rebound in size and significant improvement in porosity, which is likely due to the occurrence of fibrous morphologies rather than to the formation of more porous cores. Under such conditions, the micelles and silica condensation nuclei are likely to be fully coordinated by cyclic carbonate molecules. At the micelle level, the cyclic carbonate may even assist CTAB dissolution in the aqueous solvent by coordination (Figure c), leading to a drop in available nucleation sites.

12.

(a) Typical mechanism of MSN formation in traditional water/alcohol media. Proposed mechanism of MSN formation in the presence of a cyclic carbonate cosolvent: (b) low cyclic carbonate concentration and (c) high cyclic carbonate concentration.

These effects are expected to lower the MSN yield while allowing the formation of fewer and slightly larger particles by decreasing the number of available seeds (micelles and aggregation nuclei). In the presence of high concentrations of unreacted TEOS, such slightly larger particles allowed the further condensation of silica in the form of fibers or shafts radiating from available spaces on the surface core. Interestingly, the proposed mechanism justifies the difference in behavior between the less polar cosolvents that led to morphologies similar to those observed without cosolvents and cyclic carbonates that can provide strong coordination at the embryonic stage of particle growth. Through a main interaction between the micelle or condensation nuclei and the polar carbonyl group, the proposed mechanism also justifies the apparent lack of difference in behavior between glycerol carbonate (a cosolvent containing a hydroxy group) and propylene carbonate, which, different from the other media, does not include any hydroxy group.

Conclusions

Mesoporous silica nanoparticles (MSNs) are a well-established class of materials for various technological applications. In particular, MSNs with regular morphology and size in the 30–70 nm range are highly attractive. Current methods to synthesize such particles generally rely on organic bases such as TEA that, as volatile, toxic, and flammable chemicals, pose significant safety and health risks. In this work, we show that MSNs with tunable size in the 40–90 nm range can be accessed in the presence of well-defined amounts of cyclic carbonate cosolvents using NaOH as a nonvolatile inorganic base. The cyclic carbonate-based MSNs displayed a starburst morphology with silica spikes radiating from a central core, giving rise to microporous and mesoporous channels. In particular, monodisperse MSNs prepared using propylene carbonate as the cosolvent were obtained in moderate yields and with a tunable size in the 40–60 nm range. A systematic comparison of cosolvents including several organic alcohols showed that, when using the same range of cosolvent concentrations, only cyclic carbonates could produce MSNs with a starburst morphology and particle sizes below 80 nm. The reason for this observation was attributed to the strong interaction between micelles and/or condensation nuclei with highly polar cyclic carbonates partially inhibiting excessive growth and aggregation, especially for higher cyclic carbonate concentrations. The formation of the starburst morphology is likely due to the growth of silica from the surface of preformed cores from sites with a lower cyclic carbonate coverage. Importantly, particles with a fibrous morphology have been successfully employed for drug and gene delivery. , Their dendritic morphology allowed for high-pore accessibility and efficient payload delivery. However, such studies generally involved particles with relatively large sizes (d = 500 nm or higher), while there is general consensus that the optimal particle size for drug delivery in terms of efficiency and bioavailability is below 100 nm and often in the 30–70 nm range. , The particle internalization mechanism is also size-dependent, with faster adsorption for smaller particle sizes. , Therefore, MSNs prepared in this work using cyclic carbonate cosolvents combine fibrous starburst morphology with a suitable size range (d = 40–60 nm) and could serve as attractive vectors for drug delivery. Besides, a careful comparative study including several cosolvents carried out in this work led to the observation that the addition of small stoichiometric amounts of glycerol, a waste product of the biodiesel industry, boosted the MSN surface area to over 1300 m²/g. Finally, it is worth noting that cyclic carbonates are nonvolatile compounds that are regarded as green and sustainable chemicals on the basis of their CO₂-based footprint. , In the present study, the use of cyclic carbonates as cosolvents not only allowed for the preparation of functional nanomaterials but also expanded the portfolio of applications available to CO₂-derived materials as a strategy to promote CO₂ utilization. ,

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c03488.

• General information on materials and characterization methods, supporting FTIR spectra (Figure S1), XRD diagrams (Figure S2), particle size histograms for each type of MSN (Figures S3–S7), BET adsorption isotherms (Figure S8), pore size distributions of MSNs (Figure S9), and comparative table against literature-reported MSNs (Table S1) (PDF)

The authors declare no competing financial interest.