Letter pubs.acs.org/NanoLett
Nucleation and Growth of Monodisperse Silica Nanoparticles Camille C. M. C. Carcoueẗ ,†,¶ Marcel W. P. van de Put,†,‡,¶ Brahim Mezari,§ Pieter C. M. M. Magusin,§,⊥ Jozua Laven,† Paul H. H. Bomans,†,‡ Heiner Friedrich,†,‡ A. Catarina C. Esteves,† Nico A. J. M. Sommerdijk,†,‡ Rolf A. T. M. van Benthem,†,∥ and Gijsbertus de With*,†,‡ †
Laboratory of Materials and Interface Chemistry, ‡Soft Matter Cryo-TEM Research Unit, and §Laboratory of Inorganic Materials Chemistry, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands ∥ DSM Ahead Performance Materials B.V., P.O. Box 18, 6160 MD Geleen, The Netherlands S Supporting Information *
ABSTRACT: Although monodisperse amorphous silica nanoparticles have been widely investigated, their formation mechanism is still a topic of debate. Here, we demonstrate the formation of monodisperse nanoparticles from colloidally stabilized primary particles, which at a critical concentration undergo a concerted association process, concomitant with a morphological and structural collapse. The formed assemblies grow further by addition of primary particles onto their surface. The presented mechanism, consistent with previously reported observations, reconciles the different theories proposed to date. KEYWORDS: Monodisperse silica particles, nucleation and growth, silica chemistry, cryo-TEM, solid state NMR
T
including the two opposite ones, that is, growth by monomer addition17 or by aggregation,18 comprehensively explains all experimental observations. Here, we demonstrate using high-resolution cryogenic transmission electron microscopy (cryo-TEM), pH monitoring, and 29Si magic angle spinning nuclear magnetic resonance spectroscopy (29Si MAS NMR) that the formation of monodisperse silica nanoparticles proceeds through the synchronized association of uniform 2.3 nm primary particles with exceptional colloidal stability to larger units, labeled from now on as assemblies. We show that growth of the assemblies occurs through the continuous addition of primary particles to their surface and that the primary particles collapse upon association, both in terms of morphology and chemistry. These results prompt us to propose a formation mechanism for monodisperse silica nanoparticles that is consistent with previously reported observations in various experimental systems and reconciles the different theories proposed to date. Monodisperse silica particles were synthesized according to the procedure of Yokoi et al.,10 which involves the slow diffusion of tetraethyl orthosilicate (TEOS, final silica concentration: 16 mg/mL) into a stirred aqueous solution of lysine (see Materials and Methods in the Supporting Information). After 24 h, cryo-TEM showed the formation of highly monodisperse, near-spherical assemblies with irregular surfaces (diameters of ∼24 ± 2.8 nm) consisting of small 1.3 ±
he nucleation of solids from solution is of great importance in many geological, biological, and industrial processes and has been a topic of investigation for many decades. Classical nucleation theory describes the stochastic and dynamic association of ions or molecules forming initial nuclei which above a critical size can grow out to form the solid phase. For several systems, though, evidence has been presented for multistep nucleation mechanisms involving nanometer-sized species identified as prenucleation clusters, prenucleation complexes, or primary particles.1−5 However, the investigation into the details of these processes has mainly focused on the formation of crystalline solids. Amorphous silica (SiO2) is one of the most abundant materials in nature, produced in many forms by sponges, diatoms, and plants6 and is widely used in industrial applications, for example, in coatings, catalyst supports, abrasive agents, and so forth.7 While in nature silica is formed via the condensation of silicic acid, most synthetic silica is produced through the hydrolysis of alkoxysilanes and the concomitant condensation of the hydrolysis products.8 Of particular interest are the highly monodisperse particles that can be obtained by the so-called Stöber method9 as well as by other procedures.10 However, despite many years of investigation the mechanisms of nucleation and growth that lead to formation of particles with such narrow size distribution have still not been fully resolved. Scattering experiments11−14 and simulations15,16 have convincingly shown that the early stages of formation involve primary particles of ∼2 nm in diameter,8 but no consensus has been reached on the mechanistic steps that lead to mature particles. To date, none of the different mechanisms proposed, © XXXX American Chemical Society
Received: December 9, 2013 Revised: January 31, 2014
A
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 1. Morphology of the silica assemblies. Cryo-TEM images of the suspension taken after (a) 1 day and (b) 1 year aging. (c) Conventional TEM image of an assembly showing the structural integrity of the assembly upon drying, despite further densification of the bound particles. The white arrows point at some bound primary particles inside the structure of the larger assemblies. (d,e) Numerical cross sections through an assembly in the dry state obtained by electron tomography. Scale bars, 10 nm; inset in (a), 100 nm.
0.3 nm primary particles (Figure 1a). The assemblies remained stable in time (up to 18 months) with no observable changes in their internal morphologies, that is, primary particles remain clearly recognizable inside the assembly (Figure 1b). Conventional TEM showed only minor morphological changes in the assemblies upon drying, apart from a densification of the primary particles, underlining their structural integrity (Figure 1c, Supporting Information Tables S1, S2). 3D analysis of the assemblies by electron tomography confirmed that they consist of primary particles (Figure 1d-e, Supporting Information Movie S1). The progress of the reaction was monitored by measuring the pH of the solution (Figure 2a) and by cryo-TEM imaging at different time points. The initial pH of 9.1 measured directly after the addition of TEOS rapidly dropped to 8.6 due to its conversion into silicic acid. Concomitant with this pH decrease, cryo-TEM revealed the formation of monodisperse particles with diameters of 2.3 ± 0.5 nm, which increased in number but not in size (Figure 2c-d), as to which we refer as primary particles. After 40 min, the pH suddenly increased and over the course of 15 min reached pH 8.7 after which it slowly decreased again to show a second, more gradual, increase at 15 h, stabilizing at pH ∼ 8.5 (Figure 2a). Cryo-TEM showed that the pH jump at 40 min corresponds to the sudden association of the primary particles to form assemblies with diameters of 5.1 ± 1.3 nm on average (with the smallest observable ones being 3.5 nm). The assemblies grew further in time with “bound” primary particles still being discerned within their structure, although their diameter had decreased to 1.3 ± 0.3 nm (Figure 2e-h, Supporting Information Table S1). The presence of “free” primary particles in solution in all stages implies that they are produced throughout the entire process and are the precursors
of the growing assemblies, while the size difference between free and bound primary particles indicates a collapse of their structure upon association. The second increase in pH after 15 h is ascribed to the ongoing condensation of silicic acid monomers and oligomers, only observable after the hydrolysis reaction has consumed all the TEOS in the system and no more acidic species are being generated. When the reaction was performed at a reduced silica concentration (0.54 mg/mL), a similar pH decrease was observed but no pH jump occurred. Cryo-TEM showed a decreased concentration of primary particles that were stable and could be stored for days, implying that under these conditions the critical concentration for the formation of assemblies was not reached. However, upon increasing the silica concentration of the suspension to 25 mg/mL through controlled water evaporation, formation of assemblies with diameters of ∼7 nm was observed (Supporting Information Figure S1). The stability of the assemblies was evaluated by aging two suspensions, each containing two different size populations (5 and 15 nm, respectively, 15 and 20 nm). Cryo-TEM images indicated colloidal stability over a period of 3 months with no observable changes in morphology (Supporting Information Figure S2), indicating that exchange of primary particles, addition of soluble silicate oligomers and Ostwald ripening do not play a significant role. Reference experiments using ammonia and sodium hydroxide, also starting at pH ∼ 9.0, showed that the association process was not specific for the use of lysine (Supporting Information Figure S3) and that no specific additives are involved in the stabilization of either the primary particles or the assemblies. When the reaction was carried out in pure water B
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 2. Different stages observed during the reaction and schematic representation of the proposed formation mechanism. (a) pH monitoring of the complete reaction (final silica content: 16 mg/mL). (b) pH monitoring during primary particle formation (final silica content: 0.54 mg/mL). Cryo-TEM images of the suspension after (c) 30 min, (d) 40 min, (e) 45 min, (f) 2 h, (g) 7 h, and (h) 24 h. The white arrows in (g) points at a free primary particle. The open, diffusively charged cloudy structures densify to free primary particles (Stage I, primary particle formation) that collapse upon association into assemblies (Stage II, association of primary particles into assemblies). After association, no further new assemblies are formed and growth occurs via the addition of primary particles (Stage III, growth of assemblies). Scale bars: 10 nm.
Figure 3. Smoothening of the silica assemblies. (a,b) Cryo-TEM images and (c) conventional TEM image of the silica assemblies taken at various time points of the growth process in water. The white circles are a guide for the eye to visualize the smoothening of the assemblies while growing. (d) Conventional TEM image of silica assemblies prepared in ethanol. Scale bars: 10 nm.
of pH ∼ 6.5, similar primary particles were formed but under these conditions the primary particles were not stable and started to associate immediately; aggregates with undefined morphology resulted, demonstrating the importance of a basic pH in controlling the surface charge of the primary particles and thereby their association into monodisperse assemblies (Supporting Information Figure S3). Moreover, monodisperse assemblies (diameters ∼7 nm) were also obtained by titrating a sodium silicate solution into water of pH 8.8 (Supporting Information Figure S4). The assemblies produced in water showed a coarse surface structure compared to the smooth appearance of silica nanoparticles prepared in alcohol-based media (Stö ber method) (Figure 3). We attribute this effect to the difference in the silica condensation reactions that occur. While in water,
the condensation of two SiOH groups results in an essentially irreversibly formed Si−O−Si bond; in an alcohol-based solution and with incomplete hydrolysis, condensations are less likely to occur, and the alcoholate (SiOR) and hydroxyl (SiOH) groups can exchange reversibly,19 thereby creating a relaxation mechanism for further smoothening. Our observation of well-defined primary particles with diameters of 2.3 ± 0.5 nm is in line with the consensus that at pH > 9, the hydrolysis of TEOS and the subsequent condensation of the generated silanol groups lead to charged, highly hydrated, branched low density oligomeric structures20,21 that above a critical size become insoluble, leading to their densification through expulsion of some of the water contained in their structure (Figure 2, stage I). Their defined size suggests a local thermodynamic minimum that limits their growth. It has C
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 4. Collapse of the primary particles after association. (a) Single pulse solid state 29Si NMR on freeze-dried samples of both the primary particles (t1) and the nanoparticle assemblies just after association (t2) and at the end of the reaction (t3). (b) Cryo-TEM image of the primary particles (t1) with (c), the particle size distribution. (d) Cryo-TEM image of the final assemblies (t3) with (e), the particle size distribution of the assemblies and their constituting primary particles. Scale bars: 10 nm.
stage III) then must occur through a surface reaction limited process which allows the primary particles to preferentially react at “hot spots”, so that upon occupation of such a site by a primary particle, homogeneity as well as surface smoothness of the growing assembly increases (Figure 3). In addition and reported here for the first time, the observed difference in size between the free (2.3 ± 0.5 nm) and the bound (1.3 ± 0.3 nm) primary particles, representing a volume factor difference of at least 5, suggests a further densification of the silica network structure upon association, labeled as collapse (Figure 2e). Direct-excitation 29Si MAS NMR of freeze-dried solutions was used to provide quantitative information on the cross-linking of the silica network through the ratios between 4fold (Q4) and 3-fold (Q3) interconnected silicon atoms (Figure 4a). Comparing the Q4/Q3 ratios of the primary particles formed at a silica concentration of 0.54 mg/mL (Q4/Q3 = 0.9) with those of the assemblies obtained just after the association (Q4/Q3 = 1.4) and after 24 h (Q4/Q3 = 1.6) indicated that the association of primary particles together with their further collapse occurs concomitantly with an increase in cross-linking. Interestingly, elemental analysis performed on these samples revealed that the carbon content throughout the reaction remains the same, suggesting that the hydrolysis is almost complete at all times (Supporting Information Table S3). Further, the fact that no free primary particles of 1.3 nm diameter are observed in suspension shows that the collapse of the primary particles occurs exclusively upon their association into assemblies or during their addition onto the assembly surface. These observations imply that the collapse is driven by condensation reactions and the extra cohesive interactions (including next-nearest neighbors) upon close approach of the primary particles and must be associated with the expulsion of water from the highly hydrated free primary particles. Moreover, while the size of a free primary particle is mainly determined by surface charge, the size of a bound one is determined by the balance of network flexibility versus the extra cohesive forces given the silica content of a primary particle. Silica formation has been extensively investigated and two extreme models emerged, that is, the “monomer addition model”17 and the “aggregation growth model”.18 In the monomer addition model, identical primary particles are nucleated within a short time window after which they grow
been suggested that for SiO2 particles with a size of a few nanometers to be stable, a high surface charge is required,12,22 which leads to a high estimated surface potential ψ0 of about −120 to −170 mV. However, charge-potential considerations are usually based on the flat plate Gouy−Chapman solution of the Poisson−Boltzmann equation without considering curvature even though the latter has a strong influence.23,24 For the 2.3 nm diameter primary particles, we measured the zeta potential ζ ≅ −71 mV, while for 13 nm diameter assemblies ζ ≅ −42 mV was obtained, consistent with the aforementioned surface charge considering curvature and well within the range normally obtained at that pH25 (Supporting Information Section 1). We therefore propose that during the densification of the oligomeric structures, the majority of the charged silanolate groups become located at the particle surface. This will create thermodynamically stabilized primary particles of a defined size that balance their volumes and surface charges and have a strongly repulsive electrical double layer that maintains their stability in solution up to the point where a critical aggregation concentration (CAC) is reached and the association to assemblies occurs (Figure 2, stage II). The formation of the assemblies with increasing monodispersity (Supporting Information Section 3) and persistent internal structure implies that (i) association is confined to a small time window and (ii) the subsequent growth occurs essentially by addition of individual primary particles. For this, we propose a mechanism in which the electrostatic repulsive forces of the primary particles allow a build-up of a high concentration of primary particles in solution through the continuous generation of silica oligomers, which at nucleation rapidly drops below the CAC. After this point, the chemical potential is no longer high enough to overcome the repulsive barriers between two primary particles; however, it still allows them to add on to the less repelling surfaces of existing assemblies. From this point on, the number density of assemblies in the suspension remains constant; no new assemblies are formed. The absence of further nucleation also implies that the rate of generation of primary particles through the hydrolysis of TEOS is not higher than the rate of addition onto existing assemblies. (For further considerations for stage II, see Supporting Information Section 3). Growth (Figure 2, D
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
■
further by the addition of silicic acid monomers, preserving monodispersity throughout the whole process.17 In the contrasting aggregation growth model, primary particles with sizes ranging from 1 to 5 nm are generated in solution over a longer period, during which they coagulate to form larger, more stable particles, leading to a narrow size distribution at the later stages of the reaction.18 Other models involve a controlled aggregation process followed by growth through monomer addition, leading to smooth colloidal surfaces.26,27 Strikingly, most experimental data obtained from both ethanolic and aqueous systems on which these different models were based are in line with the mechanism proposed here. Light scattering has demonstrated that the number of silica nanoparticles present in the suspension is constant during the growth process, placing nucleation only in the early stages of the reaction.17 This matches well with SAXS data that also showed that after a dramatic decrease in scattering objects (nucleation) after the first stages of the experiment the number density of growing spheres remains constant.14,28 In addition, conductivity measurements indicated that the concentration of silicic acid was unaffected by the number density of the nanoparticles, implying that silicic acid does not directly condense onto the silica particles12 but that growth must occur through addition of primary particles. Conductivity measurements further showed two different regimes in which first a fast decrease of the conductivity was attributed to the formation of oligomeric species, while the less steep decrease after the CAC matches the continued condensation and primary particle based growth of silica in solution. Other studies26,29 analyzed the competitive growth of a dispersion of silica particles with a bimodal size distribution and showed that the particles grew at the same rate, demonstrating that the growth was surface reaction limited, as proposed for our mechanism (Supporting Information Section 2). Our high-resolution cryo-TEM observations combined with data from quantitative analysis techniques allowed us to clarify hitherto elusive aspects of silica formation and to present new insights into the formation mechanisms of monodisperse silica nanoparticles. The presence of primary particles in the early stage of the reaction has already been reported, but their behavior in solution, including their unusually stable nature, is now demonstrated. This study also enables us to discuss further several postulates previously proposed; the stability of the assemblies, including their internal morphology, over the course of a year, implies that the primary particles do not undergo any exchange or dissolution in time. In addition, the stability of aged mixtures of two size populations of particles excludes a significant role for Ostwald ripening. The collapse of the primary particles upon association, both in terms of morphology and chemistry, as well as their persistence in the internal structure of the assemblies, are new elements leading to a revised formation mechanism. Our results combined with previously reported observations indicate that this mechanism is largely independent of the reaction medium (water or alcohol) and the silica precursor (alkoxysilanes or sodium silicate) and applies for the most frequently applied approach for the preparation of uniform silica particles, the Stöber method, as well as the early stages of zeolite formation.30 More generally, our results imply that colloidally stabilized primary particles can be part of a more general phenomenon in the nucleation of well-defined inorganic nanoparticles.
Letter
ASSOCIATED CONTENT
S Supporting Information *
Additional information, figures, tables, and movie. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Tel.: +31 40 247 4947. Present Address ⊥
(P.C.C.M.M.) Center for Surface Chemistry and Catalysis, Catholic University of Leuven, Kasteelpark Arenberg 23 − box 2461, 3001 Leuven, Belgium.
Author Contributions ¶
C.C.M.C.C. and M.W.P.v.d.P. contributed equally to the research as described in this paper.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This study was partly supported by DSM Ahead Materials Performance B.V., The Netherlands.
■
REFERENCES
(1) Navrotsky, A. Proc. Natl. Acad. Sci. U.S.A. 2004, 101 (33), 12096−12101. (2) Pouget, E. M.; Bomans, P. H. H.; Goos, J. A. C. M.; Frederik, P. M.; de With, G.; Sommerdijk, N. A. J. M. Science 2009, 323 (5920), 1455−1458. (3) Habraken, W. J. E. M.; Tao, J. H.; Brylka, L. J.; Friedrich, H.; Bertinetti, L.; Schenk, A. S.; Verch, A.; Dmitrovic, V.; Bomans, P. H. H.; Frederik, P. M.; Laven, J.; van der Schoot, P.; Aichmayer, B.; de With, G.; DeYoreo, J. J.; Sommerdijk, N. A. J. M. Nat. Commun. 2013, 4, 1507. (4) Baumgartner, J.; Dey, A.; Bomans, P. H. H.; Le Coadou, C.; Fratzl, P.; Sommerdijk, N. A. J. M.; Faivre, D. Nat. Mater. 2013, 12 (4), 310−314. (5) Gebauer, D.; Volkel, A.; Cölfen, H. Science 2008, 322 (5909), 1819−1822. (6) Mann, S. Biomineralization Principles and Concepts in Bioinorganic Materials Chemistry; Oxford University Press: New York, 2001. (7) Giesche, H. Medical and technological application of monodispersed colloidal silica particles. In Medical Applications of Colloids; Matijević, E., Ed.; Springer: New York, 2008; pp 42−66. (8) Iler, R. K. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties, and Biochemistry; Wiley: New York, 1979. (9) Stöber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26 (1), 62−69. (10) Yokoi, T.; Sakamoto, Y.; Terasaki, O.; Kubota, Y.; Okubo, T.; Tatsumi, T. J. Am. Chem. Soc. 2006, 128 (42), 13664−13665. (11) Tobler, D. J.; Shaw, S.; Benning, L. G. Geochim. Cosmochim. Acta 2009, 73 (18), 5377−5393. (12) Rimer, J. D.; Lobo, R. F.; Vlachos, D. G. Langmuir 2005, 21 (19), 8960−8971. (13) de Moor, P.-P. E. A.; Beelen, T. P. M.; van Santen, R. A. J. Phys. Chem. B 1999, 103 (10), 1639−1650. (14) Pontoni, D.; Narayanan, T.; Rennie, A. R. Langmuir 2002, 18 (1), 56−59. (15) Malani, A.; Auerbach, S. M.; Monson, P. A. J. Phys. Chem. C 2011, 115 (32), 15988−16000. (16) Jin, L.; Auerbach, S. M.; Monson, P. A. J. Chem. Phys. 2011, 134 (13), 134703. (17) Matsoukas, T.; Gulari, E. J. Colloid Interface Sci. 1988, 124 (1), 252−261. E
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
(18) Bogush, G. H.; Zukoski, C. F. J. Colloid Interface Sci. 1991, 142 (1), 19−34. (19) Bogush, G. H.; Zukoski, C. F., IV J. Colloid Interface Sci. 1991, 142 (1), 1−18. (20) Fouilloux, S.; Tache, O.; Spalla, O.; Thill, A. Langmuir 2011, 27 (20), 12304−12311. (21) Boukari, H.; Lin, J. S.; Harris, M. T. Chem. Mater. 1997, 9 (11), 2376−2384. (22) Abbas, Z.; Labbez, C.; Nordholm, S.; Ahlberg, E. J. Phys. Chem. C 2008, 112 (15), 5715−5723. (23) Loeb, A. L.; ; Overbeek, J. T. G.; Wiersema, P. H. The Electrical Double Layer around a Spherical Colloid Particle; Computation of the Potential, Charge Density, and Free Energy of the Electrical Double Layer around a Spherical Colloid Particle; Massachussetts Institute of Technology: Cambridge, MA, 1961. (24) Lamm, G. Rev. Comp. Chem. 2003, 19, 147−365. (25) Kirby, B. J.; Hasselbrink, E. F. Electrophoresis 2004, 25 (2), 187− 202. (26) Van Blaaderen, A.; Van Geest, J.; Vrij, A. J. Colloid Interface Sci. 1992, 154 (2), 481−501. (27) Harris, M. T.; Brunson, R. R.; Byers, C. H. J. Non-Cryst. Solids 1990, 121 (1−3), 397−403. (28) Bertholdo, R.; dos Reis, F. V. E.; Pulcinelli, S. H.; Santilli, C. V. J. Non-Cryst. Solids 2010, 356 (44−49), 2622−2625. (29) Chen, S. L.; Dong, P.; Yang, G. H. J. Colloid Interface Sci. 1997, 189 (2), 268−272. (30) Davis, T. M.; Drews, T. O.; Ramanan, H.; He, C.; Dong, J. S.; Schnablegger, H.; Katsoulakis, M. A.; Kokkoli, E.; McCormick, A. V.; Penn, R. L.; Tsapatsis, M. Nat. Mater. 2006, 5 (5), 400−408.
F
dx.doi.org/10.1021/nl404550d | Nano Lett. XXXX, XXX, XXX−XXX
