Surfactant-free synthesis of mesoporous and hollow silica nanoparticles with an inorganic templatew Luca Bau`,* a Barbora Ba´rtova´, b Maria Arduini a and Fabrizio Mancin* a Received (in Cambridge, UK) 27th August 2009, Accepted 16th October 2009 First published as an Advance Article on the web 4th November 2009 DOI: 10.1039/b917561j A surfactant-free synthesis of mesoporous and hollow silica nanoparticles is reported in which boron acts as the templating agent. Using such a simple and mild procedure as a treatment with water, the boron-rich phase is selectively removed, affording mesoporous pure silica nanoparticles with wormhole-like pores or, depending on the synthetic conditions, silica nanoshells. Mesoporous silica has been attracting increasing attention since the synthetic procedure based on surfactant templating was devised in 1992 by Beck and co-workers. 1 By virtue of its high surface area and ease of functionalization, 2 mesoporous silica is a unique material that holds promise for a host of applications, ranging from catalysis 3 to sensing 4 and, more recently, drug delivery. 5 In recent years, considerable efforts have been devoted to the synthesis of mesoporous silica nanoparticles, combining the peculiar properties of silica nanoparticles, such as water solubility, biocompatibility, optical transparency and shorter diffusion paths, with the high loading capacity of mesoporous materials. 6 The general approach involves, as in the case of mesoporous silica, the use of a templating agent, usually a surfactant, which is removed by calcination or chemical extraction after the silica network is formed. 7 However, high temperature treatments may affect the properties of the silica particles, reducing the amount of surface silanol groups, 8 and, in general, surfactant removal procedures are often incompatible with the presence of delicate payloads (drugs or other active species) inside the particles. Avoiding the use of surfactants is hence highly desirable and would be beneficial in terms of cost, environmental impact and scale-up potential. Furthermore, most synthetic procedures based on surfactant templating lead to ordered structures with non-interconnected cylindrical pores. Though undoubtedly crucial to many applications, these features are not always needed or even desirable. Other applications, like adsorption or catalysis, would benefit from the more accessible internal volume afforded by interconnected pores. Surfactant-free synthesis of mesoporous silica can be performed using organic templates, 9 but the removal procedures are similar to those used for surfactants. A different approach may involve the use of inorganic templates, as demonstrated by Asher and co-workers, 10 who were able to tailor the morphology of growing silica nanoparticles using cadmium sulfide nanoparticles as templates. However, also in this case harsh removal conditions, namely concentrated nitric acid treatment, were required. We report here a novel surfactant-free approach to the synthesis of mesoporous nanoparticles in which the templating agent is a hydrolytically unstable inorganic phase, and the template removal treatment is a simple solvent exchange with water at room temperature. Such procedure results in the preparation of silica nanoparticles with interconnected disordered pores or, depending on the reaction conditions, silica nanoshells. It has been known for several decades that silica glasses with nanometre-sized pores can be obtained by inducing phase separation in borosilicate glasses and selectively leaching the boron oxide phase. 11 We reasoned that if a similar phase separation could be achieved in silica nanoparticles, a novel route to mesoporous nanoparticles could be envisaged. In principle, such a separation should indeed occur. Silica nanoparticles may be easily prepared, according to the Sto¨ber method, 12 by condensation of alkoxysilane derivatives in ethanol–water–ammonia solutions at room temperature. The aqueous sol–gel synthesis of borosilicates, a similar procedure, has been extensively studied by Irwin and co-workers, 13 who found the formation of borosiloxane (Si–O–B) bonds to take place only upon dehydration at high temperatures. The absence of borosiloxane bonds points to the existence of boron-rich domains that should be easily removed by hydrolysis, leaving a network of pores in their place. We set out to test this hypothesis by synthesizing borosilicate nanoparticles through a modified Sto¨ber reaction, 14 involving the co-condensation of silicon and boron oxide precursors. Tetraethoxysilane (TEOS) was used as a silicon source, while boric acid was chosen as the boron precursor in order to maximize the difference in the hydrolysis rates of the two precursors, which is known to correlate with oxide network inhomogeneity in sol–gel processes. 15 The reaction was Fig. 1 TEM bright field images of borosilicate nanoparticles (a) and the same sample after water treatment (b). a Dip. di Scienze Chimiche, Universita ` di Padova, via Marzolo 1, 35131 Padova, Italy. E-mail: luca.bau@unipd.it, fabrizio.mancin@unipd.it b CIME & LSME, EPFL, Station 12, 1015 Lausanne, Switzerland w This paper is dedicated to Professor Umberto Tonellato of University of Padova on the occasion of his retirement. 7584 | Chem. Commun., 2009, 7584–7586 This journal is  c The Royal Society of Chemistry 2009 COMMUNICATION www.rsc.org/chemcomm | ChemComm