Supraparticles with silica protection for redispersible, calcined nanoparticles

Calcination of nanoparticles is always accompanied by undesired sintering. A calcination route preventing hard-agglomeration to bulk lumps, which is transferable to almost any kind of metal oxide nanoparticle, is developed by surrounding targeted nanoparticles by silica nanoparticles within a nanostructured microparticle. After calcination, the desired nanoparticles are regained as a monodisperse sol via silica dissolution.

extremely disadvantageous as hard-agglomerates cannot be disintegrated by simple means but strong mechanical forces are required to break them down to smaller entities. 10 In order to avoid sintering, NPs can be protected by a silica coating [11][12][13][14][15] or encapsulation, 16 which is applied e.g. via a Stöber process onto the particle surfaces. 17 An alternative approach is to embed NPs in a silica-based hybrid gel. 18 In all cases, the sinter-resistant silica encages the more temperature-labile core material thus preventing its sintering. 19 This has already shown success in temperature treatments up to 800 C in the case of Pt or Pd NPs used in catalyst applications. 11,14 As a coating of the NPs can be disadvantageous for their later application, a few research groups have also shown that aer heat treatment the silica coating can be removed via etching in a strong base. 20,21 Another approach to protect NPs, such as FePt, against sintering is to intermix them with a much larger proportion of NaCl microparticles. Similar to a silica coating, the NaCl encloses the temperature-labile NPs as matrix, enabling their sinter-free annealing and their retrieval by dissolving NaCl in water. 22 Based on these insights, in this current work presented herein a route toward the fabrication of calcined NPs was developed using supraparticles 23 as intermediary structures. These micron-sized, hierarchical particles consist of silica NPs and another NP type (e.g. CaF 2 or TiO 2 ), which needs to be protected from sintering during calcination. They are assembled upon self-limited self-assembly 24,25 (SLSA) combined with subsequent spray-drying. 26 This procedure affords high exibility in terms of building block selection (comparable to a toolbox), i.e., it is possible to readily select the sizes, shapes and material types of the primary particles. It consequently enables the straight forward protection of any desired NP with silica (or other sinter-resistant) NPs in a robust, continuous and upscalable process. This displays a signicant advantage over approaches using silica coatings, which have to be developed from scratch for every single targeted NP type and are difficult to upscale. In detail, careful control of wet nanochemistry on a larger scale, e.g. aiming for the production of several hundred grams to kilograms of particles, is a true challenge. Nearly any synthesis protocol has to be modied and e.g. working with highly diluted dispersions, which becomes necessary to be able to control the process, renders it impractical and unfeasible.
As a more detailed outline of this issue would be beyond the scope of this article, the interested reader is referred to a detailed discussion on this issue, which was recently published by us. 27 The obtained supraparticles resemble raspberries in their structure 28 and allow the calcination of the desired NP type. Aer calcination of the supraparticles, the (non-silica) NPs are regained as monodisperse sol via dissolution of the silica components in a strong base. Fig. 1 depicts the principle of this fabrication route using silica NPs as protectant against sintering, which is similar to foam peanuts as protectant against mechanical damage in the packaging of fragile products. However, it should be noted here that a difference between using foam peanuts for transport (as well as the mentioned approaches of embedding NPs in a gel or a salt matrix) and the herein proposed method is the formation of supraparticles instead of simply mixing both nanoparticles types. This brings the advantage of providing a free owing micropowder sample aer calcination (Fig. S1 in the ESI †) and thus, faster dissolution of the silica components due to a better accessibility compared to bulk material.  transmission electron microscopy (TEM) images (b1 and c1, white arrows indicating CaF 2 NPs) and intensity weighted agglomerate sizes determined via dynamic light scattering (DLS, b2 and c2) before and after ultrasound-assisted SLSA of the binary NP dispersion, respectively. (d) Scanning electron microscopy after forced assembly of the pre-assembled binary NP dispersion by spray-drying.
In order to optimally protect NPs by silica within a supraparticle during calcination, the spatial separation of sinterlabile NPs from each other is a crucial parameter. This inner structural arrangement of a supraparticle during the kinetic trapping is affected by the number ratio and size ratio of silica to NPs that need to be protected, as well as by the differences in surface charge of both NP types. Several publications have already studied supraparticles made of binary NP dispersions by means of spray-drying assisted assembly. 26,[29][30][31] Based on the ndings of these studies, the conclusion can be drawn that for an efficient separation of one particle type by silica, not only silica NPs have to prevail in excess, but also both NP types need to be similar in size (Fig. S2a in the ESI †). This is at least the case for equally charged NPs displaying high surface charges and a good dispersion state in water before spray-drying 29 (Fig. S2b, case I. in the ESI †). However, comparing the surface charge, i.e., the zeta potential of silica with the one of other metal oxide NPs, it appears that the metal oxide NPs usually have their isoelectric points at higher pH values than silica (shown for the samples studied in this work in Fig. S3 in the ESI †). 32 Thus, obtaining oppositely charged binary dispersions at pH values lying between the isoelectric points of both materials is most practicable in the majority of cases. This is why the concept of so-called SLSA 24 of binary oppositely charged NP  (2), intensity weighted agglomerate sizes determined via DLS (3; with c NP indicating the sample concentration for measurements), XRD patterns (4) and nanoparticle size and size distribution obtained from TEM images measuring at least 50 particles per sample (5) as well as TEM images (6). (c) and (d) depicts the same analyses, having replaced the photoluminescence spectra by photocatalytic activity tests (2). The green background of the diagrams (2) highlights the improvement of NP properties by silica-protected calcination.
This journal is © The Royal Society of Chemistry 2019 dispersions (Fig. S2b, case II. †) was connected upstream of the forced microparticle assembly via spray-drying. If the negatively charged silica NPs self-assemble around the sinter-labile NP type, the chance of an ideal separation of these NPs from each other within the supraparticle is increased. Furthermore, differences in NP size should no longer induce a segregation of the two NP types within the spray-dried supraparticles. Thus, an efficient separation of the sinter-labile NPs is guaranteed as long as the silica displays the higher particle number and the smaller particle size in order to form the outer layer during the SLSA 33 (which is shown for CaF 2 and two sizes of silica NPs in varying weight ratios in Fig. S4 and S5 in the ESI †). The process owsheet of the continuous double assembly of Eu 3+ containing CaF 2 as exemplary sinter-labile NPs with silica NPs is shown in Fig. 2a (and Experimental details can be found in the ESI †). The aqueous silica and CaF 2 sols are pumped into a static mixer in a NP weight ratio of 9 : 1. This results in a dispersion of wellseparated silica NPs of around 20 to 30 nm in size and soagglomerates of CaF 2 NPs of around 200 nm ( Fig. 2b1 and  b2). Ultrasound application enables the SLSA of silica NPs around the CaF 2 NPs yielding so-agglomerates of around 100 nm in size besides well-separated silica NPs due to their number surplus ( Fig. 2c1 and c2). Subsequent spray-drying provides nanostructured microparticles consisting of CaF 2 NPs embedded into silica NPs (Fig. 2d). This supraparticle structure enables the calcination without hard-agglomeration to bulk lumps up to at least 800 C (Fig. S6 in the ESI † in comparison to spray-dried CaF 2 NPs without silica protection).
In order to highlight the benets of the silica-protected calcination route, the properties of the obtained Eu 3+ containing CaF 2 NPs aer calcination at 600 C were compared to the as synthesized CaF 2 crystals ( Fig. 3a and b). While the rare-earth luminescence is signicantly enhanced by the calcination, the crystallite size calculated from X-ray diffraction (XRD), using the Scherrer equation, is only slightly increased from 21 AE 3 nm to 26 AE 1 nm. TEM and DLS measurements also conrm the nanoscale sizes of the single-crystalline particle samples. While the measured NP size on TEM micrographs increases from 28 to 46 nm (Fig. S7 in the ESI †), the agglomerate size around 100 nm obtained from DLS shows the dispersion state of the samples in water and indicates so-agglomeration of the nanoparticles. This so-agglomeration also complicates the particle size determination via TEM. The increase in size shown by TEM and XRD indicates a slight sintering due to coalescence. However, the nanoscale size and nanoredispersibility of the sample are maintained.
Via transfer of the silica-protected calcination route to anatase NPs (Fig. 3c) by adaptation of the silica NP size and amount, the easy portability of this process to other NP materials could be shown. Similar to the calcined silicaprotected CaF 2 NPs, the obtained single-crystalline TiO 2 sample displays nanoparticulate size; the crystallite and TEM size only slightly increase from 6 AE 1 nm to 7 AE 1 nm and from 6 to 10 nm, respectively (Fig. 3d, while calcination without silica protection results in bulk material - Fig. S7 in the ESI †). However, the photocatalytic activity is signicantly enhanced by the calcination process.

Conclusions
In summary, a silica-protected calcination route for NPs, which prevents their hard-agglomeration to bulk lumps, using supraparticles as intermediary structures was herein presented. This procedure is straight forward, upscalable and readily transferable to almost any kind of metal oxide NP system yielding a signicant improvement of properties while maintaining their nanoscale size.

Conflicts of interest
There are no conicts to declare.