Flash microwave-assisted solvothermal (FMS) synthesis of photoactive anatase sub-microspheres with hierarchical porosity

The synthesis of nanostructured sub-microspheres of TiO2 anatase with hierarchical nano- and mesoporosity was successfully achieved by using an innovative approach that applies the principles of acidic digestion to microwave (MW) solvothermal synthesis. This process, termed flash microwave-assisted solvothermal (FMS) synthesis, facilitates the formation of spherical particles without surfactants or templating agents, exploiting the rapid reaction kinetics engendered by MW heating. Unlike many other MW-assisted solvothermal methods, the application of constant MW power leads to a rapid increase of the autogenous pressure, inducing burst-nucleation of small primary crystallites and subsequent rapid agglomeration into secondary particles, with reaction times reduced to minute-timescales. The use of non-aqueous polar solvents such as ethanol is key to the production of regular spheres with a narrow size distribution, composed of nanocrystallites. Morphology, porosity, specific surface area, phase composition, crystallite size and optical properties of the particles can be controlled via a judicious selection of physical and chemical synthesis parameters, especially precursor choice and acid concentration. The complex structure of the particles leads to surface areas of up to ca. 500 m2 g−1 with intergranular mesoporosity. The as-synthesised FMS particles show increased adsorption under dark conditions and selective de-ethylation of rhodamine B under visible light compared to a commercial photocatalyst (Degussa P25). The photodegradation mechanism hinges on the capacity of the spheres to accept electrons from the photoexcited state of molecules at the particle surface, with the large sphere surface area maximising adsorption capacity and improving the efficiency of the photocatalytic processes. The singular characteristics and properties of the particles could pave the way for further applications in water purification and optoelectronic devices.


Further details concerning the experimental methods
The size distribution of particle was calculated by measuring the particle diameters from different SEM images at different magnification. As a standard method, at least 100 particles from 5 different image frames were measured for each single sample. The calculation was performed with the software ImageJ (National Institutes of Health, US).
Raman spectroscopy was performed by using laser intensities from 1 -25 % of the maximum power in order to avoid in situ modification (e.g. heating) of the samples (e.g. phase transformations).
Raman data were analysed excluding the effect of the Rayleigh scattering and stray light, cutting the spectral region below 75 cm -1 . Baseline subtraction was then performed by using asymmetric least squares smoothing (with an asymmetric factor of 0.001, a threshold of 0.001 and a smoothing factor variable between 3 and 5 depending on the peak resolution). Baseline subtraction was performed using Origin Pro 2016. Raman peak analysis was performed by using the Fytik software (vs 0.98).
Taking the asymmetry between the two sides of the peaks into consideration, the E g I peak was fitted with a split-pseudo-Voigt function. The function provided a better fit compared to a conventional pseudo-Voigt alternative or to a combination of multiple Lorentzian-Gaussian functions. The other modes were identified through analysis of the first derivative of the spectra.
Optical band gap calculations were performed using the Kubelka-Munk function F(R): where R is the absolute reflectance and hν the photon energy. The intersection with the x-axis of the function plotted against the photon energy gives the value of the optical band gap. The exponent n depends on the nature of the optical transition, adopting n = ½ for indirect and n = 2 for direct transitions respectively. Electronic Supplementary Material (ESI) for RSC Advances. This journal is © The Royal Society of Chemistry 2020 The specific surface area (SSA) was calculated using the BET equation in the interval 0.05 ≤ (p/p0) ≤ 0.33. The pore size distribution for mesoporous samples was evaluated by implementation of the BJH equation and DFT calculations, assuming a non-linear approach on the equilibrium isotherms and assuming mixed shape for the pores on a carbonaceous matrix.
The dye concentration before and during the degradation experiment was calculated from the absorption intensity by using a calibration curve based on the linear correlation between concentration and absorption intensity described by the Beer-Lambert law. Figure S1: Example of SEM image processing for the identification of the particle size distribution by using the ImageJ software. The diameter of all circled particles is calculated from the area, assuming perfect spherical shape for each encircled particle. Grain size (nm) Figure S3: Example of the log-normal distribution of the grain size (diameter) estimated using the WPPM method for FMS-TiO 2 (2 M HNO 3 /160 mM precursor concentration; 1 min MW treatment).              Ti 3s C 1s

Figures and tables
Ti 2s Ti 2p Intensity (     HNO 3 concentration (M) Figure S18: Indirect band gap as a function of the acid concentration for the synthesis of FMS TiO 2 submicrospheres. The data are correlated with the precursor (TTIP) concentration used for the preparation of the TiO 2 particle. All samples were subjected to 1 min of MW irradiation time.