Introduction
Alumina-based ceramic membranes are used as filtration media under demanding thermal, chemical and mechanical conditions in various industrial operations. Membranes produced from nanofibre assemblies, having a structure with considerable mechanical strength, have good potential for ultra-fine scale (mechanical) filtration, in view of expected pore dimensions (in the 10-50 nm range) [1]. NAF membranes have previously shown to be promising to act as scaffolds to produce hybrid membranes, via sol-gel route by depositing silicate within them [2]. Thus membranes made of fibrous assembly can offer potential as scaffolds for the support of fine photo-active particles or coatings, and are characterised by certain parameters, such as fibre diameter, porosity, pore size distribution and fibre network architecture. For efficient water treatment, photocatalytic membranes should allow high fluxes of incoming liquid to be in contact with the photo‑active surfaces without clogging, and possess high photocatalytic effectivity with sufficient structural robustness. For example, there is interest in the photo-catalytic effect of fine titania partcles in breaking down impurities in water, using either UV or visible light. When highly dispersed in suspension, TiO2 particles can show appreciable catalytic activity. However, their subsequent removal or recovery from the treated effluent (for reuse) is difficult. Therefore, using these particles in a substrate or membrane may offer an attractive combination of separation and degradation of the pollutants and promote their flexibility of application, even though their catalytic surface area may be lower than that of the particles in suspension [3].
Fibres Used
The nanoscale alumina fibres (NAF), supplied by Metallurg Engineering (Estonia), are all single crystals of γ-alumina having a diameter of the order of 10-20 nm (Fig.1). These are produced by a novel processing route termed “Controlled Oxidation of a Melt to produce Metal Oxide Nanofibres (COMMON)”, in large quantities (kg h-1 range), thus making them attractive for membrane production.
Fig.1 (a) Photograph of as-grown slab of NAF, (b) SEM micrograph, showing individual fibre alignment, and (c) TEM micrograph of a single fibre [1].
Commercially-available Saffil® alumina fibre (SAF), produced by Saffil Limited, UK, are polycrystalline δ–alumina with diameter 3-6 μm. It is supplied in bulk, with an appearance of fine fibre wool as shown in Fig. 2.
Fig.2 (a) Appearance of as-received Saffil® wool, and (b) SEM micrograph showing fibre alignment.
Due to the availability of these fibres, especially the high production rates (kg h-1 range) and low cost of NAF, these membranes offer potential for large scale application.
Production of Membranes via SONAL Process
Membranes were produced by depositing fibres from aqueous dispersion on a mesh support, followed by compression, drying & sintering of the fibrous network, a process termed ''Sedimentation of Nanofibres from Aqueous Liquid (SONAL)'', as shown in Fig. 3. By controlling the stirring process and choice of support mesh, the extent of fibre retention could be varied and the fibre architecture within membranes could be controlled as well, from finely dispersed fibres to fibre microbundles [2].
Fig.3 An illustration of SONAL process for manufacturing membranes.
Photocatalytic Hybrid Membranes using Titania-based Sol-gel Coating
Hybrid membranes have been made by combining NAF and they were impregnated with titania-based sol-gel coatings to produce photocatalytic membranes. After heat treatment, TiO2 was found to be deposited on the fibrous network. Fig. 4 shows the steps involved in the sol-gel application.
Fig.4 Steps of sol-gel coating on substrate membrane.
As the deposition of TiO2 was quantified, a range of weight increment could be achieved by applying appropriate amount of sol to the support layer. The specific surface area and flow properties of the coated membranes were assessed and photocatalytic potential was measured by studying rates of degradation of aqueous dye solution. SEM micrographs of a hybrid membrane with sol-gel deposited TiO2 of 10% of incremental weight is shown in Fig. 5.
Fig. 5 SEM micrographs at different magnification of the same membrane having a 10% TiO2 coating, showing the ranges of pores and flow channels: (a) larger inter-microbundle and inter-layer gaps and (b) Finer inter-fibre spaces within any NAF-network.
Photocatalytic Hybrid Membranes with ‘Active Top Layer’ containing Titania Nanoparticles
Combined mixture of fibres and nanoparticles sedimented upon the pre-sedimented fibrous base layer onto the support mesh, followed by heat treatment produced hybrid membranes with ‘Photo-active’ fibrous top layer. Fig. 6 gives an illustration of the technique.
Fig.6 Designing membranes with ‘Photo-active top layer’ using TiO2-nanoparticles.
Two types of TiO2 photocatalyst were used, one a commercially available white-anatase nanopowder (25 nm) and the other silver-coated anatase produced via modification of the first [4]. Optimisation of the nanoparticle loadings was performed via assessing their photocatalytic efficiency. Specific permeability values were obtained experimentally and by prediction from the pore architecture. Fig. 7 shows the membranes with their respective SEM micrographs below their photographic images.
Fig. 7 Membranes with ‘Photo-active Top Layer’: (a) with white-anatase, (b) with Ag-coated anatase, and SEM micrographs of membranes with 20 wt% of (c) white-anatase and (d) Ag-coated anatase with an inset image (scale bar 500 nm) of dispersed nanoparticles within NAF-network.
Photocatalytic Cast Membranes
A novel form of photocatalytic membranes were produced via direct casting of a mixture of milled fibre and TiO2-based sol‑gel in circular moulds of desired diameter followed by drying and heating at higher temperature for TiO2 deposition within the membrane structure. Effects of fibre milling time and fibre to sol-gel ratio (g/ml) on their performance were studied, besides porosity and specific surface area. Their flow properties and photocatalytic efficiency were also examined.
Fig. 8 (a) Appearance of a photo-active cast membrane , and (b) SEM micrograph showing the fibre network with an inset image showing fibre surface coated with deposited TiO2 (scale bar 10 μm).
Photocatalytic Assessment of Membranes
A laboratory scale integrated cross-flow system (rig) has been designed to study the photocatalytic efficiency of the membranes, which consisted of a photo-catalytic membrane reactor or photo-reactor, a dye reservoir, a micro pump and a UV-Vis spectro-photometer. The photo-reactor had a quartz glass window on one side through which the ‘Photo-active’ side of the membranes were exposed to the light. Based on dye degradation experiments; the photocatalytic efficiency of the catalytic membranes was measured. A flow diagram of the rig is shown in Fig. 9.
Fig 9. Flow diagram of the integrated cross-flow (rig) set-up for the measurement of photocatalytic degradation, fluid (dye solution) flow or flux performance and pressure gradient through the membranes under study. (Image Courtesy: Michael Coto, Materials Chemistry Group, Department of Materials Science and Metallurgy, University of Cambridge.)
Spectroscopic assessments were done by using the ultraviolet to visible region (250-800 nm) and the kinetic efficiency of photocatalytic reactions were measured using aqueous solution of Methylene Blue dye. During dye degradation in photocatalysis, the respective absorbance value of the dye goes down with time as the concentration declines and the quantitative measurement could be done following an equation based on the Langmuir-Hinshelwood Kinetics (Eqn. 1), where C0 and C are the concentration of the dye at the beginning of the reaction and after a given time (t) and k is the pseudo first order rate constant for the photocatalytic degradation, which is often used for comparative study [5, 6].
-kt = ln (C/C0) (1)
During the photo-catalytic experiments, as the dye solution passes through the cuvette connected in the flow circuit, the absorbance value is recorded by the spectrophotometer as a function of the chosen wavelength (here, 660.032 nm) against the reference value of the solvent blank (DI water). Dye adsorption effects within the membranes which go against degradation, were taken into consideration by following the liquid recirculation under dark conditions for 4 hours. This confirmed the completion of adsorption before the experiments with illumination started for a further 7 hours of photo-degradation. The OceanView software was used to record automatically the absorbance values of the liquid at intervals of 15 minute.
Membrane Permeability
Considering the fluid flow through a porous medium to be laminar and no chemical reactions to occur, a linear relationship between the flow rate and the pressure drop is followed, known as Darcy’s law. The volumetric flow rate, Q ( in m3 s-1) is stated as,
Q = KA (dP/dx) (2)
Where A is the cross sectional area, K is the permeability (m2 Pa-1 s-1) and dP/dx is the pressure gradient across the porous medium. For most cases, this gradient can be taken as ΔP/h, h being the thickness of the medium. This is more usefully expressed as
q = (к/μ) X (dP/dx) (3)
Where к is the specific permeability (in m2), μ is the viscosity of the fluid (in Pa s) and q is flux (in m3 m-2 s-1). Using the thickness of the membrane and the viscosity of the liquid (8.9 × 10-4 Pa s) the specific permeability (using Eqn. 3) was obtained.
A multi-scale phenomenon of fluid flow is expected to occur through the hybrid membrane structures, where the specific permeability (к) of membranes is majorly contributed by the large gaps between fibre-microbundles while a homogeneous flow could take place within the fine inter-fibre spaces within any microbundle as well (Fig. 10).
Fig. 10 Schematic of fluid flow occurring through the membranes.
Acknowledgements
This work has been supported by IDB-Cambridge Commonwealth, European and International Trust, Metallurg Engineering and Fitzwilliam College. Thanks to Dr. M. Coto (The Materials Chemistry Group) for the TiO2 nanoparticles.
[1] V. Su, M. Terehov, and T. W. Clyne, "Filtration Performance of Membranes Produced Using Nanoscale Alumina Fibers (NAF)," Advanced Engineering Materials, vol. 14, pp. 1088-1096, 2012.
[2] V. M. T. Su and T. W. Clyne, "Hybrid Filtration Membranes incorporating Nanoporous Silica within a Nanoscale Alumina Fibre Scaffold," Advanced Engineering Materials, vol. 18, pp. 96-104, 2016.
[3] G. Balasubramanian, D. D. Dionysiou, M. T. Suidan, V. Subramanian, I. Baudin, and J. M. Laîné, "Titania powder modified sol-gel process for photocatalytic applications," Journal of Materials Science, vol. 38, pp. 823-831, 2003.
[4] M. Coto, G. Divitini, A. Dey, S. Krishnamurthy, N. Ullah, C. Ducati, et al., "Tuning the properties of a black TiO2-Ag visible light photocatalyst produced by a rapid one-pot chemical reduction," Materials Today Chemistry, vol. 4, pp. 142-149, 2017/06/01/ 2017.
[5] I. K. Konstantinou and T. A. Albanis, "TiO2-assisted photocatalytic degradation of azo dyes in aqueous solution: kinetic and mechanistic investigations: A review," Applied Catalysis B: Environmental, vol. 49, pp. 1-14, 2004.
[6] K. V. Kumar, K. Porkodi, and F. Rocha, "Langmuir–Hinshelwood kinetics – A theoretical study," Catalysis Communications, vol. 9, pp. 82-84, 2008.
