The scarcity of fresh water resources has been of great concern for contemporary society, particularly against the backdrop of worsening climate change impacts and a growing global population. Desalination of saline water resources is one of the most feasible and effective technologies for fresh water production. Membrane-based desalination processes such as reverse osmosis (RO) have risen to prominence owing to their reduced energy intensity and compact footprint. However, extensive pre-treatment is required to perform RO on chemically and thermal aggressive waters, opening the door for alternative methods. In particular, membrane distillation (MD), a process which combines the benefits of both membrane and thermal technologies has received significant research interest in the last decade. Up till now, most studies on MD have focussed on the effect of operating parameters by using the commercial available polymeric membranes, which are designed for ultra- or microfiltration and may not be suitable for MD. New membrane materials and morphologies have not been extensively explored. The key target of this work therefore is to design a suitable membrane for MD by exploring the potential of hybrid organic-inorganic materials, specifically using silica-based (as inorganic) and carbon-based (as organic) materials.
Silica-based materials offer good chemical and thermal resistance, high porosity, and excellent versatility in forming various nano-sized morphologies. Unfortunately, most of the membrane-related work has focused on amorphous silica, which is sensitive to steam or water vapour and yields low water fluxes making it generally unsuitable for MD. The steam degradation can be controlled at MD relevant temperatures by incorporating organic moieties into the silica network. Meanwhile, the flux issues can be addressed by increasing porosity and pore size. This work demonstrates for the first time hybrid organic-inorganic mesoporous membranes with an ordered, narrow pore size distribution was developed by using soft-templating method and successfully applied to MD under a variety of operating conditions. Despite having a hydrophilic contact angle (i.e. < 90 °) and a pore size (2 nm) larger than hydrated salt ions (0.66-0.72 nm) which intuitively may lead to pore wetting, the membrane produced pure water (up to 13 L m-2 h-1) with > 99 % salt rejection across an extreme range of salt concentrations (10-150 g L-1 NaCl) at moderate temperature (60 °C). This major finding was complimented by the fact that no concentration polarization was observed, with fluxes effectively unchanged across the entire range of salt concentrations.
Based on these results a model was proposed to explain how a hydrophilic, nanoporous membrane could operate effectively and with no observable pore wetting under vacuum MD. The model represents the second major contribution of the thesis and adapts the Lucas-Washburn equation for capillary pressure to nanopores. The model shows that the liquid/vapour interface is no longer formed in the pore entrance but shifted further into the pore channel due to the water intrusion (drawn by the capillary pressure) which balanced out by the vaporization of water from the liquid/water interface due to partial pressure difference. Crucially, fluid flow through the nanochannels experiences dramatically increased resistance, due to the sharp increase in the shear viscosity of water in nanoconfined spaces, preventing outright pore wetting. The impacts of pore size, membrane thickness, substrate thickness, concentration polarization, porosity, and contact angle on water flux and pore intrusion depth were tested using the model. The membrane hydrophilicity was found to impact on water flux and pore intrusion in a complex relationship with pore size. In order to elucidate this theory, organosilica membranes of different pore sizes and pore geometries were prepared; their performances were compared and found to be in broad agreement with the initial model.
In the second part of the thesis, a different strategy was trialled for the synthesis of hybrid organic-inorganic materials, which employs the triconstituent co-assembly method using two separate precursors for the organic and inorganic compounds and a structure directing agent. Unlike the organosilica membrane where carbon and silicon atoms are covalently bonded and homogeneously distributed in atomic scale, these new carbon-silica nanocomposites are comprised of carbon and silica networks that interact physically at the molecular level. The effects of silica content, carbonization temperature, types of surfactants and coating conditions on the formation and performance of the resulting membranes were evaluated. In the third major finding of the thesis, the results demonstrated that a carbon-silica nanocomposite could be a more economically viable and promising candidate in membrane development compared to the original organosilica membranes. Furthermore, the concept of carbon-silica nanocomposite membranes is novel in MD-based water desalination and this opens up a new development pathway for hybrid organic-inorganic membranes.