Session 6F - UF/MF 2

After 70 years, modern pressure-driven polymer membrane processes with liquids are mature and accepted in many industries due to their good performance, ease of scale-up, low energy consumption, modular compact construction, and low operating costs compared with thermal systems. Successful isothermal operation of synthetic membranes with liquids requires consideration of three critical aspects or “legs” in order of relevance: selectivity, capacity (i.e. permeation flow rate per unit area) and transport of mass and momentum comprising concentration polarization (CP) and fouling (F). Major challenges remain with respect to increasing selectivity and controlling mass transport in, to and away from membranes. Thus, prediction and control of membrane morphology and a deep understanding of the mechanism of dissolved and suspended solute transport near and in the membrane (i.e. diffusional and convective mass transport) is essential. Here, we focus on materials development to address the relatively poor selectivity of liquid membrane filtration with polymers and discuss the critical aspects of transport limitations. Relevance to nano-channels such as synthetic, biomimetic (imidazole assembly) and biological (aquaporin) will be addressed. Machine learning could help optimize membrane structure-by-design and transport conditions for improved membrane filtration performance.

Membrane technology is regarded as indispensable in this century for it can provide high grade and sustainable water supply. Poly (vinylidene fluoride) (PVDF) Ultrafiltration (UF) membranes have been widely used for various water treatments due to high chemical resistance and good filtrate water quality and hence ease of operation and maintenance. While there has been growing expectations for small substrates separation (fine separation) and high water permeability from the perspective of energy saving and cost reduction, it is generally difficult to achieve both fine separation and high water permeability at the same time, as resistance in the UF membrane structure increases as the pore diameter becomes smaller and water permeability declines.

In order to overcome this problem, we pursued the limits in terms of refinement of the size of pores on the surface of a membrane, which is necessary for fine separation, and coarsening of pores inside the membrane, necessary for high water permeability. Based on the in-situ observation of phase separation process carried out at SPring-8, a large synchrotron radiation facility, we got relationships between various process conditions and membrane structures. By leveraging the proprietary membrane process technology and precise structural design using phase separation simulation based on computational chemistry combined with the in-situ observation, we succeeded in controlling the phase separation speed and growth speed of the membrane. This made it possible to overcome the existing trade-off between pore size and permeability, and enabled the realization of nano-metric sized fine pore structure that has six times the permeability of existing membranes.

Introduction: Hydrogel facilitated phase separation (HFPS) method was recently used for the fabrication of patterned porous membranes relying on the high-water content of hydrogel to initiate the membrane formation. In this work, we used for the first time, a simple treatment method for hydrogel mold recovery, consisting of the cold and warm treatment processes to allow the continuous usage of the same mold for membrane preparation.

Methods: An ideal treatment process would fully extract the solvent diffused inside hydrogel after each membrane casting without damaging the hydrogel structure. Two methods were investigated (i) cold treatment in which the hydrogel mold was placed in a cold-water allowing for the slow diffusion of solvent out of hydrogel and, (ii) heat-treatment process in which the hydrogel mold was placed in a warm water bath to accelerate the diffusion time of solvent.

Results: Scanning electron microscope (SEM) images of 3 consecutive membranes prepared from the same hydrogel mold without treatment showed an increase in the thickness of the membrane skin layer after each replication. Therefore, the permeate flux results for these membranes significantly dropped down compared with the first prepared membrane. The heat-treatment process showed better recovery results compared with the cold-treatment process. The optimum heat-treatment procedure was found to be 10 min in a warm water bath, followed by 5 min in a cold-water bath and 4 min drying time, which resulted in 96% water flux recovery.

Discussion: The formation of HFPS membranes relies on the diffusion of organic solvent from the original membrane solution into the hydrogel mold. This changes the hydrogel contents affecting the consecutive membrane replications. We believe the best recovery achieved was due to a nearly complete organic solvent extraction from the hydrogel mold without destroying the hydrogel structure.

Membrane bioreactors (MBRs) are successfully being adopted in super-large-scale (>100,000m³/d) wastewater treatment plants (WWTP), even if loss of membrane permeability caused by fouling induces high operational costs, mainly due to aeration, chemical cleaning and excessive recirculation (Krzeminski, 2017). Modelling offers the opportunity to reduce these costs. Activated sludge models (ASMs) have been tested on large-scale MBR plants (>10,000m³/d) (Gabarrón, 2015, Sun, 2016). However, the integrated modelling of super-large-scale MBR-WWTP including complex interactions between biology, filtration, and fouling has not been done yet.

A super-large-scale MBR plant (Figure, 360,000m³/day) characterised by hollow fibre membranes (PVDF with pore size 0.4µm) with total surface area of 460,000m² and receiving wastewater from Paris region has been considered for modelling. Online operational and flow quality data were collected for three months. Furthermore, COD fractions, autotrophs (X_ANO), and heterotrophs (X_OHO) concentrations and their yield coefficients were measured. After cleaning and validation, these data were used to initialise and validate the simulations. A dynamic integrated model involving the coupling of EPS-ASM3-BioP and filtration models (Janus 2014, Rieger et al 2001) was developed and implemented in MATLAB environment. The biological part of the model considers the stoichio-kinetic activity of the biomass for the carbon, nitrogen and phosphorus removal. The filtration model covers the biomass fouling dynamics due to intermittent air scouring. The complexity of the data treatment and its simplification for modelling were discussed. Mean relative error showed that the simulated results are close to the real data. The super-large-scale MBR modelling allows optimising its functioning in terms of depollution performance and operational costs, and it could be the base of advanced control.

References

Krzeminski, et al. 2017, JMS 527:207‑27.

Janus 2014, Procedia Engineering 70:882–891

Gabarrón et al. 2015, CEJ 267:34-42.

Rieger et al. 2001, WR 16:3887‑3903

Sun et al. 2016, WR 93:205-213.

Figure: MBR Plant Schematic