M.Sc Thesis

M.Sc StudentHadas Sagi
SubjectThe Efficacy of Osmosis-Assisted Cleaning of
Scaled Nanofiltration Membranes
DepartmentDepartment of Civil and Environmental Engineering
Supervisor ASSOCIATE PROF. Guy Ramon
Full Thesis textFull thesis text - English Version


            Nanofiltration (NF) is an attractive technology for potable water treatment and wastewater reclamation and is increasingly being used in a broad range of water treatment applications. Membrane fouling and scaling is the most common problem limiting the optimal use of non-porous membranes such as NF.

            In this study, a sustainable cleaning method, based on osmotic backwashing, was investigated using NF membranes. Osmotic backwashing (OBW) refers to reversing the direction of filtration by inducing an osmotic pressure higher than the applied hydraulic pressure. In the experimental section, a bench scale system was fabricated composed of three parallel pressure vessels. After characterization of three commercial membranes - NF270, NF90 and BW30, OBW experiments were carried out in order to examine the feasibility of NaCl to maintain OBW, since it has been extensively researched as a draw solution for OBW of Reverse Osmosis membranes. However, NF membranes have low rejection of NaCl and therefore, experiments showed that NaCl had very poor ability to maintain OBW in NF membranes with an initial flux of less than 5μm/sec which lasted less than 20 seconds. Two more salts were investigated - Na2SO4 and MgSO4, which resulted in much higher initial fluxes of between 10-20 μm/sec and durations of more than 150 and 250 seconds for Na2SO4 and MgSO4, respectively, before decreasing to 10% of the initial value. These results suggest that MgSO4 is a much more promising draw solution than NaCl.

            In a second set of experiments, a comparison was made between OBW to a physical wash (PW) after fouling NF90 with CaHPO4. Results indicate that OBW is superior over PW and had a cleaning efficiency of 42.6% in a time scale as short as 5 minutes while PW was performed in a time scale of 30 minutes and resulted with an efficiency of 31.4%. It appears that, when applying OBW on a time scale of 5 minutes, efficiency results in an increase of up to 92% by shortening the fouling stage. It is assued that this result is due to a more reversible scaling layer.

            In the third section a theoretical model was solved using a finite-element package (Comsol Multiphysics), in order to investigate the influence of an unknown structure parameter β, a function of the porosity and tortuosity of the membrane support layer. Results were compared to experimental results, illustrating that β’s are possibly in the range of 0.05-0.2. Furthermore, a good correlation was observed between the experimental results to the theoretical calculation reinforcing the validity of the model and making it a useful tool to further optimize the OBW process.