Search

This dissertation addresses the role of individual fouling mechanisms on productivity decline and solute mass transport in nanofiltration (NF) of surface waters. Fouling mechanisms as well as solute mass transport mechanisms and capabilities must be understood if NF of surface waters is to be successful. Nanofiltration of surface waters was evaluated at pilot-scale in conjunction with advanced pretreatment processes selected for minimization of nanofilter fouling, which constituted several integrated membrane systems (IMSs). Membrane fouling mechanisms of concern were precipitation, adsorption, particle plugging, and attached biological growth. Fouling was addressed by addition of acid and antiscalent for control of precipitation, addition of monochloramine for control of biological growth, microfiltration (MF) or coagulation-sedimentation-filtration (CSF) for control of particle plugging, and in-line coagulation-microfiltration (C/MF) or CSF for control of organic adsorption. Surface water solutes of concern included organic solutes, pathogens, and taste and odor compounds. Solute mass transport was addressed by evaluation of total organic carbon (TOC), Bacillus subtilis endospores, gesomin (G), 2-methlyisoborneol (MIB), and threshold odor number (TON). This evaluation included modeling to determine the role of diffusion in solute mass transport including assessment of the homogeneous solution diffusion equation. A cellulose acetate (CA) NF was less susceptible to fouling than two polyamide (PA) NFs. NF fouling was minimized by the addition of monochloramine, lower flux, lower recovery, and with the use of a coagulant-based pretreatment (C/MF or CSF). NF surface characterization showed that the low fouling CA film was less rough and less negatively charged than the PA films. Thus the theory that a more negatively charged surface would incur less adsorptive fouling, due to charge repulsion, was not observed for these tests. The rougher surface of the PA films may have increased the number of sites for adsorption and offset the charge repulsion benefits of the negatively charged surface. The addition of monochloramine significantly reduced biodegradation and integrity loss of the CA membrane. PA membranes are inherently not biologically degradable due to their chemical structure. Monochloramination reduced the rate of fouling of the PA membrane but resulted in a gradual increase in water mass transfer coefficient and a decrease in TDS rejection over time, which indicated damage and loss of integrity of the PA membrane. Based on surface characterization by X-ray Photoelectron Spectroscopy (XPS) and Fourier Transform Infrared Spectrometry (FTIR), the PA membrane degradation appeared to be chemically-based and initiated with chlorination of amide nitrogen and/or aromatic rings, which ultimately resulted in disruption of membrane chemical structures. The recommended Integrated Membrane System to control fouling of a surface water nanofiltration system is CSF monochloramine/acid/antiscalent„³monochloramine-tolerant NF. This IMS, at low flux and recovery, operated with no discernable fouling and is comparable to a groundwater nanofiltration plant with cleaning frequencies of once per six months or longer. A significant portion of the organic solutes including total organic carbon (TOC) passing through the membranes was diffusion controlled. Permeate concentration increased with increasing recovery and with decreasing flux for both PA and CA membranes. The influence was diminished for the PA membrane, due to its high rejection capabilities. Total rejection of spores used as pathogen surrogates was not achieved as spores were indigenous and high spore concentrations were used in all challenge studies; however, Integrated Membrane System spore rejection exceeded credited regulatory rejection of similar sized microorganisms by conventional treatment by several logs. Spore rejection varied by NF but only slightly by MF as size-exclusion controlled. There was no difference among spore rejection of IMS with and without in-line coagulation. Consequently, these results indicate membrane configuration (Hollow fiber>Spiral Wound) and membrane film (Composite Thin Film>CA) significantly affected spore rejection. Geosmin and methylisoborneol have molecular weights of 182 and 168 respectively, and are byproducts of algal blooms, which commonly increase taste and odor as measured by the threshold odor number (TON) in drinking water. Although these molecules are neutral and were thought to pass through NFs, challenge testing of IMS unit operations found that significant removal of TON, G and MIB was achieved by membrane processes, which was far superior to conventional processes. A CA NF consistently removed 35 to 50 percent of TON, MIB, and G, but did not achieve compliance with the TON standard of 3 units. A PA NF provided over 99 percent removal of MIB and G. Challenge tests using MIB and G indicated that size-exclusion controlled mass transfer of these compounds in NF membranes.