Optimization of dead-end ultrafiltration for the tertiary treatment of produced water from oil industry

Produced water (PW), is a highly contaminated oily wastewater byproduct from oil and gas industry. With growing energy demands and more stringent oily wastewater discharge regulations, PW became one of the most complex waste streams that must be effectively treated before disposal to the environment or reuse as process-water. Generally, PW treatment comprises primary, secondary, and tertiary (polishing) treatment steps. Great efforts from researchers worldwide were dedicated to the development of effective techniques for the tertiary treatment for PW, to finally reuse or safely discharge to the environment. Pressure-driven membrane-based separation processes, e.g., microfiltration (MF) and ultrafiltration (UF), are promising due to high water recovery, simplicity, small footprint, and ease of integration into currently established treatment plants. This thesis investigated the applicability of dead-end ultrafiltration operation for the tertiary treatment of PW. Sub-micron sized emulsified oils were prepared to mimic PW characteristics in the polishing step. Emulsified oils with similar chemical compositions but produced via different emulsification methods were found to cause different membrane fouling behaviors that were related to different oil droplets size distributions. The performance of UF and MF membranes as well as the nature of fouling mechanisms were emphasized to be substantially influenced by the interrelation between oil droplet size distribution and membrane pore size distribution. Our experiments revealed that dissolved crude oil components had no significant contribution to the fouling of UF membrane by emulsified oils. Dosing anionic surfactant (sodium dodecyl sulfate, SDS) below its critical micelle concentration resulted in enhanced performance of the dead-end ultrafiltration for treating emulsified oils with oil contents up to 50 mg Total Organic Carbon (TOC) per liter. This was achieved due to a substantially enhanced hydraulic backwashing efficiency, overall permeability recovery range of 72%-90% for different emulsified oils and SDS concentrations. The promoted hydraulic reversible fouling by surfactant-enhanced ultrafiltration was attributed to the minimized hydrophobic-hydrophobic interactions, and the improved access of backwashing water via facilitated paths through the formed fouling layer. However, the method suffered from high sensitivity regarding the quality of the applied SDS. This led to a significant drop in the backwasheability of the accumulated fouling layer on the membrane during the filtration of surfactant-modified emulsified oils. Based on a scenario analysis, higher carbon footprint and operation costs were calculated for the surfactant-enhanced dead-end UF compared to the crossflow operation in most of the considered scenarios, making the first less environmentally friendly and less economically feasible than the latter. The dosage of Powdered Activated Carbon (PAC) and/or coagulants prior to membranes was also investigated. In batch adsorption experiments, it was found that significant portion of the oil components from the emulsified oils were adsorbed on the PAC (up to 75%), a certain non-adsorbable part could be noticed. Adsorption isotherms for TOC could not be clearly determined due to strong fluctuations in the elimination rate. This is attributed to the multicomponent nature of the system and the assumption that adsorbable components primarily accumulate on the external surface of the PAC particles. No significant difference could be noticed when using different PAC types made of different materials. In coagulation jar-test experiments, it was found that dosing iron-based and aluminum-based coagulants enhanced the oil elimination from the emulsions, which is expected mainly to be attributed to an increase in the oil coalescence between oil droplets. Iron-based coagulant exhibited higher elimination rates than the aluminum based one. Dosing PAC prior to capillary UF membranes, operated in In/Out mode, did not significantly reduce the fouling but dosing coagulant prior to UF membranes did reduce the fouling rate. Dosing both PAC and Coagulation prior to UF also reduced the fouling but could not outperform the coagulation-UF operation. However, neither PAC nor coagulation, or a combination of the two, could enhance backwash efficiency. Dosing PAC slightly improved the separation performance of the capillary UF membrane compared to standalone UF, on the other hand dosing coagulation alone or combined with PAC prior UF exhibited better separation performance.