Jongho Lee

Ammonia is a versatile compound for fertilizer, refrigerant, and chemical production, as well as a potential hydrogen carrier for fuel cell vehicles. Wastewater is a sustainable ammonia source which can largely substitute the conventional ammonia production via the Haber-Bosch process. Bipolar membrane electrodialysis (BMED), combined with a membrane contactor, holds promise for electrified, chemical-free ammonia recovery from wastewater. However, ammonia leakage through the ion exchange membranes (IEMs) and limited recovery rates remain among the major challenges in implementing this new approach. Our study focuses on understanding these limitations and optimizing the BMED configuration. We first examine the mode of ammonia leakage through IEMs and find that its transfer across the membranes occurs as a gas phase. Subsequently, we investigate the relation between the cross-stack electric potential, as driving force for ammonia gas production, to the extent of ammonia leakage and its recovery efficiency. We find that a lower cross-stack potential slows down the ammonia leakage at the cost of reduced ammonia recovery due to the insufficient rate of water splitting inside the bipolar membranes. Then, the impact of mass transfer resistance across IEMs on the ammonia leakage and recovery is investigated by introducing multiple layers of IEMs. We find that there is an optimal number of IEMs layers that allows for minimizing ammonia leakage as low as 2% and achieving high ammonia recovery as high as 63%, along with an energy consumption comparable to industrial Haber Bosch process (0.6 MJ/mol-N). Finally, a new BMED configuration with an additional compartment of saline solution is examined as an alternative to prevent ammonia leakage. This configuration achieves an 86% recovery rate and virtually zero ammonia leakage but requires external chemical input. Our study highlights the potential of highly efficient ammonia recovery from wastewater using BMED through optimizing BMED operating parameters and configurations.

View record

Abundant in municipal, agricultural, and industrial wastewaters, ammonia is a versatile chemical, which can be used as fertilizer or as a hydrogen carrier for the hydrogen economy. Membrane contactors (MC) are an effective membrane process for selectively recovering ammonia from a wastewater stream, but the system’s continuous acid consumption and the residual acidity in the produced stream remain a major hurdle. In this study, we propose a chemical-free, electrified process that combines MC with bipolar membrane electrodialysis (BMED) to produce ammonia as a gas from wastewater. We devise three different BMED configurations, and identify the most suitable configuration by examining the impact of different operating conditions including pH, initial ammonia concentration, and relative volume ratios on the ammonia recovery and BMED energy consumption. We demonstrate up to ~68% recovery of theoretically recoverable ammonia, and attribute the unrecovered fraction primarily to the diffusion of neutral ammonia through membranes. While the system is sub-optimal, the relative energy consumption of the system is comparable to the Haber-Bosch process for conventional ammonia production, with potentially lower energy consumption and higher ammonia recovery through higher initial ammonia concentrations, higher relative volume ratios or alternate BMED configurations.

View record

Membrane distillation (MD) is an emerging desalination technology employing a hydrophobic (water-repelling) microporous membrane that is promising for water reclamation from highly saline streams that conventional reverse osmosis (RO) cannot treat. However, conventional hydrophobic membranes are prone to wetting and fouling when treating complex waste streams, such as oil- and gas-produced waters, which limits the applications of MD. Typically, two conflicting surface properties (i.e., hydrophobic and hydrophilic) are required to mitigate pore wetting and membrane fouling, respectively. In this thesis, we develop Janus membranes comprising a hydrophilic zwitterionic polymer layer and an omniphobic (all liquid-repelling) porous substrate that simultaneously possess fouling and wetting resistances. An omniphobic membrane was first fabricated by attaching silica nanoparticles (SiNPs) to the fibers of a quartz fiber mat, creating multilevel re-entrant structures, followed by surface fluorination to reduce the surface energy. The Janus membrane was then fabricated by grafting a zwitterionic polymer brush layer via surface-initiated atom-transfer radical-polymerization (ATRP) on the omniphobic substrate. Membrane characterizations, including Fourier-transform infrared spectroscopy, fluorescence microscopy, and contact angle measurements, confirm that the surface hydrophilicity can be finely tuned by adjusting the duration of the ATRP reaction. Also, the zwitterionic polymer brush layer was confined on the top surface of the Janus membrane, rendering the surface hydrophilic, while the remaining part of the Janus membrane remained omniphobic, resisting the wicking of low-surface-tension liquids including ethanol and hexane. A static oil-fouling test showed that crude oil droplets irreversibly fouled an omniphobic membrane (without a hydrophilic top layer) in water. In contrast, oil droplets placed on the Janus membrane in air were immediately desorbed upon its immersion in water. Finally, we performed direct-contact MD (DCMD) experiments using a crude-oil-in-saline (NaCl) water emulsion as a feed solution, simulating highly saline oily wastewater. The Janus membrane exhibited superior wetting and fouling resistances, with stable water flux and nearly perfect salt rejection, while an omniphobic membrane failed in the desalination process. Our work highlights the great potential of antiwetting and antifouling Janus membranes for water reclamation from challenging industrial wastewaters through MD.

View record