Urea and ammonium rejection by an aquaporin-based hollow fiber membrane

Abstract

Urea removal from aqueous solutions is a challenge in many field applications. Most available technologies are either energy intensive or complicated. Novel membrane-based processes such as forward osmosis (FO) present potentially desirable alternatives for energy efficient urea removal. Recently, aquaporin-based membranes (ABMs) gained attention due to their high water permeability and selectivity. In this study, a hollow fiber ABM module was investigated for both urea and ammonia/ammonium rejection under different conditions such as varying water recovery rates, initial feed solute concentrations, draw solution molarities and feed solution pH. The amount of water recovery significantly impacted urea rejection as a result of feed concentration increase and draw dilution over time. While only less than 37% of urea could be rejected for a water recovery rate of 80%, the rejection of urea for water recovery rates of 25% was above 82% when rejection was calculated via overall mass balance; however ammonium rejection at lower pH did exceed 90%, even under high water recovery rates. The increase of ammonium rejection at low pH is believed to be influenced by the decrease of bidirectional diffusion of sodium and ammonium cations favored by the less negative charge of carboxyl groups dominating the membrane surface’s active layer. Since rejection for uncharged urea is rather poor, chemical processes that transform urea into ammonia/ammonium prior to the FO-step should be considered in future studies to improve rejection.

Introduction

Urea is excreted by mammals and is ubiquitous in many environments [1,2]. Urea is also a key molecule in many industrial synthesis pathways and can enter the environment via waste waters from industrial production facilities [[2], [3], [4]]. Effective strategies to remove urea from aqueous solutions such as urine are desirable, especially in exploration of remote environments. For instance, water recovery from human liquid waste is necessary during long-duration missions in space [5]. Furthermore, effective strategies to remove urea from urine are desirable to minimize human impact in natural ecosystems where nitrogenous waste needs to be removed as for example during deep exploration of pristine caves [6].

Different methods for urea removal exist, but most are still in the early stages of development [2]. Such strategies include hydrolysis, biological and enzymatic decomposition, adsorption and electrochemical oxidation [[2], [3], [4],[7], [8], [9], [10]]. Most of these technologies are either energy intense, and therefore often only considerable for applications where energy consumption is not limited, or require complex and segmental biological strategies [2]. Membrane-based filtration processes such as reverse osmosis (RO) and nano filtration (NF) are possible alternative solutions. RO as well as NF membranes have been studied regarding their rejection of small and uncharged molecules such as urea [[11], [12], [13], [14]]; however, for most RO and NF membranes tested, urea rejection rates did not exceed 50% [13,14]. Furthermore, high pressure filtration processes such as RO have several disadvantages including cost, scaling limitations and the risk of membrane fouling [15,16].

An alternative strategy to RO and NF is forward osmosis (FO). In FO, an osmotic pressure gradient between a higher concentrated draw solution and a less concentrated feed solution is utilized to facilitate water transport across a semi-permeable membrane, which separates feed from draw solution and allows water molecules to pass while other solutes are being held back. The osmotic pressure created by the draw solution, pulls water from the feed stream through the membrane, leading to the dilution of the draw solution, while solutes in the feed stream are concentrated [17]. Since the composition of the draw solution can be tailored depending on application, FO may have advantages when compared to energetically intense pressure-driven filtration processes, such as RO [16].

Different FO membranes exist and their rejection capabilities for various compounds have been studied [[18], [19], [20], [21]]. An average urea rejection rate of 94.4% was reported for a Porifera flat sheet FO membrane [22], while rejection rates of over 98% were reported in a recent study using a flat sheet cellulose triacetate (CTA) FO membrane [23]. Larger scale plate-and-frame Porifera FO membrane modules have been developed, but no data on urea rejection of those larger units is available. Furthermore, a plate-and-frame membrane configuration has several disadvantages including high cost, limited cleaning ability and size [[24], [25], [26]]. For applications that require space-saving and light-weight designs (e.g. portable FO systems) as well as large volume separations, other membrane configurations, such as hollow fiber modules, may therefore be advantageous.

A promising FO membrane, commercially available in hollow fiber configuration, is the Aquaporin Inside membrane. This aquaporin-based membrane (ABM) incorporates aquaporin (AQP) proteins into its active layer to increase water permeability [27,28]. AQPs, such as the AQPZ found in Escherichia coli, are trans-membrane proteins that are primarily permeable to water and reject most larger compounds from permeating through their pores [[29], [30], [31], [32], [33]]. AQPZ proteins have successfully been incorporated into ABMs by manufacturing techniques such as interfacial polymerization [34] or pore-spanning membrane design via AQPZ-vesicle fusion [35]. High rejection rates for small size trace organics were recently reported using small scale flat sheet ABMs [16] as well as larger scale hollow fiber ABM modules [36]. High solute rejection in combination with expected high water flux rates, therefore make ABMs an interesting candidate for exploring urea rejection capabilities.

In this study we investigated the rejection characteristics of a 2.3 m² Aquaporin Inside hollow fiber FO module for urea. Variation of different parameters such as draw solution molarity, feed solution pH, water recovery rate and initial feed compound concentration were analyzed to elucidate the rejection mechanisms of the used membrane. We also evaluated the rejection of ammonia/ammonium (NH₃/NH₄⁺) at different feed solution pH in order to determine the ability of the membrane to reject small charged molecules. Our data suggests that higher rejection can be achieved when urea is hydrolyzed into ammonium, which may allow for a possible pre-FO strategy to enhance urea rejection in future FO studies and applications.