In this study, the efficiency of electrocoagulation (EC) with iron electrodes was applied to remove two phosphonates, 1-hydroxyethylidene-1, 1-diphosphonic acid (HEDP) and nitrilotris(methylene) triphosphonic acid (NTMP) from concentrates. This work provides a detailed description of the experimental procedure and results on phosphonate removal and recovery from different electrolytes, including synthetic and real reverse osmosis (RO) membrane concentrates. This research showed high selectivity of EC, removing 100% and 80% of the NTMP and the HEDP respectively, confirming no competition with sulfates, nitrates, or silica. When experimenting with other electrolytes, calcium showed to be critical in enhancing the flocculation process, while calcium carbonate precipitation contributed to capturing the phosphonates from the concentrate. The produced iron oxide (sludge) was confirmed as goethite and akaganéite, and finally transformed into hematite, indicating the oxidation from Fe2+ to Fe3+ during the EC process. After the iron precipitate collection, an alkaline wash of the sludge was enough to recover 100% of the initial phosphorus from the NTMP phosphonate. However, further research is needed to optimize the recovery procedure and to improve the results with the HEDP. 70 and 140 A·m-2 current densities were optimal to bring HEDP and NTMP concentrations down to 32 μM (1 mg·L-1) in only 30 and 10 min respectively. In these conditions, the operational costs, 1.10 and 0.03 €·m-3 of treated concentrate, were estimated for HEDP and NTMP respectively. Even when EC has been widely studied for phosphate removal, this technique has been barely applied to treat concentrates containing phosphonate-based antiscalants. EC opens new possibilities for phosphonates and phosphorus to be removed and recovered respectively from membrane and other concentrates.

Routine chemical cleaning with the combined use of sodium hydroxide (NaOH) and hydrochloric acid (HCl) is carried out as a means of biofouling control in reverse osmosis (RO) membranes. The novelty of the research presented herein is in the application of urea, instead of NaOH, as a chemical cleaning agent to full-scale spiral-wound RO membrane elements. A comparative study was carried out at a pilot-scale facility at the Evides Industriewater DECO water treatment plant in the Netherlands. Three fouled 8-inch diameter membrane modules were harvested from the lead position of one of the full-scale RO units treating membrane bioreactor (MBR) permeate. One membrane module was not cleaned and was assessed as the control. The second membrane module was cleaned by the standard alkali/acid cleaning protocol. The third membrane module was cleaned with concentrated urea solution followed by acid rinse. The results showed that urea cleaning is as effective as the conventional chemical cleaning with regards to restoring the normalized feed channel pressure drop, and more effective in terms of (i) improving membrane permeability, and (ii) solubilizing organic foulants and the subsequent removal of the surface fouling layer. Higher biomass removal by urea cleaning was also indicated by the fact that the total organic carbon (TOC) content in the HCl rinse solution post-urea-cleaning was an order of magnitude greater than in the HCl rinse after standard cleaning. Further optimization of urea-based membrane cleaning protocols and urea recovery and/or waste treatment methods is proposed for full-scale applications.

A vital component of spiral-wound membrane modules is the spacer mesh. It not only structurally supports the membranes but also aids in mass-transport enhancement through the membrane surface. Fundamental understanding of hydrodynamics associated with these spacer designs is critical to improve the permeate flux performance by decreasing concentration polarization and minimizing (bio)fouling, as well as minimizing the axial pressure drop. In the present study, time and space resolved Direct Numerical Simulations (DNS) were performed for a commercial spacer geometry. The spacer geometry was reconstructed by measurements using Scanning Electron Microscopy (SEM). Computations were performed for three spacer cells, allowing elimination of stream-wise periodicity that was a major bottleneck in earlier studies. The numerical solver was well checked in terms of boundary layer profiles obtained from Particle Image Velocimetry (PIV) data and with pressure measurements corresponding to various flow channel velocities. Non-dimensional computations were performed for Reynolds Numbers (Re) ranging from 73 to 375 (inlet channel velocity of 0.073–0.375 m/s) covering the flow transition dynamics regime. Results indicate that flow transition from steady to unsteady regime occurs for Re > 250. The flow transition could be primarily attributed to the interaction between vortices attached to the spacer filaments and the screw-vortex that originates along the diagonal of the spacer cells. No turbulent transition was observed even at the highest investigated velocity (Re = 375). The frequency spectra of time-varying velocity signal shows that at Re > 350 a sudden shift of frequency spectra occurs from discrete to continuous mode indicating the onset of advanced instability. Spacer design criteria in terms of maximum principal stress is also proposed, which can potentially aid in minimizing biofilm seeding.

The use of optical coherence tomography (OCT) to investigate biomass in membrane systems has increased with time. OCT is able to characterize the biomass in-situ and non-destructively. In this study, a novel approach to process three-dimensional (3D) OCT scans is proposed. The approach allows obtaining spatially-resolved detailed structural biomass information. The 3D biomass reconstruction enables analysis of the biomass only, obtained by subtracting the time zero scan to all images. A 3D time series analysis of biomass development in a spacer filled channel under representative conditions (cross flow velocity) for a spiral wound membrane element was performed. The flow cell was operated for five days with monitoring of ultrafiltration membrane performance: feed channel pressure drop and permeate flux. The biomass development in the flow cell was detected by OCT before a performance decline was observed. Feed channel pressure drop continuously increased with increasing biomass volume, while flux decline was mainly affected in the initial phase of biomass accumulation. The novel OCT imaging approach enabled the assessment of spatial biomass distribution in the flow cell, discriminating the total biomass volume between the membrane, feed spacer and glass window. Biomass accumulation was stronger on the feed spacer during the early stage of biofouling, impacting the feed channel pressure drop stronger than permeate flux.

Aquaporin plays a promising role in fabricating high performance biomimetic forward osmosis (FO) membranes. However, aquaporin as a protein also has a risk of denaturation caused by various chemicals, resulting in a possible decay of membrane performance. The present study tested a novel aquaporin based biomimetic membrane in simulated membrane cleaning processes. The effects of cleaning agents on water flux and salt rejection were evaluated. The membrane showed a good resistance to the chemical agents. The water flux after chemical cleaning showed significant increases, particularly after cleaning with NaOCl and Alconox. Changes in the membrane structure and increased hydrophilicity in the surrounding areas of the aquaporin may be accountable for the increase in water permeability. The membrane shows stable salt rejection up to 99% after all cleaning agents were tested. A 15-day experiment with secondary wastewater effluent as the feed solution and seawater as the draw solution showed a stable flux and high salt rejection. The average rejection of the dissolved organic carbon from wastewater after the 15-day test was 90%. The results demonstrated that the aquaporin based biomimetic FO membrane exhibits chemical resistance for most agents used in membrane cleaning procedures, maintaining a stable flux and high salt rejection.