A reverse osmosis (RO) membrane followed by a packed tower aerator and rapid sand filter are used by an agricultural company in Emmen to treat brackish groundwater to supply water for growing crops and cleaning purposes. The iron removal by the treatment plant is sufficient most of the time throughout the year. However, the farmer periodically (once per year) reported yellowish treated water, indicating insufficient iron removal. Nevertheless, it is still not clear what is the cause of the insufficiency of the iron removal. This thesis aimed to investigate the iron removal in the water treatment plant in Emmen and propose a solution to improve the iron removal. The iron removal of the treatment plant was investigated through a set of batch iron oxidation, flocculation, and filtration experiments. The pH of the water was the main parameter that influenced the oxidation of Fe(II). It was found that within the retention time that was available in the treatment plant’s tower aerator and the rapid sand filter (20 minutes), the Fe(II) was not fully oxidized and flocculated. At pH of the permeate in the range of 6 – 7, only <11.5% of the initial Fe(II) was oxidized in 20 minutes. Increasing the pH to 8 accelerated the oxidation of Fe(II), and the Fe(II) was completely oxidized within 30 minutes.Although removal of iron that is dominated by floc filtration was not achieved, the treatment plant also removed the iron through adsorption on iron hydroxide deposit in the packed tower aerator and sand particle. However, the Fe(II) adsorption capacity of adsorbents was low at low pH. The regeneration of the adsorption capacity was achieved through oxidation of the adsorbed Fe(II) which is also influenced by pH. COMSOL model showed that without regeneration, the adsorption capacity of new sand and iron oxide-coated sand was exhausted after 24 hours and 230 hours, respectively.When the concentration of CO2 in the permeate is in equilibrium with air, the pH of the permeate should be in the range of 7.9 – 8. However, the maximum pH of the permeate was 7 after aeration because the contact time of the tower aerator was not sufficient to completely strip the CO2. The tower aerator was also clogged by iron deposits that reduces the airflow and causes short-circuiting that decrease the CO2 stripping efficiency over time.Installation of a bubble column reactor was proposed to improve the CO2 stripping and increase the pH of the permeate. The Phreeqc model showed that the CO2 could be stripped until approx. 1 mg/L within 4 minutes, and the pH also increased to approx. 7.9. Moreover, the bubble column reactor will provide additional retention time for Fe(II) oxidation, and approx. 20 – 30% of the initial Fe(II) concentration can be oxidized within 4 minutes. The bubble column was considered preferable compared to the packed tower aerator because it was not susceptible to clogging by iron deposits and requires lower maintenance.

With the growing population and industrial development, there is more stress on natural water resources. Additionally, environmental laws make the disposal of waste streams from industries difficult. In this scenario, it is crucial to treat waste streams to recover water and possibly minerals, for reuse in the industry, reducing the dependency on new resources. Reverse osmosis (RO) is a membrane-filtration technology that has been in use for decades. RO systems can waste up to 30% of the feed water through the production of brine, containing the rejected minerals. This rejection is high when dealing with saline waters. Treatment of this brine stream for reclamation of water would increase the overall efficiency of RO systems. In this research, the treatment of RO brine by a closed-circuit configuration of RO – closed-circuit desalination (CCD) - is investigated. In CCD, filtration is done in batches, with the recirculation of the concentrate stream back into the feed stream. The high cross-flow velocity and short filtration batches result in prevention of scaling in these systems.
This research focused on proving the resilience of the CCD system to silica scaling despite supersaturations of silica. Lab-scale experiments were performed on a single-element CCD system, specially built with relevant measurement instruments. Silica scaling was monitored by mass transfer coefficient (MTC) calculations, silica mass balance calculations, and membrane autopsies. Synthetic feed water was used, containing only NaCl (10 mg/L) and varying concentrations of silica (70 – 120 mg/L as SiO2). No anti-scalants were used. Longer experiments were performed, i.e. 20 consecutive cycles of 1 hour each with 120 mg/L of silica in the feed. The cycle duration was increased, 2 cycles of 3-hour each were performed, with 120 mg/L of silica in the feed.
The MTC curves in all the experiments had a gradual decline with the progression of each cycle, but always recovered at the beginning of the next cycle. This decline was probably due to the increasing osmotic pressure in the recirculation loop. There was no permanent decline in the MTC, leading to the observation that there was no scaling in the system. This was supported by the silica in the mass balance calculations, that showed there was no loss of silica from the brine. The reactive silica concentration in the brine was as high as 1800 mg/L in the 3-hour cycle, attaining a recovery of 93%, without signs of scaling. The membrane autopsy showed that the membranes used for the experiments with 120mg/l silica in the feed had higher silicon content compared to the blanks. However, there were too few samples to compare against and make strong conclusions whether there was scaling.
Thus, the designed CCD system was resistant to silica scaling in these conditions, of high pH and in the absence of other components such as iron, aluminium, calcium and magnesium. The results of this study proved that, despite high concentrations of silica in the feed, CCD can improve the total efficiency of RO systems (with regard to the water wastage) by recovering water from the brine produced by RO installations. The extremely high recoveries attained by the system would result in small volumes of very concentrated brine, making the extraction of minerals more cost-effective, because lesser volumes must be treated to obtain the same amount of minerals.

Water scarcity emphasizes the necessity of water recovery from saline sources, which are plentifully available. Nanofiltration (NF) is widely implemented in water reclamation treatment particularly in industrial field. However, the presence of silica hinders the possibility of obtaining a high recovery in a NF membrane system. In this work, the effect of silica concentration on the scaling of a NF membrane was investigated. Three types of feed water (demineralized water, water with 100 mg L-1 and 300 mg L-1 of SiO2) were fed to a single stage nanofiltration element. All cases were carried out at room temperature, pH values ranged 7 - 8 and a constant recovery of 11±1% for 24 hours continuously. The mass transfer coefficient (MTC) was used as method to determine the scaling of the membrane module. Results showed that no silica scaling occurred during the system runtime. Nonetheless, for all cases, an initial decrease in MTC could be detected. On the one hand, based on some literature reports, it is likely that a part of silica (mainly colloid) had precipitated on the membrane surface forming a more porous layer than the precipitation of polymerized monomeric silica. But to verify this argument, a further autopsy method is demanded. On the other hand, concentration polarization may also contribute to the initial decline in mass transfer coefficient. But the reason for the decrease for demi-water case remains unknown. In brief, Nanofiltration membrane can operate at a silica concentration up to 300 mg L-1 for 24 hours in lab-scale. The guideline (120 mg SiO2 L-1) suggested in the industrial standard (ASTM, D4933-08) was judged to be relatively conservative. For more practical purpose, further study is needed.

Worldwide fresh water scarcity is nowadays becoming a reality rather than a futuristic concern. Non-conventional sources of wastewater are being considered as a solution for this problem, since fresh water sources will not be sufficient to satisfy the world’s water demand in the following years. Extensive research is being done on different wastewater treatment technologies and methodologies, and lately, reverse osmosis (RO) membrane filtration has caught more attention as a solution for reclaiming and reusing wastewater. Nevertheless, the biggest problem RO membranes face is fouling, which increases energy consumption, and therefore, costs. Organic/particulate fouling affects the mass transfer of an RO membrane, but it can also contribute to biofouling, worsening the mass transfer and causing pressure drop on the RO system. For this study, a RO crossflow cell was used in order to investigate the effect of crossflow reversal on particulate fouling removal by testing two different feed spacers: (a) 1.25-mm high cavity spacer and (b) 0.71-mm high zigzag spacer. The experiments were carried out with an average crossflow velocity of 0.36 m/s, an average permeate flux of 20 L/m2 hr (LMH), an average feed electroconductivity (EC) of 14.90 mS/cm, and a feed temperature in the range of 21-25 C. Four runs were done, each with an average runtime of 1050 minutes. Two runs tested the cavity spacer, one with a crossflow reversal done at minute 806, and the other run with a crossflow reversal done twice every working day. The last two runs follow the same methodology but using the zigzag spacer. Each run used a new membrane and spacer. The mass transfer was graphed as a Mass Transfer Coefficient (MTC), and its development with time was analysed for identifying fouling, as well as particulate fouling when applying the crossflow reversal method. Crossflow reversal proved to be a reliable method for removing organic/particulate fouling. The cavity spacer caused a more severe organic/particulate fouling, but allowed it to be removed by crossflow reversal. The cavity spacer also proved to be less energy-consuming if applied to a full-scale RO facility. On the other hand, the zigzag showed to have a better effect on flow mixing and concentration polarization disruption inside the feed channel, thus keeping a higher mass transfer for a longer period of time. A more frequent crossflow reversal showed was not more efficient in removing organic/particulate fouling.

Reverse Osmosis is a possible process production of drinking water from saline or freshwater, in which water is forced through a semi permeable membrane resulting in production of pure water with very low ion concentrations. This water is then remineralized to prevent corrosion and make it suitable for consumer usage. It is reported that at higher temperatures, for instance in boilers, scaling occurs in an earlier stage in drinking water produced by RO than water produced with conventional water treatment plant. Presumed is that the absence of certain materials removed by RO during filtration is the cause of this early scaling. One of these substances is humic acids. In this research the effect of addition of several concentration of HA on scaling is measure at high temperature and calcium concentration of approximately 20mg/l and 40mg/l. The results show that dosing of at least 3 mg/L of humic acids can significantly reduce scaling in RO permeate water. In water with lower Ca2+ concentrations no scaling was measured thus the experiment was repeated with a higher initial Ca2+ concentration.

Een tekort aan zoet water zal in de toekomst een groot probleem zijn. Daarom zijn er diverse
technologieën, die het mogelijk maken om zout of brak water naar zoet water om te zetten. Eén van deze technologieën is omgekeerde osmose (Reverse Osmosis (RO)). Aan deze technologie kleven echter nog nadelen. Een van deze nadelen is de kalkaanslag (scaling) die bij het product van dit proces sneller optreedt dan bij water geproduceerd met conventionele technieken. In dit rapport wordt geprobeerd een oplossing voor dit probleem te vinden. Het doel van dit rapport is om beter te begrijpen waarom er een verschil in scaling is tussen water geproduceerd met RO -en kraanwater. Er zal worden onderzocht of dit gerelateerd is aan het tekort aan fosfaten en humuszuren in RO water. Dit zal gedaan worden door een experimentopstelling te maken. Het RO water, met en zonder de fosfaten en/of humuszuren, zal gedurende een week onderzocht worden op de snelheid en hoeveelheid van scaling.