D. Tal*, Y. Schnurmacher**, S. Lauz, S. Regev, R. Levi******
* ROTEC by WFI Group, Process and Technology Department, Caesarea Israel
** ROTEC by WFI Group, Process and Technology Department, Caesarea Israel
Type of presentation: Oral
Theme: the paper is to be submitted for: Theme 3A: Effective and Efficient Wastewater Management (Treatment)
Flow Reversal Reverse Osmosis (FR-RO) technology was tested for the purpose of treating wastewater at a semiconductor plant, to increase RO recovery beyond 75% and enhance water reuse. The study determined the ability to stabilize operation with recovery rate ranges between 88% and 90%, aiming for low maintenance and scaling cleaning in place (CIP) events.
Performance was evaluated in terms of detectable factors which predict mineral scaling over membrane systems such as ΔP, feed pressure and system permeate flux during step-up changes in recovery rate. Set points for recovery rates were adjusted in steps between 70% and 90%. After each step, the system stability was evaluated before moving to the next set point.
The system demonstrated the ability to work at high recovery rates with stable operation thanks to the FR-RO technology implementation.
Flow-Reversal, FR-RO, High Recovery, Membrane Technology, Mineral Fouling, Mineral Scaling, Pilot Study, Semiconductors, Semiconductor Industry, Water and Wastewater Treatment, Industrial Wastewater.
Since desalination processes generate considerable amounts of reject brine, the industry has adopted numerous disposal options that usually depend on the location of the desalination plant and type of process used. These options include discharge to surface water or wastewater treatment plants; deep well injection; land disposal; evaporation ponds; and mechanical/thermal evaporation. Management of reject brine has recently become an increasingly difficult challenge due to many factors (Schorr et al., 2011).
The reject stream contains chemicals used in treatment process (e.g., HCl, H2SO4, NaOH, Antiscaling agents) which impact the environment if not well managed (Panagopoulos and Charalambous., 2020). In addition, in some cases Zero Liquid Discharge (ZLD) is required.
ZLD consumes large amounts of energy, leading to significant emission of greenhouse gases (GHG) and extremely high costs as explained by Tong and Elimelech (2016). Thus, high recovery rates together with lower chemical consumption are required.
The presence of ionic components such as calcium, magnesium, carbonate, sulphate, phosphate and silicate, and components such as soluble silica (silicic acid), in the feed water of an RO process can limit the achievable water recovery by forming insoluble scale that hinders membrane filtration (Sanciolo et al., 2014).
In this study we focus on Flow Reversal Reverse Osmosis (FR-RO) technology, which was used to treat industrial semiconductors wastewater, and is based on the principle of periodically reversing the feed flow direction within the pressure vessel (flow tangential to the membrane surface along the membrane element), to reset the mineral salt “induction clock” and thus prevent mineral scaling. Permeate flushing of the RO membranes is used when the feed water is supersaturated (Gilron et al., 2006).
The goal of this study was to determine the feasibility of recovering more water for reuse in a challenging water profile environment.
Recovery Rate
The fraction of feed water that leaves the membrane device as permeate is referred to as the membrane recovery rate (%RR): 
Induction Time
The induction time can be estimated using the following equation: 
wherein:
A and B are constants related to the salt, Constant A includes the effect of the surface energy at the nucleating surface and molar volume of the salt. B includes the frequency factor for the nucleation rate.
Values of A and B have already been determined and may also be readily obtained experimentally (Hasson et al., 2001)
Water Profile
Raw water source was industrial wastewater of a semiconductor plant with cartridge filters (<5µm pore size).
Pilot System Description
The operational system worked as one pass with a 3-stage configuration (Figure 1) using 4″ brackish water membranes.
Figure 1 – Operational FR-RO pilot system
The pressure vessels were divided into 6 blocks (A to F) which held the ability to switch positions between them as stage 1, 2 or 3 by demand.
As induction time of CaCO3 or SiO2 exceeded, block rotation occurred. Brine side became feed side and vice versa.
Operation
During initial startup, recovery rate set-point was 70% for several days. When the system appeared to be stable (i.e., no change in differential pressure, flux or feed pressure), recovery rate set-point was set to 75%. After several days of stable operation, recovery rate set-point was set to 80% and so on until recovery rate set-point reached 90%.
Since operation in recovery rates of 70% and 75% could work with a conventional RO system without scaling risks (according to membrane and antiscalant suppliers), operation in those recovery rates was only applied for a short period of time.
Brine Saturation
Concentration side saturation is calculated following a concentration increase along with Recovery Rate adjustment during the experiment (from RR%=70% to RR%=90%).
Saturation indexes (e.g., CaCO3 LSI and Silica SI%) used for this study on concentrate flow were above as suggested by most membrane and antiscaling agents’ manufacturers with high recovery rates and would normally predict short and unstable operation for conventional reverse osmosis systems due to scaling resulting from supersaturated brine flow.
Stable Operation
To ensure stable operation without scaling, during operation in each recovery rate, data was collected regarding pressures, fluxes, and differentials pressures (Figures 2.2 and 2.2).
Figure 2.1 – Fluxes and feed pressure in different recovery rate operations
Figure 2.2 – differential pressure (ΔP) in different recovery rate operations
Reversing the flow direction of undersaturated flow from the feed side to the brine side as induction time was achieved, allows scaling to re-form as saturation index enables minerals to shift back to ionic form.
Fluxes, feed pressure and ΔP remain stable operating for days except for a few minor cases of biofouling that was treated before re-operating normally.
FR-RO technology showed stable operation in a high recovery system with a challenging water profile. Reversing the flow direction introduced undersaturated feed water to RO system’s brine side and performed as cleaning/flushing allowing stable operation without mineral scaling formation due to induction time limitation of minerals forming on the ionic phase. Further study regarding the abilities of reversed flow over COD/BOD and biological fouling should be conducted.