A novel antifouling technique for the crossflow filtration using porous membranes: Experimental and CFD investigations of the periodic feed pressure technique
Graphical abstract
Introduction
Produced water generated from the petroleum industry is an important waste stream. Its management is a significant challenge due to its highly pollutant nature and the large volume of wastewater generated. Other industries contributing to the production of produced water include food processing, metal fabrication, power plants, pharmaceutical, and several other industries, Mondal (2016). Finding effective ways to treat and reuse produced water can lead to financial sustainability of the industry and generate a water source for positive use such as irrigation or reuse by the industry itself. Membrane filtration technology is leading the way for oily wastewater treatment because of its many advantages. They are relatively cheap, easy to operate and maintain, consume less power, etc., Fakhru'l-Razi et al. (2009), Padaki et al. (2015). However, membrane technology also suffers from an unavoidable problem that results in the deterioration of its performance with time. This is called membrane fouling and is related to the inevitable accumulation of dispersed materials at the surface of the membrane, Padaki et al. (2015), Zhang et al. (2012). Such accumulation of materials blocks the openings of the membrane pores, reducing thereby the area available for filtration. The selectivity feature of a membrane controls how the membrane responds to the incoming flux of materials. It is also an important design parameter that needs to be optimized. The selectivity function of a membrane depends on many things including factors unique to the membrane surface and its properties, the properties of the feed stream and the operating conditions of the filtration unit, Fakhru'l-Razi et al. (2009), Bilad et al. (2012), Padaki et al. (2015). Membrane surface properties are known to play a role in flux control and rate of fouling development. Examples of these properties include pore sizes, surface roughness, surface electrical charge and affinity to the dispersed materials, Dickhout et al. (2017), Carpintero-Tepole et al. (2017). In particular, pore sizes and hydrophilicity character of the membranes are known to play a key role in filtration process and the development of fouling, Wang and Lin (2017). Agrwalt et al. (2013) studied the effect of membrane surface properties on separation performance of membranes where they highlighted that an ideal membrane is the one that has minimal roughness and is superhydrophilic. The importance of surface charge and contact angle was considered in the work of Essafi et al. (2005) and Dresselhuis et al. (2008). Dresselhuis et al. (2008) suggested that a higher ionic strength of the emulsion leads to more spreading of the oil droplet along the membrane surface, particularly when oil content is large. Membrane surface properties is determined by the type of the materials used to fabricate the membrane.
Ceramic and polymeric membranes have both been used for produced water treatment. Ceramic membranes are made of inorganic materials that are inert to harsh conditions such as solvents and extreme temperatures, Dickhout et al. (2017). Polymeric membranes, on the other hand, are made of organic materials and allow for a more selective separation; a key feature unfound in ceramic membranes, Dickhout et al. (2017). Membrane modification can lead to the enhancement of the surface properties to optimize the flux and minimize fouling, Khulbe et al. (2010). Produced water is considered an oil-in-water emulsion where the oil phase is dispersed in the aqueous phase, Dickhout et al. (2017). The exact composition of produced water can vary based on its source, however the common components include dissolved inorganic and organic material, dispersed oil and suspended solids mixed with chemical stabilizers and surfactants, Sheng (2011). At relatively higher concentrations the stability of the emulsion is influenced by the repulsive electrostatic forces between the droplets and the attractive van der Waals forces. These forces interact with the inherent charge of the membrane surface influencing filtration, rejection and fouling rates, Dickhout et al. (2017). Another key player in the rate of membrane fouling and function is the operational conditions. Operational conditions including crossflow velocity (CFV) and transmembrane pressure (TMP) are strongly related to the flux and fouling rates of the membrane Carpintero-Tepole et al. (2017), Zoubeik et al. (2017). There are critical TMP and CFV that need to be estimated to ensure optimal operation of the membrane. For example, when oil droplets attach to the surface of the membrane they form interfaces that prevent pinned droplets from permeation unless a threshold pressure is surpassed. Therefore, adjusting the TMP such that the threshold pressure is not exceeded ensures that most of the dispersed oil droplets do not permeate. On the other hand, it is required that the CFV be larger than the critical velocity associated with each size of oil droplets such that pinned droplets may easily detach by the crossflow field. However, as the filtration process starts, accumulation of oil droplets and debris occurs along the surface of the membrane and within the membrane pores leading to fouling. As this layer develops, more and more oil droplets and debris adhere to form a gel-like layer.
The literature contains many studies that have looked at ways to combat the development of fouling via membrane features' and properties’ modifications, Ochoa et al. (2003), Xie et al. (2017), Liu et al. (2017), Wang and Lin (2017). Others have considered intervening with the hydrodynamic forces affecting the stability of pinned droplets by methods like back flushing, Qaisrani and Samhaber (2011), Pourbozorg et al. (2016), Vrouwenvelder et al. (2010), ultrasound vibration, Ahmad (2012), Hengl (2014), pulsating flow, Rodgers and Sparks (1992) and crossflow filtration, Ozaki and Yamamoto (2001). Crossflow filtration, in particular, has gained recognition because of its effectiveness in reducing the problem of fouling. When the crossflow rate achieves critical velocity, pinned oil droplets along the membrane surface are flushed away, Tanudjaja et al. (2017). The critical CFV needed is dependent on the droplet and pore size ranges in addition to interfacial properties of the oil. Therefore, for any given emulsion that is composed of a range of oil droplets and a membrane that has a range of pore sizes, the critical CFV that is needed should be maximized to clear off most pinned droplets. Nevertheless, crossflow will not be able to detach all pinned droplets. Thus, accumulations of pinned droplets will occur despite the efforts of the crossflow field leading to the formation of oil chunks that grow to cover the surface of the membrane. The point in filtration time where the oil residues start to wet the membrane surface marks the point of irreversible fouling. Therefore, for an antifouling intervention to be effective, it must be applied before the point of membrane wetting, Liu et al. (2017). Once irreversible fouling occurs hydraulic cleaning is no longer effective. In order to better understand the point of occurrence of irreversible fouling, recent studies have implemented new visualization methods in order to monitor the fate of oil droplets during filtration. From these studies three stages of oil droplet behavior have been identified. These include populating the surface of the membrane with oil droplets, clustering and/or coalescence of incoming droplets with pinned ones, and the growth of coalesced oil droplets to form oil chunks or laminates, Aimar and Bacchin (2010), Tummons et al. (2016). To minimize the problem of fouling, one needs to interfere with the previously mentioned mechanisms in attempts to prevent, eliminate and minimize their evolution. One way has been to introduce chemical materials (e.g., surfactants) to the feed mixture to alter the surface tension properties of the oil-in-water emulsion, Agarwal et al. (2013), Zhu et al. (2017). Such alteration can affect the affinity properties of the oil, interfering, thereby with the pinning characteristics of oil droplets. However, as this method involves the addition of chemicals, it poses hazards to the environment and is, therefore not favorable. In this regard, physical methods are generally cleaner. One such physical methods has been to periodically clean the membrane surface by back flushing, which provides an interesting cleaning procedure by alternating the permeation flux direction across the membrane. In a predetermined pattern of the alternating pressure field between the feed and the permeate sides, it is possible, by reversing the flow to open blocked channels and partially restore the performance of the membrane, Zhu et al. (2017). However, this process is only effective in, partially, opening blocked pores while the rest of the gel layer continues to stick to the surface. Over these attached layers other incoming oil droplets attach and extend to relock opened pores. Furthermore, back flushing is not recommended for some setups due to technical difficulties. It is not recommended by the industry for membranes with external support layers (e.g., polymeric membranes). In back flushing the pressure in the permeate exceeds the pressure in the feed, in this case not only can the membrane delaminate (active layer lifted off of the support structure), the glue lines sealing the membrane envelopes in spiral wound elements can also be pulled apart. As noted, either occurrence destroys the membrane (https://dowac.custhelp.com/app/answers/detail/a_id/971). Alternatively, instead of reversing the direction of permeation flux, another method based on pulsating flow, has been proposed. In pulsating flow methods both the crossflow and permeation flux stop and resume following certain frequency. The idea has been to promote instabilities in the flow field to allow for ease dislodgment of pinned droplets. However, this technique may be effective at the very short transient period between stopping and resuming the flow. Once both the CFV and permeation flux resume, fouling likewise resumes. It is believed that the best technique to minimizing the problem of fouling is to dislodge pinned oil droplets before they start to make the gel layer. In this work, we introduce another technique that is somewhat located between the two previously mentioned techniques; namely, back flushing and pulsating flow. We call this technique the periodic feed pressure technique (PFPT) in which the crossflow field never stops, rather the permeation flux is the one that stops at regular intervals. The PFPT has the advantage that it shortens the residence time of pinned oil droplets at the surface of the membrane. This is done by periodically changing the pressure of the feed channel according to particular pattern, as will be discussed in the next section. So, when the pressure is high, the permeation of water brings with it oil droplets to the surface of the membrane, when the pressure is lowered, the crossflow will be able to dislodge attached oil droplets. An experimental work is conducted to confirm the effectiveness of the PFPT and to provide a proof of concept of the antifouling characteristics of the PFPT. We also provide a computational study using CFD to investigate the fate of oil droplets in the feed stream and at the surface of the membrane when the TMP is set to zero. This CFD study highlights why this technique works by analyzing the behavior of pinned oil droplets when the pressure is set to zero.
Section snippets
The periodic feed pressure technique, PFPT
In order to determine why the periodic feed pressure technique works in eliminating the problem of fouling, it is important to highlight the different hydrodynamic forces that determine the fate of oil droplets at the surface of the membrane, Nazzal and Wiesner (1996), Tummons et al. (2016), Salama et al. (2017, 2018), Salama, 2018-a, Salama, 2018-b, Salama, 2018-c, Salama, 2018-d. A droplet in the feed stream is dragged towards the membrane by the permeation flux. Therefore, if the permeation
Experimental setup
The filtration system used in this experimental work is provided by Sterlitech Corporation, USA. Fig. 3 shows a schematic diagram of the setup used throughout the course of this work. As depicted in Fig. 3, the feed is stored in a stainless-steel tank of 5-gallon capacity; from there it is pumped into the stainless-steel membrane cell (CF042 316 from Sterlitech Corporation) at high speed generated by a Hydracell CC pump with speed control by Emerson controller. There are multiple control valves
Operational procedures
The membranes were first treated by an overnight soak in DI water before placement in the filtration cell. The operating parameters were adjusted by the control valves to the desired settings. The filtration unit was then initiated and once the permeation flux stabilizes, the mass of permeate collected in grams every 3 s was recorded by the computer system. A small amount of the permeate was collected for analysis after a few minutes of operation. The performance of the membrane at the
Results and discussion
In order to highlight the significance of the periodic feed pressure technique, the results are categorized in three parts. In the first, the behavior of the membrane to the different feed streams is investigated considering normal filtration cycle (i.e., no periodicity in the feed side pressure). Then we apply the new technique and compare the flux decline with time for the three feeds. Second, we compare the total amount of permeate at the end of the experiments to investigate the effect of
CFD investigation of the microfiltration problem
The computational investigation of the behavior of an oil droplet at the membrane surface involving the different fates is conducted using the volume of fluid (VOF) two-phase framework, Enright et al. (2002), Hirt and Nichols, 1981 The study of multiphase flows requires the need to capture the interface as the oil droplet is transported along the surface or penetrating through the pore opening. There are generally two basic categories when modeling a two-phase system involving interface
The computational model
In this section, we provide reasoning why the PFPT technique works in eliminating the problem of fouling by conducting a computational study on microfiltration using CFD. The purpose is to highlight the behavior of oil droplets undergoing different fates when the TMP is set to zero. The computational domain chosen for this study is very much like that considered in Darvishzadeh and Priezjev (2012). This provides a framework for comparisons with their work to build confidence in our numerical
Validation of the computational model
In order to build confidence in our modeling approach, we begin by replicating the computational work found in Darvishzadeh and Priezjev (2012) and Darvishzadeh et al. (2013) who looked at the effects of the transmembrane pressure and the shear rate on the behavior of an oil droplet upon encountering a pore opening. The simulation continued until the droplet either reaches the outlet, penetrates the pore or breaks up. Four cases were considered that represent four scenarios including rejection,
CFD modeling using the periodic feed pressure technique
In the previous discussion, it was mentioned that an effective way to minimize the problem of fouling may be to interfere with the different hydrodynamic forces that lead to the deposition of oil droplets at the surface of the membrane. The idea has been to interfere with the previously mentioned mechanisms that affect the stability of oil droplets at the surface of the membrane. It has been determined that the TMP represents a key parameter that plays a significant role in all the scenarios
Conclusion
The introduced PFPT is very effective in eliminating the development of fouling. This technique is based on alternating the pressure in the feed channel between the regular TMP and zero. Therefore, in part of the cycle where the pressure is set to the TMP, permeation occurs and in the second part when the pressure is zero, cleaning of the surface occurs. This novel work shows the development of a new filtration setup whereby fouling is minimized without impairing the overall filtration process.
Acknowledgment
M.Z. thanks the Faculty of Graduate Studies and Research at the University of Regina, CBIE (Canadian Bureau of International Education), and the Libyan government for financial support.
References (45)
- et al.
Role of surface wettability and roughness in emulsion separation
Separ. Purif. Technol.
(2013) - et al.
Slow colloidal aggregation and membrane fouling
J. Membr. Sci.
(2010) - et al.
Flow split characterization of two immiscible phases with different wettability scenarios: a numerical investigation using a coupled Cahn–Hilliard and Navier–Stokes system
Int. J. Multiphas. Flow
(2018) - et al.
Novel magnetically induced membrane vibration (MMV) for fouling control in membrane bioreactors
Water Res.
(2012) - et al.
Microfiltration of oil in water emulsions: effect of membrane microstructure and surface properties
Chem. Eng. Res. Des.
(2017) - et al.
Effects of crossflow velocity and transmembrane pressure on microfiltration of oil-in-water emulsions
J. Membr. Sci.
(2012) - et al.
Oil dropletbehavior at a pore entrance in the presence of crossflow: implications for microfiltration of oil–water dispersion
J. Membr. Sci.
(2013) - et al.
Produced water treatment by membranes: a review from a colloidal perspective
J. Colloid Interface Sci.
(2017) - et al.
A hybrid particle level set method for improved interface capturing
J. Comput. Phys.
(2002) - et al.
Oil coating of hydrophobic surfaces from aqueous media: formation and kinetic study
J. Colloid Interface Sci.
(2005)
Review of technologies for oil and gas produced water treatment
J. Hazard Mater.
A new way to apply ultrasound in cross-flow ultrafiltration: application to colloidal suspensions
Ultrason. Sonochem.
Volume of fluid (VOF) method for the dynamics of free boundaries
J. Comput. Phys.
Theoretical analysis of particle trajectories and sieving in a two-dimensional cross-flow filtration system
J. Membr. Sci.
Effects of spacer orientations on the cake formation during membrane fouling: quantitative analysis based on 3D OCT imaging
Water Res.
Effect of hydrophilicity on fouling of an emulsified oil wastewater with PVDF/PMMA membranes
J. Membr. Sci.
Hydraulic effects on sludge accumulation on membrane surface in crossflow filtration
Water Res.
Membrane technology enhancement in oil-water separation. A review
Desalination
Effect of turbulence on fouling control of submerged hollow fiber membrane filtration
Water Res.
Impact of gas bubbling and backflushing on fouling control and membrane cleaning
Desalination
Effect of transmembrane pressure pulsing on concentration polarization
J. Membr. Sci.
Investigation of the problem of oily-wastewater filtration using rotating porous membrane: a multicontinuum study
Colloids Surf., A
Cited by (35)
A superhydrophilic battery sponge with positive and negative electrodes for the demulsification of oil-in-water emulsions
2024, Separation and Purification TechnologyA complete quasistatic model for the permeation of a droplet between two reservoirs
2024, Colloids and Surfaces A: Physicochemical and Engineering AspectsA comprehensive review on the behavior and evolution of oil droplets during oil/water separation by membranes
2023, Advances in Colloid and Interface ScienceInvestigation of the self-propulsion of a wetting/nonwetting ganglion in tapered capillaries with arbitrary viscosity and density contrasts
2023, Colloids and Surfaces A: Physicochemical and Engineering AspectsUltra-selective microfiltration SiO<inf>2</inf>/carbon membranes for emulsified oil-water separation
2022, Journal of Environmental Chemical EngineeringCitation Excerpt :This change tendency is in line with the hydrophilicity and average pore size of carbon membranes as shown in Fig. 9 and Fig. 6, respectively. Usually, the separation performance of carbon membranes is related to numerous factors such as gel/cake layer formation, pore structure evolution and surface properties [71,72]. When the SiO2 content is higher, it provides carbon membranes with more oil rejection due to the enhanced hydrophilic contribution and porous modulation of the membrane matrix.