Patent Application
20210354093
The invention relates to membranes for membrane distillation desalination technology. In particular, this invention is related to highly hydrophobic multi-walled carbon nanotube blended polyvinylidene fluoride (MWCNTs/PVDF) membranes. The invention also extends to the preparation process of the MWCNTs/PVDF membranes and their application in industry. The invented membranes have highly hydrophobic surface and sponge-like internal structures with interconnected crystalline nodules.
Membrane distillation (MD) is a separation technology that involves transport of water vapour via a microporous hydrophobic membrane where the driving force is the partial vapour pressure gradient generated by temperature differences between the membrane sides. The pressure gradient enables water vapours coming from warmer feed solution to migrate through the hydrophobic membrane pores to the cooler permeate side leaving behind non-volatile dissolved solutes.
Membrane distillation has shown many attractive features to the capacity of addressing the challenges that are currently facing conventional technologies for desalination such as reverse osmosis (RO), multi-stage flash distillation (MSF) and multi-effect distillation (MED). However, up to the present MD technology for desalination is not available for large commercial application. One of the major reasons being lack of specially designed membranes for MD application. Currently, MD research works, and pilot plants utilize membranes adopted from ultrafiltration and microfiltration separation technologies, which produce low water flux and are prone to fouling/scaling. It is well agreed among scientific community that new membranes need to be designed and tailor made for application in MD processes.
Membrane development approach that involve coupling membrane science and nanotechnology have attracted attention among scientist as viable direction towards solving the mystery of MD membranes. Among the different nano-materials, carbon nanotubes have a greater potential of producing unique membranes when blended within the membrane polymer matrix.
Carbon nanotubes (CNTs) are allotropic forms of carbon which consist of cylindrical sheets of graphite. Their appearance is like that of a rolled-up tube of fencing wire. There are two types of CNTs, single-walled (SWCNTs) and multi-walled (MWCNTs). The SWCNTs consists of cylindrical shape made up of one graphene shell. On the contrary, MWCNTs is made up of multiple cylindrical layers of graphene sheets.
Although CNTs have extensively been studied in the past few decades, its application in membranes water treatment, particularly desalination is still a concept. CNTs have several intrinsic properties, which make them attractive in water treatment, membrane technology in particular. The smooth and hydrophobic surfaces in the outside and inner core of their hollow tube, allows frictionless movement of water molecules with high flow rate. CNTs have the ability to prevent microbial growth on the membrane surface due to the inherent microbial activity they possess and hence can significantly prevent biofouling in membrane desalination activities. CNTs have also been reported to improve mechanical strength when blended in membranes.
Despite the promising potentials of CNTs in membrane desalination application, the difficulty in properly aligning CNTs on the thin surface layer of membranes still remain a challenge. Significant amount of work on incorporation of CNTs in membranes for water treatment is available in the literature for other membrane technologies such as reverse osmosis, ultrafiltration (UF), nanofiltration (NF) and microfiltration (MF). Less information is available on the application of CNT in membranes for MD application. Dumée et al., J. Memb. Sci. 351 (2010) 36-43 prepared CNT bulky paper membranes with surface contact angle as high as 113°, high porosity and relatively lower thermal conductivity (2.7 kW/m2 h) for direct contact membrane distillation (DCMD) desalination. However, the bulky paper membranes were reported to undergo delamination and cracking when applied in DCMD desalination. CNTs have also been reported to be immobilized into the pores of polypropylene hollow-fiber membranes using PVDF solution as gluing material (Gethard et al., ACS Appl. Mater. Interfaces 3, (2011) 110-114). These membranes were reported to have superior performance in MD application compared to their counterpart. However, the use of PVDF as a gluing solution lead to the reduction of membrane pore size as well as porosity for the CNTs immobilized membranes. Silva et al., Desalination, 357, (2015) 233-245, prepared MWCNTs blended PVDF membrane via phase inversion method for MD application. They found that PVDF membranes blended with 0.2% pristine MWCNTs were better performing in DCMD desalination than those blended with functionalized MWCNTs or commercially available membranes. Despite high membrane performance observed in DCMD, the MWCNTs blended membranes had lower surface hydrophobicity (contact angle<95°) compared to that of the reported commercial membrane (119°). Still more work needed to be done to uncover the potentials CNTs can offer in the area of membranes for MD application.
In the present invention, the advantages introduced using mixed solvent and dual coagulation bath plus the potential benefits offered by carbon nanotubes were employed to prepare composite PVDF membranes with unique properties suited for application in MD processes which addresses the shortcomings of the previously fabricated CNT composite membranes.
According to the invention, multi-walled carbon nanotube blended polyvinylidene fluoride (MWCNTs/PVDF) membranes for MD treatment of saline water are prepared using non-solvent induced phase separation (NIPS) methods. The synthesis procedure involves mixing two solvents with different solubility parameters and the use of dual coagulation bath system to control the formation of membrane pore structures and enhance surface hydrophobicity. By combining the two synthesis approaches in the presence of MWCNTs, blended PVDF membranes are produced for application in MD processes.
The method may include dispersing pristine MWCNTs (0.2 wt %) in an optimized mixture of triethyl phosphate (TEP) and N, N-dimethylacetamide (DMAC) of 1:1 mass ratio thereby producing 81 wt % of a final dope solution.
The MWCNTs dispersion may take place by sonicating the solution for not less than 1 hour.
4 wt % ethylene glycol (EG), a pore forming agent, may be added into the mixture followed by 15 wt % of PVDF.
The mixture may be mechanically stirred at 60-70° C. for up to 24 hours.
The dope solution may be left to cool at room temperature for up to 24 hours.
The prepared dope solution may be cast onto thin membrane films on a glass plate using casting knife at a thickness from 100 μm to 600 μm, typically 300 μm.
The glass plate with a nascent membrane on top may first be immersed in a coagulation bath containing 40% water and 60% mixture of TEP and DMAC (2:3) for a maximum of 30 seconds.
The glass plate with its membrane may then be dipped in a second coagulation bath containing pure water to complete the precipitation process.
Precipitated membranes may be soaked in deionized water for up to 24 hours to completely remove the solvents.
The membranes may then be dried in air at room temperature ready for characterization and application.
The invented membranes generally as described above were characterized and tested for their performance in DCMD desalination of seawater. The present invented MWCNTs/PVDF membranes were found to have the following inherent properties; Morphologically, the invented membranes were microporous with a rough membrane surface which is composed of spherical palate like nodules that are interconnected by a network of fibres (
The invented MWCNTs/PVDF membranes were tested for their performance in seawater desalination using a self-designed and fabricated laboratory DCMD unit and the results compared with that obtained in a similar way for a commercially obtained PVDF membrane and PVDF membranes synthesised without MWCNTs.
Results indicated that, the present invented MWCNTs/PVDF membranes had the highest average water flux (34.0±2.6 L/m2h) followed by PVDF membranes synthesized without MWCNTs (27.9±0.3 L/m2h). Under the same DCMD operation conditions, commercial PVDF membranes had the least average permeate flux (26.6±0.2 L/m2h). Salt rejection for all membranes were high (above 99.99%) with seawater. The high-water flux obtained from MWCNTs/PVDF membranes was contributed by the presence of MWCNTs which enhanced water vapour flow across the membrane by providing extra pathways as well as reducing friction forces on their walls.
The present invented MWCNTs/PVDF membranes have proved to be more efficient in terms of water flux in MD application compared to PVDF membranes without carbon nanotubes. The modified method of synthesizing MWCNTs/PVDF membranes have resulted into better performing membranes with respect to the commercially obtained membranes.
The unique features of the invented MWCNTs/PVDF membranes lies on the composition, morphology and methods in which the membranes were synthesized. The invented membranes are composed of PVDF polymer blended with pristine MWCNTs. The presence of MWCNTs in the PVDF membrane matrix provide extra pathways for water molecule but also, provide less flow resistance (less friction force) since the walls of MWCNTs offer friction free flow to water molecules.
The morphological properties of the invented MWCNTs/PVDF membranes are unique and are tailored to application of the membranes. The surface morphology of the invented membranes is composed of spherical-like palates interconnected with a network of fibres, which made the surfaces to look rough (
The method of synthesis of the invented MWCNTs/PVDF membrane was also unique since it combined several synthesis approaches into one. The membranes were prepared using a common non-solvent induced phase separation technique, however, a mixture of solvent (triethyl phosphate (TEP) and N, N-dimethylacetamide (DMAC)) (1:1) was used instead of pure solvent. Also, dual coagulation bath system was used where the membrane's precipitation rate was controlled by immersing the nascent membrane in a bath containing a mixture of TEP, DMAC and water. Later, a complete membrane precipitation was achieved by immersing the pre-conditioned nascent membrane into a bath containing pure water as a non-solvent. This combination of synthesis steps has not been reported before for PVDF membranes containing MWCNTs and thus makes it unique for this invention.
The invented membranes have shown to be suitable for application in membrane distillation desalination and related industry, where they provide higher water fluxes than currently available commercial PVDF membranes. Furthermore, the invented membranes have proved to be more resistant to fouling/scaling hence they can be used to treat saline water of very high concentration of dissolved solutes. As a result, the invented membranes can provide higher water recoveries than that obtained using commercial PVDF membranes. Meanwhile, salt rejections offered by the invented membranes are high even at a very high concentration of dissolved solutes in the sources/feed water.
The potential application of the invented MWCNTs/PVDF membranes is in the treatment of saline water sources to obtain pure drinking water by using membrane distillation technology. The membranes can also be used in pharmaceuticals for concentration of non-volatile drugs. The MWCNTs/PVDF membranes have additional application potential in fruit juice and dairy industries for making concentrates of their products.
In an experiment, performance in desalination by membrane distillation of multi-walled carbon nanotubes-polyvinylidene fluoride membranes was tested. The DCMD unit described here below was used to carry out membrane distillation investigation of the desalination performance of synthesized membranes.
Multi-walled carbon nanotube embedded PVDF membranes were prepared by non-solvent induced phase separation (NIPS) methods using DMAC and TEP solvent mixtures at a composition of 1:1 ratio. Optimal loading of MWCNT was used where 0.2 wt % of pristine MWCNT was dispersed in appropriate amount of DMAC and sonicated for an hour. An appropriate amount of TEP was then added into the dispersed MWCNT and sonicated for further 30 minutes. 15 wt % of PVDF polymer and 4 wt % of pore forming agent (ethylene glycol) were dissolved in the above mixture and mechanically stirred for 24 hours to ensure complete dissolution. The dope solution was allowed to cool at room temperature for 24 hours to remove air bubbles. The dope solution was cast into thin films of membranes on a smooth glass plate using a casting knife (Elcometer 3580/7) at a thickness of 300 μm. The glass plate with a nascent membrane thin film was first immersed in a coagulation bath made up of 40% water and 60% mixture of TEP/DMAC (2:3) for a maximum of 30 seconds. Thereafter the nascent membrane was transferred to a second coagulation bath containing pure water to complete the precipitation process. Fully precipitated membranes were soaked in deionized water for 24 hours to ensure that all solvents were completely removed. The membranes were then dried on air at room temperature, characterized and tested for desalination performance in MD. Another set of membranes that contained PVDF without MWCNTs (neat PVDF) were also synthesized for comparison purpose.
The dried synthesized membranes were analyzed using a range of characterization techniques to determine their physicochemical properties. An analytical scanning electron microscope (SEM), a JOEL JSM-6010PLUS/LA, was used for the analysis of surface and cross-sectional morphologies of the membranes. Gold sputtering was used for coating membrane samples and the scanning was performed at appropriate accelerating voltage and magnification.
Contact angle (CA) between membrane surface and water was measured using a Kruss DSA 30E instrument equipped with a video capture system. Sessile drop method was used where 2 μL drops of deionized water were placed on the surface of membrane samples using an automated syringe and the contact angle between water drop and membrane surface was calculated using advanced software in a computer used to control the instrument. A minimum of ten different sections on each membrane sample were analyzed where mean values were taken as the contact angle for the respective membranes. All the measurements were conducted at room temperature.
Membrane mechanical strength was analyzed on an Anton Paar SAXSpace instrument which was coupled with a TS600 tensile stage of 600 N load cell. Three samples from each membrane were analyzed at a 2.0 mm/min velocity and the mean value was considered as the strength of a membrane.
Membrane pore size and pore size distribution were determined by a capillary flow porometry method using a Quantachrome Instruments Porometer 3 Gz. The membrane sample was cut into a required size and placed into sample holder chamber. The sample was then wetted using Porofil wetting liquid and the sample chamber closed air tightly. Nitrogen gas was pressurized across the membrane and the wetting liquid forced to extrude on the other side of the membrane while opening membrane pores for the gas to flow. Large membrane pores opened first, and small pores opened last. Gas pressure and flow rate across the membrane was detected and recorded by the instrument software (3GWin V2 Porometer software) and used to determine pore size distribution of the membrane samples.
Water liquid entry pressure (LEPw), was also determined using the Quantachrome Instruments Porometer 3 Gz equipped with liquid permeability accessories. Membrane sample was placed in the sample holder and deionized water was pressurized over it. A single pressure point mode was used where a target pressure was set and allowed to hold for 10 minutes. The target pressure was increased at intervals of 10 kPa up to a point where water flow across the membrane was detected. The pressure at which water flow was observed was considered as the LEPW for the membrane.
A schematic diagram for the designed and fabricated DCMD module is presented in
The membrane cell was rectangular and was made of acrylic plastics (Perspex) to reduce heat loss to the environment. Flow channels was engraved in two acrylic blocks to make two half cells, feed flow channel half-cell and distillate flow channel half-cell. The two blocks were designed to be joined and tighten by screws while holding a flat sheet membrane in between which separated the feed and permeate flows into the cell. Each channel was 1.5 mm deep, 50 mm wide and 70 mm long to make a total membrane active surface area of 0.0035 m2 (
The DCMD module is made up of two independent water circulation loops, the raw water (feed) loop and the treated water (permeate) loop. Within the membrane cell the two circulation loops are separated by a microporous hydrophobic membrane which prevents liquid water to cross the membrane from one side to the other but allows water vapour to do so. In the fabricated DCMD module, a hydrophobic membrane is first installed in the membrane cell and the cell connected to the respective circulation loops. The membrane is installed in such a way the membrane active side faces the raw water when circulated into the cell. The raw water is filled in a feed reservoir and is heated to a desired elevated temperature. Deionized water (clean water) is filled in the treated water tank which is connected to a cooling device. The deionized water in the permeate tank is cooled to a lower temperature which causes condensation of water vapour once extracted from the raw water. Once temperatures of interest are reached in the feed tank and permeate tank, circulation pumps are switched on and the pumping speed is regulated to that of a required flow rate. Water flow meters installed in each circulation loop help to determine the flow rates of water circulated across the membrane in the membrane cell. Care is taken to make sure that no air bubbles are trapped in the membrane cell. After water circulation starts, the water level in the permeate tank decreases below the overflow arm. Additional deionized water is added into the permeate tank to a level where water start to flow through the overflowing arm. The overflow of deionized water in the permeate tank will stop after the system has stabilized. The process is then left to equilibrate for about 20-60 minutes depending on the feed water inlet temperature used. The higher the temperature the faster is the system equilibration due to increased MD driving force. During this time, the first drops of permeate which is a result of distillation will be observed in the overflow channel of the permeate tank. The permeate is then collected in a separate container from the overflow channel. The analyst starts recoding time at the beginning of the permeate collection to determine the rate of DCMD permeation (permeate flux) at a given period. Flux permeation is subject to several factors both operational and membrane based. Different membranes will have different rate of permeate production in similar operation condition due to the differences in membrane properties. Likewise, different operation conditions such as feed inlet temperature, feed solution concentration, flow rate and overall change in temperature across the membrane will result into different permeate flux for the same membrane.
Salt rejection efficiency in the DCMD desalination process is determined by measurement of electric conductivities of raw water before the beginning of the distillation process. Also, the conductivities are measured for deionized water before the process starts and for the collected permeate after the experiment.
The performance of synthesized MWCNTs/PVDF membranes in membrane distillation process was evaluated by carrying out DCMD experiments at different feed inlet temperatures using the described fabricated DCMD set-up. The feed inlet temperatures were varied from 40 to 70° C. and circulated at feed flow rate of 24 L/h. Permeate inlet temperature was maintained at 20° C. and a flow rate of 24 L/h throughout the experiments. Each experiment was allowed to run for at least 6 hours while water vapour flux and salt rejection measurement were taken after every 2 hours. After setting up each experiment, it was allowed to run and equilibrate before data recording was started. Initial and final electrical conductivities (EC) for the feed and permeate were determined using YSI Professional Plus field meter whereas volumes of collected permeate were determined by a measuring cylinder. Water vapour flux and salt rejections were later determined using Equations 1 and 2, where Cf is the concentration of feed and Cp is the concentration of permeate. Similar MD performance experiments were conducted for neat-PVDF synthesized membranes as well as commercially obtained PVDF membranes ((HVHP 29325) and the results were compared.
The results for membrane characterization experiments for contact angle, mean pore size, porosity, thickness and liquid entry pressure are presented in Table 1. Average contact angles between membrane surface and water drops of 135° were recorded for the synthesized membranes with and without MWCNTs. The higher wetting resistance of these membranes has resulted from removal of surface skin layer on the membranes, a role played by the dual coagulation bath system used.
Results obtained from MD performance experiments are as indicated in
On the other hand, all membranes were observed to have salt rejections of non-volatile dissolved solutes of higher than 99.98% across the entire feed inlet temperature range investigated (
| Number | Date | Country | Kind |
|---|---|---|---|
| 2018/06582 | Oct 2018 | ZA | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/ZA2019/050063 | 10/4/2019 | WO | 00 |
