SOLAR-DRIVEN MEMBRANE DISTILLATION OVERVIEW
Membrane distillation (MD) technology requires low-grade thermal heat and a small amount of pumping electrical energy, which could be wholly supplied by solar thermal collectors (e.g., flat plate collectors or evacuated tube collectors) and photovoltaics (PVs), respectively. This makes MD an excellent fit to be driven by solar energy. This article provides a brief overview of the different membrane distillation technological options, including module design, membrane configurations, and appropriate integration of this technology with the solar resource.
1. MEMBRANE DISTILLATION OVERVIEW
The membrane distillation process is a thermal-driven membrane-based separation process that separates salts and other impurities from a feed solution (e.g., seawater or groundwater) using the partial vapor pressure difference across the membrane. This is different from other membrane-based separation technologies, such as reverse osmosis, forward osmosis, electrodialysis, or capacitive deionization, which requires electrical energy whether to drive a high-pressure pump or to apply an electrical potential to separate the salts from the seawater. Generally, MD membranes must possess the following characteristics:
The membranes should be hydrophobic (to avoid wetting), because only vapor is to be transported across the membrane pores and no condensation should occur inside these pores.
The membranes should be inert and should not change or alter the vapor equilibrium of the feed solution.
At least one side of the membrane must always be in direct contact with the feed solution.
MD membranes must be manufactured from polymers with high permeability and low surface energies. Generally, these membranes are designed with a pore size of 10 nm–1 μm and porosity of 70–80%. The narrow pore size is critical to only allow small vapor particles while repelling large liquid particles (i.e., avoid membrane wetting). Also, the membranes need to be as thin as possible with a low pore tortuosity to avoid a long vapor transport distance.
The most common material used to manufacture MD membranes is polytetrafluoroethylene (PTFE) since it is naturally hydrophobic with good chemical resistance and high thermal stability, but they are difficult to manufacture. An alternative is to use polypropylene (PP), but as it is naturally hydrophilic and requires a additional coating to make it hydrophobic, which increases its manufacturing cost. A third common membrane material is polyvinylidene fluoride (PVDF), which is naturally hydrophobic and has high thermal stability, excellent mechanical strength, and chemical resistance. Table 1 shows a list of commercially available membranes commonly used in MD.
TABLE 1: Commercially available membranes commonly used in MD (Reprinted from Khayet and Matsuura with permission from Elsevier, Copyright 2011)
Manufacturer | Membrane Trade Name | Material | Thickness (μm) | Average Pore Size (μm) |
---|---|---|---|---|
3M | 3M# | PP | < 100 | * |
Gore | Gore | PTFE+ | < 50 | * |
PTFE+ | 400 | 2.00 | ||
Millipore | GVHP | PVDF | 110 | 0.45 |
HVHP | PVDF | 140 | 0.22 | |
Enka (Akzo) | Enka | PP | 100 | 0.20 |
PP | 140 | 0.10 | ||
Accurel # | PP (Tube) | 150 | 0.43 | |
Gelman Inst. Co. | TF# | PTFE+ | 60 | 0.20 |
PTFE+ | 60 | 0.45 | ||
PTFE+ | 60 | 1.00 | ||
Hoechst-Celanese | Celgard # | PP | 25 | 0.02 |
Celgard X-20 | PP (Tube) | 25 | 0.03 | |
*Wide range of pore sizes has been given. +PTFE membranes need fabric support that is not included in the given thickness values. |
Zooming in around the membrane, there are three key layers: (i) the hot feed solution layer, (ii) the hydrophobic permeable membrane layer, and (iii) the permeate layer, as shown in Fig. 1. The distillation process happens along the feed-membrane boundary layer, where the permeate vapor diffuses across the membrane pores to the permeate side. On the permeate side, cold water condenses the diffused permeate vapor before being collected as freshwater.
![Membrane distillation different layers schematic (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)](/content/10197/Fig1.png)
Figure 1. Membrane distillation different layers schematic (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
Among other conventional thermal-desalination processes (e.g., multi-effect distillation, multistage flashing, or mechanical vapor compression), MD requires a relatively low feed temperature, allowing it to be driven by low-grade heat, such as industrial waste heat or solar energy. In addition, its ability to treat a wide range of feedwater qualities (from brackish to seawater) with a high product water quality, makes it one of the few technologies that could achieve the zero liquid discharge model. However, treating these high salt concentration solutions can have a harmful impact on the membranes by accelerating scaling and fouling issues. Table 2 summarizes the membrane distillation typical operating conditions as well as production rates and permeate quality.
TABLE 2: Membrane distillation operating conditions
Parameter | Value |
---|---|
Operating temperature (°C) | 40–90 |
Specific energy consumption (kWh/m3) | 15–750 |
Typical system capacity (m3/day) | 1–456 |
Unit cost of water (USD/m3) | 0.5–20 |
Feedwater salinity range | Brackish to seawater |
Sensitivity to feedwater salinity | Low |
Product water quality (ppm) | 1–40 |
Pretreatment requirement | Low |
Scaling and fouling issues | Medium–High |
Maintenance and operation requirements | Low |
2. MEMBRANE DISTILLATION MODULE DESIGN
The membranes used in the MD module can be manufactured in four different configurations: (i) flat sheet membranes, (ii) tubular membranes, (iii) spiral wound membranes, and (iv) hollow fiber (capillary) membranes. Flat sheet and tubular membranes are commonly used in the MD process due to their simple operation, but they have limited productivity compared to the spiral wound and hollow fiber membranes. Table 3 compares the different membrane configurations based on their packing density, cleaning difficulty, fouling resistance, and manufacturing cost. Combining the same family of membranes into a shell constructs a module that can operate in four different operation modes: (i) direct contact membrane distillation (DCMD), (ii) air gap membrane distillation (AGMD), (iii) sweeping gas membrane distillation (SGMD), and (iv) vacuum membrane distillation (VMD), as shown in Fig. 2.
TABLE 3: Membrane configuration comparison (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
Membrane Configuration | Packing Density (m2/m3) | Cleaning Difficulty | Fouling Resistance | Manufacturing Cost |
---|---|---|---|---|
Flat sheet | 45–150 | Easy | Moderate | High |
Tubular | 6–120 | Great | High | Very high |
Spiral wound | 150–380 | Difficult | Low | Moderate |
Hollow fiber | 150–1500 | Difficult | Very Low | Low |
![]() | ![]() |
(a) | (b) |
![]() | ![]() |
(c) | (d) |
Figure 2. Membrane distillation operation modes: (a) direct contact membrane distillation; (b) air gap membrane distillation; (c) sweeping gas membrane distillation; and (d) vacuum membrane distillation (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
The DCMD operation mode is the simplest and the most used. In this configuration, the hot feed and the cold permeate are in direct contact with the membrane, where water vapor diffuses across the membrane with minimum mass resistance. This diffused permeate vapor is condensed as it gets in contact with a cold fluid (e.g., cold water) flowing on the other side of the membrane. However, the conductive heat losses are significant in this operation mode since the hot feed solution and the permeate vapor are in direct contact with the membrane.
The AGMD operation mode is like the DCMD but uses an air gap between the permeate and the cold side to mitigate the high conductive heat loss. However, this extra air gap induces a larger vapor diffusion distance that increases the mass transfer resistance.
The SGMD operation mode uses an inert gas to sweep the permeate vapor out of the MD module. This reduces the conductive heat losses and keeps a short vapor diffusion distance for the permeate vapor to travel, but it requires an external condenser to condense the permeate vapor. In addition, it requires another separation process to remove the inert gas from the permeate vapor.
The VMD operation mode applies a vacuum pressure at the permeate side to accelerate the process by “pulling” the permeate vapor via an increase in the membrane's partial vapor pressure difference. This results in a significantly higher permeate flux than other operation modes but at the cost of high specific energy consumption.
3. MEMBRANE DISTILLATION MASS TRANSFER MODEL
The transport of water vapor through the membrane pores has been extensively studied, and theoretical models have been developed based on the kinetic theory of gases. The membrane distillation mass transfer model is described by dusty gas models, such as the Knudsen model, Poiseuille model, Knudsen–Poiseuille transition models, and the molecular diffusion model. The selection of an appropriate mass transfer model is important because it depends on the membrane and feed properties (Phattaranawik et al., 2001). Mostly, these models explain that the permeate mass flux (J) is a linear function based on the partial vapor pressure difference across the membrane (ΔPv) and the membrane permeability (βm), as follows:
![]() |
(1) |
It is often assumed that the membranes are made of uniform and non-interconnected cylindrical pores with mean pore size. This could oversimplify the model, requiring real-life experimental data to validate those assumptions and models. Another important parameter is estimating the amount of air present within the membrane pores. This trapped air could hinder the mass transfer and MD performance. However, the effect of the molecular diffusion resistance from the trapped air is considered negligible compared to the water vapor flux within the pores (Khayet, 2011). Thus, the molecular diffusion resistance could be neglected when estimating the membrane permeability (or membrane distillation coefficient).
The membrane permeability is a function of the membrane properties, vapor properties that are transported across the membrane, and the membrane mean temperature (T̅m) (i.e., the average temperature of the membrane between the feed and the permeate), shown as follows:
![]() |
(2) |
where r, ε, t, τ are the mean pore size, membrane porosity, membrane thickness, and membrane tortuosity, respectively. Mw and R are the water molecular mass and ideal gas constant, respectively.
4. MEMBRANE DISTILLATION HEAT TRANSFER MODEL
The temperature gradient and the nonzero velocity field of the feed relative to the membrane interface create thermal boundary layers on both sides of the membrane at the hot feed and the permeate sides, which results in the evaporation process. When the thermal boundary layers on both sides of the membrane are similar, this results in a heat transfer resistance between the liquid–vapor interface, which could be measured using the temperature polarization coefficient (Θ), as follows:
![]() |
(3) |
where Tf–m and Tp–m are the feed-membrane and permeate-membrane temperatures, respectively, and T̅f and T̅p are the average feed and permeate temperatures, respectively. As a reference, a highly efficient process would have a temperature polarization coefficient that approaches 1, indicating a good mixture of heat and low mass transfer resistance. This could be achieved by using baffles that promote turbulent flow and would result in a temperature polarization coefficient in the range of 0.90–0.97 compared to 0.5–0.76 for unbaffled modules (Tamburini et al., 2013).
Another important variable that defines the effectiveness of the membrane distillation technology is the heat transfer coefficient of its membrane. It is important to accurately measure the heat transfer coefficient because it defines the mass transfer properties (i.e., productivity) of the MD process. There are different heat transfer models that are based on the Nusselt number on the outer surface of the membrane, which can evaluate the heat transfer coefficient (hf) based on the following:
![]() |
(4) |
where dh is the hydraulic diameter of the module and kf is the feed solution thermal conductivity. To estimate the Nusselt number, heat transfer models that were developed based on experimental data and assumptions could be used to provide a reasonable estimation of the heat transfer coefficient. These correlations [Eqs. (5)–(7)] are summarized in Table 4.
TABLE 4: Membrane distillation heat transfer correlations, adapted from Phattaranawik et al. (2003)
Flow Regime | Correlations | Eq. Nos. | |
---|---|---|---|
Laminar |
|
(5) | |
|
(6) | ||
|
(7) | ||
|
(8) | ||
|
(9) | ||
|
(10) | ||
Turbulent |
|
(11) | |
|
(12) | ||
|
(13) | ||
|
(14) | ||
|
(15) | ||
|
(16) | ||
|
(17) | ||
Note: f = [0.79 ln(Re) – 1.64]–2 for all correlations |
As shown in Table 4, the Nusselt number correlations are functions of the Reynolds and Prandtl numbers, which are evaluated based on the following expressions:
![]() |
(18) |
![]() |
(19) |
where ρf, μf, and cpf are the feed density, dynamic viscosity, and specific heat at constant pressure, respectively.
As can be inferred from Eqs. (18) and (19), the feed temperature significantly affects the heat transfer coefficient, which then reflects on the MD module's permeate flux (or production). As such, it is important to operate the MD module at a high feed temperature (e.g., 60–90°C) to have a high production rate. Nevertheless, these temperatures could still be easily achieved using solar energy, making this technology a good choice to be powered by solar energy.
5. SOLAR-DRIVEN MEMBRANE DISTILLATION CONFIGURATIONS
Membrane distillation technology requires low-grade thermal energy, making it the best fit for solar integration with flat plate solar collectors (FPCs) or evacuated tube collectors (ETCs). In addition, it could be integrated with photovoltaic (PV) panels that can supply the necessary electric energy to operate the pumps. As such, there are various coupling techniques to drive a MD process using solar energy. In the former integration, the FPC or ETC can heat the feedwater to the desired temperature before entering the MD module to start the distillation process, as shown in Fig. 3.
![Solar thermal-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)](/content/10197/Fig3.png)
Figure 3. Solar thermal-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
To make the MD system operate using solar energy alone, the abovementioned design could be altered by adding PV panels to provide the electrical energy needed by the pumps, as shown in Fig. 4.
![Solar thermal + photovoltaic-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)](/content/10197/Fig4.png)
Figure 4. Solar thermal + photovoltaic-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
Another design is to couple both the thermal and electrical aspects of the solar panel by using a PV thermal collector to provide the necessary thermal energy to heat the feedwater and electrical energy to run the pumps simultaneously. A schematic of this configuration is shown in Fig. 5. This could decrease the system's capital cost since one collector provides all the necessary energy to operate the MD module.
![Photovoltaic thermal-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)](/content/10197/Fig5.png)
Figure 5. Photovoltaic thermal-driven membrane distillation configuration (Reprinted from Omar et al. with permission from Elsevier, Copyright 2022)
Another innovative idea is to submerge the membranes inside the solar tubes of the evacuated tube collector. This will reduce the pumping power necessary and grant the MD process with continuous feedwater heating along the membranes. A schematic of this process is shown in Fig. 6.
![Integrated solar-driven membrane distillation (Reprinted from Li et al. with permission from Elsevier, Copyright 2019)](/content/10197/Fig6.png)
Figure 6. Integrated solar-driven membrane distillation (Reprinted from Li et al. with permission from Elsevier, Copyright 2019)
6. CONCLUDING REMARKS
In recent years, much progress has been made in the development of membrane distillation and its integration with renewable energy, particularly solar energy. However, further work is needed to demonstrate these concepts since membrane distillation may face degradation over time and may suffer from intermittent operation (frequently switching on/off) when integrated with solar resources. Thus, robust operation and maintenance schemes need more development. In addition, low annual average thermal efficiency needs to be overcome via cost minimization of the solar energy harvested as MD has a relatively high specific energy consumption. Nevertheless, MD has been—and likely will continue to be—the subject of academic research to create compact solar-driven water treatment units for small- to medium-size applications (e.g., 1–500 m3/day). It is clear, however, that this research effort is indeed worthwhile; this technology has shown a lot of promise in becoming a pivotal solution to alleviate water-stressed regions that are coincidentally blessed with high solar resources.
REFERENCES
Khayet, M. (2011) Membranes and Theoretical Modeling of Membrane Distillation: A Review, Adv. Colloid Interface Sci., 164(1): 56–88.
Khayet, M. and Matsuura, T. (2011) Membranes Used in MD and Design, Membrane Distillation, Khayet, M. and Matsuura, T. (eds.), Amsterdam: Elsevier, pp. 17–40.
Li, Q., Beier, L.-J., Tan, J., Brown, C., Lian, B., Zhong, W., Wang, Y., Ji, C., Dai, P., Li, T., Le Clech, P., Tyagi, H., Liu, X., Leslie, G., and Taylor, R.A. (2019) An Integrated, Solar-Driven Membrane Distillation System for Water Purification and Energy Generation, Appl. Energy, 237: 534–548.
Omar, A., Li, Q., Saldivia, D., Nashed, A., and Van Dang, B. (2022) Solar-Driven Water Treatment: Generation III-Low Technology Readiness, in Solar-Driven Water Treatment: Re-Engineering and Accelerating Nature's Water Cycle, Cambridge, MA: Academic Press, pp. 201–261.
Phattaranawik, J., Jiraratananon, R., Fane, A.G., and Halim, C. (2001) Mass Flux Enhancement Using Spacer Filled Channels in Direct Contact Membrane Distillation, J. Membr. Sci., 187(1): 193–201.
Phattaranawik, J., Jiraratananon, R., and Fane, A.G. (2003) Heat Transport and Membrane Distillation Coefficients in Direct Contact Membrane Distillation, J. Membr. Sci., 212(1): 177–193.
Tamburini, A., Pitò, P., Cipollina, A., Micale, G., and Ciofalo, M. (2013) A Thermochromic Liquid Crystals Image Analysis Technique to Investigate Temperature Polarization in Spacer-Filled Channels for Membrane Distillation, J. Membr. Sci., 447: 260–273.
References
- Khayet, M. (2011) Membranes and Theoretical Modeling of Membrane Distillation: A Review, Adv. Colloid Interface Sci., 164(1): 56–88.
- Khayet, M. and Matsuura, T. (2011) Membranes Used in MD and Design, Membrane Distillation, Khayet, M. and Matsuura, T. (eds.), Amsterdam: Elsevier, pp. 17–40.
- Li, Q., Beier, L.-J., Tan, J., Brown, C., Lian, B., Zhong, W., Wang, Y., Ji, C., Dai, P., Li, T., Le Clech, P., Tyagi, H., Liu, X., Leslie, G., and Taylor, R.A. (2019) An Integrated, Solar-Driven Membrane Distillation System for Water Purification and Energy Generation, Appl. Energy, 237: 534–548.
- Omar, A., Li, Q., Saldivia, D., Nashed, A., and Van Dang, B. (2022) Solar-Driven Water Treatment: Generation III-Low Technology Readiness, in Solar-Driven Water Treatment: Re-Engineering and Accelerating Nature's Water Cycle, Cambridge, MA: Academic Press, pp. 201–261.
- Phattaranawik, J., Jiraratananon, R., Fane, A.G., and Halim, C. (2001) Mass Flux Enhancement Using Spacer Filled Channels in Direct Contact Membrane Distillation, J. Membr. Sci., 187(1): 193–201.
- Phattaranawik, J., Jiraratananon, R., and Fane, A.G. (2003) Heat Transport and Membrane Distillation Coefficients in Direct Contact Membrane Distillation, J. Membr. Sci., 212(1): 177–193.
- Tamburini, A., Pitò, P., Cipollina, A., Micale, G., and Ciofalo, M. (2013) A Thermochromic Liquid Crystals Image Analysis Technique to Investigate Temperature Polarization in Spacer-Filled Channels for Membrane Distillation, J. Membr. Sci., 447: 260–273.