Separation Principle

2.1. MD phenomenon

As shown in Fig. 1, a liquid solution with high temperature is brought into contact with one side of a porous, hydrophobic membrane. The membrane which is a porous thin flexible sheet or tube acts as a barrier to separate the warm solution (called the feed side), and the permeate in either a liquid or a gas phase enters into a cooling chamber (called the permeate side), lhe hydrophobic nature of the microporous membrane prevents liquids / solutions from entering its pores due to the surface tension forces. As a result, a fixed interface is formed at the pores entrances. Fig. 2 |2] gives a cross sectional view of supposed straight cylindrical pores in contact with an aqueous solution to show how the vapor-liquid interfaces arc supported at the pore openings. If the solution contains at least one volatile component, temperature difference at two ends of the pores produces a vapor pressure gradient within the pores. By the driving force the vapor molecules of volatile component (produced by evaporation from the feed solution at the vapor-liquid interface) migrate from the feed side to the permeate side of the membrane. At the permeate side the immigrated moleculcs (depending on the membrane configuration used) are either condensed or removed in vapor form from the membrane module. In this way the solution from the feed side is concentrated. In summary, membrane distillation is, by its nature, a combination of membrane separation and evaporation / condensation process, and a microporous hydrophobic membrane is employed to act as the supporter of vapor liquid interfaces.

Fig. 1. MD phenomenon

Fig. 2. Vapor liquid interface in MD; adapted from the source [2].

2.2. Definition of MD process

Compared with such separation processes as osmotic distillation and pervaporation, the features of MD process [3] are:

(1) The membrane should be porous;

(2) At least one side of the membrane should keep in contact with the liquid to be dealt with;

(3) The driving force for each component to pass through the membrane pores is its partial pressure gradient in the vapor phase inside the pores;

(4) The membrane shouldn't be wetted by the liquid to be dealt with;

(5) No capillary condensation of vapor takes place inside the membrane pores;

(6) The membrane doesn't alter the vapor-liquid equilibrium of different components in the feed.

It is generally thought that the term "membrane distillation" arises from the similarity of this process to the conventional ordinary distillation process. Both processes depend on vapor-liquid equilibrium as the basis for separation and latent heat should be supplied to produce vapor phase. However, the remarkable difference between these two processes is that in conventional distillation the feed solution must be heated up to the temperature of its boiling point, whereas it isn't necessary for MD. It will be mentioned afterwards that MD can be performed well with the feed temperature lower than its boiling point. Moreover, the components to be separated can have closc boiling point or form azeotrope. So MD belongs to a special distillation process.

2.3. Membrane characteristics

Fig, 3. Micrograph of flat sheet membrane (PVDF) by scanning electronic microcopy.
Fig. 4. Micrograph of hollow fiber membrane (PVDF) by scanning electronic microcopy; adapted from the source [56J

Although MD is still emphasized as a low cost, energy saving and potential alternative to traditional separation processes, special membranes only for this purpose haven't been provided yet. At present the membranes employed in MD are those made for micro filtration purposes, because most of the required specifications by MD processes are available from those membranes. Hydrophobic microporous membranes made of polypropylene (PP), polyethylene (PE), polytetrafluoro ethylene (PTFE) and polyvinyl id ene fluoride (PVDF) are the commonly used membranes in MD, These membranes are fabricated in the form of either flat sheets or hollow fibers. A micrograph of a flat sheet membrane made of PVDF is shown in Fig. 3, as well as a hollow fiber in Fig. 4. By far, there are many kinds of commercially available microporous membranes maybe feasible for MD. However, as pointed out by many researchers [4], the requirement for MD membranes is: a higher permeability, lower membrane thickness, higher liquid entry pressure and low heat conductivity.

The membrane used in MD can exert its influence on transmembrane vapor flux of volatile component in three aspects [5J. Firstly, the vapor molecules move only through the pores of the membrane so that the effective area for mass transfer through the membrane is less than the total membrane area. Secondly, for most of the practical membranes, the membrane pores don't go straight through the membrane, and thus the path for vapor transport is greater than the thickness of the membrane. Thirdly, the inside wall of pores also increases the diffusion resistance by depleting the momentum of the vapor molecules,

A microporous membrane is generally characterized by four parameters, i.e. the thickness, 5 (m), the mean pore size, diameter d or radius r (m), the porosity z (defined as the porous volume fraction relative to the total membrane volume) and the tortuosity z (defined as the ratio of pore length to membrane thickness). Each of the parameters influences the permeability of the membrane. The relationship with the permeability can be summarized as r°e permeability ac-

where a may be equal to 1 or 2, depending on the predominant mass transfer mechanism within the membrane pores (about the detail see 3. transport process). It is evident that in order to obtain a high permeability or MD productivity, the membrane acting as a supporter of the vapor-liquid interface should be as thin as possible and its surface porosity and pore size should be as large as possible. Table 1 shows the usual characteristic values of hydrophobic membranes used in MD. The porosity of a membrane mainly depends on the technology of membrane preparation. Now the membrane with the porosity up to 90% is commercially available [6], As for the thickness, there is a trade-off between high productivity and strength of the membrane. But another trade-off also exists between high productivity and high selectivity. The larger the pore size is, the more easily the pores are wetted by the feed solution.

Table 1

The usual values of characteristics of membrane used in MD

Table 1

The usual values of characteristics of membrane used in MD

Mean pore diameter




0.1 - 1.0 |am

40 - 90%

20 - 200 |4m

0 0

Post a comment