Overview

Contents

1.1 Membrane Technology - What is it?
1.2 The Development of Membrane Technology
1.3 The Driving Forces of Separation
1.4 Purification, Concentration, Fractionation
1.5 Performance Limits
1.6 Membrane Structures
1.7 Quality, Productivity, and Life

1.1 Membrane Technology - What is it?

Membrane technology is devoted to the separation of the minutiae of particles ranging from bacteria to atoms. To some people the concern is simply the removal of this detrious. To others the recovery of the inhabitants of this sub-microscopic kingdom is the essential goal. In size its constituents span some 4 orders of magnitude, and they are dominated by colloidal/molecular forces, rather than by the gravitational forces of their larger brethren. The various inlet and outlet streams can be all liquids, all gases or combinations. Not surprisingly membrane technology is not one technology but many technologies with one common aspect; the use of a membrane which separates two streams enabling materials to be selectively transported across it. As might be expected there is plenty of commonality between these various membrane processes, but, equally, the diversity and range of applications mean that there are significant differences. In recognition of these differences a classification of membrane processes has developed.

Of the various membrane technologies, the class of membrane filtration is the largest and most diverse. One of the commonest questions is where does conventional filtration end and membrane filtration begin. In a similar vein where does ultrafiltration take over from microfiltration. To answer this sort of question can be likened to defining where does the desert end and arable land begin; the two are clearly different but there is obviously some arbitrariness in defining the boundary. Nevertheless, a semantic definition provides a quick and expedient guide as to what to expect. However, to focus too heavily on the boundary is to miss the point. Customers are not interested in whether something lies on one side or other of a boundary but on what that something can do for them. The purpose of a classification is to convey the potential use.

Membrane technology is generally regarded as addressing the separation needs of sub-micron particles. Selectivity comes through the interaction between the membrane and the surrounding phases. Two factors contribute to selectivity, the partitioning of molecules and or particles between the membrane and the surrounding phase, and the relative diffusion rates of these materials once in the membrane. It is invariably the product of these two factors which contributes to the overall selectivity of the membrane.

One feature that is common to many membrane processes, though not to all, is cross-flow. Cross-flow involves moving fluid tangentially across the membrane surface (see figure 1.1) as well as normal to it. The benefit is that particles/solutes that would otherwise accumulate at the membrane surface are moved along, achieving a steady-state distribution of particles or solutes at the interface, rather than the continually developing one that is seen in conventional filtration. The consequence of cross-flow is that in continuous operation the flux through the membrane tends to a constant while in conventional filtration the flux continues to fall. If higher fluxes are desired then higher cross-flows are required.

The benefits of cross-flow do not come without a penalty, which is the energy required to move the fluid across the surface. Fortunately, the additional cost is small compared to that required in conventional filtration to push the fluid through a filter cake. A key factor in this effect is the ratio of the cross-flow to the permeate flow. Not surprisingly, this ratio is a key aspect underlying the design of membrane elements, and selecting optimal operating conditions.

Another consequence of cross-flow is that the system is basically designed to remove only a small proportion of the feed. Thus a feature of most membrane plants is how to design systems to overcome this limitation (see Chapter 8).

In the last few years the boundary between conventional filtration and membrane filtration has been further blurred with the development of hybrid processes. These processes allows some-build up of material at the membrane surface but then the material is dislodged by passing water or air back through the membrane. By repeating this process at frequent intervals (circa 15 min) a reasonable flux through the membrane can be maintained. In this way the deposits on the surface have limited effect and the membrane remains the controlling factor.

1.2 The Development of Membrane Technology

Membrane technology grew out of a 19th century endeavour to investigate a kingdom of particles too small to be seen. With no way of seeing these sub-microscopic constituents, membranes proved to be a useful tool to probe these invisible components. The resulting exploration that ensued provided key ingredients in the development of molecular theory of matter, which burst onto the scene at the start of the 20th century. In contrast it took nearly a 100 years to engineer membranes from a scientific tool to an industrial tool.

The Early Years - A Scientific Tool

A significant contributor in these early years was Thomas Graham, a Scottish chemical physicist and Master of the Mint. In 1861 he discovered that substances like salt and sugar rapidly passed through parchment, whereas material like gum arabic and gelatin would not pass. Materials that permeated he called crystalloids, since these materials could easily be crystallised. Those materials which did not pass, typified by glues, which at the time he believed did not crystallise, he called colloids after the Greek word for glue (Kolla). Graham showed how colloidal material could be purified from crystalloid contamination by putting the colloid in a porous container which was then placed in running water. The crystalloids pass through and the colloids remain. This process he called dialysis and the transport through - osmosis.

Thomas Graham made another important contribution as a result of studying the diffusion of gases through flat rubber membranes. In explaining his results he regarded the rubber as a liquid in which the gas dissolves and then diffuses due to a concentration gradient. This is the so called solution-diffusion mechanism which is an important element in the molecular theory of transport in some of the membrane technologies.

Another early contributor was Thomas Fick, of Fick's law fame. In 1855 he made a membrane by dissolving collodion (cellulose nitrate) in ether/alcohol solution which he then coated onto a ceramic thimble. This enabled him to dialyse biological fluids.

Membranes Coming of Age - A subject of scientific investigation

The first half of the twentieth century saw membranes themselves become the topic of investigation. Bechold provided the first systematic study, and coined the term "ultrafiltration", He pointed out that in addition to particle size effects adsorption processes play a role in the degree of separation that is achieved. This was perhaps the first clear recognition that membrane filters frequently involve more than a mechanical basis of separation i.e. one depending purely on size. In 1911 Donnan published his work on the distribution of charged species across a semi-permeable membrane. Teorell and Meyers and Sievers were able to build on this and provide a model for the behaviour of charged membranes which is the basis of much of our understanding of electrodialysis membranes.

By 1927 membranes were in sufficient demand for Sartorius to start selling ultrafiltration and microfiltration membranes. This commercial reality though was largely aimed at those who used membranes as a laboratory tool rather than an industrial tool.

Research into the nature of the microporous structure of membranes was severely hampered by a lack of tools to investigate these structural aspects. A significant development came in the 60's with the application of electron microscopy which allowed an understanding of the relationship between manufacturing variables and membrane morphology. At last the science that underpinned the empirical development of membrane manufacturing processes became understood, and meant that new manufacturing processes could be quickly developed the new generation of synthetic polymers such as the polysulphones could be exploited.

The Development of Membrane Technology - Commercialisation

Large scale commercial application of membrane technology started in the 50's, with the development of electrodialysis membranes for the desalination of brackish water. The next major development was by Loeb and Sourirajan who successfully modified an ultrafiltration cellulosic membrane to create a viable reverse osmosis membrane for desalination of brackish water. This opened the door, and by the mid 60's a number of companies had developed systems. Most notably to General Atomic (now Fluid Systems) who by 1965 had manufactured and built the first large scale reverse osmosis plant. This industrialisation catalysed other membrane applications and developments. In particular it spurred on the development of ultrafiltration membranes for industrial usage, with applications like paint recovery in the electrocoat process. A process which is now used throughout the automotive industry. Another development of the 60's was Nafion. As part of a study into fuel cell technology by NASA, DuPont developed a hydrophilic type of PTFE by grafting onto the extremely hydrophobic polyfluoroethylene backbone, side chains with charged groups. It was quickly recognised that this material could be exploited in the extremely challenging application of the production of chlorine and caustic from salt. The 70's and 80's saw a number of chemical companies trying to use their more advanced synthetic polymers and skills to enter the membrane market. One chemical company which had an initial success was Monsanto who developed the Prism membrane, based on polysulphone, for gas separation. The interest of chemical companies waned in the late 80's and many who had entered in the 70's and 80's exited in the 90's as they sought to streamline their businesses.

A major development of the late 70's was the development of the composite membrane by Cadotte et al (see ref [10] for history of development). They had recognised that conventional reverse osmosis membranes were limited because different regions of the membrane had to carry-out the duties of mechanical support, and separation. They reasoned that if the separation layer and the mechanical support could be manufactured from different materials and tuned to the demands of each function, it should be possible to create a higher performing membrane. After many false starts they eventually succeeded in generating a good interfacial composite membrane that surpassed many others and laid the foundations for Film Tec which was later bought up by Dow in the late 80's.

As products have become established, suppliers have tried to open up new markets with varying degrees of success. The 90's brought a new factor into the equation, that of the environment. This has impacted on both the waste and supply side. Perhaps the largest single development has been the developing ultrafiltration and microfiltration technology for use in the municipal production of potable water to deal with cysts, bacteria, and viruses. What characterises many of these developments is not the universality of the technology but how the technology has to be developed for each application segment.

As table 1.2.2 highlights different applications demand different membrane technologies (see table 1.2.3). Sometimes different membrane technologies can be used to solve the same problem. For example both reverse osmosis and electrodialysis can be used to produce potable water from sea water. In the former water is passed through the membrane, while in the latter the salts are removed. A comparison to determine which is best inevitably depends on the customers requirements, and circumstances. Despite the obvious differences in the various membrane technologies there are many common features at a fundamental level (see Chapter 2 and 3).

1.3 The Driving Forces of Separation

For processes like crystallisation, distillation, adsorption, the separation achieved is related to the thermodynamic stability, with kinetics serving to dictate the time and size of plant required. In contrast for membrane processes separation is determined by the relative kinetics of permeation, with thermodynamics providing the time-scale and size of plant required.

Irreversible thermodynamics provide the framework for understanding membrane separations. The driving force for separation comes from gradients in thermodynamic variables. Commercial separation processes are governed by the differences in 1 or more of four thermodynamic factors

* Pressure

* Concentration

* Electric Potential

* Temperature

that exist between two phases being separated by the membrane. In response to these forces there are flows of mass, heat, electricity. At a local level the relationship between the forces, Xi and fluxes, Jj is a linear one of the general form/

Ji = ΣjLijXj (1.3.1)

where the Lij. are phenomenological coefficients to be provided either by experimentation or molecular theories. These coefficients occur in a variety of problems and many have been given names (see table 1.3.1). In the application of these principles to membranes the problem frequently becomes more complex in that the coupled sets of equations have to be solved over regions. Nevertheless the linear nature of the equations means that the fluxes are in general related to the differences in the thermodynamic properties of the two phases on either side of the membrane.

These thermodynamic forces are exploited in a number of different ways to give rise to a wide variety of membrane processes (see Table 1.3.2).

For the most part the different processes arise out of using different membranes to meet the different needs of applications. Some work has gone into membrane processes which used combined fields; most notably the use of potential fields to control fouling in pressure driven processes. To date though these process have not proved sufficiently attractive for large-scale development.

While thermodynamics provides a framework for separation it does not provide a mechanism for separation. Even in filtration processes it was recognised that separation was more than that of sieving. In general the relationship of

* Size

* Charge

* Affinity

between the membrane and the feed all play a role in determining the selectivity. The relationship between the size of particulates in the feed and the pores of a microporous membrane is a basis for separation. However, for large polymeric molecules conformational fluctuations can allow them to slip through pores much smaller than their radius of gyration might suggest. One can express this relationship as a solubility with size exclusion being an extreme example of the relationship. Size and conformation are not the only factors that determine solubility. Most colloidal materials carry some charge. If the charge on the colloid and the membrane are similar then there will be a tendency to exclude the colloidal material. Another fundamental factor that influences the separation is how fast the molecule or particle diffuses. At a fundamental level a recurring theme in membrane separation is that the power to separate two constituents is given by the the ratio, α, of the solubility, s, times diffusivity, D, for each component

αi/j [equivalent to] si Di/sj Dj (1.3.2)

from which it can be seen that the separation power is a product of a thermodynamic factor (relative solubility) and a kinetic factor (relative diffusivity).

A general equation which embodies the issues discussed above is

Solute Flux = Concentration * Mobility * Force (1.3.3)

One further point which is worthy of attention is that the thermodynamic force is the gradient of the thermodynamic parameter. In many cases this can be approximated by the difference in the thermodynamic variable across the membrane divided by its thickness. Thus, the flux can be increased by decreasing the film thickness or increasing the thermodynamic difference. Films as thin as 500 nm have been manufactured. Such thin films create many practical issues of how to handle and support them, without damaging them. For other reasons flux values are usually set by the application and other operating conditions, so reductions in thickness are usually reflected in reductions in the applied force.