Membranes are a permeable or semi-permeable matrix made of metal, polymer or other materials. These materials are capable of separating micron and sub-micron size particles from liquids and gases by retaining particles larger than the pores on the surface of the membrane. Ultimately, a membrane is a structure, or selective barrier by which mass transfer occurs by way of some driving force.
Before membranes were used commercially, there was a considerable amount of research conducted by scientists who utilized membranes as tools purely for scientific study. Research and testing was conducted on membranes to study and establish scientific principles, particularly in the fields of osmosis and diffusion. In 1748, Jean-Antoine Nollet performed the first documented study involving natural membranes to observe osmosis and membrane’s effect on osmosis.
All of the studies conducted initially containing membranes used natural membranes, which were constructed from animal bladders. Moritz Traube was the first person to successfully create an artificial membrane by combining copper sulfate and potassium ferrocyanide solutions. Adolf Fick then further expanded on the concept of artificial membranes to study the properties of diffusion, leading to Fick’s laws of diffusion.
Commercial investment and involvement in the membrane industry began when Sartorius started selling synthetic, nitrocellulose membranes for laboratory use in the 1920s. In the 1950s, there was a growing interest in the commercial potential of membranes and explorations into possible industrial applications of membranes increased.
During this time period, W.L. Gore developed and began selling Gore-Tex, an expanded PTFE membrane. In addition, developments in the use of membranes to produce safe drinking water helped increase interest in the market.
In 1960, Sid Loeb and Srinivasa Sourirajan engineered a break-through membrane that made it an increasingly practical option for industrial use. They created an anisotropic, reverse osmosis (RO) membrane that demonstrated considerably greater flux than other RO membranes. The membrane developed by Loeb and Sourirajan and subsequent innovations caused the commercial membrane industry to grow and expand considerably.
The 1970s and 80s saw the membrane market diversify and produce new membrane types, structures, and geometries to better serve different industrial applications. The membrane industry began looking at different polymers (polyamides, polyethylene, etc.) for possible membrane compositions as opposed to simply using cellulose acetate.
This eventually turned into an investigation of other materials - for example ceramic membranes became very commercially popular - in addition to possible surface modifications in order to continually improve or customize membranes.
Although there are many different types of membranes currently available in the market, the basic purpose of a membrane is to successfully control the separation of specific molecules without fundamentally changing or damaging the particles. Therefore, permeability is an essential characteristic to consider when evaluating membranes.
Depending on the intended end-use, membrane structures with varying pore sizes and distributions can be engineered from different materials to control selectivity.
Hydrophobicity/licity is a major characteristic that is used to determine a membrane’s application. A hydrophobic membrane is one that will repel water from its surface, while a hydrophilic membrane will exhibit an affinity for water and will readily adsorb the liquid.
Membranes that exhibit hydrophilic properties will have a higher surface tension value compared to its hydrophobic equivalent. Hydrophobicity/licity is particularly important in liquid filtration because it will drastically change how the membrane is interacting with the fluid.
In most liquid filtration applications, the membrane will be hydrophilic, allowing the permeate to pass through the filter material. Membrane hydrophobicity is desired for gaseous, venting, and low surface tension solvent applications, because the liquid that is upstream to the filter is trying to be contained or will damage a product if passed through the membrane.
Membranes can also be separated into categories based on their pore sizes.
Reverse Osmosis (RO)
RO membranes contain pores that are typically <1 nm in diameter.
Nanofiltration membranes are a newer membrane category and as such, is less defined. However, it is commonly understood that NF membranes contain pore sizes that range from 0.1 to 10 nm and exist between RO and UF membranes. They are used in water treatment and the food and beverage industry.
Ultrafiltration (UF) membranes are <2 to 100 nm and are used in pharmaceutical processing, the production of potable water, and the beverage industry.
Microfiltration membranes are 0.1 to 3 um and are normally used in the biotech/pharmaceutical, food and beverage, and semiconductor industries.
Membranes can be manufactured from both inorganic and organic polymers. Inorganic membranes include ceramic and metallic membranes.
Ceramic membranes, valued for their chemical inertness and longevity, can be manufactured from oxides (titanium, zirconium, aluminum). Metallic membranes can be manufactured from metals such as stainless steel, tungsten, or palladium (in powder form).
In addition to inorganic polymers, there are a variety of organic polymers frequently used for membranes, including polytetrafluoroethylene (PTFE), polyvinylidene fluoride (PVDF), polypropylene (PP), and cellulose acetate (CA).
PTFE and PVDF membranes are naturally hydrophobic and valued for their chemical compatibility. PP membranes are particularly suited for bioprocessing due to their ability to be autoclaved. They are often chosen for their hydrophilicity and chemical compatibility. Finally, CA is an organic polymer that had previously been widely used to manufacture RO membranes that are naturally hydrophilic.
There are two types of basic membrane structures: anisotropic and isotropic.
Anisotropic or asymmetric membranes do not contain a uniform and homogeneous structure. There is usually varying pore sizes and distribution throughout the membrane. Anisotropic membranes are composed of a thin surface layer on top of a thicker, more permeable sub layer. The purpose of the thicker layer is purely to act as structural support for the surface layer (to provide mechanical stability). The thinner surface layer is what determines the separation and permeability characteristics of the overall membrane.
There are several different types of anisotropic membranes. Composite membranes are anisotropic membranes that contain a surface and support layer of different polymers while an integrally skinned membrane is composed of the same polymer for both layers. Anisotropic membranes contain high fluxes.
Isotropic or symmetric membranes have a uniform structure and composition. Permeability characteristics can be controlled by altering the thickness of the overall membrane structure; the thicker the isotropic membrane, the less permeable it is.
There are several different types of isotropic membranes. Microporous membranes are a type of isotropic membrane that can have interconnected, small pores. Another type of isotropic membranes, nonporous membranes, are dense films that implement solution-diffusion separation mechanisms.
Isotropic membranes can be produced several ways including through stretching, phase inversion, and track etching. In cast solutions that use reverse phase inversion, the pore size is developed by controlling the evaporation of the solvent or removal of the complex solvent system. However, although all isotropic membranes have defined pore sizes, track-etched membranes provide the most precise and high quality pore dimensions.
There are several different manufacturing methods used to create membranes which can then be further modified using a variety of modification techniques.
The phase inversion process, which is the method used by Loeb and Sourirajan to create their membrane, is a commonly used membrane manufacturing process. During phase inversion, a single phase (homogeneous) solution is separated into two phases, a liquid phase with a low concentration of polymer and a solid phase with a high concentration. The solid phase will eventually be used to form the permeable membrane structure while the liquid phase will be in the pores themselves.
Thermal Phase Inversion (TIPS)
There are several membrane formation techniques that utilize the phase inversion process. One technique involving phase inversion is known as thermal phase inversion or thermally induced phase separation.
During TIPS, the temperature of the homogeneous solution is lowered to a point where the one phase solution separates into two phases (a polymer-rich solid phase and a polymer-poor liquid phase) and a membrane structure is produced.
Diffusion Induced Phase Separation (DIPS)
The phase inversion technique is also used in diffusion induced phase separation to produce membranes. During DIPS, the polymer is dissolved with a solvent to form a solution from which a film is cast. The film is then submerged in a solventless solution, which promotes diffusional exchange between the non-solvent and solvent, ultimately forming the solid and liquid phases of the membrane structure.
The last method of inducing phase separation to form membranes involves a process similar to DIPS, where polymer is dissolved in a solution and cast into a film. However, instead of submerging it in a non-solvent to promote diffusional exchange, the original solvent is evaporated to induce the formation of the two phases.
Another membrane manufacturing method is track etching. Producing membranes by track etching allows for accurate, uniform pore size and distribution. Track-etch membranes offer distinct advantages over conventional membranes due to their precise structure.
During track etching, thin films are exposed to radioactive, charged particles that create “tracks” in the film. The film is then treated with a solution that dissolves the membrane material along the tracks created, forming pores. Thus, the pore size and distribution can be regulated by controlling both the initial intensity of the charged particles and the contact time and concentration of the solution to the membrane surface.
Sintering is a technique used particularly for the formation of ceramic membranes. During this process, ceramic particles/powder, binder, and solvent are combined to form a mixture. The mixture is poured into a tubular mold and exposed to high temperatures to produce a porous membrane. The final membrane properties can be determined by both the original particles chosen (particle size and shape) and through controlling the temperature at which the mixture is sintered.
There are several different techniques to modify a membrane in order to further customize it.
One membrane modification technique is stretching, in which the membrane is physically stretched (either uniaxially or biaxially) to produce membranes with elongated pores. For instance, membrane stretching can produce expanded PTFE (ePTFE) membranes.
Another technique is chemical modification, which can be used to alter the membrane surface by introducing chemicals that either physically change the geometry of the surface or chemically oxidizes the surface.
Plasma & Graft Polymerization
In addition, techniques such as plasma and graft polymerization are potential methods to alter membrane surfaces and increase hydrophilicity. During plasma polymerization, the membrane is exposed to plasma that chemically changes the membrane surface. During graft polymerization, radiation is used to create active sites on the membrane surface which monomers can be grafted onto. These techniques can be used to improve the hydrophilicity of the membrane by producing surface changes that makes the membrane more susceptible to bond to hydrophilic monomers.
Due to their fragile nature, membranes are typically housed inside a module to maintain their structural integrity. There are several types of designs that can be utilized depending on whether the membranes are flat sheet or cylindrical.
Flat sheet membranes can be implemented into a plate and frame or spiral module, or undergo a pleating process, whereas cylindrical membranes are manufactured into tubular and hollow modules.
Plate and Frame
The plate and frame system consists of a membrane sheet laid in between support plates. This structure is no longer common due to its high cost.
Spiral wound structures are composed of a perforated tube (also known as a permeate collection tube) that has repeating layers of flat-sheet membrane and feed spacer (a layer whose function is to provide separation between membrane layers) wrapped around it.
Spiral wound structures are often chosen due to their ability to be customized; different membrane and feed spacer materials can be selected depending on the specific need. However, since there are more layers involved, spiral wound membranes can be more susceptible to fouling.
A pleated sheet module is comprised of a flat sheet membrane that has been pleated. The concentration and size of the pleats can be controlled to further customize the final structure.
A tubular membrane system is created when membranes are cast onto the inside of a tubular structure, which acts as mechanical support for the membrane. Tubular membranes can then be grouped together into a larger tube.
Similarly, hollow fiber modules are composed of membranes bundled into a larger tube. However, hollow fiber modules utilized longer, finer tubes. Membranes with diameters ranging from < 0.1 um to 2 mm are bundled into a larger tubular structure.
Due to the small diameter of the membrane fibers, a large amount of them can be packed into a tube, thus increasing recovery rates. Depending on the function of the membrane, the diameter of the membrane can be adjusted accordingly.
For example, a capillary membrane - a subtype of hollow fiber membranes consisting of membranes with a diameter of 1 to 2 mm - decreases the risk of fouling but also decreases the recovery rate. The main advantage of these tubular modules is the high degree of flexibility allowed and the high recovery rates due to the density of the overall system. Conversely, due to the degree of flexibility allowed, the thin fibers are fragile and susceptible to damage.
There are two main types of filtration modes: direct and cross-flow. Direct flow (dead-end) occurs when fluid is pushed through the membrane perpendicular to the membrane surface and cross-flow occurs when fluid is pushed through the membrane parallel to its surface. In direct flow filtration, the buildup occurs on the membrane surface whereas in cross flow filtration it occurs only at the end of the membrane. Thus, cross-flow filtration structures naturally last longer due to the decreased surface area available for fouling.
Fouling is an inevitable side-effect of membrane filtration that frequently damages and prevents a membrane from functioning properly. Fouling occurs when particles or dissolved solids build-up in the membrane’s pores or surface (adsorption of particles either onto the membrane surface or within the pores), impeding the membrane’s intended function.
Fouling can be categorized as being either reversible or irreversible depending on the severity and strength of the particle adsorption. If the fouling is determined to be reversible, there are several different types of removal techniques: physical and chemical or biological.
Physical methods used to combat fouling generally involve a technique called back-flush or back-walk. Back-flushing addresses the issue of surface build-up by pushing a stream of air or water backwards through the module to remove any particle accumulation. Chemical/biological methods of eliminating fouling includes using chemical (acidic, neutral, alkaline) and biological (biocides) solutions that are selected depending on the specific type of fouling that has occurred.
Membrane costs can span a wide range depending on the type of membrane and the size or geometry needed. For example, ceramic membranes are typically more expensive due to their robustness and longevity, while spiral wound structures are less expensive, but more likely to experience fouling.