By Peter S. Cartwright, P.E.

System Performance
The vast majority of membrane system failures occur as the result of membrane fouling. This fouling is usually caused by one or more of the following mechanisms:

  • Suspended solids in the feed stream resulting from incomplete feed water filtration
  • Precipitation of insoluble salts or oxides resulting from concentration effects within the membrane device
  • Biofilms resulting from microbiological activity

These mechanisms cause the membrane surface to become coated with fouling materials that build up in layers. As the layer thickness increases, the flow rate and resulting turbulence across the membrane surface (and immediately adjacent to it) is reduced, thereby encouraging more settling of suspended solids and increasing the fouling layer thickness which further slows the rate of permeate flow through the membrane – a vicious cycle.

With NF and RO membranes which reject ionic contaminants, fouling usually creates a phenomenon known as concentration polarization. The fouling layers inhibit the free movement of contaminants in the feed stream away from the membrane surface via turbulent flow and as salts are rejected from the membrane, their concentration at the surface is higher than in the bulk solution (that portion above the fouling layer). Since ionic rejection is always a percentage of the salts concentration at the surface of the membrane, the permeate quality decreases as a result of concentration polarization and this phenomenon may actually indicate the presence of foulants before a reduction in permeate rate is detected. The increased salts concentration at the membrane surface also promotes precipitation of those salts whose solubility limit is exceeded.

Figure 7 is a schematic of a complete membrane processing system (or a single membrane element).

Note that the feed stream enters the system (or membrane element) and as the stream passes along and parallel to the surface of the membrane under pressure, a percentage of the water is forced through the membrane polymer producing the permeate stream (percent recovery). Contaminants are prevented from passing through the membrane based on the polymer characteristics. This contaminant-laden stream exits the membrane system (or element) as the concentrate stream also known as brine or reject.

The permeate rate of a given membrane element cannot be changed without varying the applied pressure or temperature. Recovery, however, can be easily changed by varying the feed flow rate to the element and this is opne of the variables that is controlled by the system designer.

The effect of recovery on system performance is important. As recovery is increased, the flow rate of the concentrate stream diminishes; all contaminants that are rejected by the membrane in the concentrate stream become more concentrated. The effect of increasing recovery upon the concentration of contaminants in the concentrate stream is illustrated in Table III and Figure 8.

For wastewater treatment and water reuse applications, the minimum recovery is usually at least 80 percent.

One way to understand concentrate factor is to think about the evaporation or distillation process. If half of a given volume of water is distilled and the condensate recovered as pure water (permeate), this is the same concept as operating a membrane system at 50 percent recovery. Evaporating three-fourths of the water is 75 percent recovery and so on.

The advantage of operating systems at high recoveries is that the volume of concentrate is small and the flow rate of the feed pump is lower; the potential disadvantages are significant:

  • Higher concentration of contaminants is likely to result in fouling. In NF and RO applications, the concentrated salts solution results in high osmotic pressure, requiring a higher-pressure pump and a more-pressure-tolerant system.
  • As higher recoveries reduce the quantity of concentrate to be discharged, the higher concentration of the concentrate stream may present discharge problems.

The selection of recovery is definitely application-specific: most water purification applications – those treating raw water to be purified for some downstream application (drinking, product manufacturing, rinsing, etc.) – generally operate at relatively low recoveries, not exceeding 85 percent, even for the largest applications. In general, most water purification applications involve feed water conductivities that are relatively low; the one exception is seawater desalination. Usually, the larger the system, the higher its recovery.

It is possible to completely close off the concentrate line, through the use of a valve, thereby using the membrane as a conventional or ‘dead-end’ filter, forcing 100 percent of the water through the membrane, with occasional periods when the concentrate valve is opened to allow the crossflow feature to ‘flush out’ the contaminants at the surface of the membrane. Some membrane elements of tubular, capillary fiber or plate and frame can also be ‘backwashed’, which involves running permeate (or another high quality water supply) backwards into the element to dislodge suspended materials from the surface of the membrane.

MBR technology
As the newest membrane technology application and one with huge potential, membrane bioreactor (MBR) technology justifies special mention.

For wastewaters containing biodegradable contaminants, the traditional treatment method is to encourage the use of bacteria to break down the contaminant (bioremediation). This encouragement can take the form of adding air or oxygen (in the case of aerobic treatment); some sort of a mechanical matrix (for bacterial attachment); mixing; and other approaches intended to maximize the metabolic activity of these microorganisms.

MBR involves utilizing an MF or UF membrane to filter the treated water to remove particles and microorganisms. The permeate is recovered and, if necessary, further purified with RO or another polishing process. The concentrate stream is returned to the bioremediation tank. Compared to other biological processes, MBR offers the following advantages:

  • High-quality effluent, almost free from solids;
  • The ability to disinfect without the need for chemicals;
  • Complete independent control of hydraulic retention time and sludge retention time (HRT and SRT);
  • Reduced sludge production;
  • Process intensification through high biomass concentrations;
  • Treatment of recalcitrant organic fractions and improved stability of processes such as nitrification;
  • Ability to treat high-strength wastes.

The device configurations most commonly used today are capillary fiber and plate and frame, although tubular devices are also available.

The most common biological treatment is aerobic and, typically, air is bubbled into the treatment tank. A very popular approach is to immerse the membrane element in the treatment tank and either allow the hydrostatic head of the solution to provide the driving force, or to use a pump to pull the permeate through the membrane (or both). In this case, air bubbles are also directed over the surface of the membrane (air scouring) from below, in an effort to increase turbulence and reduce fouling.

Another design involves pumping the water through the membrane system external to the treatment tank and yet another uses a separate tank for membrane processing downstream of the biological treatment tank. Additional designs and configurations are sure to appear as MBR technology becomes more widely used.

Figure 9 illustrates aerobic MBR applications for both immersed and external designs.

In general, every stream must be tested to identify the following design parameters:

  • Feed stream chemistry
  • Optimum membrane element configuration
  • Total membrane area
  • Specific membrane polymer
  • Optimum pressure
  • Maximum system recovery
  • Flow Conditions
  • Membrane element array
  • Pretreatment requirements

Specific parameters are discussed below.

Feed stream chemistry

The chemical composition of the feed stream can affect the membrane element in a number of ways. The polymer itself or components of the membrane element can be degraded by certain chemicals. For example, cellulosic membrane polymers are subject to hydrolysis by high pH; thin film composite polymers are degraded by oxidizing agents such as chlorine; and most of the polymers are affected by chlorinated hydrocarbon solvents in concentrations above five percent. Water chemistry can also contribute to fouling problems, the bane of all membrane systems. Certainly, suspended solids of any kind represent a potential problem and the configuration of the membrane element plays a major role in its ability to resist fouling.

  • Stream chemistry
  1. Total solids content
  2. Suspended (TSS)
  3. Dissolved organic (ROC, MBSS, COD, BOD)
  4. Dissolved inorganic (TDS)
  5. PH
  • Chemicals of concern
  1. Oxidizing chemicals
  2. Organic solvents
  3. Saturated solutes
  • Operating temperature
  • Osmotic pressure as a function of system recovery
  • Variation in chemistry as a function of time

Membrane element design
The particular way that the membrane polymer is configured in an element has a direct bearing on the resistance of the membrane to fouling. It is desirable to pack as much membrane area into the device as possible without it becoming too large of heavy. Unfortunately, the element designs that provide the greatest ‘packing density’ also have the lowest resistance to fouling as a result of the close spacing required to accomplish the high packing density.

Membrane area
In general, the great the membrane area, the higher the permeate rate (all else being equal).

Membrane polymer
As indicated in Table I, many different membrane polymers are now on the market, with new ones frequently becoming available. Obviously, these are the key components with regard to effecting separation and each polymer has it particular strengths and weaknesses; none is perfect. It is essential that the design engineer understand the particular characteristics of each polymer well enough to select the one(s) most appropriate for the testing.

Because of its lower viscosity, warm water will flow more readily through a membrane than cold water; therefore, as temperature is increased, permeate rate increases. Unfortunately, most membrane polymers are thermoplastic and become more compressible when warmed. The combination of temperature and pressure can cause irreversible compaction in some polymers, resulting on premature failure. Certain plastic polymers as well as ceramic and metallic membranes exhibit excellent thermal stability and offer significant promise in those applications where it is considered desirable to process a stream at elevated temperatures. Each membrane element manufacturer provides data on temperature limits and the relationship of temperature to permeate rate for its products.

Applied pressure
In general, the permeate rate of a membrane element is directlyl proportional to the Net Driving Pressure. Net Driving Pressure is defined as the total pump pressure minus the osmotic pressure minus any back pressure in the permeate line. Net Driving Pressures range from as low as two bar (30 psi) for MF systems to over 68 bar (1,000 psi) for RO systems.

Of critical importance to solute concentration is the recovery. As defined, it is the percentage of feed water flow that exits the elements or system as permeate. The relationship between concentration factor and recovery is illustrated in Table III and Figure 8. Because osmotic pressure is directly proportional to concentrate concentration, this effect becomes critically important in many effluent treatment and seawater desalination applications.

Flow conditions
It has been shown that membrane elements are much less susceptible to fouling from suspended or precipitated solids if all of the flows through the element are turbulent. This is indicated by the term Reynolds Number, which is a dimensionless number defined as:

Diameter x mass velocity divided by Viscosity

By definition, Reynolds Numbers above 4,000 indicate turbulent flow and those below 2,000 indicate laminar flow. Each membrane element manufacturer has determined the flow requirements for their elements which result in turbulent flow and all systems should be designed using these data.

Membrane element array
In order to maintain turbulent flow conditions, elements must be grouped according to specific criteria. Typically, the permeate from each element (or presume vessel in the case of spiral wound configurations) is collected separately in a manifold, whereas the concentrate from bone becomes the feed from the next element (or pressure vessel). Obviously, as permeate is removed from a feed stream the total volume of concentrate available as feed to the next bank of elements decreases, so the total number of elements in parallel in each successive bank must decrease. With this design, turbulent flow conditions are maintained through the entire system.

Testing options
To generate the necessary design data, several testing options are available.

Cell Testing

Insert Figure 10, cell test unit

Cell test devices are available for purchase (or through a consulting engineering firm skilled in the art), which evaluate small sheets of membranes on the stream to be processed. Typically, the sheet is placed between two stainless steel plates and the test stream pumped across the membrane surface at a selected pressure and flow rate. The permeate is collected and analyzed for the degree of separation, possible effect of the stream on the test membrane and other properties.

The cell test offers a number of advantages:

  • Unit is very simple to operate.
  • Only small quantities of membranes are needed; excellent for screening membrane polymer candidates.
  • Can be run on small volumes of test stream
  • Takes very little time.

The disadvantages of this testing approach are:

  • Cannot obtain engineering design data
  • Cannot be used for long-term fouling study.
  • Is only useful with membranes available as a flat sheet.

The cell test approach is useful as an initial step, primarily to select one or more membrane candidates for further testing.

Figure 10: Applications test schematic

Applications testing
Applications testing utilizes a full-sized membrane element in a test unit capable of operating similar to a production unit. Since the data from this testing will be used to scale up the design to full size, it is essential that the membrane element manufacturer supplies an element capable of this scale-up.

The applications test equipment should be designed so that the very high recoveries can be achieved without compromising the flow rates required to produce turbulent flow, for example. This requires that the pump be capable of not only producing the desired pressure, but also flow rate to accomplish the minimum crossflow velocity across the membrane surface.

Because the system must also be capable of testing at very high recoveries, the concentrate valving must be adjustable to accurately product extremely low flow rates. This typically involves the assembly of a ‘valve nest’ using micrometer valves. Additionally, the recycle line should be equipped with a diaphragm valve for adjustment of flow and pressure.

The most important feature for application testing equipment is versatility. Different membrane elements have very specific operating parameters and the equipment must accommodate these. To cover the entire gamut of membrane technologies, two different pieces of application testing equipment are generally required: one for MF and UF and the other NF and RO.

The latter must be capable of pressures up to 1,0000 psi (68 bar) and it is virtually impossible to find a single pump capable of supplying the flows and pressure requires for all four technologies. For MF and UF applications, a variable speed drive centrifugal pump works fine, although the variable speed feature makes it expensive.

Materials of construction are an important consideration; 316L stainless steel is essential for applications requiring pressures in excess of 60 psi (four bar); below that, schedule 80 PVC is sufficient.

Applications testing is capable of generating complete design data for the full size system. An applications test can be run on as little as 50 gallons (200L) of test stream and after setup, can be completed in one hour or less for each membrane element tested.

A typical applications test is run as follows:

  1. Run high quality (tap water or water treated with RO or DI) into the system at low recovery to minimize any possible contaminant concentration effects. Take date (see below).
  2. Feed water is then run into the unit set at low recovery and after stabilization (usually less than five minutes), the following data are taken: pressures, primary (feed) and final. Flow (recycle, permeate and concentrate). Temperature (recycle). Quality (conductivity)–feed, permeate, concentrate.
  3. The system recovery is then increased incrementally while adjusting the recycle valve to ensure that the correct crossflow velocity is maintained.

At each recovery, in addition to the collection of flow and pressure data, analytical samples should be taken for performance evaluation. Of course, the choice of parameters to be measured depends upon the separation goals of the test. It is unusual for system recoveries to exceed 95 percent; however, that also depends upon the goals for the testing and it is possible to run a well-designed test unit up to 99 percent recovery.

Once the optimum conditions have been established, such as operating pressure and maximum system recovery, the normalized performance data will enable the test engineer to determine the total membrane area required for the full-sized system.

Application testing provides the following advantages and disadvantages:

Advantages:  fast; provides scale-up data (flox rate, element efficiency, osmotic pressure as a function of recovery, pressure requirements, etc.); can provide an indication of membrane stability.

Disadvantages: does not reveal long-term chemical effects; does not provide data on long-term fouling effects.

Pilot testing
Usually this involves placing a test machine (such as that used for the applications test) in the process and operating continuously on a ‘side stream’ for a minimum of 30 days.

The advantage is that pilot testing accomplishes all of the functions of the applications test palus provides long-term membrane fouling and stability data; the disadvantage is that pilot testing is expensive in terms of monitoring and time requirements.

Slowly but surely, the mindset of water recovery and reuse is reaching global proportions. Membrane technologies represent some of the most successful processes to accomplish this goal. As they become better understood and the costs decrease, these technologies will be come the lynchpins of water reuse and recovery. For these applications, it is imperative that any candidate stream be thoroughly tested. This requires a combination of knowledgeable, experienced personnel to design, run and interpret testing on well-designed testing equipment.

About the author
Peter S. Cartwright, President of Cartwright Consulting Co., Minneapolis, is a registered Professional Engineer in Minnesota. He has been in the water treatment industry since 1974, has authored almost 100 articles, presented over 125 lectures in conferences around the world and has been awarded three patents. Cartwright has chaired several WQA committees and task forces and has received the organization’s Award of Merit. A member of the WC&P Technical Review Committee since 1996, his expertise includes such high technology separation processes as reverse osmosis, ultrafiltration, microfiltration, electrodialysis, deionization, carbon adsorption, ozonation and distillation. He can be reached at (952) 854-4911; fax (952) 854-6964; email:, or website:



Comments are closed.