INTRODUCTION

If one were to select a single word describing the most important factor when applying membrane technology to a wastewater application, it would be chemistry. That is not to say that other factors are not important; certainly fluid dynamics and other physical science factors play a major role as well. But it is the chemistry that is so important to success.

And it is the chemistry that is so often overlooked.

All too often reverse osmosis, or any membrane system for that matter, is considered the quintessential "black box", being viewed as merely an arrangement of pipes and membrane elements that will take a dirty wastewater and "clean it up".

This line of thinking frequently ends up as being a prescription for disaster. Since a membrane system is a dynamic operation, one can not view the system as being some sort of static instrument. And it is the chemistry that is the most dynamic of all.

CONCENTRATION FACTOR

The chemistry plays such an important role because the system is constantly concentrating contaminants. In fact, one shouldn't think of a membrane system for wastewater processing as a "filtration system", or a "purifier", but, rather as a "waste concentrator", constantly increasing the contaminant concentrations, producing purified water as a byproduct.

This is important since it is the "concentration" aspect that is so often taken lightly. And it can be the key to the successful design of a wastewater membrane application. This concentration "factor"- that is the degree of concentration - relates to both soluble solids and insoluble solids alike. Clearly, if one concentrates insoluble solids to a point where they tend to accumulate within a system, problems will develop.

Concentration of dirt is fairly straightforward. What often does not appear to be all that clear are potential problems harbored within the make-up of the soluble solids. As they become concentrated, certain species can surpass their limit of solubility, creating a problematic insoluble solid.

If one does not understand the chemistry of the soluble solids (or the insoluble solids, for that matter) membrane fouling problems typically result.

Concentration factor is a simple physical phenomenon - as one removes purified water from a aqueous solution, the solute remaining becomes more concentrated. If, for example, one removes fifty percent of the water, the remaining solids are concentrated by a factor of two; if 75% of the water is removed, the factor is four, and so on.

The membrane industry has applied the term "recovery" to this "degree of water removal" and this is directly proportional to the concentration factor (assuming zero passage) as depicted in the following chart (see Figure 1).

FIGURE 1

FLUID DYNAMICS

Closely following chemistry is the importance of fluid dynamics. The capability of a membrane system to deal successfully with any precipitation or dirt and debris that does end up in the system is closely related to the flow characteristics of the system.

For example, it is important to operate a membrane system with a high degree of "cross flow" velocity in an effort to create turbulence and to prevent the destruction of the "pure water layer". This is the relatively narrow band of "purified" water that becomes created as one passes water over the surface of an RO membrane.

If not designed properly -- where the ratio of crossflow velocity versus permeate rate is inadequate -- this pure water layer will cease to exist and the results will become readily apparent in the quality of the permeate. Higher passage of solutes will occur under such conditions.

It is equally important, however, to stress that the simple velocity across the membrane must be maintained at adequate levels in order to generate "turbulence". Thus, under cases where the permeate rate is relatively low (under high osmotic pressure conditions, for example), where the ratio of crossflow velocity versus permeate rate may be adequate, the turbulence may not be adequate.

Just what velocities are required for specific systems are dependent on the nature of the material to be processed, the system concentration factor (recovery) and the flow channel above the membrane (design, cross-sectional area, etc.) Unfortunately, one can not cite a certain flow rate that is applicable to all needs.

Pilot testing or working from established engineering data is the only certain method of establishing the proper criteria.

MEMBRANE PACKAGES

Flow designs are dependent on the type of membrane package used. And there are several different membrane packages which have been developed over the past thirty years. One sort of design over another may be proven to be more appropriate  for a given application and economic consideration.

There is probably little dispute that the "spiral-wound" package has come to dominate the fresh water purifying applications. And, this device has found many applications in the wastewater areas as well.

The success of the spiral-wound device has been related to the capability to package a relatively large amount of membrane surface area into a relatively small element while maintaining a certain degree of "cleanability", which, of course, is particularly important for wastewater processes.

Other designs have also been used for wastewater applications - and many of these have been applied very successfully. Examples are "plate and frame" type designs, tubular membranes and various forms of fiber type membranes.

In all cases, including the spiral-wound form, the success of the application has to do with the most economical approach while adequately dealing with the "cleanability" demands of a particular application.

In certain cases, a tubular type design or a fiber type may be the best approach when dealing with relatively high levels of insoluble solids, for example.

However, due to attractive economic features and the "turbulence promotion" created by the "net" spacer material, the spiral-wound design has found its way into wastewater applications to an ever increasing degree... to a point where it is now probably the most common selection for a given wastewater application.

Several unique designs of this type of package are available.

For example, one may want to select a spiral-wound design using different spacer material thickness. A design using a "tubular" type spacer as opposed to the "netting" type, gaining tolerance to higher insoluble solids loading at the expense of higher velocities may be found attractive in certain cases. Or, an "unbound", or Full Fit^™ type design, lacking a ridged outer cover and consequently negating the need for a perimeter seal (brine seal) may be best suited for a particular application.

DESIGN CRITERIA

In summary, the following is a listing of the design criteria that is important to address when considering the application of a spiral-wound system to a wastewater application:

1). Chemistry of the soluble solids. One must make certain that precipitation will not occur after concentration of the feed supply soluble solids. Operation on a stream containing 500 ppm calcium sulfate, for example, at a l0X factor (90%) recovery will not work in the long run, for example, since calcium sulfate precipitates at around 2000 ppm.

2.) Filtration of insoluble solids. The more insoluble solids removed the better the operation of a membrane system. Generally, if one can achieve filtration to five micron, chances of success are greatly improved. It should be noted here that simple filtration using cartridge filters alone is generally not sufficient. Typically, a more extensive prefiltration followed by 5 micron cartridge filtration is the most workable design.

3.) Recovery per membrane element. The lower the percent recovery per any given membrane element within the system, the greater the chances for success. Of course, this must be balanced against the additional expense in pumping power for lower percent recoveries per element.  Twenty percent recovery per element, for example, is too high. A good design should be well below that level.

4.) Flow velocity across the membrane surface. Irrespective of meeting the percent recovery per membrane element needs, the absolute velocity across the membrane surface must be considered. Using a typical wastewater design spacer material, a flow of 20 gpm (75.7 Lpm) exiting from any given eight inch diameter membrane element in a system is not great enough.

CASE HISTORY

In 1988, Pepsi-Cola moved into a new world headquarters in Westchester County, NY in a very picturesque setting within the New York City Reservoir Protection District. As one could imagine, zoning requirements called for very tight wastewater controls.

This presented many problems for building such a facility, one of which was the problem of cooling the building in the summer months and disposing of the "blowdown" from cooling tower.

Quite simply, the blowdown was required to be hauled off-site at a very high cost, making "water recovery" for re-use very attractive. Reverse osmosis was chosen because of its simplicity of design, installation and operation.

Two Osmonics 43 Series design RO units were installed in 1987 on the cooling tower blowdown coming from their evaporative condenser type towers, processing a total of 13.6 gpm (51.5 Lpm) blowdown, returning 10 gpm (37.9 Lpm) as purified water. Hence, 70% of the cooling tower blowdown is recovered for re-use.

The chemistry of the system is outlined in Table 1

As one can see, caution was given to avoiding the solubility limits of species such as calcium sulfate, calcium carbonate, silica and other potential problems. At higher recoveries, this would not necessarily be the case.

The system (see Figure 2) consists of backwashable dual media filtration designed for less than 5 gpm (18.9 Lpm) per square foot service flow. Media is manganese greensand/anthracite followed by the RO units which are designed using Osmo 416T-ST10 sepralators. These are 4 inch dia. x 40 inch long spiral-wound membrane elements specially designed for wastewater processing. The membrane is Sepa CA cellulose acetate.

FIGURE 2

The percent recovery per membrane element does not exceed 10% on any membrane element in the system and the flow velocities are all above 7.0 gpm (26.5 Lpm), exiting from any given membrane element.

Operation of the RO is from May through October each year, and generally requires cleaning once per week which is very simply done using a built-in CIP (Clean-In-Place) system.

Overall, some 1,000,000 gallons of water are recovered each year, translating into a pay-back of 3-4 months for the total system... not counting the value of 1 million gallons of fresh water that is no longer required for cooling tower make-up.