Using Nanofiltration in Beverage Production
The following article is based on a paper presented in August, 1990 at the 21st Annual Meeting of the Fine Particle Society in San Diego.
Soft drink manufacturers, not to mention manufacturers of other beverage products, have always had a keen interest in the water that goes into their products. After all, it is the water that typically makes up 80%-90% of the product.
In the past the water has been treated by a variety of methods, depending on the type of water supply and the type of drink being produced. Or, often the water is treated in a manner which is perceived as necessary based on experiences in other areas, etc.
What are we really after when we treat the water?
One objective that we can identify with a reasonable degree of certainty is that of alkalinity. It is probably safe to say that high alkalinity imparts a poor taste to water and products made from high-alkalinity water. What may not be so certain is what level of alkalinity is necessary.
A second requirement is having a low total dissolved solid, most notably tolerable levels of chloride and sulfate anions. Typically 500 ppm TDS is thought to be adequate and chloride and sulfate levels of below 250 ppm, as their respective ion, are thought to be acceptable.
Since the vast majority of water supplies in the U.S. and even throughout the world meet these criteria, treatment requirements are relatively limited. Of course, other inorganic contaminant levels, such as lead and other heavy metals, are of equal if not greater concern. But inordinate levels in the typical American water supply, for example, are by far the exception not the rule.
Beyond alkalinity, TDS, chloride and sulfate, the remaining physical/chemical concern is the level of dissolved organic compounds and general levels of dirt, debris and particles.
The organic contamination concerns may range from general contamination by organics such as humic acids, lignins, etc. - which impart a color to the water supply and can often cause staining of containers, etc. - to suspected cancer causing compounds such as trihalomethanes (THMs) and their precursors (organic compounds which may breakdown or transform into cancer causing agents).
With alkalinity and other inorganic contaminants within guidelines, it is common to see water supplies treated simply by backwashable multimedia sand filtration followed by chlorine treatment for microbiological control.
Generally, this process filters the water to 5-10 micron and the super chlorination steps are effective at controlling microbiological contamination. The process does, however, have the potential to create THMs via chlorination of organics that pass through the backwashable filter.
Perhaps the most common method of water treatment in use in the soft drink industry is the addition of 'lime softening/ferrous sulfate flocculation' to the above scheme. This not only can reduce alkalinity - if that is a problem - but it also employs the ferrous sulfate oxidation to ferric sulfate reaction as a means of aiding precipitation, removing particles, larger complex organics and the like via 'flocculation' and precipitation.
This is a relatively effective means of water treatment and serves as the standard at soft drink facilities throughout the world.
Recently, more advance forms of filtration and contaminant removal have been introduced to the soft drink industry using very fine filtration in the form of polymeric (plastic) membranes to remove not only particles, but also dissolved inorganic compounds and most dissolved organics as well.
Membranes, both reverse osmosis (RO) and ultrafiltration (UF) have been used extensively in the past to reduce dissolved and insoluble solids levels in water supplies, most notably in seawater desalination. More recently, membranes have been used widely for purification of numerous freshwater supplies.
These two processes, UF (filtration to 10 Angstroms for dissolved organic removal) and RO (filtration to 5 Angstroms for dissolved organic and inorganic removal) are relatively new to the beverage industry. Several RO systems have been installed over the past five years and a few UF units have been installed where only filtration and organic removal have been necessary.
RO can address the alkalinity, TDS and fine filtration issues, but it is the most costly ($0.75-$1.20/1000 gallons) of the membrane systems to operate, due primarily to the higher pressures and resulting higher energy costs associated with producing those pressures.
UF, on the other hand, is a very low cost ($0.30 - $0.50/1000 gallons) method of water treatment. And it is a very effective means of fine filtration, including removing most organics precursors to THMs. However, UF does not affect the dissolved inorganic levels, making it suitable only for water supplies that are within the inorganics guidelines not uncommon for U.S. water supplies.
It is now becoming apparent, however, that a process in-between RO and UF, or 'nanofiltration (NF),' may prove to meet the greatest needs of the beverage industry which is fine filtration of particles and only a slight reduction of dissolved contaminants.
UF can be thought to be 'under-engineered' where the potential exists for alkalinity or other inorganic contamination to rise above their respective limitations. RO, on the other hand, may be thought to be over-engineered' in that it provides a water quality many magnitudes beyond known soft drink requirements, at a higher cost.
Nanofiltration provides the advantage of lower cost operation ($0.75- $0.95/1000 gallons), plus some other advantages.
One major advantage is the lower discharge (mass-wise) of dissolved materials to the waste stream (concentrate), since it inherently passes more of the dissolved solids to the permeate (pure water product) compared to RO.
The advantage is that it allows flexibility in allowing a lower TDS discharge where such an issue is critical. Otherwise, it allows one to operate the system at a higher 'recovery' (ratio of feed water to permeate) since there is a less critical concern over surpassing solubility limits of some of the dissolved chemical species. This even further reduces the operating costs, by sending a smaller percentage of water to drain.
A case in point is a recently built soft drink facility operating on Colorado River water in Riverside, California. Being a surface water, the potential for high levels of dirt and debris is always a possibility and the dissolved solids level is well-known as being relatively high (500-700 ppm).
The objective was to reduce the insoluble solids content of the water, at least to a point comparable to that of a conventional lime softening/ferrous sulfate coagulation system, and to reduce the dissolved solids to the point where they are within specification.
The system consists of an Osmonics®' Osmo® 43B-PR(CA) 288IVDLX Nanofiltration (NF) unit pretreated by chlorine injection and backwashable dual media filtration (anthracite/manganese greensand). Following the NF unit, the permeate is post-treated by additional chlorination and activated carbon before being sent to the can line and blending areas.
The Riverside NF unit reduces the dissolved solids level by about 70% and brings all of the inorganic levels well within the criteria for soft drink production water.
This was all expected. What was not necessarily predetermined was the level of filtration the system provided, particularly in contrast to that of a more 'conventional' ferrous sulfate/lime type system.
While it was expected that NF, having a pore size somewhere in the 8-Angstrom range, would be better than simple flocculation followed by sand filtration, it was difficult to necessarily predetermine the submicron filtration capabilities. Since operation began in late 1989, the opportunity was available to measure the filtration capability by conducting submicron particle testing.