New Water Purification Systems Provide Increasingly Clean Solutions

Twenty-five years ago, water was considered "pure" enough for laboratory use if it would resist electrical current fairly well, suggesting it was relatively free of conductive ions. A new generation of highly sensitive analytical instruments-including high-performance liquid chromatography (HPLC) and inductively coupled plasma mass spectrometry (ICP-MS)- demand ultrapure water. Today's purification systems eliminate most contaminants, delivering water with total organic carbon (TOC) levels lowe

Sep 2, 1996
Ginger Pinholster

Twenty-five years ago, water was considered "pure" enough for laboratory use if it would resist electrical current fairly well, suggesting it was relatively free of conductive ions. A new generation of highly sensitive analytical instruments-including high-performance liquid chromatography (HPLC) and inductively coupled plasma mass spectrometry (ICP-MS)- demand ultrapure water. Today's purification systems eliminate most contaminants, delivering water with total organic carbon (TOC) levels lower than a few parts per billion, inorganic contamination in the 50 parts-per-trillion range, and resistivity better than 18 megohms per centimeter.


ADVANCED LINE: The WaterPro Reverse Osmosis system, distributed by Labconco.
The most advanced water purification systems also eliminate bacteria, bacterial fragments known as pyrogens, and even enzymes such as RNase. At the Fairfax Medical Lab in Chantilly, Va., for example, researcher Tony Cooper feeds ultrapure water into a chemistry analyzer whenever he measures liver enzymes, electrolytes, proteins, and other constituents in blood samples. Cooper, who has a bachelor's degree in medical technology, relies on a WaterPro Reverse Osmosis system distributed by Kansas City, Mo.-based Labconco Corp. to scrub bacteria and pyrogens from sample water.

"Our ultimate goal is to provide clients with analytical precision," he explains. "If these contaminants are in the water, it affects our reagents, and our results won't be valid." Pyrogens, the long-chain polysaccharides resulting from bacterial degradation, can sabotage results by prompting an immune system response and elevated temperatures in tissue or blood samples, adds Robert Moreland, an assistant research professor of urology and physiology at Boston University School of Medicine. Similarly, RNase can destroy data by "chewing up the RNA you're studying," Moreland says. "It's definitely a bad thing in reagent water in the lab."

While technological advancements in the life sciences are creating a need for water free of biological contaminants, research in engineering fields recently has resulted in purification systems for removing inorganics. In the semiconductor industry, for instance, trace metals will quickly foul molecular-scale computer components. That's why researchers at Hewlett-Packard Co.'s Little Falls site in Wilmington, Del., route tap water through a customized Milli-Q system before subjecting samples to ICP-MS, says applications chemist Joseph L. Hedrick. Incorporating reverse osmosis, ion-exchange technology, and ultra-violet radiation, the Milli-Q purification system-made by Millipore Corp. of Bedford, Mass.-easily provides water with fewer than 50 parts per trillion (ppt) of boron, Hedrick reports.

Water-purity specifications established by a number of professional organizations minimize data errors and ensure high lab standards. For example, the National Committee for Clinical Laboratory Standards has decreed that ultrapure, or "Type I," water must be clean enough to prevent interference with atomic absorption, flame emission spectrometry, and various other analytical techniques. In response to these recommendations, the water-purification industry is scrambling to develop more durable reverse osmosis membranes capable of reducing water waste and ultrafilters with smaller pore sizes for capturing a larger percentage of biological contaminants. Companies like Millipore and Barnstead/Thermolyne Corp. of Dubuque, Iowa, already market "RNase scrubbers." Millipore also is trumpeting its new "electrodeionization" technology as a more convenient way to remove ions from water-without the need for chemical regeneration of ion-exchange resins.

In a bid to give major market players a run for their money, Pittsburgh-based Fisher Scientific Co. recently introduced Elga, a popular European system, to researchers in the United States.

Meanwhile, distributors like U.S. Filter Corp. of Lowell, Mass., are promoting the user-friendliness and price-point value of their water purification systems.


REVERSE OSMOSIS: Applications chemist Joseph Hedrick of Hewlett-Packard uses a Millipore system to transform tap water into Type I water.
When Jeff Hurst first began working in a lab 20 years ago, he simply distilled his feed water twice and sometimes passed it through a few polishing cartridges before conducting basic chromatographic analysis. "Ignorance was bliss," recalls Hurst, now a senior staff scientist at Hershey Food Corp. of Hershey, Pa. "You couldn't use river water, of course, but contaminants were less of an issue than they are now." These days, Hurst needs an HPLC-grade Elga water purification system, incorporating pretreatment by reverse osmosis, to support highly sensitive chromatography and capillary electrophoresis instruments.

Distillation used to be a scientist's only option for achieving pure water in the lab. Like a teakettle, a distillation system boils and then condenses water, leaving most contaminants in liquid phase. The process is elegantly simple and surprisingly effective for removing contaminants with boiling points higher than 100°C, notes Byron Stewart, a product manager for Millipore. Unfortunately, he adds, other contaminants can become concentrated in the still, and highly acidic cleaners are needed to remove mineral scaling. Glass tubes, heaters, timers, and other mechanical devices on distillation systems require weekly maintenance, and energy costs are high, adds Bob Applequist, a technical product specialist at Labconco.

Enter reverse osmosis, or RO technology. The first RO membranes were developed in the late 1950s, primarily to process salt water for drinking water supplies. Under natural osmotic pressures, semipermeable membranes dilute "dirty water" by introducing "clean water," Applequist explains. However, osmotic pressures can be manipulated to reverse the natural flow of contaminants, thereby producing a constant flow of purified water. Many thin-film RO membranes will reject 95 percent to 99 percent of all contaminants, he says.

Compared with distillation, RO can be significantly cheaper in terms of energy use and maintenance costs, Stewart notes, but up-front RO equipment costs may be slightly higher. To prolong the life of a $600 to $800 point-of-use RO membrane, Applequist points out, a carbon adsorption filter to remove chlorine and a micron particulate filter are placed ahead of the membrane.

Harriet Mattox, a microbiologist at the Ed Love Water Treatment Plant in Tuscaloosa, Ala., hasn't had to replace her Labconco RO membrane since it was installed two years ago. As with any thin-film RO membrane, however, "you have to deal with the constant replacement of all other filters that go before it," she reports.

Since a 12-liter-per-hour distillation system may suck up 10,000 watts of energy, RO membranes are still a cost-effective way to pretreat Type I water before feeding it into a deionization system, Stewart says. Nevertheless, distillation systems remain an attractive pretreatment option for scientists accustomed to the technology. "It depends on whether you pay your own electricity bill or not," he comments.

Reverse osmosis technology has come a long way in the past 40 years. Unfortunately, according to Applequist, even the best RO membranes still use lots of water. "In general, for every 7 liters of water you put in an RO unit," he says, "1 liter of permeate purified water will be produced. The rest is going to go out as concentrate or reject waste water." In selecting an RO membrane, he suggests, researchers should always ask about water-production capacity.

Many scientists, like Joe Hedrick at Hewlett-Packard, use only a liter of Type I water per day and therefore don't need high-capacity RO membranes. Manufacturers often carry two or more systems that offer different water-production capacities. Barnstead/Thermolyne, for example, markets the NANOpure system, which produces 1.5 liters of ultrapure water per minute.

It also offers the EASYpure model for researchers who use less water. Ray Johnson, vice president of product development at Chromagen Inc. in San Diego, currently uses 10 liters of ultrapure water per day while developing DNA probe diagnostics. The NANOpure system, he says, "fulfills all of our water requirements."

Modern, lab-scale RO membranes may be made of several different types of material: cellulose acetate/triacetate; thin-film composites (which may contain polysulfone and polyamide); or polysulfone, notes Sue Bagrowski, a lab water technical service representative for Millipore. For lab systems, thin-film composites are perhaps the most common type of RO membranes because they're durable, economical, and effective over a fairly wide pH range (4 to 11), she says. By comparison, cellulose acetate RO membranes tolerate a pH range of only 4 to 7.5, she notes. However, the material is about 15 percent cheaper than thin-film membranes, and it can be used in conjunction with chlorine-unlike expendable carbon adsorption filters. Polysulfone RO (PSRO) membranes tolerate the widest pH range, from 3 to 11, she adds.

OPTIONS: Barnstead/Thermolyne's NANOpure system, left, offers 1.5 liters of ultrapure water per minute, while the company's EASYpure model, right, is designed for researchers who use less water
Because thin-film RO membranes have a limited tolerance for chlorine, researchers have to place carbon adsorption filters and other scrubbers upstream, explains Rick Passanisi, a senior product manager for Barnstead/Thermolyne. Polysulfone membranes "can be a little more expensive," but researchers may steam-clean them, says Rick Mahoney, the Topsfield, Mass.-based director of sales and marketing for Elga Inc. Many companies, such as Osmonics Inc. of Minnetonka, Minn., offer free water analysis to help potential customers select the most appropriate type of RO membrane and other purification apparatus.

As a chemistry professor at Emory University, Vincent Conticello uses proteins to create biomaterials. Ultimately, the research may result in replacement tissue for burn victims and cancer patients. Bacteria, pyrogens, and other biological contaminants in lab water could jeopardize Conticello's efforts. To protect his data, the university installed a Purelab system, distributed by U.S. Filter. Combining RO pretreatment and deionization technology with UV radiation for bacteria removal, the Purelab system is typical of many existing water-purification products for lab use. To improve resistivity and capture the remaining 1 percent to 5 percent of contaminants left behind by RO membranes, Conticello relies on resin-based ion-exchange beads, which trade either hydrogen ions for cations or hydroxyl ions for anions.


PROTECTIVE MEASURE: The Purelab system, distributed by U.S. Filter

Deionization is the workhorse of the water-purification industry, and DI-based systems are available for reasonable up-front costs-generally in the ballpark of $3,000 to $8,000, depending on pretreatment and polishing options. But resin DI beads must be regenerated when all hydrogen or hydroxyl ions have been exhausted, and they also need periodic cleaning to prevent bacterial buildup. Consequently, maintaining a standard RO/DI system with a UV lamp and ultrafilter may cost about $800 per year, Mahoney estimates.

]At the Cole-Parmer Instrument Co., a Vernon Hills, Ill.-based distributor for Barnstead/Thermolyne, product manager Kelly McCollum says customers should consider the cost of DI replacement cartridges before selecting a system. "The majority of our business comes from selling the cartridges," she adds.

Researchers may change DI resin cartridges in-house. Alternatively, they can pay a water-purification company to service the system. A number of distributors, including Solution Consultants Inc. of Jasper, Ga., sell replacement cartridges for popular systems such as Millipore and Barnstead/Thermolyne products.

Full-service resin maintenance contracts are more expensive, of course, but some researchers say technical support is well worth the price. Carla Mueller, a lab technician at the William Wrigley Jr. Co. in Chicago, bought a resin-maintenance contract from Fisher Scientific to make sure her Elga purification system always complies with the company's stringent standards. "We're a corporate quality-assurance lab," Mueller explains. "We're doing HPLC at 210 nanometers, and we're also doing microscopy . . . . The preventive maintenance contract was costly, but it was comprehensive. We want to make sure the numbers we send out to our subsidiaries around the world are as accurate as possible." A DI system may involve "mixed-bed" or "separate-bed" reactions, in which the cation and anion reactions take place in two distinct resin layers, Rick Passanisi points out. A dual-bed DI system can generate water more quickly, but resistivity may be low, compared with that of a mixed-bed reactor, which completes the ion-exchange process. Mixed-bed reactors should include a 1:1 ratio of cation-to-anion resins, he says.

Challenging traditional DI technology, Millipore recently began marketing "electrodionization" systems-combining ion-exchange with electrodialysis-as a way to avoid resin-regeneration costs and down-time. Coupled with RO pretreatment and Milli-Q final polishing systems, Millipore's ELIX electrodeionization systems let researchers generate Type I water without stopping the system to add regeneration chemicals, Byron Stewart says. Instead, resins are regenerated by electrical current. Compared with those of traditional DI systems serviced by a distributor, "initial capital costs for ELIX systems may be slightly higher," Stewart maintains, but "running costs are so much less." Also, he says, electrodeionization may provide added protection from biological contaminants.

Because he grows primary cultures of human vascular smooth muscle cells to study the molecular pathology of impotence at the Boston University School of Medicine, Robert Moreland swears by his Milli-Q PF water polishing system. In fact, Moreland coauthored a 1995 paper in BioTechniques (Y.-H. Huang et al., 19[4]:656) to test the ultrafilter's effectiveness for removing RNase. Prevalent in the lab environment, RNase traditionally is scrubbed from solution, glassware, or plastic by autoclaving with a volatile, expensive liquid known as diethyl pyrocarbonate (DEPC). Because the Milli-Q capillary fiber ultrafiltration device has a 5,000-Dalton molecular weight cutoff, Moreland says, it eliminated all detectable RNase-without DEPC treatment.

and Supplies
Barnstead / Thermolyne Corp.
Cold-Parmer Instrument Co.
Corning Inc. -- Separations Division
Elga Inc.
Fisher Scientific Co.
Labconco Corp.
Millipore Corp
Osmonics Inc.
PGC Scientific Corp
Solution Consultants Inc.
U.S. Filter Corp
Barnstead/Thermolyne's EASYpure and NANOpure systems-sold by a variety of distributors, including PGC Scientific Corp. of Frederick, Md.-feature a 10,000-Dalton spiral-wound ultrafilter, plus UV radiation and oxidation. According to Passanisi, the systems eliminate RNase while maintaining 18.3 megohms resistivity and TOC as low as 1 part per billion. Customers can expect even better performance in the future, he says. "I'm not at liberty to talk about some of the things we're working on," he states, "but we're investigating the oxidation effectiveness of various materials. Currently, for example, we're using mercury lamps, generally at wavelengths of 185 and 254."

Efforts to achieve lower and lower TOC levels are fueling marketing wars by major water-purification manufacturers. Like Passanisi, however, Martha Shadan, director for laboratory products at U.S. Filter, declines to divulge potential technology solutions for the future. "We're looking at TOCs in the very low parts-per-billion range now, and soon we may even be concerned about the parts-per-trillion range," she says. "No one has been able to achieve that yet, but we're moving in that direction."

Applequist cautions, however, that such low TOC levels may not be necessary in every lab setting. "It's been my experience that when a customer says, "'I'm doing HPLC analysis,' and I ask them how low they want to go with TOCs, they simply say, 'As low as possible.' It's like asking, 'How high do you want to fly?'" Consequently, Applequist urges researchers to "quantify the levels of performance you really need" before paying top dollar for a water-purification system that overpurifies water for users' specific applications.

Also in the near future, microporous filters for final polishing will probably offer increased durability and even smaller pore sizes. Already, companies like Corning Inc. Separations Division of Cambridge, Mass., are selling final polishers with pores as small as 0.05 micron. Since the smallest bacteria are 0.3 micron in size, "fundamentally you'll remove everything" with Corning's top-of-the-line final polishers (VTEC and QR Track-Etched Filter Cartridges), claims David Springett, director of new products and business development.

Additional advancements in water purification will include delivery systems that don't contribute any trace metals or other inorganics to water. "We're at 0.5 parts per billion for inorganics," Springett says. "That's 50 parts per trillion. That's the state of the art right now."


Ginger Pinholster is a freelance science writer based in Wilmington, Del.