How Clean is Clean?

by Dana Mackenzie

      The Clean Room in Russell Flegal's laboratory is a plastic manufacturer's vision of heaven: plastic containers, plastic tubing, plastic gloves, plastic hood, plastic walls. Dust and dirt don't have a chance to sneak in. Positive air pressure keeps out airborne dust, and a sticky mat just inside the door traps any particles that you might have tracked in on your plastic booties.

      Welcome to the high-tech world of water-quality measurement. In labs like this one, scientists routinely tease out the one molecule per million in a water sample that can make the difference between a safe drink and a contaminated one. In the words of one researcher, "parts per million can kill." But this molecular sleuthing is tricky work. A stray speck of dust in the wrong place can impersonate a killer, and incriminate a stream or well that isn't a threat to human health.

      Flegal, a professor of geosciences at the University of California at Santa Cruz, and graduate student Carol Creasey believe commercial environmental consultants have been sounding a lot of false alarms along with the real ones. At two wells in central California, they found that concentrations of seven metals - lead, cadmium, chromium, copper, nickel, silver and zinc - were from 2 to 1000 times lower than reported by a consulting company that used EPA-mandated procedures. The higher levels that the consultants saw, according to Creasey and Flegal, came either from contamination in their own labs or from pumping the water out of the wells too rapidly.

      The "trace-metal clean" methods Flegal uses were pioneered by oceanographer Ken Bruland of UCSC 30 years ago. "He showed that if you didn't use these techniques, it was like weighing yourself with a lot of rocks in your pockets and saying that you weigh 10,000 pounds," Flegal says.While the news that our water may be cleaner than we thought might sound like cause for celebration, water researchers say it's not that simple.

      With Federal laws requiring cleanup of sites that exceed certain levels of contamination, mistakes can be expensive. "The millions of dollars we've spent monitoring the environment are going down a rathole," says Kenneth Coale, a former collaborator of Flegal. ( go to Environmental Protection Agency data on the amount of money spent on cleaning up pollution since 1981.) Coale, a chemist at the Mos Landing Marine Labs, argues that commercial testers have an economic incentive to use tests that are only as sensitive as the EPA's Maximum Contaminant Levels. "If the limit for tin is 600 parts per million, they develop a detection method that detects 600 parts per million." But when a lab is testing for levels at the limit of its detection abilities - or possibly much lower, as Flegal and Creasey found - it is more likely to make mistakes.

      On the other hand, Flegal says, "Consultants are not 'bad guys.' They are doing what has been previously recognized as correct." Since Creasey announced their work at a meeting of the American Geophysical Union in December, he says, numerous consulting firms have called to inquire about their methods.

      It may sound like the easiest thing in the world to dip a bottle in a well and come out with a pure sample of well water--until you hear all the ways your sample can be contaminated.

      Let's start with that bottle. Is it made out of glass? "One of the dirtiest materials you can use" for measuring metals, Coale says. "The only glass suitable for trace metal analysis is quartz, which is really expensive." All other glasses contain metals that may leach into your sample. Flegal's lab uses only plastic containers.

      Suppose your bottle is plastic. Did you use it straight out of the manufacturer's sterile packaging? If so, it is far too dirty. Flegal's lab washes all the containers at least five times, using two different kinds of acid baths and "ultrapure" water that has been through six chemical filters. The process of cleaning the bottles can literally take months.

      How did you keep your bottle from floating on the surface of the well? Maybe you used a lead weight? Too bad: if even an imperceptible fleck of lead breaks off that weight, it can ruin your sample.

      In reality, hydrologists don’t get their samples just by dipping a flask in a well: they pump the water out of the well. That poses additional problems: if the water ever comes in contact with the metal of the pump, then it may become contaminated. Flegal and Creasey use a peristaltic pump ( go to Perfect Pump commercial site) instead, an ingenious contraption in which the water never touches the pump, only the plastic tubing. The pump operates on a similar principle to a straw. If you squeeze a tube filled with water between your fingers, and then move your fingers along the tube, the water will follow along. Replace your fingers with a rotating head, and you have a peristaltic pump.

      Still, collecting the sample is a messy business. Soil and water make mud, and mud is the last thing you want in your sample. So it takes two people to collect the sample: the "dirty hands" person who operates the pump, and the "clean hands" person (usually Creasey) inside the van. Each sample is collected in two bags, one inside the other. The "dirty hands" person is only allowed to touch the outer bag, the "clean hands" person is only allowed to touch the inner one.

      "Contamination consciousness is more important than any procedure," says Charles Alpers, a chemist with the United States Geological Survey. "If you're the clean-hands person, and you've got to scratch your nose, you've got to ask the other guy to do it."

      The rate of pumping is also critical. Creasey and Flegal pump the water out of the well at its natural rate of seepage - perhaps a cup a minute. Commercial consultants typically pump 5 to 10 gallons a minute. "When you do that, you're re-suspending things that have drifted to the bottom," Creasey says. You may also be fracturing the rocks around the well, causing more metal to leach into your sample.

      How did you drive your sample back to the lab? "We used to use mercury to preserve nutrient samples. So we have traces of mercury all over our vehicles," says Alpers. "We have to be aware of this possibility if we're testing for mercury." Transportation is a weak link in commercial monitoring of contaminated sites. Commercial testers usually ship their samples to an outside lab after collecting them, so the samples pass through three sets of hands: the collector, the courier, and the lab technician.

      Back in the lab, Creasey and Flegal analyze the samples in the plastic paradise of the "Clean Room" or the "Ultra-Clean Room," which are designed to keep all metals away. By contrast, a lab whose primary business is monitoring organic chemicals, like benzene, might not have such facilities available. "When you measure benzene, problems of contamination are inconsequential," Flegal says. "You just don’t have benzene in the lab."

      Perhaps most important, the workers in Flegal's lab spend at least half of their time analyzing blanks - containers of ultrapure water that are brought along on the collection trips, and treated in exactly the same way as the containers of well water. No matter how many precautions Flegal and Creasey take, there is still the possibility of contamination from unexpected sources. By measuring the blanks, they can determine just how much contamination occurred, and compensate for it.

      Finally, Creasey and Flegal take each measurement three separate times, so that they can get an estimate of their experimental error. Commercial labs usually skip this step. The words "repeat three times" are easy to say, but they mean three times as much work for each result.

      The reward for this obsessive attention to cleanliness is a measurement 1000 times more precise than the figures produced by commercial consultants. Creasey, who used to work for one of those consultants, knows the corners they cut - not cleaning their containers enough, using only one person to collect the sample, running a blank sample only once every three months instead of every time they do an analysis.

      The recent report on well water is not the first time that Flegal has challenged the conventional wisdom among hydrologists. In the late 1980s, the United States Geological Survey reported that lead levels in the nation's rivers were declining. The decline correlated well with the removal of lead from the nation's gasoline beginning in the early 1970s. The rivers were getting cleaner - or so it seemed.

      Flegal and Coale didn't think so. In fact, the more they looked at the USGS data, the more they began to think that it didn't make any sense at all. With the trace-metal clean techniques Bruland had developed for measuring lead in ocean water, they found that the amount of lead in river water in the Great Lakes basin was ten times lower than the USGS hydrologists were reporting. But if the U. S. Geological Survey wasn't measuring the lead in the rivers, what was it measuring? Contaminants accidentally added by the USGS itself, said Flegal and Coale.

      "We really pissed off the Survey," Flegal says. Most groundwater experts, including those at the USGS, had assumed that such painstaking methods were not needed for groundwater, which they expected to contain much higher concentrations of trace metals. "We showed that they were generating bogus data," Flegal says. "But in the last couple years, they've acknowledged that they need to change their methodology." Two USGS scientists are studying with him now.

      But Arthur Horowitz, who wrote the new USGS sampling protocols (which have been adopted by the EPA as well), says that cleaner sampling does not necessarily guarantee more meaningful results. For example, the low-flow methods Flegal and Creasey used would be appropriate for a well that was drilled to monitor a contaminated site, but would not be appropriate for a drinking well. "You want to sample it under the conditions the well is used," Horowitz says. If the normal use of the well involves using a pump with metal parts and stirring up the sediments in the bottom, that's the way it should be tested.

      Horowitz believes that too many researchers focus on a single number, the concentration of a chemical, when that number can have different meanings under different circumstances. He has seen some bodies of water that pass all the tests for dissolved contaminants, yet are devoid of life. "How many dead cows or tundra swans would convince you that metals are a problem?" he asks.

      The great unknown in research on water quality goes by the name of "bioavailability" - how many contaminants actually get into the plants, animals or people that depend on the water, and what effect they have once they get there. A piece of contaminated sediment might pass right through a cow's stomach. The same amount of metal suspended in the water might get digested and lodge in the cow's fatty tissue, where it is almost impossible to get rid of. Or, before the cow drinks the tainted water, a friendly bacterium might get to the metal first and remove it from the environment.

      "The more we learn, the more we realize we don't know," Alpers says. "We're just beginning to understand natural systems."

      In Flegal's view, biology holds the key to the difference between contamination and pollution. Contamination, he says, simply means that toxins are "above natural levels." He has no doubt that the well water he measured is contaminated. Pollution, on the other hand, means "enough to cause adverse effects on life." If we ever find out how much metal it takes to pollute a well, it would be appropriate for the answer to come from the cleanest lab in town.