Technology Institute Is Miles Ahead in Making Fine-Scale Measurements
In September, the food scare was about dioxin in milk. Earlier this year, it was Alar in apples. Before that, infinitesimal amounts of drugs used to fatten farm animals were found in meat headed to market.
When a government agency announces, as the Food and Drug Administration did in September, that its scientists have detected less than one part dioxin in every trillion parts of milk in paper cartons, many Americans shudder.
Dioxin is a potent chemical that causes cancer in animals, and many experts say even the tiniest amount is too risky for humans to consume.
But in a world where such minuscule amounts can cause immense anguish, another, more important question is often raised: How on Earth can anyone measure something as small as one in a billion or one in a trillion?
The amounts are so tiny they defy comprehension. Detecting one part in a billion is like finding proof that a packet of sugar has dissolved in an Olympic-sized swimming pool. One in a trillion--a quantity only 1/1,000th as large--would be like detecting less than a single grain of sugar in the same pool.
Another way to look at it is through an analogy with time. One part per million is one second in about 11 1/2 days; one part per billion would be one second in 31 1/2 years; and a part per trillion would equal to one second in 31,500 years. But that is nothing: Right now, the most accurate clock in the world will lose or gain no more than one second every 370,000 years.
With modern technology and properly trained scientists, such extremely fine-scale measurements have become as easy as weighing a baby on a bathroom scale. Scientists can now make such precise measurements of individual atoms that they have to take into account the atom’s temperature--a warm atom measures bigger than a cool one.
“When I came here in the early 1970s we were straining to measure one to 10 parts per million,” said Harry S. Hertz, director of the center for analytical chemistry at the National Institute of Standards and Technology, which provides the United States its ultimate standards for measuring everything: the flow in a gasoline pump, the quality of steel in automobiles, the way atoms resonate in the most precise clocks.
Ten years ago, scientists at the National Bureau of Standards, as NIST was called then, could measure about 10 parts per billion. Now they routinely find and measure substances in concentrations far less than one part per billion.
“As our technology of measurement gets better, you can almost find some small amount of anything in everything,” Hertz said.
The most accurate way to measure the amount of most substances is by comparison, using methods of separation called gas-liquid chromatography. There are several techniques, but essentially all put a mixture of unknown substances--the sample being tested--into conditions that cause them to separate and emerge from the process in single file.
Aflatoxin, for example, is a cancer-causing chemical that can be found in minute amounts in peanut butter. To find out how much there is in, say, a tablespoon of peanut butter would not be hard even though the aflatoxin would not cover the point of a pin.
First, scientists dissolve the peanut butter in a solvent such as acetone. Then they slowly inject the solution into the chromatography machine for 10 minutes to half an hour.
The sample flows into a tube packed with porous material to which substances in the sample stick with differing affinities. A liquid or a carrier gas keeps flowing through the tube, maintaining steady pressure on the molecules to move along.
The ingredients in the sample are, in effect, running a race. Some will move quickly through the tube--or column, as it is called--because they lack the chemical or physical ability to stick to the material in the column. They slip right through. Others will be slowed because they stick to the packing or dissolve in the liquid that may coat the packing material.
Each ingredient, depending on the properties of its molecules, will emerge from the column in its own good time. The ingredients then enter a detector that identifies each and checks the quantity.
Some detectors, for example, shine ultraviolet light through a glass tube carrying the ingredient. A sensor on the other side registers how the light has been changed as it passes through the substance. The readout is compared with a catalogue of characteristic changes recorded from tests of known samples.
Gas chromatographs can also measure quantity by keeping the sample’s ingredients in vapor form, imparting an electrical charge to each molecule and allowing the charged molecules to land on a sensor. The more molecules that land, the bigger the sample. Again, known quantities are measured first to obtain a reference point.
Modern chromatography machines will work on samples as small as 1/50th of a drop. Millions of measurable molecules of a contaminant are still present in that droplet even at concentrations of just a few parts per billion.
Just because scientists at the nation’s laboratory for standards can measure such minute amounts, however, does not mean that everyone else can.
A few years ago, when national clamor over cholesterol began in earnest, it would have been possible to send a blood sample to five labs and get five different readings.
NIST has tried to help by developing standards other labs can use to calibrate their equipment. Doctors concerned about the accuracy of their cholesterol testing machines can purchase the standard freeze-dried human serum reference material NIST prepares: blood that has been measured precisely in every way.
Once reconstituted with the special water NIST also provides, the blood can be run through the doctor’s machine. If the cholesterol or some other reading doesn’t match, the physician will know the machine has problems.
NIST aims to measure everything 10 times more precisely than anybody else. So, although the standard atomic clock is now based on the vibration of a cesium atom and is accurate to one second in every 370,000 years, researchers are trying to develop a clock based on movement of the mercury ion, which they hope will lead to a clock that loses no more than one second in 30 billion years.
By comparison the average quartz watch loses a second a month or nearly a thousand years over the same period of time.
“In order to be able to make anything properly you have to control it,” said Dennis A. Swyt, chief of the precision engineering division of the agency. “And in order to control it you have to be able to measure it.
