Serendipitous Radiation Monitors

Robert L Fleischer. American Scientist. Volume 90, Issue 4. Jul/Aug 2002.

On the last day of September 1999, harried technicians at the Japan Nuclear Fuel Conversion Company in Tokaimura made a grave mistake: They inadvertently let a uranium-rich solution they were processing attain critical mass. The nuclear chain reaction that ensued released a flood of gamma rays and neutrons on the unprepared workers. Three were hospitalized; two died within weeks. Fortunately, other plant personnel were subjected to much lower doses-none of them received more than the allowable annual limit. But does this really mean that these people are free from worry? Just how much nuclear radiation can a person receive before it begins to pose a hazard to his or her health?

The answer depends on the form of radiation. The degree of danger also hinges on a host of other factors, including the kind of tissue exposed, and a person’s age and genetic makeup. But even if all these complications could be taken into account, it would still be difficult to calculate risk from first principles of physics and biology. So setting regulatory standards for radiation exposure is a difficult business.

The most common approach is to consider people exposed to high levels of radiation who suffered in obvious ways, and then, through extrapolation, deduce the health consequences of lower exposures. Studying the victims of the two atomic bomb blasts of the Second World War has, for example, been a vital component in setting modern exposure standards. The experience of Hiroshima residents is most relevant for neutrons, because the uranium-based bomb dropped there released a far higher proportion of its radiation as neutrons than did the plutonium design first tested near Alamogordo, New Mexico, and used on Nagasaki.

The Hiroshima bomb was unique. No others of its type were ever assembled and tested. So no direct measurements of neutron radiation are available, and physicists have been forced to estimate the doses people experienced at ground level in Hiroshima when the atomic age burst into public consciousness on August 6, 1945.

A pioneer in this effort was Sakae Shimizu, a physicist from Kyoto University and one of the first scientists to visit the devastated city. The strategy he pursued was inspired. Shimizu collected samples of fused sulfur, used commonly at the time on utility poles to mount glass insulators on wooden pegs. He realized that the neutrons released by the bomb would have converted some of the normal sulfur-34 to radioactive sulfur-35 and that he could use this phenomenon to estimate the intensity of neutron radiation at various sites. Shimizu acted quickly enough to make the appropriate measurements (the halflife of sulfur-35 is 87 days). Sadly, an official of the U.S. occupation force confiscated his records of these measurements, which were ultimately lost.

In more recent times, Shimizu and many other investigators have attempted to gauge the neutron radiation from the Hiroshima explosion using a variety of other radiochemical measurements or from theoretical calculations. The radiochemical results differ, disturbingly, by as much as a factor of five from the detailed calculations. So better numbers are sorely needed. Fortunately in this case-and some others for which radiation exposure must be estimated indirectly-common materials can serve as radiation detectors because they retain subtle traces of the speeding particles that pass within them. These features are called nuclear radiation-damage tracks. The use of such particle tracks to monitor radiation is an excellent example of the unexpected benefits that continue to accrue from a field of study I helped to pioneer in the early 1960s. I described many of the varied uses for particle tracks in an article for American Scientist in 1979. Here I concentrate on just one: retrospective radiation dosimetry.

Tracking Down Evidence

Some 15 years ago, I realized that ordinary glass from Hiroshima could serve as a dosimeter. The idea sprang naturally enough from my early work in nuclear-track etching at the General Electric Research Laboratory (now called the GE Global Research Center) in Schenectady, New York. To those not already familiar with these processes, the generation and revelation of particle tracks in everyday materials may come as a surprise. In most insulators, and in certain other solids, fast-moving particles produce needle-shaped paths of local atomic disorder. These corridors of disruption dissolve preferentially when exposed to corrosive liquids. A soaking with a strong acid or base etches the tiny particle tracks, enlarging them enough to be seen and studied using ordinary optical microscopes. In certain cases, etched particle tracks can become large enough to be visible to the naked eye. They range in shape from tunnels of nearly uniform width to conical pits, but sufficiently long tracks will have more complicated shapes.

For such features to form, the intensity of disturbance to the normal atomic arrangement must exceed a threshold value, which is different for each substance. Above this critical level, the vulnerability of the material to chemical attack along the track increases with the degree of damage done by the passing particle. These two properties, plus the mere existence of tracks, make possible many valuable applications, of which retrospective neutron dosimetry using glass is one recent example.

Actually, neutrons do not themselves leave particle tracks in glass, but their passing can nevertheless be readily discerned. Why? Most glass contains trace amounts of uranium, and when hit by a neutron, a uranium nucleus may break apart. The resulting nuclear fragments have considerable energy and so create etchable tracks as they move through the glass. This process has long been exploited to measure neutron radiation within nuclear reactors. Most such tracks are from slow (thermal) neutrons fissioning the isotope uranium-235.

When I proposed using this technique to sense neutrons from the Hiroshima blast, I speculated that locating appropriate samples of glass would be difficult-as indeed it has been. In 1994, I described the problem and the requirements for appropriate glass samples in Physics Today and other places, and I wrote some 30 Japanese scientists directly asking for help. But the pieces of glass submitted to me as a result had either been too distant from the explosion to serve as neutron monitors or had been so directly exposed to the heating from the flash of the detonation as to become partially melted. The intense heating created distinctive souvenirs, but it completely erased any bomb-produced nuclear tracks.

As it happens, I was later invited to Hiroshima to give a talk on a different subject, and there I met Shoichiro Fujita, a scientist with the Radiation Effects Research Foundation. Reminded of my interest in neutron dosimetry, he shortly sent me a decorative glass object–thought to have been a paperweight–unearthed about 200 meters from ground zero during preparations for building the Hiroshima Peace Memorial Museum in the early 1950s.

Together with Fujita and Masaharu Hoshi (of Hiroshima University), I set out to determine the dose of neutrons this button-shaped piece of glass had received. I treated a portion of the surface with a strong acid and then examined about 4 square centimeters of it under 500-power magnification, scanning in all some 10,000 fields of view. Enduring much eye strain, I counted 28 etched tracks-sites where slowed-down neutrons from the devastating fission explosion overhead struck uranium nuclei in the glass and caused tiny fission reactions of their own. To determine how much trace uranium was in the glass, we subjected it to a known quantity of neutrons in a research reactor and counted the newly formed tracks. Combining these observations, we calculated the dose of neutrons the glass button experienced back in 1945.

In this case, as is usual in science, the initial simple description gives less than the full story. We still needed to consider whether some of the tracks created more than a half-century ago could have faded over time. So we compared the sizes of the old tracks with those of the newly induced ones: Because their diameters did not differ appreciably, we could be confident that fading was not a significant problem. But a second key question-whether the sample was shielded from neutrons at the time of the explosion-was harder to address quantitatively. Some shielding must have been present, otherwise the button would have been heated enough to erase the tracks. But that was all we knew; the glass itself did not hold any further clues to where exactly it sat at the time of the blast. So my Japanese colleagues and I were forced to estimate the amount of shielding from knowledge of typical walls found in homes of the type that had been present in the neighborhood where the glass was later found. In doing so, we relied on experimental work by John Auxier of Oak Ridge National Laboratory, who described measurements of neutron penetration of simulated Japanese-style houses built in the United States for these tests. After applying this correction, we concluded that an exposed object at the location of the button experienced some 1.5 x 1012 neutrons per square centimeter-a value somewhat lower than prior calculations for this distance from the blast but consistent with radiochemical measurements. More recently, Jonathan MacDonald (my colleague at Union College), Fujita, Hoshi and I found bomb-induced tracks in the porcelain glazes of two shards recovered from the same area at Hiroshima. The glaze on porcelain is the thin glass overlayer. The doses measured are 3 x 1012 and 1.1 x 1013 neutrons per square centimeter. The highest of the three values exceeds both the theoretical calculations and prior radiochemical results. Presumably, it comes from the least shielded of the samples. These results are important to setting standards, because if it indeed took more neutrons than formerly thought to produce the radiation-related sicknesses observed in Hiroshima, neutrons must be less damaging than most experts now believe.

The Home Front

Although our analysis of a glass artifact from Hiroshima contributes directly to the task of establishing safety standards for neutrons, such regulations affect relatively few people, chiefly those working in the nuclear industry. Retrospective dosimetry does, however, have much broader application: protecting the public from the effects of radon gas, a ubiquitous, natural source of radiation.

Radon, through its decay products, is the primary cause of occupational lung cancer among underground mine workers. The risk to miners has been documented in many studies, but the relation between the dose of radon and the likelihood of developing lung cancer is yet to be demonstrated clearly for residential indoor exposure. The two most complete studies (by R. William Field and his coworkers at the University of Iowa and by Bernard L. Cohen at the University of Pittsburgh) disagree strikingly. Further measurements of radon exposure from indoor air might thus aid in resolving this issue.

But what exactly should be measured? Although one casually speaks of the “hazards of radon,” the true troublemakers are its prompt decay products: polonium-218, lead-214, bismuth-214 and polonium-214, each of which has a half-life of less than 30 minutes. When you inhale air containing radon, only a tiny fraction of this gas will undergo radioactive decay in your lungs before you exhale. But you will also be breathing in the radioactive daughter products of this radon, either as free atoms or attached to aerosol particles. These materials cling to lung tissue, and most will decay there before they can be cleared.

It would seem to make sense then to measure the real culprits, the four daughter products, which are the direct threats to health. Unfortunately that job is tricky: For reliable results, one needs a pumped system to gather, separate and measure the different radioisotopes. The alternative-passive collection-is plagued with other uncertainties. For example, deposition on a piece of furniture situated close to the sampling location will reduce the concentration of the decay products in the nearby air.

So as a practical matter, one usually ignores the decay products and measures radon, specifically the isotope radon-222, relying on the observation that the air in most homes contains about half of the decay products that would be present at equilibrium. Radon concentration, being a single parameter, is easy and cheap to measure, and the uncertainty in the amount of its decay products actually in the air is normally less than would have been encountered in trying to measure those products directly.

The real difficulty is that radon concentration in a typical home varies erratically over time. Radon levels might be relatively low, for example, during good weather when windows are left open. Conversely, radon builds up when the occupants are on vacation and the house is tightly shut. Thus a single measurement might not be even close to representative. Although it is not technically difficult to measure the long-term average concentration of radon, the task requires a lot of waiting. Also, there is no guarantee that the current long-term average corresponds to previous conditions: The installation of an air conditioning system or a change in the occupants’ habits can easily alter how much radon accumulates indoors.

For this reason, epidemiologists seek ways to measure past radon levels in homes. One promising technique makes use of lead-210, a product of the prompt daughter product of radon, polonium-214. Like this polonium isotope, lead-210 is radioactive, but it has a much longer half-life, 22 years, and thus remains long enough to document decades of radon exposure. The problem is that much of the lead-210 produced inside a home is swept away (literally)or perhaps mopped or vacuumed away.

Retrospective dosimetry is, however, possible, because most residences steadily accumulate lead-210 on or just beneath the surfaces of glass objects. The process begins when an atom of radioactive polonium settles on a piece of glass, perhaps the glass of a framed picture, a mirror or a windowpane. When this atom decays, it releases an energetic alpha particle (made up of two protons and two neutrons) and turns into an atom of lead-210. If the alpha particle shoots into the air, the recoil energy is sufficient to embed the lead-210 a tiny distance into the glass, about 25 nanometers (about 100 atomic diameters). Here it is safe from easy mechanical removal and can be readily measured by placing a strip of plastic against the glass. (Etchable tracks form in the plastic because alpha particles are given off by polonium-210, the granddaughter of lead-210. Counting tracks in the plastic detector thus provides a convenient measure of lead-210.) The other pertinent number-the length of time the glass was exposed– must also be ascertained. For windowpanes, it usually equals the age of the house. And glass on a framed photograph or print may have been displayed in a series of residences. Thus it would have recorded radon still further back into time.

Richard S. Lively of the Minnesota Geological Survey and Edward P. Ney of the University of Minnesota explored this technique in U.S. homes some 15 years ago. About the same time, Christer Samuelsson of Lunds University did the same in Sweden, a country with a significant radon problem. More recently, Daniel J. Steck of St. John’s University in Minnesota and William Field used this method to assess the lead-210 embedded at more than 1,000 positions on glass in some 500 homes. They also measured the radon concentration in these homes using standard procedures. This study revealed a clear correlation between measured radon levels and lead-210 on glass surfaces, but the scatter in the data is quite high-approximately a factor of 10.

Significant improvements are clearly needed. So last year, Robert H. Doremus of Rensselaer Polytechnic Institute and I investigated possible sources of error and ways to reduce them. One potential problem is leaching by water during washing, which can remove a fraction of the embedded lead-210 and thus suggest that past radon levels were less than they really were. Another complication is the formation of a hydration layer, a thin stratum of chemical alteration at the surface of the glass. As with leaching, the easy diffusion of lead-210 from a porous hydration layer would make a retrospective radon measurement appear too low. Such biases can, however, be reduced or eliminated by avoiding glasses that are low in aluminum oxide, a component that slows diffusion. We are continuing to study the effects of leaching and glass composition in an effort to improve this technique for gauging past radon levels in homes. But even if this source of error can be accounted for, there seems little doubt that some uncertainty will remain, given that the deposition of daughter products on surfaces is a strong function of air motion and thus highly variable.

Eye Spy

Although cancer epidemiologists are eager enough to make use of embedded lead-210 as a measure of residential radon levels, what they really wish is to assess a person’s overall exposure-at home, at work, at school, wherever. Perhaps they’ll soon have the means to do so. Stephen Hadley, Nicholas Meyer, Jonathan MacDonald (my colleagues in the Department of Geology at Union College), Alfred Cavallo (of the Department of Energy’s Environmental Measurement Laboratory in New York City) and I recently developed a novel way to measure personal radon exposure. The key ingredient, it turns out, had been right in front of our eyes.

Lenses for eyeglasses in the United States are most commonly made of CR39 (allyl diglycol carbonate), a plastic known to record reliably the passage of alpha particles. Counting the accumulated alpha tracks from radon and its prompt daughter products thus provides a measure of the wearer’s exposure to these potential carcinogens. Unlike the use of embedded lead in glass, this method provides a yardstick for exposure that is weighted automatically for the different air a person breathes, place by place and day by day, up to the time the glasses are abandoned.

Our simple idea-to count etched particle tracks and infer radon exposure-is another example of a seemingly straightforward plan that turns out to have complications. One is that CR-39 varies in its etching behavior from one supplier to another, and normally the manufacturing history for a particular pair of eyeglasses is inaccessible. As a result, the calibration for a set of lenses must be determined individually. In our trials, we measured relative sensitivities by exposing part of each lens to a standard dose of alpha particles and then counting the newly created tracks. We performed a number of absolute calibrations too, subjecting eyeglasses to known concentrations of radon in sealed chambers.

Another problem we have had to grapple with is that the decay products of radon do not deposit themselves uniformly over an eyeglass lens. We can, however, minimize such confounding effects by counting tracks on the inward-facing surface of the lens: In this sheltered region of quiet air, the radon daughters move primarily by diffusion, which should in principle provide for more reproducible results. Conveniently, it is the inner side of the lens that is cut and polished just prior to use, thereby switching on the dosimeter. And because most eyeglasses are fabricated according to an optometrist’s prescription, one can normally obtain good documentation of when the lenses were prepared for use. Thus total exposure time is usually known precisely.

So far we have tested this method with only about 20 different sets of eyeglasses. Most of our results have given undramatic values for average radon concentration (below 3 picocuries per liter, or equivalently, 110 becquerels per cubic meter, which is typical for indoor air). But one set of lenses provided a notable exception: It gave a reading of at 175 picocuries per liter, more than 40 times the EPA “action level.” These eyeglasses were worn by one of the operators of the Free Enterprise Radon Health Mine at Boulder, Montana-an establishment that advertises its radonrich atmosphere as therapy for an assortment of ailments including arthritis, bursitis, gout, emphysema and skin disorders.

We did not, of course, measure this set of eyeglasses by accident. Rather, we sought them out to test our method, knowing that radon measurements in the mine had shown the average concentration to be about 1,300 picocuries per liter. In the Free Enterprise office, the average concentration is 500 picocuries per liter. The owner of these glasses reported that she spends roughly 175 hours a year within the mine and about 2,400 to 3,000 hours a year at the office. So the personal radon average we obtained from her eyeglasses, elevated as it is, could have been easily attained from her exposure to these two sources alone. Incidentally, her family has operated the mine for many years, and, remarkably, none of them has had lung cancer. They apparently have the right genes to operate such a facility.

A Compact Vehicle

Our preliminary work gives us great hope that counting etched tracks on eyeglass lenses will soon provide an accurate way to determine personal radon exposures. But other plastic objects can serve as well. One particularly promising possibility is the ubiquitous compact disk, which is made of polycarbonate, a plastic that, like CR-39, is an alpha detector (but a less sensitive one). Because radon is soluble in polycarbonate, the interior of a CD contains a record of long-term radon exposure where it was stored.

Dobromir Pressyanov of St. Kliment Ohridski University in Sofia, along with Jozef Buysse, Annick Van Deynese, Andre Poffijn and Geert Meesen from the University of Ghent, recently explored this technique, measuring track density in CDs as a function of the radon concentration they were subjected to. Their results showed an impressively good correlation.

Like most things in life, this approach has pluses and minuses. One advantage over using eyeglass lenses is that the decay products are immobilized within the plastic. Thus counting the tracks formed below the surface of the plastic and dividing by the age of the CD gives a direct measure of radon concentration-uncomplicated by worries as to whether all of the radioactive daughters were around to make particle tracks. The fact that CDs are usually stored in boxes will not matter unless they are sealed radontight. The biggest obstacle to the use of CDs is the low rate of track accumulation: At typical indoor radon levels, say I picocurie per liter, only a few tracks will accumulate in each square centimeter each year.

Although I was able to find rare tracks in the glass button from Hiroshima, scanning 10,000 fields of view under the microscope is not something one wants to do routinely. Counting such sparse features normally demands electrochemical etching, a technique that subjects the object to a high voltage while it is immersed in an electrolyte. The resulting electrical discharges enlarge the tracks but leave scars that deprive the etched tracks of their clean, distinctive geometryincreasing the danger that random defects in the plastic will be mistakenly counted as particle tracks. Perhaps procedures can be established to overcome this hurdle: This technique is in its infancy, and it is really too soon to say whether counting particle tracks within CDs will ultimately prove of widespread utility

What Went Through Their Minds

Compact disks, eyeglass lenses, picture glass, what next? I hesitate to make any specific predictions: Research on etched particle tracks has too many twists and turn-is, as evidenced by my own investigations over the years. For example, my current work with eyeglasses grew in a straightforward way out of studies I did decades ago, estimating (after the fact) the exposure of Apollo astronauts to the damaging heavy atomic nuclei in cosmic radiation. But I stumbled on that possibility by a lucky accident.

One day in 1969 1 visited the Manned Spacecraft Center in Houston to attend a briefing for the scientists who were to study lunar samples. To impress us with some of the difficulties that the astronauts would face collecting rocks from the moon, NASA officials described the complex construction of their space suits. I was on the verge of falling asleep in my chair when the speaker casually mentioned that the space helmets were made of a clear plastic called Lexan polycarbonate. A bell went off in my head.

At the time, I was one of a limited number of people in the world who was aware that this material was one of the best track detectors available for heavy ions. Soon, my associates at General Electric and I arranged to borrow several space helmets for study. Examination of etched particle tracks in these helmets, which had flown on different Apollo missions, provided an elegant means to measure the heavy ions in cosmic radiation-some of which had demonstrably passed through the astronauts’ heads! We also got good results from some Lexan parts they took into space for a biology experiment.

Finding particle tracks in the plastic from the Apollo missions provided an unexpected bounty. It led to the discovery of a simple method for identifying the elemental composition of individual nuclei. All that is needed is to measure the taper of the etched tracks as a function of the distance along them. (This technique works because the taper depends inversely on the rate of etching along the track, which in turn reflects the characteristic ionization rate of the particle as it penetrates the material.) Indeed, the study of etched particle tracks has a 40-year history full of just such surprises many of which have been exploited commercially.

No one anticipated, for example, that etched particle tracks could make uniformly sized holes for filters. Yet, within a few years of when my colleagues and I came to that realization, General Electric had set up an operation to manufacture “Nuclepore” filters. By one estimate, the market for such exacting filters now exceeds $30 million a year. The study of etched particle tracks has also proved a boon for the petroleum industry because heating reduces the lengths of fossil tracks in minerals. This phenomenon thus provides a thermometer to determine how much heating a geologic formation experienced in the remote past. This information is key in deciding whether gas or oil might either be formed by moderate heating or dissociated by severe heating.

Serendipity in scientific research is inherently unpredictable. But I venture a prophesy: Investigations of etched particle tracks will lead to many more unanticipated rewards.