WHEN President Clinton and other world leaders signed the landmark Comprehensive Nuclear Test Ban Treaty last September, they served notice that any signatory nation trying to conceal an underground nuclear test would have to elude a vigorous international verification program armed with the latest monitoring technologies. Thanks to the work of a multidisciplinary Lawrence Livermore team, the international community now has a powerful new forensic tool to help enforce the treaty by detecting even deeply buried clandestine nuclear tests. Under the terms of the treaty, which bans all nuclear weapons test explosions, a system of verification and inspection will be administered by the Comprehensive Test Ban Treaty Organization in Vienna, Austria. Lawrence Livermore scientists have long played an important role in providing monitoring technologies in support of nuclear treaty verification and on-site inspection. The latest Livermore technology is based on the discovery that minute amounts of rare, radioactive gases generated in underground nuclear detonations will migrate toward the surface along natural fault lines and earth fissures. Livermore geophysicist Charles Carrigan led the team that included physicists Ray Heinle, Bryant Hudson, and John Nitao and geophysicist Jay Zucca. With the help of results from earlier studies, they theorized that highly sensitive instruments might detect telltale radioactive gases rising during periods of barometric low pressure through natural fissures in the ground above the blast. To test the hypothesis, the team obtained two gases, 0.2 kilograms (7 ounces) of helium-3 and 50 kilograms (110 pounds) of sulfur hexafluoride, as tracers. These nonradioactive gases are ideal tracers because they are present in very low quantities in the natural environment. As the photo below shows, the bottles containing the gases were placed with a 1.3-kiloton charge of chemical explosives into a mined cavity that was 15 meters (50 feet) in diameter and 5 meters (17 feet) high. The cavity was located 400 meters (1,300 feet) below the surface, two to three times deeper than that required for a similar sized underground nuclear test. A somewhat shallower detonation, says Carrigan, might have produced a collapse crater or extensive fractures connecting the cavity with the surface, both telltale signs of an underground explosion. Hence, clandestine tests would very likely be conducted at the greater depth to avoid easy detection of treaty violations.

Simulating a Nuclear Test The detonation, known as the Non-Proliferation Experiment, occurred on September 22, 1993, in the rocky Rainier Mesa of the Nevada Test Site, where some of the nation's nuclear tests were conducted until a testing moratorium went into effect in 1992. The chemical explosion simulated a 1-kiloton underground nuclear detonation, which, as expected, did not produce any visible new cracks in the Earth. Over the year and a half following the blast, team members, including technical support personnel from Test Site contractors EG&G and REECo, collected nearly 200 samples of subsoil gases for measurement. At some sampling stations, sampling tubes were driven into the ground to depths of 1.5 to 5 meters (5 to 16 feet) along fractures and faults. At other stations, tubes were simply placed beneath plastic sheeting that was spread on the ground to trap rising soil gases and to limit atmospheric infiltration (see photo below). The first positive finding came 50 days after the explosion, when sulfur hexafluoride was detected in fractures along a fault. Interestingly, the much lighter helium-3 showed up 375 days--more than a year--following the explosion. Both gases were first detected along the same natural fissure within 550 meters (1,800 feet) of the blast site.

Over the course of the extended sampling period, virtually all the samples yielding concentrations of the two tracers appeared along natural faults and fractures in the mesa during periods of low atmospheric pressure, mainly at the beginning of storms. The low pressure accompanying storms, says Carrigan, makes it possible for the gases to move toward the surface along the faults. Although over the course of a year the number of low-pressure days equal the number of high-pressure days, the gases are eventually drawn upward. "There's a ratcheting effect," he explains. "The gases don't go back down as much as they go up." (See the simulation below.) Carrigan notes that it is counterintuitive that helium-3 takes so much longer to make its way up natural fissures than sulfur hexafluoride, which is 50 times heavier. Computer models developed at Livermore showed that this result occurred because most of the heavier sulfur hexafluoride gas moved directly up the rock fractures. In contrast, the helium-3 diffused readily into the porous walls of the rocks as it slowly moved upward toward the soil surface. Critical to determining why helium-3 behaved as it did was Bryant Hudson's analysis of helium-3 in Livermore's noble gas laboratory, where he used mass spectrometry to measure the presence of helium-3 in soil-gas samples down to parts per trillion.

Modeling the Detonation Carrigan and Nitao modeled the experiment using a porous-flow simulation software called NUFT (Non-Isothermal Unsaturated Flow and Transport) developed at LLNL by Nitao. In attempting to make the simulation as realistic as possible, the team used actual barometric pressure variation data from the Rainier Mesa weather station. The simulation showed the two gases moving at different rates toward the surface following the detonation. The calculated arrival times at the surface for both tracers were in excellent agreement with the data. Given the good agreement between the computer model and the observations, the team then used NUFT to simulate the gases released from an underground 1-kiloton nuclear test under atmospheric conditions similar to those that followed the 1993 Non-Proliferation Experiment. The software was used to predict the arrival of detectable concentrations of the rare gases argon-37 and xenon-133 at 50 and 80 days, respectively, after the detonation. These two isotopes are ideal indicators of nuclear explosions because they are not produced naturally in significant quantities; thus, background levels are extremely low. Also, their short half-lives of 34.8 days and 5.2 days can be used to infer how recently an event had occurred. Other, more long-lived isotopes might still be present in the environment from decades-old tests and would tend to muddy the conclusions of investigators trying to determine whether a clandestine test had recently occurred. The successful confirmation of the experiment by computer simulation implies that sampling of soil gases for rare, explosion-produced radioactive tracer gases at the surface near a suspected underground test can be an extremely sensitive way to detect nearby underground nuclear explosions that do not fracture the surface. As a result, says Carrigan, an on-site inspection has a good chance of finding conclusive evidence for a clandestine nuclear explosion for several months afterward. Putting Treaty Evaders on Notice "If detected, the radioisotope signals would be unequivocal," according to Bryant Hudson. "They would put treaty evaders on notice that they risk detection if they try to explode a nuclear device underground. We can't absolutely guarantee there won't be cheating, but we've made it more difficult." Carrigan points out that because of political considerations, it may take some time to get a country to agree to an on-site inspection under the terms of the test ban treaty. The thinking of many experts has been that such inspections need to be conducted within a few days to capture evidence of a test. The Livermore team's work, however, shows that waiting weeks or even months to detect rare gases is not a problem and may well be advantageous, because the gases need time to arrive at the surface. Team members caution that searching for tracer gases is only one of many detection tools. Other methods that might be used at a suspected test site include analyzing the printouts of seismographs for aftershocks from an explosion, looking for explosion-induced stress in plants and trees, drilling for explosion debris, examining the earth for fractures and craters, and searching for pipes and cables leading underground. In discussing the work of the team, Carrigan attributes its accomplishments to a confluence of Lawrence Livermore strengths in computer simulation, geophysical theory, nuclear test containment, and radiochemistry. "Interdisciplinary collaboration made this work possible," he says.

--Arnie Heller

Key Words: Comprehensive Nuclear Test Ban Treaty, nuclear proliferation, nuclear treaty verification, NUFT (Non-Isothermal Unsaturated Flow and Transport).

For further information contact Charles Carrigan (510) 422-3941 (carrigan1@llnl.gov).