Bomb testers beware: trace gases linger.

An underground blast at the Nevada Test Site on Sept. 22, 1993, heaved Rainier Mesa up a meter or more, then let the mountain fall back to its resting place.

The only casualties were some trees that lost their leaves unusually early that year.

Scientists triggered this explosion, the start of the Non-Proliferation Experiment (NPE), to evaluate a sensitive new method for detecting clandestine nuclear tests. A ground inspection team armed with tubes for boring could find telltale radioactive gases from weeks to a year or more after an underground test, reports a group from the Lawrence Livermore (Calif.) National Laboratory in the Aug. 8 Nature.

Lars-Erik De Geer of Sweden's National Defence Research Establishment notes in an accompanying commentary that the experiment proves "an on-site inspection has a good chance of finding conclusive evidence for a 'well- contained'" nuclear blast-the kind of test a cheater nation might stage. The success of the NPE thus bolsters the prospect of rigorously enforcing the Comprehensive Test Ban Treaty (CTBT), now being negotiated. The treaty would ban all nuclear testing, above and below ground (SN: 5/11/96, p. 298).

The NPE blast was not triggered by a nuclear device. In a chamber 400 meters underground, U.S. Department of Energy technicians rigged up more than a million kilograms of chemical explosives, comparable to a small, 1- kiloton nuclear bomb. The Lawrence Livermore team added two nonradioactive trace gases, helium-3 and sulfur hexafluoride, to substitute for the rare xenon-133 and argon-37 produced by a nuclear bomb.

Since the experimental blast, the group has tested hundreds of samples collected from tubes inserted 1 or 2 meters into the ground. They were able to detect helium-3 after 375 days and sulfur hexafluoride after only 50 days.

Because of its small size, the helium molecule tends to pass into surrounding rock rather than rise directly to the surface, the report notes.

"I really think we have a breakthrough here," says lead researcher Charles R.

Carrigan of Lawrence Livermore. Similar techniques could detect argon-37 from a nuclear bomb, even though a mere 15 cubic centimeters of the gas- "the volume of a Ping-Pong ball"- would be produced per kiloton of explosive yield, he explains.

Although the scientists had expected to detect the trace gases, they were surprised by the route those gases took to the surface. Carrigan thought the test explosion would produce fissures reaching up to the surface, but the blast was completely contained. So instead of traveling through fissures created by the blast, the gases rose through preexisting faults.

Even if an "evader" nation managed to test in a relatively fissurefree zone, Carrigan says, "all we need is one fracture" to send the gas to the surface.

De Geer points out that the method can't guarantee detection, but "it certainly highlights a problem and a great source of uncertainty for the cheat."

Low barometric pressure, associated with storms, is needed to pull gases toward the surface, so weather conditions would influence the gases' arrival time. Diffusion alone, they calculate, would require "tens to hundreds of years" to produce detectable quantities.

Carrigan calculates that under the same weather conditions, a team monitoring a nuclear blast would have detected radioactive xenon-133 after 50 days and the lighter argon-37 after 80 days. The argon should remain detectable for more than a year, he adds.

The new sampling technique will almost certainly become a part of the official detection network established by the CTBT, says Steven R. Bratt, director of the U.S. Nuclear Treaty Programs Office in Arlington, Va.