MIT physicists have shown that a possible new route to nuclear fusion reported nearly three years ago is not the phenomenon it was heralded to be, but is based on contaminants in the experimental system.
The MIT work was published in the March 30 issue of Physical Review Letters by Daniel H. Lo, a nuclear engineering graduate student at the Plasma Fusion Center (PFC); Richard D. Petrasso, principal research scientist at the PFC; and Kevin W. Wenzel, a postdoctoral fellow at the PFC. It was based on an analysis of the original data for the phenomenon published by scientists at Brookhaven National Laboratory (BNL).
In the same issue of Physical Review Letters, the BNL scientists themselves published an erratum saying that their original conclusion was wrong, and contaminants were probably responsible for the nuclear fusion they had observed. Dr. Petrasso notes that he gave a draft of the MIT paper to the BNL scientists last fall, but as late as February they firmly disagreed with it.
In 1989 Robert Beuhler, Lewis Friedman and Gerhart Friedlander of BNL described a reaction they dubbed "cluster impact fusion." Using an accelerator, the scientists shot clusters of heavy water molecules at a target loaded with deuterium, or heavy hydrogen, and found nuclear fusion occurring at vastly higher levels than expected.
The BNL group believed that when the clusters of 100 to about 1,000 heavy water molecules hit the target, the energy of the collision caused some of the deuterium atoms in the cluster/target system to fuse via a hitherto unknown process.
In their original paper in Physical Review Letters, the BNL scientists suggested that this might open a new route to nuclear fusion. For the last two decades scientists have been pursuing only two routes to produce a net energy gain from fusion-one uses large magnets (like the Alcator tokamak at the PFC), the other powerful lasers. The BNL results were achieved in a small accelerator that is much less complicated.
But cluster fusion "would have been extraordinary in itself, even if it never made a watt of power," Dr. Petrasso said, because there are no current laws of physics to explain it. Given today's understanding of nuclear fusion and how particles interact collectively, such clusters of heavy water should not have had enough energy to cause deuterium atoms to fuse upon impact.
More specifically, "the energy of each individual deuteron (D) in a cluster containing 100 to 1,000 molecules of heavy water (D2O) was far too small to cause the number of fusion reactions [the BNL scientists] observed, unless there were some incredibly new physics going on," Dr. Petrasso said.
However, the MIT scientists found that the number of fusion reactions in the BNL experiment could be explained by contaminants-perhaps charged individual deuterons (D+) or a group of two or three (D2+ or D3+). "They're much smaller [than a cluster], so they'd have enough energy to cause significant numbers of nuclear reactions," Dr. Petrasso said.
As a rough analogy, he said, imagine throwing a cluster of baseballs (bound together) at a target, then throwing a single baseball at the same target from the same distance and with the same energy. The single ball would hit the target with much more energy than the individual balls in the cluster.
Dr. Petrasso notes that "it's very hard to get rid of contaminants [in a system like that used by the BNL team] unless you use a magnetic field to deflect them away from the target." The BNL scientists published their erratum after doing just that and finding that the fusion rate dropped a hundredfold.
The MIT group actually didn't start seriously investigating cluster fusion until last November, just over two years after it was first reported. (The MIT work was supported by the Department of Energy and Lawrence Livermore National Lab.)
"We waited because we'd just come off cold fusion. Also, we didn't really believe the BNL results," Dr. Petrasso said. (In May of 1989 Drs. Petrasso, Wenzel, and colleagues at the PFC wrote an article for Nature that was the first to point out what they believe are fatal errors in Drs. Pons and Fleischmann's nuclear data on cold fusion.)
However, a story on cluster fusion in the October 25 issue of Science last year piqued their interest. "In that article we learned that Steven Koonin of the California Institute of Technology, an MIT graduate [PhD physics, 1975] and extremely well respected theoretical physicist, was now taking the BNL measurements at face value," Dr. Petrasso said.
"When a hard-nosed skeptic like Steve Koonin begins to question whether his own calculations have included all the proper physics-since they were inconsistent with the BNL interpretation-you think, `maybe it's time to pay attention.'"
(Dr. Koonin will be on campus this Thursday, May 14, to give a talk on "Supercomputer Visualization of Nuclear Collisions" for the 46th anniversary of MIT's Laboratory for Nuclear Science. The talk, which is free and open to the public, will be at 4:10pm in Kresge.)
So how did the Brookhaven scientists and others misinterpret the data? Dr. Petrasso notes that the number of fusion reactions the BNL team measured was still fairly small. "And when you're measuring things at levels much smaller than you're used to, all those effects that you'd ordinarily discount, or those that you'd never even thought about, come into play and mislead you." Put another way, he said, "Murphy's Law got them."
A version of this article appeared in the May 13, 1992 issue of MIT Tech Talk (Volume 36, Number 30).