Sensitive to tiniest hint of terror

The Forensic Explosives Laboratory (FEL) in Fort Halstead in Sevenoaks, Kent, has one of the most sensitive detection systems in the world. It can detect explosives residues as tiny as one thousand millionth of a gram - 100,000 times lighter than a grain of sand. In prosecuting an alleged terrorist, the presence of even the tiniest amount of RDX (a component of Semtex) or PETN (used in detonators) would be presented as proof positive of the accused's involvement with explosives.

But on 14 March, an accident in the laboratory revealed that one of the key pieces of equipment - a centrifuge - was contaminated with 30 millionths of a gram of RDX. Furthermore, it seems plausible that this could in turn have contaminated some of the case samples.

The samples to be tested arrive in the form of cotton swabs from a crime scene. The staff take every precaution against contamination. In the laboratory they may not wear watches or jewellery, and must wear disposable oversuits, overshoes and gloves. Samples are handled with glassware that is used only once.

The preparation of samples for final analysis by gas chromatograph (which identifies chemicals) would often use the centrifuge for purification. First, the swabs are soaked in a solvent, chosen to dissolve RDX or PETN. This mixture of liquid and solid is put in a test-tube and slotted into the centrifuge. Solids are separated from the liquid which is drawn off and refined for testing.

The contamination found yesterday was in one of eight rubber bungs which pad the test-tubes. The bungs do not come into direct contact with samples. But after the accident, contamination was also found on the body of the centrifuge itself. This suggests some transfer of RDX did occur from the bung.

The report released yesterday says that the bung had enough RDX to contaminate many samples. The key question though is whether any tests were falsely contaminated. The forensic tests always included a "control" sample which was known not to contain any explosives. But Dr Marshall, the FEL's head, admits that if a control sample tested positive, that test would be abandoned. This means that only the tests which were obviously wrong were rejected. But it is not yet known how many control samples tested negative, while the case sample wrongly tested positive.