Researchers have identified the cause of a hiccup in the construction of the world's next top particle smasher, the Large Hadron Collider (LHC). During stress tests last week at the European Laboratory for Particle Physics (CERN), a support structure tore loose from the housing of a key ultracold magnet because it simply was not strong enough for the job, according to a statement released today. CERN will decide later this month whether upgrades will set back the start date of the LHC, which is set to switch on in November at the CERN facility near Geneva, Switzerland.

The magnet is part of a system for aiming the LHC's two counter-rotating beams of high-speed protons so that they collide into each other, liberating rare particles in a burst of energy. Researchers expect that these collisions will, at a minimum, allow them to cap off the vaunted Standard Model of particle physics, which neatly describes the forces at work inside atoms.

As protons emerge from the machine's 27-kilometer circular track, a sequence of three "inner triplet" magnets will focus them to collide with protons moving in the opposite direction. The superconducting helium that encircles and cools the magnets will occasionally release tremendous energy, which will exert pressure along the length of the magnets. But on March 27, when CERN engineers applied 20 atmospheres of pressure to one of the inner triplets in the LHC's underground tunnel, a support snapped in the last of the three magnets, the new statement confirms.

The supports were constructed from a fiberglass-epoxy laminate, which is not strong enough to withstand those pressures, according to independent reviews by CERN and Fermilab in Batavia, Ill., which designed and built the magnets.

"What the analysis shows so far is that something extraordinarily simple was missed in the design," namely, that the force on the magnets could be stronger in one direction than another, Fermilab's director Pier Oddone said in an accompanying statement. "[E]ven though every magnet was thoroughly tested individually, they were never tested with the exact configuration that they would have when installed at CERN—thus missing the opportunity to discover the problem sooner."

Fermilab engineers believe that workers can replace the faulty supports in all eight of the installed triplets—two per collision experiment—without delaying start-up of the LHC, says Judy Jackson, a Fermilab spokesperson. But she notes that CERN will revisit the issue later this month to make sure there will not be any holdups.