Mostly recovered, the LHC readies for restart

Just nine days after it was first turned on in September 2008, the Large Hadron Collider at CERN was brought to a standstill when an improperly soldered splice connecting two superconducting cables failed (see PHYSICS TODAY, November 2008, page 24 ). When it starts back up in November, the plan is to test and debug both the LHC and its experiments for a while at 3.5 TeV per proton beam, then to ramp up to perhaps 5 TeV, hopefully sometime in 2010. Some lingering problems still need to be solved to make it safe to go to the design goal of 7 TeV, or 14-TeV center-of-mass collision energy.

“The first job is to rediscover the standard model,” says LHC project leader Lyn Evans. “Produce the top quark, measure its mass, and use the data to calibrate all the physics we know from the Tevatron,” which produces 2-TeV collisions at Fermilab near Chicago. Adds Tejinder Virdee, spokesman for the LHC’s Compact Muon Solenoid experiment, “It’s important to get going.”

In preparation for the collider’s restart, three of the four LHC experiments have checked their instruments using cosmic rays. “This has allowed us to measure the performance parameters of all the component detectors of CMS, such as resolution in space and time, alignment, calibration, energy and momentum resolution, and efficiency,” Virdee says.

Even at 3.5 TeV, he says, “we do expect new physics, because the standard model breaks down.” At that and higher energies, he adds, the LHC may reveal the origin of mass—“the Higgs is the favored mechanism, but we do experiments to see what nature has done. There may be new symmetries. There may be new forces. There may be extra dimensions. We are going to unexplored territory.”

New potential problem

In last year’s accident, says Evans, “it was a simple splice that gave way, but it did a lot of collateral damage.” Due to heat not being dissipated speedily enough, and to a pressure wave generated by overheated helium, 53 magnets had to be removed and the electrical, vacuum, and cryogenic connections between them redone. Spare magnets were used, and the ones that were removed are now being fixed to be spares. The total cost of the repairs, not including CERN staff salaries, was 40 million Swiss francs (roughly $37 million).

Since the mishap, scientists and engineers have tested some 20 000 splices in the LHC. They fixed two faulty ones in the machine and one from the spare lot, added pressure relief valves, and made changes so that in the event of a quench—when a superconducting magnet goes normal—the energy will be removed quickly. “The worst case we can imagine is the quenching of a magnet adjacent to a bad joint,” says Evans. “Even then, we will have no problem with the machine with the modifications [we have made to] the speed of descent of the power supply.”

Although the splices are no longer a threat, another problem emerged during the thorough going-over of the machine: The resistance of some of the busbars that connect magnets is too high. Copper sits next to superconducting cable in the busbars and will carry the current if the cable quenches. The higher the beam energy, the lower the busbar resistance needs to be—for 7 TeV, only about 15 µΩ can be tolerated, but resistances up to 70 µΩ have been measured.

Magnet memory loss

Also limiting the LHC energy are the magnets themselves. “The surprise we got,” says Evans, “is a loss of memory in the magnets from one of the manufacturers.” Retraining those 400 magnets—a third of the total number in the LHC—involves increasing the current through them until they can tolerate the required fields. At present, Evans says, no magnet will quench at currents needed to reach 5 TeV. Getting up to 6.5 TeV will go quickly, he adds. “But from previous experience we know that from about 6.6 TeV to 7 TeV will be trickier.”

“It’s not a big issue, but it’s annoying,” Evans says. Fingers are not pointing as to the cause of the memory loss, but among the theories are that one of the suppliers is to blame or that the magnets were damaged by sitting outside in a CERN parking lot for a couple of years. The plan is to retrain the magnets to 7-TeV beam energies after the 5-TeV run. And in 2014, an upgrade will take the machine to beyond its design luminosity of 10³⁴ cm^-1s^-1.

CERN experiments usually don’t run in the winter, and running the LHC beginning next month will add about €15 million ($21.5 million) in electricity costs. “If you polled anyone involved,” says Jeff Spalding, a Fermilab researcher who works on CMS at CERN, “they would say the most important goal is to complete commissioning the accelerator with beam. There is a lot to be learned.” In the grand scheme of things, he adds, a year’s delay doesn’t hurt science. And “an initial period at 3.5 TeV is not a loss. But the delay can be serious on a personal level for young scientists who need publications.” The LHC’s delay does keep the Tevatron in the spotlight, and if, as Fermilab now hopes, it gets funding to stay on through 2011, it would double the Tevatron’s data set.