International Linear Collider gets reference design and cost estimate

For more than five years now, a linear electron–positron collider big enough to explore the so-called tera-scale (collision energies of order 10¹² electron volts or 1 TeV) has topped the wish list of the international community of particle physicists (see Physics Today, September 2004, page 49 ). Given the present state of accelerator technology, the collider’s two face-to-face linacs would need a combined length of about 30 km to achieve a first-phase collision energy of 0.5 TeV. The cost of such a gargantuan facility dictates that the undertaking—from R&D and design, to construction, to operation—be thoroughly international from the start. Appropriately, the project carries the name International Linear Collider. The ILC is regarded as an essential complement to the Large Hadron Collider (LHC) ring at CERN, which should begin providing 14-TeV proton–proton collisions next year.

Now the ILC has its first estimated price tag, based on a reference design prepared over the past two years by the Global Design Effort, a 60-member team headed by Barry Barish of Caltech. GDE’s report of its design and cost estimate (http://media.linearcollider.org/rdr_draft_v1.pdf ) was released at the February meeting in Beijing of the International Committee for Future Accelerators, GDE’s parent organization. Although the report doesn’t give the estimated total cost as a straightforward sum, it comes to roughly $7.5 billion in 2007 US dollars.

Sample sites

Because it will be several years before a site is chosen for the ILC, the reference design and cost estimate are not site-specific. But civil-engineering cost estimates are included for three sample sites: in the mountains west of Tokyo, near CERN on the Swiss–French border, and near Fermilab in Illinois. Despite the obvious geological contrasts, it turns out that the tunneling and other civil-engineering costs for the underground machine, about $1.8 billion, are much the same for the three sites. That’s because each site has different difficulties and compensating advantages. The problems posed by the mountainous terrain of Honshu, for example, are balanced against the virtues of horizontal access and a granite substrate that, unlike the Illinois prairie or the Rhone valley, requires no concrete lining of tunnel walls.

A possible site near the DESY laboratory in Hamburg was much discussed in previous years when DESY pioneered the superconducting RF acceleration technology that was selected in 2004 for the ILC. But Hamburg was not included among the sample sites because a machine there could not sit nearly as deep as at the other three sites. That would require significant changes in the reference design. Furthermore, if the collision point were at DESY itself, the Elbe river would obstruct the ILC’s eventual extension to 50 km for 1-TeV collisions in a later upgrade foreseen in the reference design.

What about the cost of acquiring and installing the machine’s high-tech components? Because different governments have different accounting systems for costing the creation of a major facility, the GDE report separates the cost of acquiring the injectors, RF structures, focusing magnets, and other high-tech components from the labor required to install and test them. The latter, including administration and other personnel support, is estimated in the GDE report to be 13 000 person-years over the seven years of the construction phase.

The report estimates the cost of acquiring the high-tech accelerator components (excluding the detectors) through worldwide competitive bidding to be $4.9 billion. If one then adds a rough dollar estimate for the unspecified labor cost, the machine’s total $7.5 billion price tag is not unlike those of comparable international megaprojects like ITER and the LHC. The latter comparison works only if one adds the civil-engineering cost of the LHC’s preexisting 27-km-circumference tunnel, left over from the earlier LEP electron–positron collider ring.

Modifications for thrift

Just last July, GDE’s preliminary estimate of the sum of the civil-engineering and high-tech-acquisition costs, now quoted as $6.7 billion, was roughly $9 billion. “So we spent the next six months,” says Barish, “scrutinizing the design to find modifications that would yield significant savings without hurting the physics.” And the team did indeed identify 10 such modifications and incorporate them into the reference design.

The modifications yielded a gratifying 25% reduction in the ILC’s estimated cost. Two were particularly significant: Instead of separate 7-km-circumference tunnels at opposite ends of the ILC for the electron and positron beam-damping rings, the reference design calls for a single tunnel near the center to house both damping rings. Also, the design now calls for a single beam-collision point—into and out of which the ILC’s two large detector complexes will be moved alternately at intervals of a few weeks.

The preliminary design had called for a Y junction in each beam so that the beams could be directed alternately at two collision points, one at the center of each permanently situated detector. The expectation is that moving the detectors is not significantly slower than redirecting and refocusing the beams. But it is certainly cheaper than providing bending and focusing magnets for two collision points.

The reference design and its cost estimate are meant to set the stage for the next phases of the project. Over the next three years, R&D will proceed at various laboratories worldwide in tandem with the preparation of a detailed engineering design. An important R&D goal, for example, will be optimization of design and fabrication details for the machine’s 16 000 superconducting niobium RF acceleration cavities. “Without such optimization of high-tech components,” says Barish, “we’d be handing the eventual mass production off to industry with too many risky question marks.”

What about site-specific details? If negotiations haven’t begun to converge on a specific site by the time the engineering design is due in 2010, “then the design will be incomplete,” says Barish. “We could design to several sites, but that’s wasteful.” Even if a site is provisionally selected by 2010 and incorporated into the engineering design, further international negotiations would be required to tie down funding commitments.

Sharing the cost

It is expected that the host country would bear the civil-engineering cost, which represents about a quarter of the total. The remaining three-quarters would be divided equally among the three participating regions: Europe, Asia, and the Americas. Because Europe includes many countries with big particle-physics programs, a European host could expect to bear not much more than 30% of the ILC’s total cost. But the US or Japan as host would bear closer to half the total. It’s generally felt, however, that no one country’s share should exceed 50%. “People don’t want a majority stockholder,” explains Barish.

By the time the engineering design is ready, theorists expect, the LHC will already have given convincing evidence of the much-sought-after Higgs boson, with a mass somewhere between 100 and 200 GeV. There’s also much anticipation of possible supersymmetric particles accessible at terascale energies. Pinning down the mass of the Higgs boson at the LHC would reassure physicists and funders of the adequacy of the ILC’s 500-GeV first-phase collision energy. And discovering supersymmetric particles or additional Higgs states at the LHC with masses near or above 500 GeV would argue the urgency of proceeding to the linear collider’s 1-TeV upgrade. “But there’s little chance of getting approval for the ILC before the LHC has seen something interesting,” says the University of Chicago’s Melvyn Shochet, chair of the High Energy Physics Advisory Panel (HEPAP) to DOE and NSF.

Schedule realities

If all goes well, including adequate and timely funding, the GDE report concludes that ILC construction could begin in 2012 and be completed by 2019. But at a HEPAP meeting in Washington, DC, just one week after the report’s release, DOE Undersecretary for Science Raymond Orbach urged a more realistic view of the schedule and worried aloud about its consequences.

“Negotiating an international structure, selecting a site, obtaining firm financial commitments, and building the machine,” Orbach warned, “could take us well into the mid-2020s, if not later.” Fermilab’s Tevatron and SLAC’s B factory are scheduled for shutdown by 2010. So Orbach’s prognosis would leave the US particle-physics community without a collider for at least 15 years. In that light, he urged the community to come up with a productive program of lesser initatives to fill the uncomfortably long gap.

Barish looks at Orbach’s protracted schedule as a “useful kick in the pants.” He responds that “finishing the ILC before the end of the next decade will require, in parallel with the engineering design work, a major effort to organize the international collaboration, divide up responsibilities, and get commitments from governments.”