Never before have the questions we can ask about the universe by means of accelerator-based experiments been more compelling (see figure 1). Yet the rate of progress in this field is lagging. It is a truism that progress in particle physics is paced in large measure by progress in accelerators. Until about 10 years ago, the rate of advance in making ever higher collision energies available to experimenters was spectacular. Center-of-mass interaction energies were increasing steadily by almost a factor of 100 every 25 years. That the pace has now slowed dramatically can be seen in figure 2, the traditional Livingston plot that records the evolution of accelerator facilities and the consequent advance of the energy frontier.
The falloff in the energy frontier’s rate of advance might inspire the reader to ask whether we are approaching some inherent physical limit to the capability of accelerators, or perhaps some other limit. The answer is complex, but one thing is clear: We are not approaching a technical limit to the energies that can be achieved in the laboratory.
There are certainly technical limits, but they are relatively soft and far away. They have to do largely with electrical input power requirements, and they differ for different types of accelerators. 1 For lepton colliders, the input power limits do not come into play for particle energies below tens of TeV. For proton accelerators, they do not arise below hundreds of TeV. One TeV (1012 electron volts) is the energy to which protons and antiprotons are accelerated in the Fermilab Tevatron, the highest energy particle accelerator built thus far.
Because we find ourselves in the middle of this historical development, it is probably not possible to have a detailed understanding of the factors and events that have influenced where we are. But one can say with confidence that both economic and social factors are at work. We begin with the economic factors, because they are easier to deal with conceptually, if not in the laboratory.
Let me put it baldly and simplistically: If the cost of the ill-fated Superconducting Super Collider had been on the order of $1 billion rather than many times that sum, it might well not have been canceled. If the cost of CERN’s Large Hadron Collider (LHC), scheduled for completion in 2005, were only of order $1 billion, it might well have been completed by now. If any of the TeV electron–positron linear collider schemes now being studied around the world could be built for something like $1 billion, one or more of them might already be under construction.
The present state of affairs should have come as no surprise. Let me quote from the cover letter transmitting the 1980 report of the Department of Energy’s high-energy physics advisory subpanel on accelerator research and development:
You will note that, in the 50-odd years of American accelerator science associated with particle physics research, enormous strides in increasing particle beam energies and in decreasing unit costs have been and are being made. … Our primary conclusion is that, despite spectacular past and present accomplishments of the field, we must redouble our efforts to improve cost effectiveness of our accelerators if the needs of US particle physics are to be met in the resource-limited situation in which we find ourselves. This needed improvement will entail continual and vigorous improvement of existing technologies as well as the discovery and reduction to practice of new acceleration methods and technologies. If we are to meet this challenge we will need to devote more of our intellectual and monetary resources to high energy accelerator R&D than is now our practice. 2
These words are ever so much truer today than they were 20 years ago. To be semi-quantitative in addressing this challenge, I note that, in 1980, about 1% of the resources available to high-energy physics in the US were being devoted to what might be called advanced accelerator research and development (AARD), laying the conceptual and technological foundations for future frontier facilities. We should distinguish this from research and development applied to projects (RDAP), that is to say, development directed at already highly developed facility concepts.
Currently, AARD is faring better, with more than 3% of the available resources. Half of that 3% goes to support accelerator work at universities. In the past, the RDAP allocation has been as high as 10% or even more when big projects were in the offing. Today RDAP amounts to about 4%. It might be useful in future to think in terms of an intermediate R&D category in which the overall concept of a facility has been widely accepted but subsystem concepts are explored in parallel. The focus would be on finding the most economical means of accomplishing the final goal. On the worldwide stage, evolution in this direction is already upon us.
Cheaper but more expensive
It is, of course, a gross oversimplification to measure the economic challenge strictly in dollars per GeV of acceleration. But it’s a good place to start. Over the past 40 years, the cost (not corrected for inflation) per center-of-mass GeV for proton accelerators has decreased by about a factor of 10. That is remarkable indeed, considering that inflation tacks on an additional order of magnitude. In that 40-year time span, however, the energy frontier has risen by a factor of almost 2000. Therefore, despite the great improvement in GeV per dollar, the nominal cost of new facilities—the dollars that folks actually spend—has risen about 200 times. That has landed accelerator facility costs in the realm of high political visibility, with all its attendant challenges.
We can gain some insight into the evolution of costs by examining the causes of both the decreasing and increasing costs. There are several elements in the decreasing cost per GeV. One important factor was the great and continued advance in understanding the underlying accelerator physics that followed the invention of strong focusing and the high-level language it brought to the analysis of complex beam systems. That spawned an elaboration of the division of labor out of which came the “accelerator physicists.” Most of them, in the beginning, were particle physicists who turned their talents to elaborating the new accelerator theory and the technology that went with it.
Another major factor has been the continual improvement in materials and manufacturing processes that can be applied to accelerator components and subsystems. Normal (nonsuperconducting) magnet materials and construction techniques, for example, have advanced so much that the cost per pound of high-quality magnet steel has remained at about $1 per kilogram for many years, seemingly oblivious of inflation. Likewise, the nominal cost of punching high-precision laminations from this material has remained unchanged despite inflation. To cite a more high-tech example, fine-filament superconducting NbTi wire cost about $3 per kiloamp meter in 1982. Now it costs only $1/(kA m). The inflation-adjusted cost has fallen more than fivefold. (For comparison, copper wire nowadays costs $10/[kAm].) The combination of improved physics understanding and improved materials and processes has been doubly beneficial. It has let us simplify and use less of the input material for a given function.
Many components indispensable to modern accelerators—for example, electronic instruments, computer technology, and control systems—have been taken straight from other technological sectors. Continued improvements in technological productivity and in materials, devices, and processes are manifest in the world at large. We can expect to harness them as we have done in the past, hoping to do even better as necessity presses in upon us.
On the other side of the ledger is the fact that the bar is constantly being raised. Just making raw high energy isn’t good enough. As the energy frontier advances, the beam intensities (or luminosities) required to keep the important event rates adequate have to increase as the square of the collision energy in order to compensate for reaction cross sections that plummet with increasing energy. At the energy frontier, one commonly speaks of cross sections in picobarns (10−36 cm2) or even femtobarns (10−39 cm2). A collider’s luminosity is the measure of its event rate per unit reaction cross section.
In earlier days, just getting some beam at full energy was usually sufficient to produce acceptable data rates. Today storage-ring devices carry gigawatt circulating-beam energies, and the TeV linear accelerators under discussion will dissipate tens of megawatts of beam power continuously. Furthermore, to be successful at all, these beams must be of very high quality. They must be confined to a very small phase-space volume so that they can be brought to ultrafine focus at the point where the counterpropagating beams collide. All this has required not only much better understanding of accelerator physics but also much greater complexity, and thus cost, of the accelerator systems—particularly vacuum, radio-frequency, and control systems.
So how can we make significant strides toward lowering cost per unit energy while maintaining scientifically useful event rates? One possible approach would be to implement some substantially new acceleration schemes, such as lasers with plasma mode conversion, 3 to give the transverse electromagnetic field of the laser beam a longitudinal accelerating component. Another idea would be to push some parameter of present methods into a new regime—for example, operating linacs at millimeter wavelengths 4 or magnets at tens of tesla using some high-temperature superconducting material yet to be developed. 5
A third approach might be to work very hard at doing much better things that we already do quite well—for example, building storage rings, linacs, synchrotrons, or some others of the classical accelerator types that have proven so successful in recent decades. Finally, to circumvent some of the cost constraints that confront attempts to push the traditional schemes, we might consider accelerating some particle species other than the usual protons, electrons, and their antiparticles. (See the article on muon and photon colliders by Andrew Sessler in Physics Today, March 1998, page 48.)
Work is actually now going on in all of these areas. So why has this broad approach been insufficient, thus far? The short answer is that more new ideas are needed, for which we need more intellectual resources and support for them. Of course even that may not be enough, but we sure haven’t tried hard enough yet.
Social and cultural factors
Bringing to bear resources that match the challenge will require a culture change in more than one community. In the early days, accelerators were treated more or less as part of the general apparatus of nuclear and particle physics, and the important inventions in the art were made by the stakeholders themselves. The cyclotron, the synchrotron, and the proton linac are but three examples (see the article by Valentine Telegdi in Physics Today, October 2000, page 25). But as accelerators became larger and more complex and the theoretical understanding became more sophisticated, a more-or-less spontaneous division of labor spawned a breed of accelerator specialists.
This specialization permitted rapid exploitation of the new and deeper understanding of the classical acceleration methods. But ultimately it lulled us into thinking that these new specialists would suffice to carry the accelerator ball forever. We completely neglected the rapidly mounting challenges that had to be faced. This notion of the accelerator specialist has permeated the culture of our national labs and university physics departments to such an extent that in only very few physics departments is work on accelerators considered respectable physics. The government funding agencies do tend to take a supportive view of frontier facility development. But by their nature, these agencies cannot depart very far from the cultural norms of particle physics.
Another significant factor is the illusion of complexity: How can an ordinary physicist contribute to this business when there are experts at the national labs who know accelerators inside out? But I argue that many of the challenges involve very basic physics, instrumentation, and a need for clever ideas about how to build things. And these are just exactly the talents that it takes to do experimental physics!
The most difficult challenge of all may come from how our patrons view the nature of the field. The need to be forever preparing for the future has the feel of an enduring commitment of public resources. The great questions that physics addresses are not going to be laid to rest tomorrow. We’re talking about an eternal conversation concerning things that matter in the physical world, a conversation that evolves to ever-new dimensions, contributing all the while to the general culture and to the economy.
Put this way, our quest sounds open-ended. And, as we know well, there are few things that the keepers of the purse like less than open-ended commitments. To surmount this obstacle, we will need, at the very least, to show the public that we have a plan that addresses frontiers of important and exciting knowledge, and that the plan is responsible rather than megalomaniac.
To be more specific about the technical details of current frontier challenges, I mention a list of the facilities that have been under discussion in the world particle-physics community. None of these have yet attained the status of “project.” All will need great attention to make realization possible. At the “sensitivity frontier,” where the principal issue is event rate rather than maximum energy, there are proposals to build a tau–charm factory for producing τ leptons and charmed mesons in profusion; a very high-luminosity b physics machine to study the physics of the b (bottom) quark well beyond the projected capability of the current B-meson factories in the US and Japan; a neutrino factory; and a Z boson factory with luminosity many times that of LEP, the large electron–positron collider at CERN that is now being dismantled to make way for the LHC. At the energy frontier are proposals for a TeV electron–positron linear collider, a muon collider, and the Very Large Hadron Collider (VLHC), a successor to the LHC. The charm and b factories might be national efforts. But the other, larger undertakings would probably have to be international facilities.
The “natural” time scales (R&D plus construction) for all of these frontier accelerators are long—at least 5 years for the most modest, and perhaps 20 years for the most difficult. For it to be possible, even in principle, to have frontier facilities ready when they are needed, the preparatory work must proceed in parallel. We will need a disciplined approach, in which the world community selects a realistic subset of the facilities we envision.
There are needs and opportunities for willing experimentalists to make key contributions to all of these potential projects. The clearest example is that of the neutrino factory or the muon collider described in Sessler’s Physics Today article and shown in figure 3. These two are closely related in their technical challenges. For a neutrino factory, the decay of the muons in a storage ring would provide a well-characterized and controlled beam of neutrinos. An important rationale for the muon collider is that the muon—a cousin of the electron, but with about 200 times its mass—dissipates far less energy by synchrotron radiation when its trajectory bends. That would permit muon machines with much stronger bending and focusing magnets than an electron machine could tolerate.
A neutrino factory based on decaying muons would look much like half of a muon collider, although the two are quantitatively different. Each would start with an intense proton beam bombarding a pion-production target in unprecedentedly short nanosecond bunches. That would yield a very diffuse beam that decays into muons, all of which must be captured in a very high magnetic field to preserve as many as possible. One must then somehow manipulate and cool the muon beam to shrink its phase-space volume so dramatically that it can fit into something like an ordinary transport line for acceleration and storage.
The technical challenges here are many and varied. A few of them are standard accelerator problems; others are quite new. The beams have energy spreads of perhaps factors of two rather than the usual 1% or less that is common in most accelerator applications. No cooling process now in use can do the needed phase-space condensation on the time scale required here. Remember that a muon’s life span is measured in microseconds.
Thus the whole idea of ionization cooling must be developed in some form or other, or we must invent some very different way of cooling beams. Well-grounded experimental physicists are as capable of grappling with many of these challenges as any accelerator expert.
Hope for the future
There are signs that this need for a major change in the way we do business is penetrating the minds of the worldwide community of particle physicists. Consider four examples:
▹ The formation of the Neutrino Factory and Muon Collider Collaboration (MC) is very interesting. This is a group of about 150 physicists representing many institutions, the majority of which are universities. 7 Some individual members are accelerator experts, but many are experimental particle physicists interested in someday using frontier facilities based on muon beams. Recently the MC participated in a feasibility study for an entry-level neutrino factory under the aegis of Fermilab. The study’s author list covers 43 institutions in the US and abroad. The MC has its own spokesperson and project manager, and a budget from DOE. Collaboration members are engaged in a wide variety of conceptual and R&D activities. So here we have stakeholders outside the national labs, in on the ground floor of a potentially important frontier accelerator facility. Naturally, to have some of the needed infrastructure available for the R&D work, the collaboration must work in closely with national laboratories. So far this is happening, but we’re not yet far enough along to see the ultimate outcome.
▹ Another possibility for university physicists is the Orion project at the Stanford Linear Accelerator Center. 8 A facility has been set up at SLAC with powerful infrastructure that lets university groups with ideas for advanced accelerators involving electron beams and lasers do experiments even though they may have only modest resources of their own.
Orion and MC might loosely be termed intra-regional attempts to involve the stakeholders in developing the facilities required to push back the frontiers of elementary particle science. The two other examples are more global in their geographic scope:
▹ The worldwide electron–positron “linear-collider derby” is now quite competitive, with separate entrants based in Europe, Japan, and the US. Each is based at a national or regional international laboratory, but each involves collaborators from other nations or regions. Despite the competitive nature of this worldwide activity, the participants have made a remarkable achievement in voluntary collaboration to the extent that they have agreed to cast their developing designs in an elaborate system of common terms of reference, so that detailed comparisons can be made. 9 As a result, all of the concepts have improved enormously over the original offerings.
Each linear-collider collaboration has by now constructed test facilities for demonstrating important features of its particular approach. There is a growing feeling in the world particle-physics community that only one such facility may be possible, and that inter-regional collaboration may be essential for its final stages of R&D, design, financing, and even its operation. With this understanding comes the realization that, to make such a project happen, the world community of stakeholders will themselves have to take the initiative in selecting the best concept, leaving the eventual siting to some international political process yet to be devised. Whether such an unusual social advance can be achieved is questionable. That question is very much on the minds of the various international physics communities as they address their futures and try to discern the best way to address the compelling physics questions we can now ask. It is interesting to note that similar world-scale collaborative/competitive activity can now be discerned both in the neutrino-factory and muon-collider efforts.
▹ Our last example is the concept of the Global Accelerator Network. 10 At a time when there will be only a few frontier facilities big enough to seek out the ultimate nature of the physical world, how can the cultural benefits of that adventure continue to be shared widely? The vision embodied in the Global Accelerator Network is to create a means for the currently involved institutions across the globe to share actively in the conception, development, construction, and eventual operation of those few facilities from their own home sites, no matter where the actual physical facilities might be placed.
Realizing such a vision will be very complex, both technically and socially—the latter, no doubt, overshadowing the former. The International Committee on Future Accelerators (ICFA), a longstanding organization of the world community, has taken up the challenge and has formed two international working groups to address it. The viability of this vision may be tested soon enough, as one or another of the strong national centers puts forward a proposal for a new major frontier facility.
The bottom line
If we are to continue progress with accelerator-based particle physics, we will have to mount much more effective efforts in the technical aspects of accelerator development—with a strong focus on economy. Such efforts will probably not suffice to hold constant the cost of new facilities in the face of the ever more demanding joint challenges of higher collision energy and higher luminosity. Making available the necessary intellectual and financial resources to sustain progress would seem to require social advances of unprecedented scope in resource management and collaboration.
Significant steps are already being taken in this direction. For example, the world community of physicists interested in neutrino factories and the eventual possibility of colliding muons has already begun to organize effectively and the various electron–positron linear-collider proponents have put considerable effort into establishing real international collaborations and cooperation. A crucial step not yet taken—but now needed—is that of collaboration in setting priorities on a world scale. That will take considerable courage and insight. But there seems to be a growing will to find the way. Only time will tell.
MAURY TIGNER is director of the Laboratory for Nuclear Studies and professor emeritus of physics at Cornell University in Ithaca, New York. He led the Central Design Group for the Superconducting Super Collider.