Strong focusing involves shaping the magnetic fields that guide the particles so that they are held much closer to an ideal orbit. The key idea is to machine the pole pieces into appropriate curves so that the magnetic forces on the particle generate rapid oscillations with tiny amplitudes around the ideal orbit. That is stability. Before strong focusing, the doughnut-shaped vacuum chambers had to be 20 to 40 inches wide, requiring magnet poles of similar sizes. The Brookhaven breakthrough permitted reduction in the size of the magnet's vacuum chamber to 3 to 5 inches. The result? A huge savings in cost per MeV of accelerated energy.
Strong focusing changed the economics and, early on, made it thinkable to build a synchrotron with a radius of almost 200 feet. Later we'll talk about the other parameter; the strength of the magnetic field. As long as iron is used for guiding the particles, this is limited to 2 tesla, the strongest magnetic field that iron can support without turning purple. Breakthrough is a correct description of strong focusing. Its first application was a 1 GeV electron machine built by Robert Wilson the Quick at Cornell. Brookhaven's proposal to the AEC to build a strong-focusing proton machine was said to have been a two-page letter! (Here we can lament the growth of bureaucracy but it would do no good.) This was approved, and the result was the 30 GeV machine known as AGS, completed at Brookhaven in 1960. CERN scrapped its plans for a 10 GeV weak-focusing machine and used the Brookhaven strong-focusing idea to build a 25 GeV strong-focusing accelerator for the same price. They turned it on in 1959.
By the late 1960s, the idea of using tortured pole pieces to achieve strong focusing had given way to a separated function concept. One installs a "perfect" dipole guide magnet and segregates the focusing function in a quadrupole magnet symmetrically arrayed around the beam pipe.
Using mathematics, physicists learned how complex magnetic fields direct and focus particles; magnets with larger numbers of north and south poles—sextupoles, octupoles, decapóles—became components of sophisticated accelerator systems designed to exercise precise control over the particle orbits. From the 1960s on, computers were more and more important in operating and controlling the currents, voltages, pressures, and temperatures in the machines. Strong focusing magnets and computer automation made possible the remarkable machines that were built in the 1960s and '70s.
The first GeV (billion-electron-volt) machine was the modestly named Cosmotron, which began operation at Brookhaven in 1952. Cornell followed with a 1.2 GeV machine. Here are the other stars of that era...
ACCELERATOR ENERGY LOCATION YEAR
Bevatron 6 GeV Berkeley 1954
AGS 30 GeV Brookhaven 1960
ZGS 12.5 GeV Argonne (Chicago) 1964
The "200" 200 GeV Fermilab 1972 (upgraded to 400 GeV in 1974)
Tevatron 900 GeV Fermilab 1983
Elsewhere in the world there were the Saturne (France, 3 GeV), Nimrod (England, 10 GeV), Dubna (USSR, 10 GeV), KEK PS (Japan, 13 GeV), PS (CERN/Geneva, 25 GeV), Serpuhkov (USSR, 70 GeV), SPS (CERN/Geneva, 400 GeV).
The third breakthrough was cascade acceleration, a concept attributed to Cal Tech physicist Matt Sands. Sands decided that, when one is going for high energy, it is inefficient to do it all in one machine. He envisioned a sequence of different accelerators, each optimized for a particular energy interval, say 0 to 1 MeV, 1 to 100 MeV, and so on. The various stages can be compared to gears on a sports car, with each gear designed to raise the speed to the next level in the optimal manner. As the energy increases, the accelerated beam gets tighter. At the higher energy stages, the smaller transverse dimensions thus require smaller and cheaper magnets. The cascade idea has dominated all machines since the 1960s. Its highest exemplars are the Tevatron (five stages) and the Super Collider under construction in Texas (six stages).
IS BIGGER BETTER?
A point that may have been lost in the preceding discussion of technical considerations is why it helps to make cyclotrons and synchrotrons big. Wideröe and Lawrence demonstrated that one doesn't have to produce enormous voltages, as earlier pioneers believed, to accelerate particles to high energies. One just sends the particles through a series of gaps, or designs a circular orbit so that one gap can be reused. Thus in circular machines there are but two parameters: magnet strength and the radius of the orbiting particles. Accelerator builders adjust these two factors to get the energy they want. The radius is limited by money, mostly. Magnet strength is limited by technology. If we can't boost the magnetic field, 'we make the circle bigger to increase the energy. In the Super Collider we know that we want to produce 20 TeV in each beam. And we know (or we think we know) how strong a magnet we can build. From that we can extrapolate how big around the tube must be: 53 miles.
A FOURTH BREAKTHROUGH: SUPERCONDUCTIVITY
Back in 1911 a. Dutch physicist discovered that certain metals, when cooled to extremely low temperatures—just a few degrees above absolute zero on the Kelvin scale (−273 degrees centigrade)—lose all their resistance to electricity. A loop of wire at that temperature would carry a current forever with no use of energy.
In your house, electrical power is supplied via copper wires from the friendly power company. The wires get warm because of the frictional resistance they offer to the flow of current. This waste heat uses power and adds to your bill. In conventional electromagnets for motors, generators, and accelerators, copper wires carry currents that produce magnetic fields. In a motor the magnetic field turns bundles of current-carrying wires. Feel the warm motor. In an accelerator the magnetic field steers and focuses the particles. The magnet's copper wires get hot and are cooled by a powerful flow of water, usually through holes in the thick copper windings. To give you some idea of where the money goes, the 1975 electric bill for the Fermilab accelerator was about $15 million, some 90 percent of which was for the power used in running the magnets for the 400 GeV main ring.
Early in the 1960s a technical breakthrough took place. New alloys of exotic metals were able to maintain the fragile state of superconductivity while conducting huge currents and producing high magnetic fields. All of this at the more civilized temperatures of 5 to 10 degrees above absolute zero rather than the very difficult 1 to 2 degrees required for common metals. Helium is a true liquid at 5 degrees (everything else solidifies at this temperature), so the possibility of practical superconductivity emerged. Most of the large laboratories began working with wire made of such alloys as niobium-titanium or niobium 3-tin in place of copper and surrounding the wires with liquid helium to cool them to superconducting temperatures.
Large magnets using the new alloys were built for particle detectors—for example, to surround a bubble chamber—but not for accelerators, which required that magnetic fields increase in strength as the particles gain energy. The changing currents in the magnets generate frictional effects (eddy currents) that normally destroy the superconducting state. Much research was addressed to this problem in the 1960s and '70s, with Fermilab, under Robert Wilson, serving as a leader in the field. Wilson's team began R&D in superconducting magnets in 1973, shortly after the original "200" accelerator began operating. One motivation was the exploding costs of electrical power due to the oil crisis of that era. The other was competition from the European consortium, CERN, based in Geneva.
The 1970s were lean years for research funds in the United States. After World War II the world leadership in research had been solidly in this country, as the rest of the world labored to rebuild war-shattered economies and scientific infrastructures. By the late 1970s, balance had begun to be restored. The Europeans were building a 400 GeV machine, the Super Proton Synchrotron (SPS), which was better funded and better supplied with the expensive detectors that determine the quality of the research. (This machine marked the beginning of another cycle in international collaboration and competition. In the 1990s Europe and Japan remain ahead of the United States in some research fields and not far behind in most others.)
Wilson's idea was that if one could solve the problem of varying magnetic f
ields, a superconducting ring would save an enormous amount of electrical power while producing more powerful magnetic fields, which for a given radius would translate to higher energy. Aided by Alvin Tollestrup, a Cal Tech professor spending a sabbatical year at Fermilab (he eventually extended this to permanence), Wilson studied in great detail how changing currents and fields create local heating. Research going on in other labs, especially the Rutherford Lab in England, helped the Fermilab group build hundreds of models. They worked with metallurgists and materials scientists and, between 1973 and 1977, succeeded in solving the problem. One could ramp the model magnets from zero current to 5,000 amperes in 10 seconds, and the superconductivity persisted. In 1978–79 a production line began producing twenty-one-foot magnets with excellent properties, and in 1983 the Tevatron began operating as a superconducting "afterburner" at the Fermilab complex. The energy went from 400 GeV to 900 GeV, and the power consumption was reduced from 60 megawatts to 20 megawatts, with most of that used to produce liquid helium.
When Wilson began his R&D program in 1973, the annual production of superconducting material in the United States was a few hundred pounds. Fermilab's consumption of 125,000 pounds of superconducting material stimulated producers and radically changed the posture of the industry. Today the biggest customers are firms that make magnetic resonance imaging (MRI) devices, for medical diagnosis. Fermilab can take a modicum of credit for this $500-million-a-year industry.
THE COWBOY LAB DIRECTOR
The man who deserves much of the credit for Fermilab itself is our first director artist/cowboy/machine designer Robert Rathbun Wilson. Talk about charisma! Wilson grew up in Wyoming, where he rode horseback and studied hard at school, winning a scholarship to Berkeley. There he was a student of E. O. Lawrence's.
I have already described the architectural feats of this Renaissance man in building Fermilab, but he was technologically sophisticated as well. Wilson became the founding director of Fermilab in 1967 and received an allocation of $250 million to build (so said the specifications) a 200 GeV machine with seven beam lines. Construction, started in 1968, was to take five years, but Wilson completed the machine ahead of schedule in 1972. By 1974 it was working steadily at 400 GeV with fourteen beam lines and with $10 million left over from the original allocation—all this with the most splendid architecture ever seen in a U.S. government installation. I recently calculated that had Wilson been in charge of our defense budget over the past fifteen years with the same skills, the United States would now be enjoying a tidy annual budget surplus and our tanks would be the talk of the art world.
One story has it that Fermilab first sprang into Wilson's mind in the early 1960s in Paris, where he was an exchange professor. One day he found himself sketching a beautiful, curvaceous nude model with a group of other artists in a public drawing session at the Grande Chaumière. The "200" was being discussed in the United States, and Wilson didn't like what he read in his mail. While others drew breasts, Wilson drew circles for beam tubes and adorned them with calculations. This is dedication.
Wilson wasn't perfect. He took short cuts when building Fermilab, and not all were successful. He complained bitterly that one blooper cost him a year (he would have finished in 1971), and an extra $10 million. He also gets mad, and in 1978, disgusted with the slow pace of federal funding for his superconducting work, he quit. When I was asked to become his successor I went to see him. He threatened to haunt me if I didn't take the job, and that did it. The prospect of being haunted by Wilson on his horse was too much. So I took the job and prepared three envelopes.
A DAY IN THE LIFE OF A PROTON
We can illustrate everything that has been explained in this chapter by describing Fermilab's cascade accelerator which has five sequential machines (seven if you want to count the two rings in which we make antimatter). Fermilab is a complex choreography of five different accelerators, each a step up in energy and sophistication, like ontogeny recapitulating phylogeny (or whatever it recapitulates).
First we need something to accelerate. We run over to Ace Hardware and buy a pressurized bottle of hydrogen gas. The hydrogen atom consists of an electron and a simple nucleus of one proton. There are enough protons in this bottle to run Fermilab for a year. Cost: about twenty dollars if you return the bottle. The first machine in the cascade is nothing less than a Cockcroft-Walton electrostatic accelerator, 1930s design. Although it is the most ancient of the Fermilab series of accelerators, it is the most futuristic looking, adorned with very large and shiny balls and doughnutlike rings that photographers like to shoot. In the Cockcroft-Walton a spark strips the electron away from the atom, leaving a positively charged proton essentially at rest. The machine then accelerates the protons, creating a 750 KeV beam aimed at the entrance to the next machine, which is a linear accelerator, or linac. The linac sends the protons down a 500-foot-long series of radio-frequency cavities (gaps) to bring them to 200 MeV.
At this respectable energy they are transferred via magnetic steering and focusing to the "booster," a synchrotron, which whirls the protons around and raises their energy to 8 GeV. Just think: at this point we've produced higher energies than the Berkeley Bevatron, the first GeV accelerator, and we have two rings yet to go. This load of protons is then injected into the main ring, the almost-four-mile-around "200" machine, which in the years 1974–1982 worked at 400 GeV, twice the official energy it was designed for. The main ring was the workhorse of the Fermilab complex.
After the Tevatron came on-line in 1983, the main ring began taking life a little easier. Now it takes the protons up to only 150 GeV and then transfers them to the superconducting Tevatron ring, which is exactly the same size as the main ring and is just a few feet beneath it. In the conventional application of the Tevatron, the superconducting magnets carry the 150 GeV particle around and around, 50,000 circuits per second, gaining about 700 KeV per turn until, after about 25 seconds, they reach 900 GeV. By this time the magnets, powered by currents of 5,000 amperes, have increased their field strength to 4.1 tesla, more than twice the field that the old iron magnets could provide. And the energy required to maintain the 5,000 amperes is approximately zero! The technology of superconducting alloys is continually improving. By 1990 the 1980 Tevatron technology had been improved so that the Super Collider will use fields of 6.5 tesla, and CERN is working hard to push the technology to what may be a limit for niobium alloys—to 10 tesla. In 1987 a new kind of superconductor was discovered based on ceramic materials that require only liquid nitrogen cooling. Hopes were raised that a cost breakthrough was imminent, but the requisite strong magnetic fields are not there yet, and no one can estimate when and if these new materials will ever replace niobium titanium.
At the Tevatron, 4.1 tesla is the limit, and now the protons are kicked by electromagnetic forces into an orbit that brings them out of the machine into a tunnel, where they are divided up among some fourteen beam lines. Here experimental teams provide targets and detectors to do their experiments. Some thousand physicists work in the fixed-target program. The machine operates in cycles. It takes about 30 seconds to do all the acceleration. The beam is spilled out over another 20 seconds so as not to crowd the experimenters with too high a rate of particles for their experiments. This cycle is repeated every minute.
The external beam line is very tightly focused. My colleagues and I set up an experiment in "Proton Center," where a beam of protons is extracted, focused, and steered for about 8,000 feet onto a target 0.01 inches wide, the width of a razor blade. The protons collide with the thin edge. Every minute, day after day for weeks, a burst of protons strikes this target, never shifting by more than a small fraction of its width.
The other mode of using the Tevatron, the collider mode, is quite different, and we will discuss it in detail. In this mode, the injected protons coast around in the Tevatron at 150 GeV waiting for antiprotons, which in due course are delivered from the p-bar source and sent around the ring in the opposite direction. When both beams are
in the Tevatron, we begin ramping up the magnets and accelerating both beams. (More about how this works in a moment.)
At every phase of the sequence, computers control the magnets and radio-frequency systems, keeping the protons tightly bunched and under control. Sensors give information on currents, voltages, pressures, temperatures, the location of the protons, and the latest Dow Jones averages. A malfunction could send the beam careening out of its vacuum pipe and through the enveloping magnet structure, boring a very neat and very expensive hole. This has never happened—at least not yet.
DECISIONS, DECISIONS: PROTONS VS. ELECTRONS
We've been talking a lot about proton machines here, but protons aren't the only way to go. The nice thing about protons is that they are relatively inexpensive to accelerate. We can accelerate them to thousands of billions of electron volts. The Super Collider will accelerate protons to 20 trillion electron volts. In fact, there may be no theoretical limit to what we can do. On the other hand, protons are full of other particles—quarks and gluons. This makes the collisions messy and complicated. That's why some physicists prefer to accelerate electrons, which are pointlike, a-tomlike. Because they are points, their collisions are cleaner than with protons. The downside is that they are low in mass, so they are difficult and expensive to accelerate. Their low mass results in a large amount of electromagnetic radiation when steered around a circle. Much more power must be put in to make up for the radiation loss. While this radiation is a waste from the point of view of acceleration, it's a spinoff boon to some researchers because it is very intense and of very short wavelength. Many circular electron accelerators are actually devoted to producing this synchrotron radiation. Customers include biologists who use the intense photon beams to study the structure of huge molecules, electronic chip makers who do x-ray lithography, condensed-matter scientists, who study the structure of materials, and many other practical types.
The God Particle: If the Universe Is the Answer, What Is the Question? Page 30
