How to MakeAntiprotons
Rare particles reside in the realm of everyday reality at Fermilab.
by Mike Perricone
Antimatter is the stuff that science fiction dreams are made of: use it as a power source, and you can embark on your own "Star Trek." Or so the movies tell us.
Antimatter will remain the fuel of dreams for the foreseeable future. But antimatter in the form of antiprotonsthe mirror image of protons, carrying the same mass but an opposite (negative) chargeis a fact of everyday life at Fermilab, and nowhere is the production process better known or documented.
Collecting just 19 antiprotons requires a million collisions between protons and a fixed target. There are about 2x1025 antiprotons to the ounce; it would take about 20 billion years to produce a full ounce of antiprotons at Fermilab. As rare as they are, antiprotons fuel hopes for new discoveries. The observation of the top quark in 1997, for example, resulted from the collision of protons and antiprotons.
The Antiproton Source, the triangular-shaped ring wedged between the Booster and the new Main Injector, has three main components: a target, where antiprotons are produced; and two storage rings. The Debuncher accepts pulses of antiprotons and begins cooling them into a beam; and the Accumulator refines that beam into a dense core and stores it.
Antiprotons originate as a beam of 120-GeV protons extracted from the Main Injector and transferred to the target, a drum consisting of sections of nickel. The proton beam is directed not toward the end of the drum, but through the side. The drum is rotated after each "hit," so the proton beam does not keep landing in the same place.
"If we hit the same spot all the time, we would destroy the target," explained Dave McGinnis of the Antiproton Source.
The particles coming off the target are splayed in all directions, but they must be channeled through a beam pipe of limited diameter. But conventional magnetic focusing, using two separated quadrupole magnets, like those in the Main Injector, requires too much space and allows the beam to spread out too wide to fit the downstream beampipe. Each of the quadrupoles focuses the beam in just one plane, horizontally or vertically. The solution is a magnetic lens focusing in both planes simultaneously.
"The only way to do that," said McGinnis, "is to take a (tubular) hunk of metal and run a ton of current down through it, so that the magnetic field provides focusing that is radially inward no matter what plane youre in."
The lens is made of lithium, a metal light enough to prevent the antiprotons from scattering. The lens is surrounded by a transformer to supply the necessary current of 650,000 amps. The current, magnetic field and radiation subject the lithium lens to tremendous stresses.
"Were on our 20th lithium lens," McGinnis said. "They cost about $100,000 each. We try to get 10 million pulses from a lens."
After the jumble of particles is focused through the lithium lens, a pulsed magnetic field kicks out positively charged particles and bends negatively charged particles farther along the beam path. There are other negatively charged particles in addition to the antiprotons; but these, primarily pions, decay quickly. Whats left is a beam of antiprotons.
The antiprotons come off the target in bunches 20 nanoseconds apart, bound for the Debuncher Ring. Using radio frequency accelerating cavities, the Debuncher eliminates the bunch structure to reduce the beams very large energy spread. This process takes only 40 milliseconds, with a two-second wait before the Main Injector accelerates another batch of protons to make antiprotons. The Debuncher uses this extra time to "pre-cool" the beam.
The antiprotons produced at the target station form a diffuse (or random) beam that would not be useful for colliding beam physics. Stochastic cooling removes the randomness of the particles. Since heat can be defined as randomness of motion, like the randomness of motion of a gas in a balloon, a less-random beam is said to be cooled.
"In stochastic cooling," McGinnis explained, "we go to every little antiproton in the beam and say, Are you random? Are you moving around? If you are, please stop!"
Stochastic cooling also dictates the distinctive triangular shape of the Debuncher and Accumulator.
"The Debuncher would love to be a circular ring," McGinnis said. "But we want to have it in the same tunnel as the Accumulator, which is a triangle to help in stochastic coolingour bread and butter. Most of the Antiproton Source has been designed to be a very good stochastic cooling machine."
Actually, the Accumulator is a pseudo-triangle with flat sides and rounded corners. The corners act as a prism, separating out particles with different momenta; as they go around the corner, particles with low momenta go to the inside of the track, while particles with high momenta go to the outside of the track.
To resolve or "see" antiprotons, the stochastic cooling systems must have wide bandwidths on the order of several billion Hertz (gigahertz). Microwave phased-array antennas (pickups) that are placed on the walls of the beam pipe detect the motion of the antiproton beam. These pickups intercept the electromagnetic wake of the high-speed antiproton beamlike detecting the motion of a boat traveling on a river by recording the size of the wake that hits the riverbank.
Even with such huge bandwidths, the stochastic cooling systems cannot resolve the motion of a single antiproton: theres too much noise from all the other antiprotons passing over the pickup at the same time. Only a phenomenon called "mixing" makes cooling possible.
Mixing is just what it says: particles with different momenta take different times to travel around the ring, and get spread out over the beam. After a few turns around the ring, the initial "noise" signal is replaced by a weaker "noise" signal from a more diffuse set of background antiprotons. Eventually the noise averages to zero and the cooling system can resolve a single antiproton.
The signal detected by the pickup is next amplified by a factor of 1015(picowatts to kilowatts), then filtered and applied to the kicker array. Because mixing can quickly re-jumble the antiprotons, the pickup signal must travel to the kicker array in less than one half of a millionth of a second. Since an antiproton takes about 1.6 millionths of a second to make the trip around the Accumulator or Debuncher, the pickup signal must take an underground short-cut across the rings to arrive at the kicker array in time. The kicker array, very similar to the pickup array, transmits electromagnetic waves to the center of the beampipe which deflect (or kick) the antiproton beam in the right direction.
After pre-cooling in the Debuncher, the beam is transferred to the Accumulator and merged with antiprotons created during previous Main Injector cycles. Stochastic cooling increases the density of the beam by a factor of about 18,000 to make it useful for colliding beam physics. In the Debuncher, where there are about 50 million antiprotons, the cooling process takes seconds. But in the Accumulator, with as many as a trillion antiprotons, cooling takes tens of minutes.
With so few antiprotons in the Debuncher, the thermal noise generated within the pickups themselves can overwhelm the antiproton signal. Among the major upgrades for Run II, the pickup temperature will be lowered from -320 degrees Fahrenheit to -440 degrees, reducing that noise. The frequency range of the cooling systems will grow from 2-4 GHz to 4-8 GHz. Also, the bandwidth of the main Accumulator cooling system will grow from 1-2 GHz to 2-4 GHz to increase the rate for accumulating antiprotons. To stabilize the cooling system with this large increase, the Accumulator "mixing" was modified by changing the strength and location of several quadrupole magnets.
Once the 8-GeV antiproton beam is packed up nice and tight, its ready for transfer back to the Main Injector, where it will be accelerated to 150 GeV and handed over to the Tevatron.
"It used to take 20 to 24 hours to build up enough antiprotons to supply the Tevatron," McGinnis said. "Now it will take us about 12 hours. With our upgrades, the process will go much faster and well produce a much denser beam of particles."
And use that beam to fuel the hopes of new discoveries in particle physics.
|last modified 03/05/1999 email Fermilab|