The God Particle

by Leon Lederman

  P-bars don't grow on trees, and you can't buy them at Ace Hardware. In the 1990s Fermilab is the world's largest repository of antiprotons, which are stored in a magnetic ring. A futuristic study by the U.S. Air Force and the Rand Corporation has determined that one milligram (one thousandth of a gram) of antiprotons would be an ideal rocket fuel, for it would contain the energy equivalent of about two tons of oil. Since Fermilab is the world leader in antiproton production (1010 per hour), how long would it take to make a milligram? At the present rate, a few million years of twenty-four-hour operation. Some incredibly optimistic extrapolations of technology might reduce this to a few thousand years. So my advice is not to invest in Fidelity's P-Bar Mutual Fund.

  The Fermilab collider scheme works as follows. The old 400 GeV accelerator (the main ring), operating at 120 GeV, throws protons against a target every two seconds. Each collision of about 1012 protons makes some 10 million antiprotons heading in the right direction with the right energy. With each p-bar there are thousands of unwanted pions, kaons, and other debris, but these are all unstable and they go away sooner or later. The p-bars are focused into a magnetic ring called the debuncher ring, where they are processed, organized, and compressed, then transferred to the accumulator ring. Both rings are about 500 feet around and store p-bars at 8 GeV, the same energy as the booster accelerator. It takes five to ten hours to accumulate enough p-bars to inject back into the accelerator complex. Storage is a delicate affair as all of our equipment is made out of matter (what else?), and the p-bars are antimatter. If they come into contact with matter—annihilation. So we must be fastidious in keeping the p-bars orbiting near the center of the vacuum tube. And the quality of the vacuum must be extraordinary—the best "nothing" that technology can buy.

  After accumulation and continued compression for about ten hours, we are ready to inject the p-bars back into the accelerator from whence they came. In a procedure reminiscent of a NASA launch, a tense countdown ensues to make sure that every voltage, every current, every magnet, and every switch is correct. The p-bars are zapped into the main ring, where they circulate counterclockwise because of their negative charge. They are accelerated to 150 GeV and transferred, again by magnetic legerdemain, to the Tevatron superconducting ring. Here the protons, recently injected from the booster via the main ring, have been patiently waiting, circulating tirelessly in the customary clockwise direction. Now we have two beams, running in opposite directions around the four-mile ring. Each beam is composed of six bunches of particles, with about 1012 protons and a somewhat smaller number of p-bars per bunch.

  Both beams are accelerated from 150 GeV, the energy imparted to them from the main ring, to the full Tevatron energy of 900 GeV. The final step is "squeeze." Because the beams are counter-circulating in the same small tube, they inevitably have been crossing each other during the acceleration phase. However, their density is so low that very few collisions between particles occur. "Squeeze" energizes special superconducting quadrupole magnets that compress beam diameters from soda straws (a few millimeters) to human hairs (microns). This increases the density of particles enormously. Now, when the beams cross, there is at least one collision per crossing. Magnets are tweaked to make sure the collisions take place at the center of the detectors. The rest is up to them.

  Once we have established stable operation, the detectors turn on and begin collecting data. Typically this continues for ten to twenty hours while more p-bars are being accumulated with the help of the old main ring. In time the proton and antiproton bunches become depleted and more diffuse, cutting the event rate. When the luminosity (the number of collisions per second) has gone down to about 30 percent, and if there are enough new p-bars stored in the accumulator ring, the beams are dumped and another NASA countdown ensues. It takes about a half hour to refill the Tevatron collider. About 200 billion antiprotons is considered an okay number to inject. More is better. These face some 500 billion protons, far easier to come by, to produce about 100,000 collisions per second. Improvements to all of this, designed for installation in the 1990s, will increase these numbers by about a factor of ten.

  In 1990 the CERN p/p-bar collider retired, leaving the field to the Fermilab facility with its two powerful detectors.

  WATCHING THE BLACK BOX: THE DETECTORS

  We learn about the subnuclear domain by observing, measuring, and analyzing the collisions induced by high-energy particles. Ernest Rutherford locked his team up in a dark room so they could see and count the scintillations generated by alpha particles hitting zinc sulfide screens. Our techniques of particle counting have evolved considerably since then, especially in the post-World War II period.

  Prior to World War II the cloud chamber was a major tool. Anderson used it to discover the positron, and it was found in cosmic ray laboratories around the world. My assignment at Columbia was to build a cloud chamber to operate with the Nevis cyclotron. As an absolutely green graduate student, I was unaware of the subtleties of cloud chambers and was competing with experts at Berkeley, Cal Tech, Rochester, and other such places. Cloud chambers are finicky devices, susceptible to "poisoning"—impurities that create unwanted droplets, which compete with those that delineate the particle tracks. No one at Columbia had any experience with these loathsome detectors. I read all the literature and adopted all the superstitions: clean the glass with sodium hydroxide and wash with triple distilled water; boil the rubber diaphragm in 100 percent methyl alcohol; mutter the right incantations ... A little prayer can't hurt.

  In desperation, I tried to get a rabbi to bless my cloud chamber. Unfortunately, I picked the wrong rabbi. He was Orthodox, very religious, and when I asked him to say a brucha (Hebrew: "blessing") for my cloud chamber, he demanded to know what a cloud chamber was. I showed him a photograph, and he was furious at my suggested sacrilege. The next guy I tried, a Conservative rabbi, upon seeing the picture, asked how the cloud chamber worked. I explained. He listened, nodded, stroked his beard, and finally said sadly that he just couldn't do it. "The law..." So I went to the Reform rabbi. He was just getting out of his Jaguar XKE when I came to his house. "Rabbi, can you say a brucha for my cloud chamber?" I pleaded. "Brucha?" he responded. "What's a brucha?" So I was worried.

  Finally I was ready for the big test. At this point everything should have worked, but each time the chamber was operated, I got dense, white smoke. At this stage Gilberto Bernardini, a true expert, arrived at Columbia and began looking over my shoulder.

  "Whatsa de brass rod, poking into de chamber?" he asked.

  "That's my radioactive source," I said, "to give tracks. But all I get it white smoke."

  "Tay-ka id oud."

  "Take it out?"

  "Si, si, oud!"

  So take it out I did, and a few minutes later ... tracks! Beautiful wavy threads of tiny droplets tearing through the chamber. The most beautiful sight I'd ever seen. What happened was that my millicurie source was so strong it was filling the chamber with ions, and each grew its own drop. The result: dense, white smoke. I didn't need a radioactive source. Cosmic rays, omnipresent in the space around us, kindly provide enough radiation. Ecco!

  The cloud chamber turned out to be a very productive instrument because one could photograph the trail of tiny droplets formed along the track of particles passing through it. Equipping it with a magnetic Held caused the tracks to curve, and measuring the radius of this curvature gave us the momentum of the particles. The closer the tracks were to being straight (less curvature) the more energetic the particles. (Remember the protons in Lawrence's cyclotron, which gained momentum and then described larger circles.) We took thousands of pictures that revealed a variety of data on the properties of pions and muons. The cloud chamber—looked at as an instrument rather than as a source of my Ph.D. and tenure—allowed us to observe some dozen tracks in each photograph. The pions take about a billionth of a second to pass through the chamber. We can provide a dense plate of material in which a collision can take place, which happens perhap
s once in every hundred photographs. Because pictures can be taken only about one per minute, the data accumulation rate is further limited.

  BUBBLE, BUBBLE, TOIL AND TROUBLE

  The next advance was the bubble chamber, invented in the mid-1950s by Donald Glaser, then at the University of Michigan. The first bubble chamber was a little thimble of liquid ether. The evolution of liquid hydrogen chambers up to the 15-foot monster, retired from Fermilab in 1987, was led by the famed Luis Alvarez at die University of California.

  In a chamber filled with liquid, often liquefied hydrogen, tiny bubbles form along the trail of particles passing through. The bubbles indicate the onset of boiling due to a sudden deliberate lowering of the pressure in the liquid. What this does is put the liquid above the boiling point, which depends on both temperature and pressure. (You may have experienced the difficulty of cooking an egg in your mountain chalet. At the low pressure of mountaintops, water boils well below 100 degrees C.) A clean liquid, no matter how hot, will resist boiling. For example, if you heat some oil in a deep pot above its normal boiling temperature, and if everything is really clean, it won't boil. But toss in a single piece of potato, and explosive boiling takes place. So to produce bubbles, two things are required: temperature above the boiling point and some kind of impurity to encourage the formation of a bubble. In the bubble chamber, the liquid is superheated by the sudden decrease in pressure. The charged particle, in its numerous gentle collisions with atoms of the liquid, leaves a trail of excited atoms that, after the pressure is lowered, are ideal for nudeating the bubbles. If a collision occurs between the incident particle and a proton (nucleus of hydrogen) in the vat, all the emerging charged products are also rendered visible. Since the medium is a liquid, dense plates are not necessary, and the collision point can be seen clearly. Researchers around the world took millions of photographs of collisions in bubble chambers, their analysis aided by automated scanners.

  So here is how it works. The accelerator shoots a beam of particles toward the bubble chamber. If this is a charged particle beam, ten or twenty tracks begin to crowd the chamber. Within a millisecond or so after the passage of the particles, a piston is rapidly moved, lowering the pressure and thereby beginning the formation of bubbles. After another millisecond or so of growth time, a light is flashed, film is moved, and we are ready for another cycle.

  It is said that Glaser (who won the Nobel Prize for his bubble chamber and promptly became a biologist) got his idea for nucleation of bubbles by studying the trick of increasing the head on a glass of beer by adding salt. The bars of Ann Arbor, Michigan, thus spawned one of the more successful instruments used to track the God Particle.

  There are two keys to collision analysis: space and time. We would like to record a particle's trajectory in space and its precise time of passage. For example, a particle comes into the detector, stops, decays, and gives rise to a secondary particle. A good example of a stopping particle is a muon, which can decay into an electron, separated in time by a millionth or so of a second from the stopping event. The more precise your detector, the more information. Bubble chambers are excellent for space analysis of the event. The particles leave tracks, and in bubble chambers we can locate points on those tracks to an accuracy of about 1 millimeter. But they provide no time information.

  Scintillation counters can locate particles in both space and time. Made of special plastics, they produce a flash of light when struck by a charged particle. The counters are wrapped in light-tight black plastic, and each tiny light flash is funneled to an electronic photomultiplier that converts the signal, indicating passage of a panicle, into a sharply defined electronic pulse. When this pulse is superimposed on an electronic train of clock pulses, the arrival of a particle can be recorded to a precision of a few billionths of a second. If a number of scintillation strips are used, a particle will strike several in succession, leaving a series of pulses that describe its path in space. The space location depends on the size of the counter, which typically establishes the location to a precision of a few inches.

  A major breakthrough was the proportional wire chamber (PWC), the invention of a prolific Frenchman working at CERN, Georges Charpak. A World War II hero of the Resistance and a concentration camp prisoner, Charpak became the preeminent inventor of particle detector devices. In his PWC, an ingenious, "simple" device, a number of fine wires, only a few tenths of an inch apart, are stretched across a frame. Typically the frame is two feet by four feet, with a few hundred two-foot-long wires strung across the four-foot span. Voltages are organized so that when a particle passes near a wire, it generates an electrical pulse in the wire, and the pulse is recorded. The accurately surveyed location of the struck wire locates one point on the trajectory. The time of the pulse is obtained by comparison with an electronic clock. By further refinements, the space and time definition can be pinpointed to approximately 0.1 millimeters and 10−8 seconds. With many such planes stacked in an airtight box filled with an appropriate gas, one can precisely define the trajectories of particles. Because the chamber is active for only a short interval of time, random background events are suppressed and very intense beams can be used. Charpak's PWCs have been a part of every major particle physics experiment since about 1970. In 1992 Charpak was awarded the Nobel Prize (alone!) for his invention.

  All of these different particle sensors and more were incorporated into the sophisticated detectors of the 1980s. The CDF detector at Fermilab is typical of one of the most complex systems. Three stories high, weighing 5,000 tons, and built at a cost of $60 million, it is designed to observe the head-on collisions of protons and antiprotons in the Tevatron. Here some 100,000 sensors, which include scintillation counters and wires in exquisitely designed configurations, feed streams of information in the form of electronic pulses to a system that organizes, filters, and finally records data for future analysis.

  As in all such detectors, there is too much information to handle in real time—that is, immediately—so the data are encoded in digital form and organized for recording on magnetic tape. The computer must decide which collisions are "interesting" and which are not, since there are over 100,000 collisions per second in the Tevatron, and this is expected to increase in the early 1990s to one million collisions per second. Now, most of these collisions are of no interest. The jewels are those in which a quark in one proton really smacks an antiquark or even a gluon in the p-bar. These hard collisions are rare.

  The information-handling system has less than a millionth of a second to examine a particular collision and make a fateful decision: is this event interesting? To a human this is mind-boggling speed, but not to a computer. It is all relative. In one of the big cities, a turtle was attacked and robbed by a gang of snails. When later questioned by the police, the turtle said: "I don't know. Everything happened so fast!"

  To alleviate the electronic decision making, a system of sequential levels of event selection has evolved. The experimenters program the computers with various "triggers," indicators that tell the system which events to record. For example, a common trigger would be an event that discharges a large amount of energy into the detector, for new phenomena are most likely to occur at high rather than low energies. The setting of triggers is a sweaty-palm business. Make them too loose, and you overwhelm the capability and logic of the recording technology. Set them too tight, and you may miss some new physics, or you may have done the entire experiment for nothing. Some triggers will flip "on" when an energetic electron is detected emerging from the collision. Another trigger will be convinced by the narrowness of a jet of particles, and so on. Typically there are ten to twenty different configurations of collision events that are allowed to set off a trigger. The total number of events passed by these triggers may be 5,000 to 10,000 in a second, but now the event rate is low enough (one every ten-thousandth of a second) to "think" and examine—er, have the computer examine—the candidates more carefully. Do you really want to record this event? The screening goes on
through four or five levels until it gets down to about ten events per second.

  Each of these events is recorded on magnetic tape in full detail. Often, at the stages where we are rejecting events, a sampling of, say, one in a hundred is recorded for future study to determine if important information is being lost. The entire data acquisition system (DAQ) is made possible by an unholy alliance of physicists who think they know what they want, clever electronic engineers who try hard to please, and, oh yes, a revolution in commercial microelectronics based on the semiconductor.

  The geniuses in all of this technology are too numerous to list, but in my subjective view, one of the leading innovators was a shy electronics engineer who functioned in a garret at Columbia's Nevis Lab, where I grew up. William Sippach was way ahead of his physicist controllers. We specified; he designed and built the DAQ. Time and again I would telephone him at three in the morning crying that we'd come up against a serious limitation in his (it was always his when we had trouble) electronics. He would listen quietly and ask a question: "Do you see a microswitch inside the cover plate of rack sixteen? Activate it and your problem will be solved. Good night." Sippach's fame spread, and in a typical week, visitors from New Haven, Palo Alto, Geneva, and Novosibirsk would drop in to talk to Bill.