The God Particle: If the Universe Is the Answer, What Is the Question?

by Leon Lederman

  ***

  The next to last course that Friday was a large carp braised in black bean sauce with scallions and leeks. It was during this serving that Lee reiterated the key information: the effect Wu was observing was large, more than ten times larger than expected. The data were rumored, tentative, and therefore very preliminary but (T. D. served me the fish head, knowing I liked the cheeks) if the effect was that large, it was just what we would expect if neutrinos were two-component ... I lost the rest of his excitement because an idea had started growing in my own mind.

  After lunch there was a seminar, some departmental meetings, a social tea, and a colloquium. In all of these activities I was distracted, bugged by the notion that Wu was seeing a "big effect." From Lee's talk at Brookhaven in August I remembered that the effects produced by the suggested violations of parity when pions and muons decayed were assumed to be minuscule.

  Big effect? I had looked briefly in August at the "pi-mu" (pion-muon) chain of decays and had realized that to design a reasonable experiment one would need to have parity violation in two sequential reactions. I kept recalling the calculations we had done in August before deciding that the experiment was borderline or less in chances of success. However, if the effect was large...

  By 6 P.M. I was in my car heading north to dinner at home in Dobbs Ferry and then to a quiet evening shift with my graduate student at the nearby Nevis Lab in Irvington-on-Hudson. The Nevis 400 MeV accelerator was a workhorse for producing and studying the properties of mesons, relatively new particles in the 1950s. In those happy days, there were very few mesons to worry about, and Nevis worried about pions as well as muons.

  At Nevis we had intense beams of pions coming off a target bombarded by protons. The pions were unstable, and during their flight from the target, out of the accelerator, through the shielding wall, and out into the experimental hall, some 20 percent would undergo the weak decay into a muon and a neutrino.

  π → μ + ν (in flight)

  The muons generally traveled in the same direction as the parent pion. If the parity law was violated, there would be an excess of muons with spin axis aligned in the direction of the motion of the muon over muons with spin axes pointing, say, opposite to the flight. If the effect was large, nature could be providing us with a sample of particles all spinning in the same sense. This is the situation Wu had to organize by cooling cobalt-60 to extremely low temperatures in a magnetic field. The key was to watch those muons whose direction of spin axis was known to decay into an electron plus some neutrinos.

  THE EXPERIMENT

  The heavy traffic on the drive north on the Saw Mill River Parkway on Friday evening tends to obscure the lovely forested hills that line this road, which winds along the Hudson River, past Riverdale, Yonkers, and points north. It was somewhere on this road that the implications of the "big effect" possibility dawned upon me. In the case of a spinning object, if any direction of the spin axis is favored in the decay, that is the effect. A small effect might be 1,030 electrons emitted in one direction relative to the spin axis versus 970 in the other, and this would be very difficult to determine. But a big effect, say 1,500 versus 500, would be much easier to find, and the same fortunate bigness would help in organizing the spins of the muons. To do the experiment, we need a sample of muons all spinning in the same direction. Since they will be moving from the cyclotron to our apparatus, the direction of motion of the muons becomes a reference for the muon spin. We need most of the muons to be right-handed (or left, it doesn't matter), now using the direction of motion as a "thumb." Muons will arrive, pass through a few counters, and stop in a carbon block. Then we count how many electrons are emerging in the direction in which the muons were moving against how many are emerging in the opposite direction. A significant difference would be proof of parity violation. Fame and fortune!

  Suddenly, my usual Friday night calm was destroyed by the thought that we could trivially do the experiment. My graduate student, Marcel Weinrich, had been working on an experiment involving muons. His setup, with simple modifications, could be used to look for a big effect. I reviewed the way muons were created in the Columbia accelerator. In this I was a sort of expert, having worked with John Tinlot on the design of external pion and muon beams some years ago when I was a brash graduate student and the machine was brand new.

  In my mind I visualized the entire process: the accelerator, a 4,000-ton magnet with circular pole pieces about twenty feet in diameter sandwiches a large stainless steel evacuated box, the vacuum chamber. A stream of protons is injected via a tiny tube in the center of the magnet. The protons spiral outward as strong radio-frequency voltages kick them on each turn. Near the end of their spiral trip, the particles have an energy of 400 MeV. Near the edge of the chamber almost at the place where we would run out of magnet, a small rod carrying a piece of graphite waits to be bombarded by the energetic protons. Their 400 million volts is enough to create new particles—pions—as they collide with a carbon nucleus in the graphite target.

  In my mind's eye I could see the pions spewing forward from the momentum of the proton's impact. Born between the poles of the powerful cyclotron magnet, they sweep in a gradual arc toward the outside of the accelerator and do their dance of disappearance; muons appear in their place, sharing the original motion of the pions. The rapidly vanishing magnetic field outside the pole pieces helps to sweep the muons through a channel in a ten-foot-thick concrete shield and into the experimental hall where we are waiting.

  In the experiment Marcel had been setting up, muons would be slowed down in a three-inch-thick filter and then be brought to rest in one-inch-thick blocks of various elements. The muons would lose their energy via gentle collisions with the atoms in the material and, being negative, would finally be captured by the positive nuclei. Since we did not want anything to influence the muons' direction of spin, capture into orbits could be fatal, so we switched to positive muons. What would positively charged muons do? Probably just sit there in the block spinning quietly until their time came to decay. The material of the block would have to be chosen carefully, and carbon seemed appropriate.

  Now comes the key thought of the driver heading north on a Friday in January. If all (or almost all) of the muons, born in the decay of pions, could somehow have their spins aligned in the same direction, it would mean that parity is violated in the pion-to-muon reaction and violated strongly. A big effect! Now suppose the axis of spin remained parallel to the direction of motion of the muon as it swept through its graceful arc to the outside of the machine, through the channel. (If g is close to 2, this is exactly what happens.) Suppose further that the innumerable gentle collisions with carbon atoms, which gradually slowed the muon, did not disturb this relationship of spin and direction. If all this were indeed to happen, mirabile dictu! I would have a sample of muons coming to rest in a block all spinning in the same direction!

  The muon's lifetime of two microseconds was convenient. Our experiment was already set up to detect the electrons that emerge from the decaying muons. We could try to see if equal numbers of electrons emerged in the two directions defined by the spin axis. The mirror symmetry test. If the numbers are not equal, parity is dead! And I killed it! Arggghh!

  It looked as if a confluence of miracles would be needed for a successful experiment. Indeed, it was just this sequence that had discouraged us in August when Lee and Yang read their paper, which implied small effects. One small effect can be overcome with patience, but two sequential small effects—say, one percent of one percent—would make the experiment hopeless. Why two sequential small effects? Remember, nature has to provide pions that decay into muons, mostly spinning in the same sense (miracle number one). And the muons have to decay into electrons with an observable asymmetry relative to the muon spin axis (miracle number two).

  By the Yonkers toll booth (1957, toll five cents) I was quite excited. I felt pretty sure that if the parity violation was large, the muons would be polarized (spins all pointing in th
e same direction). I also knew that the magnetic properties of the muon's spin were such as to "clamp" the spin in the direction of the particle's motion under the influence of the magnetic field. I was less certain of what happens when the muon enters the energy-absorbing graphite. If I was wrong, the muon spin axis would be twisted in a wide assortment of directions. If that happened there would be no way to observe the emission of electrons relative to the spin axis.

  Let's go over that again. The decay of pions generates muons that spin in the direction in which they are moving. This is part of the miracle. Now we have to stop the muons so we can observe the direction of the electrons they emit upon decay. Since we know the direction of motion just before they hit the block of carbon, if nothing screws them up we know the spin direction when they stop and when they decay. Now all we have to do is rotate our electron detection arm about the block where the muons are at rest to check for mirror symmetry.

  My palms started to sweat as I reviewed what we had to do. The counters all existed. The electronics that signaled the arrival of the high-energy muon and the entrance into the graphite block of the now slowed muon were already in place and well tested. A "telescope" of four counters for detecting the electron that emerged after muon decay also existed. All we had to do was mount these on a board of some sort that we could pivot around the center of the stopping block. One or two hours' work. Wow! I decided that it would be a long night.

  When I stopped at home for a quick dinner and some bantering with the kids, a telephone call came from Richard Garwin, a physicist with IBM. Garwin was doing research in atomic processes at the IBM research labs, which were then just off the Columbia campus. Dick hung around the Physics Department a lot, but he had missed the Chinese lunch and wanted to know the latest on Wu's experiment.

  "Hey, Dick, I've got a great idea on how we can test for parity violation in the simplest way you can imagine." I explained hastily and said, "Why don't you drive over to the lab and give us a hand?" Dick lived nearby in Scarsdale. By 8 P.M. we were disassembling the apparatus of one very confused and upset graduate student. Marcel saw his Ph.D. thesis experiment being taken apart! Dick was assigned the job of thinking through the problem of rotating the electron telescope so we could determine the distribution of electrons around the assumed spin axis. This wasn't a trivial problem, since wrestling the telescope around could change the distance to the muons and thus alter the yield of detected electrons.

  It was then that the second key idea was invented, by Dick Garwin. Look, he said, instead of moving this heavy platform of counters around, let's leave it in place and turn the muons in a magnet. I gasped as the simplicity and elegance of the idea penetrated. Of course! A spinning charged particle is a tiny magnet and will turn like a compass needle in a magnetic field, except that the mechanical forces acting on the muon-magnet make it rotate continuously. The idea was so simple it was profound.

  It was a piece of cake to calculate the value of the magnetic field needed to turn the muons through 360 degrees in a reasonable time. What is a reasonable time to a muon? Well, the muons are decaying into electrons and neutrinos with a half-life of 1.5 microseconds. That is, half of the muons have given their all in 1.5 microseconds. If we turned the muons too slowly, say 1 degree per microsecond, most of the muons would have disappeared after being rotated through a few degrees and we wouldn't be able to compare the zero-degree and 180-degree yield—that is, the number of electrons emitted from the "top" of the muon as opposed to the "bottom," the whole point of our experiment. If we increased the turning rate to, say, 1,000 degrees per microsecond by applying a strong magnetic field, the distribution would whiz past the detector so fast we would have a blurred-out result. We decided that the ideal rate of turning would be about 45 degrees per microsecond.

  We were able to obtain the required magnetic field by winding a few hundred turns of copper wire on a cylinder and running a current of a few amperes through the wire. We found a Lucite tube, sent Marcel to the stockroom for wire, cut the graphite stopping block down so it could be wedged inside the cylinder, and hooked the wires to a power supply that could be controlled remotely (there was one on the shelf). In a blur of late-night activity, we had everything ready by midnight. We were in a hurry because the accelerator was always turned off at 8 A.M. Saturday for maintenance and repairs.

  By 1 A.M. the counters were recording data; accumulation registers recorded the number of electrons emitted at various directions. But remember with Garwin's scheme, we didn't measure these angles directly. The electron telescope remained stationary while the muons or, rather, their spin axis vectors, were rotated in a magnetic field. So the electrons' time of arrival now corresponded to their direction. By recording the rime, we were recording the direction. Of course, we had lots of problems. We badgered the accelerator operators to give us as many protons hitting the target as possible. All the counters registering the muons coming in and stopping had to be adjusted. The control of the small magnetic field applied to the muons had to be checked.

  After a few hours of data taking, we saw a remarkable difference in the counts of electrons emitted at zero degrees and those emitted at 180 degrees relative to the spin. The data were very crude, and we mixed excited optimism with skepticism. When we examined the data at eight the next morning, our skepticism was confirmed. The data now were much less convincing, not really inconsistent with the hypothesis that all directions of emission were equivalent—a predictor of mirror symmetry. We had pleaded with the accelerator operators to give us an additional four hours, but to no avail. Schedules are schedules. Discouraged, we walked down to the accelerator room, where the apparatus was set up. There we were greeted by a small catastrophe. The Lucite cylinder on which we had wound the wire had become warped due to the heat produced by the current in the wires. This warping had permitted the stopping block to fall. Obviously, the muons were no longer in the magnetic field we had designed for them. After some recriminations (blame the graduate student!) we cheered up. Our original impression might still be correct!

  We made a plan for the weekend. Design a proper magnetic field. Think about increasing the data rate by increasing the number of muons stopping and the fraction of the decay electrons counted. Think about what happens to the positively charged muons in their collisions on the way down to rest and in the microseconds in which they sit in the lattice of carbon atoms. After all, if a positive muon managed to capture one of the many electrons that are free to move about in graphite, the electron could easily depolarize (mess up the spin of) the muon so that they would not all be doing the same thing in lockstep.

  The three of us went home to sleep for a few hours before reassembling at 2 P.M. We worked through the weekend, each at an assigned task. I managed to recalculate the motion of the muon from birth as it is kicked forward by its decaying pion parent, through its sweep toward the channel and through the concrete wall into our apparatus. I kept track of spin and direction. I assumed maximum violation of mirror symmetry so that all the muons would be spinning precisely along in the direction of their motion. Everything indicated that if the violation was large, even half of maximum, we should see an oscillating curve. This not only would prove parity violation but would give us a numerical result as to how much parity was violated, from 100 percent down to (no! no!) zero. Anyone who tells you that scientists are dispassionate and coldly objective is crazy. We desperately yearned to see parity violated. Parity was not a young lady, and we weren't teenagers, but we lusted to make a discovery. The test of scientific objectivity is not to let the passion influence the methodology and the self-criticism.

  Eschewing the Lucite cylinder Garwin wound a coil directly on a new piece of graphite and tested the system at currents twice as high as we would need. Marcel rearranged the counters, improved the alignment, moved the electron telescope closer to the stopping block, tested, and improved the efficiency of all counters, all the while praying that something publishable would come out of this frantic activity
.

  The work went slowly. By Monday morning, some news of our intense activity had leaked out to the operator crew and to some of our colleagues. The accelerator maintenance gang found some serious problems in the machine, so Monday was out—no beam until Tuesday, 8 A.M. at best. Okay, more time to fume, fuss, check. Colleagues from the Columbia campus arrived at Nevis, curious as to what we were up to. One clever young man who had been at the Chinese lunch asked a few questions and, by my disingenuous answers, deduced that we were trying the parity experiment.

  "It'll never work," he assured me. "The muons will depolarize as they lose energy in the graphite filter." I was easily depressed but not discouraged. I remembered my mentor, the great Columbia savant 1.1. Rabi, telling us: spin is a very slippery thing.

  About 6 P.M. on Monday, ahead of schedule, the machine began to show signs of life. We hastened our preparations, checking all the devices and arrangements. I noticed that the target with its elegant copper wire wrapping, positioned on a four-inch slab, looked a bit low. Some squinting through a surveying scope convinced me, and I looked for something that would raise it an inch or so. Over in the corner I saw a Maxwell House coffee can partially filled with wood screws, and I substituted it for the four-inch slab. Perfect! (When the Smithsonian Institution later wanted the coffee can in order to replicate the experiment, we couldn't find it.)