Shifting Standards: Experiments in Particle Physics in the Twentieth Century

Conclusion

• 221 It is clear that there have been significant changes in both experiments and in the reporting of experimental results during the twentieth century . I caution, however, that any conclusions drawn apply only to experiments on elementary particles and their properties and to experiments in high-energy physics, which were the only kinds of experiments discussed.1 Although there have also been significant changes in experiments in other fields of physics, such as condensed matter physics, plasma physics, and atomic, molecular, and optical physics, I have restricted my study to elementary particle physics because issues such as scale and statistics are more apparent in those areas and also because it is the field of physics I know best. Scale I began working in experimental high-energy physics in the early 1960s and the change since that time that most impresses me is that of scale. I have already noted that the size of the experimental apparatus has increased from the approximately 1 m3 oil-drop apparatus used by Robert Millikan to the 4,000 m3 Compact Muon Solenoid (CMS) at the Large Hadron Collider (LHC). This does not include the size of the accelerators needed to perform the experiments. Initially, in particle physics experiments experimenters used X-rays or radioactive sources, which could easily fit on a tabletop, or waited for events caused by cosmic rays. In the 1930s we saw the advent of particle accelerators, which not only allowed the control of the particle beam but also allowed higher energies and beam intensities . E. O. Lawrence’s first cyclotron, built in 1931, had a diameter of about 11.5 centimeters, whereas the LHC has a circumference of 27 kilometers (an approximate diameter of 8.6 kilometers), an increase in size by a factor of almost 75,000.2 The increase in energy has been even larger. Lawrence’s Conclusion 222 • Conclusion first cyclotron had an energy of 80 keV, although Lawrence’s 11-inch cyclotron , as well as the Cockcroft-Walton accelerator and the Van de Graaff generator, soon reached an energy of 106 eV (1 MeV). The LHC is currently running with two beams, each with an energy of 3.5 TeV (3.5 × 1012 eV), or a total energy of 7 TeV, an increase in energy by a factor of 7 million.3, 4 The number of events in an experiment has also increased considerably (table C.1). Millikan (see chapter 3) took data on 175 oil drops, of which he published 58, although he used only 23 for his determination of e, the charge of the electron. By 2009, the BaBar collaboration had an experimental run in which they observed 467 × 106 B meson pairs (see chapter 18). The CMS collaboration recently reported 6.5 million K0 S mesons. Consider the evidence for CP violation.5 In 1964 James Cronin, Val Fitch, and their collaborators reported a total of 45 ± 9 events above background that showed K0 L decay into two charged pions, which was forbidden by CP conservation (Christenson et al. 1964), a discovery for which Cronin and Fitch later won the Nobel Prize. The 1999 KTeV experiment (Alavi-Harati et al. 1999) found a net yield of 2,607,274 examples of K0 L decay into two pions.6 There have also been significant changes in data analysis. As we have seen, data analysis is essential in producing an experimental result. A result is not given by merely perusing the data. Robert Millikan made all of his timing measurements by hand, observing the rise and fall of the oil drops and recording the values in a laboratory notebook. The most advanced computing technique he used was a table of logarithms. I have already discussed the fact that in my own 1965 dissertation experiment we projected the photographs of the optical spark chambers onto graph paper, wrote down the numbers, recorded them on punched cards, and then ran them through a computer. In the 1967 Ke2 + experiment (see chapter 10), the positions of the sparks were recorded by human scanners using machine digitizers that then recorded the digitization on magnetic tapes, which were then read into a computer for analysis. Obviously such techniques cannot be used when one is processing tens of millions, or even Table C.1. Number of events in experiments Author Number of Events Millikan (1913) 175 Leighton (1953) 74 Ke2 + (1967) 16,965 FOCUS (1996–97) 20 × 109 triggers 6.5 × 109 photon events BaBar (2009) 467 × 106 BB̄ pairs CMS...