Last weekend I joined Fran Day, Tessa Baker, Ken Peach and Andrei Seryi on a discussion panel on the topic of “Mistakes in Physics” as part of the APPEAL teachers’ conference in Oxford. I was happy to do this, confident that this was one area where I had significant experience to contribute. After thinking this, I did pause to wonder why they had chosen to ask me.
What has been the biggest recent mistake in physics? turned out to be a very interesting question to ponder. First of all you have choose how to interpret it—what actually counts as a mistake. There are of course some obvious case studies such as the non-discovery of cold fusion by Fleischman and Pons in the 1980s, yet there are also things like theories, such as the steady state universe, which turned out not to lead anywhere. Do these count as mistakes, when they were perfectly legitimate avenues to explore in their time, and until they were investigated who was to know? Physics would be a much more boring subject if we never made any mistakes. Of course we don’t know what mistakes we are making now. Popular ideas like dark matter and supersymmetry may, in due course, turn out to lead nowhere.
Experimental particle physics has seen some quite spectacular blunders. In 2008 the Large Hadron Collider was turned on, with the world’s media watching closely. Nine days later a manufacturing fault caused a superconducting wire in a magnet assembly to become non-superconducting. As 13,000 amps of current suddenly encountered resistance, the temperature shot up, an electric arc damaged the liquid helium enclosure, and as two tonnes of liquid helium became non-liquid the resulting explosion damaged 53 magnets and delayed the project for another year while this was repaired. Oops.
Yet, nine years on, we can look back on this as a minor set-back as the LHC recovered, and went on to collide proton beams at energies up to 13TeV and discover the Higgs Boson in 2012.
Another recent ‘mistake’ in our field would be the apparent observation in 2011 of neutrinos travelling faster than the speed of light. A result which no physicist dared believe, but which would have been so so cool if it had been true that we avidly followed every detail as the story unfolded. It must be stressed that the OPERA collaboration never claimed to believe it, and reported the anomaly in the hope that the community could help identify the origin. It was just a bit embarrassing for them that it turned not to be a subtle technical effect involving general relativity or other advanced physics, but a faulty electronic module and optical fibre not properly plugged in.
In my career, the most newsworthy mistake I have been involved with was the non-detection of a dark matter signal. I was a member of the CRESST experiment, based at the same underground laboratory as the OPERA neutrino detector, which searches for signs of the hypothetical weakly interacting massive particles believed to make up most of the mass of the galaxy. If this turned out to be real it would have been big. The mystery dark matter—what makes up the missing mass of the Universe—is perhaps the biggest unsolved mystery in science. The experiment was a super sensitive underground particle detector, with sophisticated techniques to veto any radioactive background.
In 2011 the project briefly made the news with an exciting claim. They had seen an excess of events, more than the predicted background – so dark matter?
Having worked on the project, I knew many of the things that could go wrong. The Oxford group thought this was too strong a claim. We just didn’t understand all the background processes well enough to be sure that it was a dark matter signal. So, after a lot of discussion with our international collaborators, who decided the all-important phrasing of the claim, in the end we declined to be authors on that paper.
It turned out that what was happening was that the surface of the crystal detectors was contaminated – probably due to exposure to radon gas. When a radioactive atom on the surface decayed, it could sometimes kick an alpha particle away from the crystal, which was not seen, and the atomic nucleus would recoil into the detector and look exactly like it had been hit by a dark matter particle. This was a known problem. The collaboration thought they could account for it but it turned out to be more complicated than first thought. A year or two later the experiment released some more data, which did not show this dark matter signal.
CRESST was not the first experiment to report a false dark matter signal. It is unlikely to be the last. This episode highlighted the need to take a sceptical attitude when analysing the data. When you see a signal that you’ve been searching for years, it is tempting to cry “Yes!” straight away, when you should really carefully consider all the other effects which could mimic such a signal.
Thanks to Fran Day and the organisers of the APPEAL event for giving me such an interesting topic to write about.