I mentioned, as I had in my talk, that the CMS results had shown two bumps, not just the one close to the bump found at the ATLAS detector. Monika smiled and said that the CMS group had more analysts working on the data than the ATLAS group, of which she was part, and that they had devoted more time to their analysis, which could explain why there was more structure in their data. I also questioned her about the fact that, in the ATLAS data, the excess of events, or signal strength, for the Higgs boson decaying into two photons was tiny, and would be hidden inside the background data, whereas the two-bin excess of events in the CMS for the same channel was significantly larger than what the standard model predicted. She said yes, this was a troublesome issue, possibly suggesting that the excess of events was a statistical artifact or fluctuation. The three of us stared at the plot on her laptop, and Philip nodded in agreement; there was something unsatisfactory about all this.
At a later coffee break, I got hold of David Cline again and told him what the ATLAS analyst had said, particularly about the size of the predicted bump for the decay of the Higgs into two photons compared with the size of the observed CMS excess in the data at about 126 GeV. David said that he had talked to a colleague who suggested that the branching ratio for the Higgs decay into two photons should be multiplied by a factor of 10 or more—that is, that the predicted signal strength in the standard model could be wrong by this amount. I said I considered this highly unlikely. Moreover, I pointed out, as in my talk, that the bumps in the CMS and the ATLAS results did not coincide in energy, but differed by as much as 2 to 3 GeV. Considering that the width of the Higgs bump, if it was real, would be less than 1 GeV, then this difference of at least 2 GeV in the positions of the bumps could be significant, and one should be careful about claiming that they are at the same energy. He admitted that the results could be due to statistical artifacts or fluctuations in the data. Indeed, he said that if the experiment provided just a smooth curve without fluctuations, then one could not believe the results, because fluctuations always occur in statistical analyses of data.
Regarding whether there was financial pressure from CERN to announce a discovery, David said that the younger generation of physicists at CERN was naive and did not understand the politics of the situation. He said that CERN had to guarantee that the funding of the machine would continue, particularly in view of the current financial crisis in Europe. Therefore, it was important to make some kind of announcement at the press conference to ensure that funding would continue for the next few years.
At this interesting point in our discussion, we were joined by Lars Brink and Françoise Englert. After filling them in on the main points, I said that I certainly agreed with David on the need for the CERN machine to continue to be funded. It would be a disaster for physics if they cut the funding as the United States had with the supercollider in Texas during the 1990s. However, we must not lose our scientific integrity. I was not willing to compromise and make claims that could eventually turn out to be invalid. Brink and Englert listened impassively.
Later, at another coffee break, I spoke to the ATLAS analyst Monika Wielers again and mentioned that David Cline had suggested that the theoretical calculation of the size of the bump in the Higgs boson into the two-photon decay channel could be increased by a factor of 10 or more. She huffily dismissed this, saying it simply could not be correct. The error in the theoretical calculation could only be about 15 percent, she said, and one could not entertain such a large error. I said it was anticipated that next year, in 2012, the integrated luminosity for the LHC proton–proton collisions would increase to maybe 10 inverse femtobarns, and this could definitely verify or rule out the Higgs boson.4 She said she was not convinced by this. She felt that they might have to go to 20 or 30 inverse femtobarns before they could reach the magic 5-sigma statistical significance for a bump, which is necessary to claim a real discovery of a particle. Such a statistical significance would amount to saying that there was only a chance of one part in about two million of the signal being a random fluctuation. The statistical significance of the recent LHC results was between 2 sigma and 3 sigma, and therefore could not meet any reasonable statistical requirement of claiming that it was a real signal of the Higgs boson. It was a hint, nothing more.5
THE EXPERIMENTALISTS HAVE THEIR SAY
Later on the Saturday of the conference, it was the experimentalists’ turn to present their data. Yasar Onel, one of the experimentalists in the CMS group, reported their results. First he discussed the CMS results that appeared to exclude supersymmetric particles, extra dimensions, and other exotica such as mini black holes. Next, Onel presented the CMS results for the Higgs boson search. He went through the data for each decay channel of the Higgs, and finally singled out the decay of the Higgs into two photons. These data produced two bumps, one at 123.5 GeV and one at 137 GeV, and there were also hints of a possible third bump at 119 GeV. The ATLAS data had only revealed one bump in this energy range. I remembered that Monika had said that the number of experimentalists analyzing the CMS data was larger than the number analyzing the ATLAS data and that the CMS analysts were taking longer to analyze the data, and were analyzing it in greater detail. She had suggested that this was the reason that more structure was seen in the CMS data compared with the ATLAS data.
Onel continued that it was important to note that the energies of the bumps in ATLAS and CMS did not coincide or overlap within 2 sigma and therefore it could not be claimed that both occurred at the same energy, leaving open the possibility that the bumps were artifacts or statistical fluctuations. Another important issue, he said, was that the bump in the CMS data at around 123 to 124 GeV had a statistical significance of 1.9 sigma after the look-elsewhere effect was taken into account. This again suggested that far more accurate data were needed to claim that the CMS bumps were a real signal of a discovered particle.
Monika Wielers then presented the most recent results from the ATLAS detector at the LHC. She described how there have been significant exclusions of superpartners in supersymmetry models. Indeed, large swaths of the parameter space of the simplest supersymmetry model of particle physics, called the constrained minimal supersymmetric model, had been excluded up to an energy of almost 1 TeV. Moreover, there were new exclusions of extra dimensions of space predicted by various higher dimensional theories such as string theory up to 2 to 3 TeV. Thus, both detectors were, in effect, knocking the ground out from under hundreds of theoretical physicists’ feet, in saying that they had not found supersymmetric particles or the higher dimensions of string theory or mini black holes.
Last, Monika described the results of the search for the Higgs boson at the ATLAS detector. Like Onel, she presented the data for the various possible decay channels of the Higgs boson. The two golden channels—namely, the Higgs decaying into ZZ and into two photons—were of particular significance because the decay of the Higgs into a more dominant channel, WW, had serious background problems and the number of data events was not significant. The most significant data was the bump in the diphoton decay channel, which had a statistical significance, or sigma, of about 2.3 standard deviations occurring at an energy of 126 GeV. She showed the important plot of the expected probability of the decay of the Higgs into two photons versus the horizontal axis depicting the Higgs mass—namely, the mass associated with the two photons (see Figure 6.3).
Figure 6.3 Number of events per billions of electron volts for the Higgs decay into two photons. © CERN for the benefit of the ATLAS Collaboration
Two issues of importance appeared from this plot. There was, indeed, a two-bin excess of data events at about 126 GeV that could be the signal of a particle. However, at 120 GeV, there was a little pimple drawn in dashed red depicting the probability of the Higgs boson appearing at that energy level as calculated from the standard Weinberg–Salam model. It was clear from Monika’s graphs that, as I had explained in my talk, the observed two-bin events at 126 GeV were much in excess of what could be expected of a standard
-model Higgs boson, as shown in the pimple at 120 GeV. I raised my hand and asked her, “Could this not indicate that the excess seen is a statistical artifact, such as a fluctuation in the data?”
Monika walked up close to where I was sitting near the front of the auditorium, smiled, and said, “Yes, I agree with that.”
I then suggested that the fact that it was only a slightly more than 2-sigma effect, and that the LEE would reduce its statistical significance to 1.9 sigma, meant that the existence of a Higgs boson could not be confirmed at 126 GeV. I suddenly said in exasperation, “Why is it that this long-sought-after particle, after 40 years, insists on hiding itself in the most difficult place for the accelerators to discover it!”
Monika laughed and said, “I don’t know the answer to that question.” Returning to address the audience, she then stated that it was too early to draw any conclusions about whether they had seen the Higgs boson in the ATLAS data. Perhaps next year, with a larger luminosity in the proton–proton collisions, and more data, we would be able to decide whether the bump at 126 GeV would increase in significance or just go away.
I pointed out, too, in the discussions after Monika’s presentation, that at the July conference in Grenoble, the experimentalists who presented the ATLAS data up to two inverse femtobarns showed a bump with a statistical significance of 2.8 sigma at about 140 GeV. However, this statistical significance was reduced by the LEE. I suggested that the conclusion to all of this next year might be a repeat of last July’s ATLAS story. On the other hand, if more data increased the significance of the bump, then we could begin to anticipate that a new particle had been discovered. Time would tell.
Although the Tevatron collider had shut down officially in September 2011, the experimental group was still busy analyzing its data on the search for the Higgs boson and other exotic particles. Two speakers presented the Tevatron data. The first, Todd Adams, concentrated on the results for tests to observe charge-parity (CP) violation—namely, the violation of charge and parity symmetries in bottom quark decays. The standard model did not predict any violation of charge and parity symmetries. The charge quantum number of the quark results from replacing the quark with an antiquark, whereas parity is seeing a particle as an exact mirror image. The CP symmetry consists of making a charge conjugation transformation on a quark (i.e., replacing it with an antiquark) followed by a parity transformation. Adams left the discussion of the Tevatron Higgs search to the next speaker, Florencia Canelli.
Canelli showed the results for the Higgs search by describing the data from different decay channels, such as the decays into WW, ZZ, and two photons. Nowhere was there a significant excess of events constituting bumps, as had been seen in the CMS and ATLAS data for the two-photon decay channel. I raised my hand and said that, looking at the data, there seemed to be no excess of events above one standard deviation, particularly in the critical region of 115 to 145 GeV. Moreover, this was particularly true of the crucial energy range between 120 GeV and 130 GeV, in which the putative signal of a Higgs boson was claimed by CMS and ATLAS. She nodded in agreement, and said that she didn’t believe that the CERN results were conclusive. This raised the ire of the CMS representative, Yasar Onel. There was a sharp exchange between the two, with Onel directing some of his heated comments at me, sitting several rows in front of him.
I then asked the question: “When can we expect to combine all three experiments—namely, the ATLAS, CMS, and Tevatron results?” It seemed to me that this was necessary to reach a final decision on the existence of the Higgs boson, because at this point the “score” was CMS two or three bumps; ATLAS, one; and Tevatron, zero in the critical energy range of 115 to 145 GeV. “Can we expect that the continued analysis of the Tevatron data could produce stronger results?” I asked.
Canelli said that they were indeed continuing the analysis, and she hoped that finally, when all the data were analyzed, they would have a strong enough result to be combined with the data from the two LHC detectors.
HEDGING BETS ON THE HIGGS BOSON
As a member of the audience, I was impressed by the integrity and show of caution on the part of the speakers presenting the LHC and Tevatron data. Even though so much was riding on them finding the Higgs boson—fame, fortune, Nobel Prizes, and, in the case of the LHC, a guarantee of continued funding—the speakers did not exaggerate the possibility of having discovered the long-sought Higgs boson.
In view of the preliminary presentation of the data at the CERN meeting on December 13, I anticipated that there would be a rash of theoretical papers claiming to have predicted a Higgs boson at about 126 GeV. Indeed, among what turned out to be many such papers, one on the electronic archive by physicists at Michigan University claimed that a complicated model based on string theory and supersymmetry predicted a Higgs boson at about 125 GeV.
During the past three decades, there has been quite a history of “observed” new particles; research groups at colliders would report the discovery of a new particle, and subsequent data would eventually show that these new particles did not exist after all. A striking example comes from CERN. Shortly after Carlo Rubbia discovered the W and Z particles in 1983, he announced that he had detected the top quark at 40 GeV. Soon after that, he also announced the discovery of a supersymmetric particle. Further data analysis eventually proved that these two “discoveries” were incorrect, and the evidence for these particles faded away. Much later, the top quark was truly discovered at Fermilab at an energy of 173 GeV. No supersymmetric particles have yet been discovered.
In another instance, in 1984, the experimental group working with the Crystal Ball experiment at SLAC announced the discovery of the Higgs boson at 1.5 GeV. Indeed, ignoring the difference in energy levels, the SLAC plot of the number of events for the decay of the Higgs into two photons looks uncannily similar to the equivalent plots for CMS and ATLAS results presented at the CERN meeting on December 13, 2011. Subsequent additional data and analysis made this 1984 “Higgs discovery” disappear.
A more recent famous example of false discoveries is the finding of a new particle called the pentaquark in experiments performed at accelerator laboratories in the mid-2000s. The pentaquark supposedly consists of four quarks and an antiquark bound together—a much different particle than the three quarks making up the proton and neutron. Viewing the plots that were published in the literature between 2003 and 2005, we see a significant excess of events for the pentaquark, reaching almost 5 sigma—the gold-plated standard for the confirmation of new particles. Yet, further data and more detailed analysis again made the new particle disappear. The claimed discovery of the pentaquark resonance was greeted with skepticism, in any case, by the theoretical physicists working on quark models, because the particle did not fit in with the standard, nonrelativistic quark model.
With the advent of more data in 2012, and after the startup of the higher energy LHC in 2015, the bumps in the ATLAS and CMS experiments might similarly disappear. If so, the Higgs boson will be ruled out as an elementary particle up to an energy so high that the standard Weinberg–Salam model would no longer remain viable. In that case, one would expect a surge of theoretical papers claiming, on the basis of one model or another, that the authors had predicted this outcome all along, and we would then enter a long period of theoretical explanations as to why the Higgs boson had not been seen.
Indeed, prior to the December 13, 2011, CERN presentation, papers had already appeared on the electronic archive hedging their bets, claiming that the Higgs boson could be made invisible through various mechanisms, such as the Higgs only decaying into dark-matter particles. It is a truism in physics that theoretical physicists are allowed to speculate and make mistakes in their prognoses of physics. However, experimentalists cannot make mistakes in their claims of discoveries. False claims can seriously damage reputations and careers.
7
Trying to Identify the 125-GeV Bump
On July 4, 2012, CERN announced that the CMS and ATLAS groups had di
scovered a Higgs-like boson at 125 to 126 GeV. This could turn out to be the standard-model Higgs boson, in which case the standard model of elementary particles will have been completed.
However, if the new boson turns out not to be the standard-model Higgs boson, an alternative to the standard electroweak model must be found. There are already several competing models, and we will enter a phase with the experimental program at the LHC when physicists will be attempting to decide which possible alternative is most viable. One model, my quarkonium resonance model, is based only on the already observed particles in the standard model—namely, the 12 quarks and leptons, the two W bosons, the Z boson, the photon, and the gluon. This model identifies the new boson as a spin-0 resonance composed of a quark and an antiquark.
Another scenario, if the new particle turns out not to be the expected Higgs boson, is that when the energy of the LHC is increased to its maximum, 14 TeV, and the intensity of the proton–proton collisions is increased significantly, then perhaps new particles, including the Higgs, will still be discovered at these higher energies. However, in this case, because of the fundamental problems associated with the Higgs boson, such as the Higgs mass hierarchy problem, new physics beyond the standard model will still have to be found to ameliorate the difficulties with the Higgs boson.
INTERPRETING A BUMP
Because the Higgs boson is so short-lived, we cannot detect it “directly” at the LHC. Its presence can only be inferred by observing the lower-energy particles that it can decay into—the so-called decay product channels. Detecting a Higgs–like signal within the large backgrounds is like the proverbial search for a needle in a haystack.
Cracking the Particle Code of the Universe Page 15