Pontecorvo and neutrino physics

In connection with the centenary of Bruno Pontecorvo’s birth, CERN Courier (September 2013 p78) published an article "Pontecorvo and neutrino physics" by Vadim Bednyakov, a collaborator of Pontecorvo’s in his later years.

Pontecorvo, given his remarkable imagination, concentration and insight, made several important contributions to progress in particle physics in the 20th century. However, in his later years, he also imagined some contributions that he did not make, and this involves, to some extent, some of my own work. In addition, the article largely credits Pontecorvo with the foundation of our understanding of neutrino oscillations. I have tried to understand this history and will make some comments. Much credit is given to Pontecorvo for the theoretical prognosis of neutrino oscillations, but as best I can determine, Pontecorvo’s contributions here are much less clear than generally suggested (Pontecorvo 1957, 1959, 1967). I believe that these questions are of interest for understanding the history of the physics involving neutrinos, so as to justify this review.

Universality of the Fermi interaction. I came to know Pontecorvo in 1947 and 1948, at the time when he was a physicist at the nuclear reactor at Chalk River, Canada, and when he came, I think twice, to Chicago to see Fermi, who had been his teacher in Rome and who was then my thesis adviser. One or two years later, Pontecorvo emigrated to the Soviet Union. At the time of his visits, he was involved at Chalk River with E P Hincks, perhaps part time, in a cosmic-ray experiment on stopped "mesotrons" (now known as muons) using Geiger counters. This had similarities to my thesis experiment, which showed that the energy spectrum of the electrons in mesotron decay is continuous (Steinberger 1949). This turned out to be a much more interesting result than Fermi and I realized. Immediately three papers, one by T D Lee, M Rosenbluth and C N Yang, another by J Tiomno and A Wheeler, and the third by G Puppi, pointed out that this showed that, in addition to the Fermi nucleon and electron-neutrino currents, there was a third current, a muon-neutrino current, of equal amplitude. At the time this was called the "Puppi triangle". It demonstrated the universality of the Fermi interaction. The Hincks–Pontecorvo experiments were much smaller and the geometries quite different (Hincks and Pontecorvo 1947, 1948, 1950). Their results did not permit conclusions about the continuity of the decay-electron spectrum, and at the time no such claim was made. Nevertheless, in his later years, Pontecorvo claimed that these experiments permitted the conclusion that the muon-decay spectrum is continuous, as repeated in the CERN Courier article.

It is interesting to recall that Pontecorvo had anticipated the universality of the Fermi interaction already in 1947, in a very different way. Following the publication in 1946 by Conversi, Pancini and Piccioni of the experiment that showed that the nuclear interaction of the mesotron is too weak for it to be the particle proposed by Yukawa as responsible for nuclear forces, in a letter to Physical Review Pontecorvo noted that the result of that experiment – namely that negative mesons stopped in carbon absorbers decayed normally, whereas those stopped in iron absorbers are captured by the nucleus before their decay, within the uncertainty factor (ZFe/ZC)2, that is, about 20 – showed that the nuclear interaction amplitude of the mesotron is the same as that of the Fermi interaction in β decay, and that therefore there might be a universal Fermi interaction (Pontecorvo 1947). This, now clearly correct, idea was universally rejected by the physics community at the time. The notion of a parallel between the electron, the particle in atoms, and the mesotron of cosmic rays was just too imaginative. Fermi clearly rejected it. A year later, after my thesis experiment, the universality of the Fermi interaction was accepted by everyone, but Pontecorvo’s 1947 letter is not remembered. It might be of interest to note that probably Pontecorvo himself was not convinced of the correctness of the idea, or he would have proposed the obvious way to test it – that is by measuring the capture rates in elements with atomic number between those of carbon and iron.

Two neutrinos. I am personally indebted to Pontecorvo for proposing, in 1959, to check experimentally if the neutrinos associated with muons in pion and kaon decay are the same, or not, as those in β decay, and that the higher energy accelerators, then under construction at Brookhaven and CERN, would permit neutrino beams of energy high enough to allow such an experiment (Pontecorvo 1959) – the experiment for which M Schwartz, L Lederman and I later shared the Nobel prize (Danby et al. 1962). Independently, Schwartz had proposed that neutrino beams would permit the study of weak interactions at higher energy, but he did not consider the particular question of the possible inequality of the two neutrinos, proposed by Pontecorvo (Schwartz 1960).

Pontecorvo and neutrino oscillations. Pontecorvo’s work on what he called neutrino "oscillations" dates back to the 1950s and 1960s, but the processes that he then referred to as oscillations, such as neutrinos changing to (μ+e), or (μe+), or νμ → νe, or ν → ν cannot be related to neutrino oscillations as we now observe and understand them (Pontecorvo 1957a, 1957b, 1967, Gribov and Pontecorvo 1969). Neutrino oscillations were discovered, serendipitously, by M Koshiba and collaborators, in the 1990s, in the Kamiokande and later Super-Kamiokande detectors, which had been built to look for proton decay (Nakahata et al. 1986, Hirata et al. 1988, Fukuda et al. 1994, Fukuda et al. 1998). The 1969 paper by Gribov and Pontecorvo is, to my knowledge, the first work that imagined mass–flavour mixing, which, as we now know, is the mechanism of neutrino oscillations. At the time (1969), it was an imaginative, mathematical invention, without observational motivation, and was ignored. The actual neutrino oscillation mechanism, that is, the mass–flavour eigenstate mixing, was understood in the mid-1980s, with the important help of the work of Eliezer and Ross (Eliezer and Ross 1974).

Jack Steinberger, CERN.

Neutrinoless double beta decay

In the November 2013 issue of CERN Courier, p42, a letter by H V Klapdor-Kleingrothaus puts forward the question about progress in the field of neutrinoless beta decay (0νββ). In contrast to his doubts, the GERDA collaboration would like to answer this question positively.

There has been tremendous progress in the field since his own last measurement – much beyond "some fresh breeze" – first by the number of experiments investigating different isotopes and different methods, and second by experimental measures to improve the background compared to the precursor 76Ge experiments, Heidelberg-Moscow (HdM) and IGEX. Moreover, it is surprising to see that the letter quotes conference contributions of GERDA only, but not our refereed articles.

The GERDA design follows the idea of Gerd Heusser (1995 Ann. Rev. Nucl. Part. Sci. 45 543). This idea was tested with liquid nitrogen in the GENIUS test facility without success by Klapdor-Kleingrothaus (2008 Int. J. Mod. Phys. E 17 505). The GERDA collaboration refurbished the same germanium crystals and achieved a stable operation in liquid argon over a programmed period of 1.5 years. With this technique, the background was reduced by about one order of magnitude with respect to the historical HdM and IGEX experiments.

The GERDA collaboration is well aware of the 2006 article by Klapdor-Kleingrothaus (Mod. Phys. Lett. A21 1547). The reasons why we do not compare to this result – as requested in his letter to CERN Courier – are mentioned in our publication on 0νββ results (2013 Phys. Rev. Lett. 111 122503) and discussed by B Schwingenheuer (2013 Ann. Physik 525 269). The statistical error on the signal counts given is too small, the significance of the peak is not estimated correctly and no efficiency correction is applied in the calculation of the half-life time. For this reason, GERDA compares its result to the claim published in 2004 (Nucl. Instr. Methods A522 371 and Phys. Letts. B 586 198).

In the above-mentioned 2006 paper, the peak reconstructs 2 keV below the Q-value of the decay. Double escape-peak events in 208Tl or 56Co decays are proxies for the 0νββ signal and do not show such a shift in energy with respect to calibration lines at full energy. Therefore, the quoted ballistic deficit is not relevant in the energy reconstruction of the GERDA data and the signal line should appear at the Q-value.

The GERDA analysis followed a rigorous analysis path, where parameters for background and pulse shape have been fixed prior to the "unblinding" and are documented in papers (arXiv:1306.5084, accepted by EPJC, and 2013 EPJC 73 2583, respectively) prior to the publication of the physics results (2013 Phys. Rev. Lett. 111 122503). Owing to the reduced background, we reach a higher sensitivity than HdM after 21 kg yr exposure and exclude the claim of 2004 with 99% probability.

The GERDA collaboration.

• All GERDA publications are available at www.mpi-hd.mpg.de/gerda/public/index.html.