The appearance of the tau-neutrino
Y. Fukuda et al., Phys. Rev. Lett.
0 DOI: 10.1051/epn/2010604
On August 22nd 2009 a very special event was recorded by the OPERA experiment at the INFN Gran Sasso Laboratory produced by a neutrino coming from CERN. Born as mu-neutrino, it had transformed into a tau-neutrino on its 730 km long route. This is a long-awaited achievement of an enterprise started 30 years ago.
Neutrinos change family times, neutrinos only by their lepton flavours, which are
e spin ½ elementary particles come in three different conserved in the interactions. Neutrinos are produced
groups, called families, of identical structure. Each in states of definite flavour, as νe , νμ or ντ , each with an
family is made of two quarks, of charge +2/3 q and – antiparticle of the same flavour, its antineutrino or e+, µ +
1/3 q (q is the proton charge), and two leptons, one of or τ+, respectively. When a neutrino of a certain flavour
charge –q and one neutral. e neutral leptons are col- interacts with a target producing a charged lepton, the
lectively called neutrinos, but are three different latter always has its flavour; for example a νμ produces a
particles, distinguished by an additive quantum number µ –, never an electron or a tau.
called “lepton flavour”. e electron (e–) and the elec- Experiments in underground laboratories have shown that
tron-neutrino (νe) have one unit of electronic flavour neutrinos do not behave as assumed in the Standard Model
(–1 their antiparticles); the muon (µ –) and the muon- (SM): they do change,“oscillate”, between one flavour and
neutrino (νμ) have one unit of muon-flavour, and another. e evidence has gradually grown in the last four
similarly for the tau (τ –) and the tau-neutrino (ντ). decades, by studying the νes produced by the fusion
reace charged leptons are distinguished by their different tions in the core of the Sun and the νμs indirectly produced
masses (increasing with the family number) and life- by the cosmic rays collisions in the atmosphere.
by defining its initial state. In each mode the two
pendulums oscillate harmonically and synchronically with
(proper) frequencies, say, ω1 and ω2. e first mode is
obtained by letting them go from the same initial
elongation with zero initial velocities. e squared
frequency is ω 12=g/l. In the initial state of the second
mode the two elongations are equal and opposite and
the velocities are zero. e squared frequency is
ω 22 =g/l+2k/m. Considering for the moment a world
with only two neutrinos, the analogues of the modes are
ν1 and ν2 and the analogues of the frequencies are the
masses m1 and m2.
e analogue of a neutrino of flavour a is the initial
state with pendulum a abandoned with zero velocity at
a certain initial elongation and pendulum b in its at rest
position. Vice versa for flavour b. One observes that the
motion does not have a definite frequency. Rather, the
FIG. 1: Geographical sketch of the neutrino beam from CERN to Gran Sasso by A. amplitude of the vibrations of a gradually decreases
Zichichi (1979) down to zero, while that of those of b increases up to a
maximum. en the evolution inverts. Hence, the
energy “oscillates” periodically from a to b with a
freNeutrino oscillations quency ω2-ω1. In the quantum system the analogue of
Neutrino oscillations happen because neutrinos of energy is the probability to observe the flavour, because
definite flavour, νe, νμ and ντ, are not stationary states, both are proportional to the square of the amplitude. To
i.e., states of definite mass, but linear combinations of observe the phenomenon one needs to leave a time long
those. We call the latter ν1, ν2 and ν3 and m1, m2 and enough to allow for its development. In practice this
m3 their masses. means having the detector far enough from the source.
Oscillations phenomena are common in physics. Consi- en there are two alternatives: to look for the
disapder for example two identical pendulums, a and b, of pearance of the initial flavour (looking at a) or for the
mass m and length l weakly coupled by a spring of appearance of the new one (looking at b).
constant k. e system has two normal modes, obtained However, there are three neutrinos. e analogue is a
system of three coupled pendulums. Moreover, the
pendulums may not be equal. Clearly the phenomenology
FIG. 2: A. Bettini (left) and L. Maiani (right), together with a representative of local becomes much richer.
S1w2/i1s0s/a2u0t0h0o.rPithieostoceClEeRbNrating at the “ground breaking” event for the CNGS project on In 1968 R. Davis  published his measurement,
started in 1964, of the flux of neutrinos from the Sun with
an experiment sensitive to νe. The flux was smaller,
about 1/3, than the theoretical calculations by
J. Bachall . When the preliminary results became
known, B. Pontecorvo  claimed that the
disappearance might be due to neutrinos changing flavour
through oscillation, a concept he had introduced in
1957. Further νe disappearance experiments, in
different energy ranges, KAMIOKANDE  and Super
KAMIOKANDE  in Japan, GALLEX/GNO  at
Gran Sasso and SAGE  at Baksan, had established
by 1998 that the Pontecorvo idea was right. The final
proof came from the SNO  in Canada in 2002. All
neutrino flavours can interact with nuclei producing a
neutrino in the final state, instead of a charged lepton.
The final neutrino cannot be detected, but the effect
on the nucleus can. This was done by SNO, proving
that the total neutrino flux interacting in this way
corresponded exactly to the missing νe flux. A
confirmation came from the KamLAND experiment in
Japan that observed the oscillation in anti-electron and P. Strolin . e study of the proposals led to
neutrinos from nuclear power plants. a common decision by the CERN Director General,
By 1998 the Super-KAMIOKANDE  experiment had L. Maiani, the INFN
Premeasured, with high statistics and high accuracy, the sident, E. Iarocci, and the
dependence of the atmospheric νμ flux on the flight- LNGS Director, myself,
length and on energy. e shape of this function proved for the more risky (but
the oscillation phenomenon and made it possible to much more rewarding if
determine the corresponding oscillation parameters. successful) appearance
Also in this case, confirmation of the oscillation inter- experiments. e project
pretation came from disappearance experiments with was approved by the
artificial neutrinos, νμ produced at the KEK accelerators INFN and CERN
Counin Japan (K2K experiment) and later at the Fermilab in cils in 1999. It was funded with ad hoc contributions
the USA (MINOS experiment). mainly from Italy and from several other countries.
e civil engineering works at CERN and the
construction of the beam took place between Autumn 2000 (see
Fig. 2) and Summer 2004. e subsequent delicate and
complex phases of testing and commissioning were
completed by the Spring 2006. In August of the same
year the large detectors at LNGS, LVD, OPERA and
BOREXINO detected the first events produced by the
neutrino beam. In the same period the OPERA and
ICARUS experiments were developed.
Neutrinos do not behave
as assumed in the Standard
Model (SM): they do change,
“oscillate”, between one
flavour and another
From the third family to CNGS
As mentioned above, ντ and its charged partner τ are
part of the third “family”. As a matter of fact, the
existence of the third family, and the concept of family itself,
was experimentally established in the lepton sector
much earlier than in the quark sector. We briefly recall
that the neutrino was introduced as a “desperate
hypothesis” by Pauli in 1930, when only the electron was
known, to explain the apparent non-conservation of
energy in beta decay. It was discovered in 1956, when
also the muon had been discovered, by F. Raines . It
was in 1962 that an experiment at Brookhaven 
established that neutrinos were two, νe and νμ. Not much
later, the idea of the possible existence of a third lepton
family, called “heavy lepton” Hl and its neutrino νHl was
introduced by A. Zichichi. e idea was to search for
lepton pairs, which, in the case of e–µ + would be a clear
signature of the Hl. e search started at CERN in
1963, with the PAPLEP experiment, and continued at
the e+e– collider ADONE  at Frascati in 1967. e Hl
did indeed exist, but was found, and called τ, only in
1975 by M. Perl  and collaborators at the SPEAR
collider, which, differently from ADONE, had enough
energy to produce it.
irty years ago, A. Zichichi, then president of INFN
(Istituto Nationale di Fisica Nucleare), succeeded in
having approved by the Italian Parliament the Gran
Sasso project, to build a large, technologically advanced,
laboratory under the Gran Sasso massive. e laboratory
halls were oriented, in particular, toward CERN, in order
to be able in a future to host experiments on a neutrino
beam from CERN. e dra presented by Zichichi to
the Parliament is shown in Fig. 1.
e vision started to become reality around 1997.
Recalling that accelerators produce (almost pure) νμ
beams, the alternative νμ disappearance vs. ντ
appearance was open. Notice that they require different
characteristics both for the beam and the experiments.
Vivid discussions started in the community leading to
proposals for both. In particular, the OPERA
experiment was proposed in that year by A. Ereditato, K. Niwa
The first appearance
As mentioned, the beam produced at CERN is mainly
composed of νμ with no ντ . Consequently, the
observation of any ντ at LNGS must be due to the appearance in
the oscillation phenomenon. Experimentally, the two
types of neutrinos can be distinguished, when they
produce a charged lepton: a νμ produces a μ and a ντ
produces a τ. e principle is shown in Fig. 3.
In a tracking detector the μ appears as a long almost ECC are, however, only a component of OPERA, which
straight track, while the τ, which has a very short life- is a very complex structure: It includes different
time, picoseconds, decays already aer few hundreds tracking elements and magnetic spectrometers,
develomicrometres. Consequently, to observe the τ track a ped by the collaborators, both INFN groups and from
micrometre scale spatial resolution is necessary, some- other countries. e detection principle is shown in
thing that only nuclear emulsions can provide. Fig. 4. Further fundamental elements are the automatic
However, there is a further problem. Indeed, even a scanning and measuring microscope systems that are
distance of 730 km is small for the oscillation pheno- needed to extract the information from the emulsion.
menon, because the corresponding 2.5 ms flight time is A big effort was invested in increasing by an order of
only a small fraction of the oscillation period. Conse- magnitude the speed of these devices to cope with the
quently, only a very small fraction of neutrinos, 1-2%, above mentioned enormous emulsion area.
are expected to “oscillate”. Considering in addition the In 2008 – 2009 OPERA has collected about 1/5 of the
very small neutrino cross section the conclusion is rea- total foreseen data; of these about 35% have been
anached that the detector target mass needs to be lysed and the first ντ candidate event has been already
considerably larger than 1000 tons. found. It is shown in Fig. 5. The τ lepton is the
is is why the “Emulsion Cloud Chamber” (ECC) short red track. It decays in a charged hadron,
presutechnique, so called because it provides images similar mably a pion and a π˚, which in turn decays into 2
to a cloud chamber, was chosen. It is based on sand- gammas, which are detected. Even if the calculated
wiches of thin (50 mµ ) emulsion sheets, providing the probability for any background to simulate a τ is only
~1 µm resolution tracking, interleaved with 1 mm 1.8%, it is too early to claim the discovery of the
thick Pb sheets, providing the mass. appearance phenomenon. But a few other similar
ECC is a rather old technique, which was continually events will hopefully lead to this long-awaited
discodeveloped in Japan, with increasing levels of automa- very in the next years.
tion of the read-out, mainly at Nagoya. Historically,
ECC detectors were successfully employed for the References
study of the high-energy cosmic-rays. Particularly
important is the discovery by Kiyoshi Niu of a meson  R. (D1a9v6i8s)J1r.2,D05.S.. Harmer and K.C. Hoffman, Phys. Rev. Lett. 20
of a new flavour (in modern terms) in 1971 . It
was the charm, three years earlier than the discovery  J.N(1.B9a6c8h)a1l2l,0N9..A. Bachall and G. Shaviv, Phys. Rev. Lett. 20
of hidden charm with the J/ψ particle by Burton  B. Pontecorvo, Zh. Eksp. Teor. Fiz. 53 (1967)1717, [Sov. Phys.
Richter and collaborators and Samuel Ting and colla- JETP 26 (1968) 984].
borators. The discovery of the ντ in 2000 by Kimio  Y. Fukuda et al., Phys. Rev. Lett. 77 (1996) 1683.
Niwa and collaborators  at Fermilab was obtained
with an evolution of the same technique. OPERA is  M.B. Smy et al., Phys. Rev. D 69 (2004) 011104.
the latest and largest chapter of this evolution, com-  W.PHhayms. Lpeetlt.eBta6l1.,6Ph(2y0s.0L5e)tt1.7B44. 47 (1999) 127; M. Altmann et al.,
posed, just to give a few numbers, of 150 000
sandwiches, called “bricks”, including about 110 000 m2  N.[AZbh.dEukrsaps.hTietoorv. Feitza.1l.,2J2.E(x2p0.0T2h)e2o1r.1P].hys. 95 (2002) 181,
emulsion films and 105 000 m2 lead plates, for a total  Q. R. Ahmad et al., Phys. Rev. Lett. 87 (2001) 071301.
of about 1250 tons.