THE NEUTRINO FACTORY History and Introduction the Neutrino

THE NEUTRINO FACTORY

F. CERN, Geneva,DYDAK Swi tzerland

The discovery of neutrino oscillations marks a major milestone in the history of neutrino physics, and opens a window to what lies beyond the Standard Model. Many current and forthcoming experiments will answer open questions however a major step forward, up to and possibly including CP violation in the neutrino-mixing matrix, will be offered by the intense and well-understood neutrino beams from a neutrino factory.

1 History and introduction

The neutrino factory is, in essence, a muon storage ring with long straight sections along which decayir1g muons produce well-defined neutrino beams. The idea is a natural offspring of the concept of a muon collider, first proposed by Budker 1 at the 1969 Yerevan Conference, and developed further by Skrinsky. 2 The firstproposal of a genuine neutrino factory was put forward in quite some detail as early as by Koshkarev. 1974 3 It was not until the nineties that the idea of the muon col!ider was again taken up, this time going well beyond the level of paper studies. The US-led Muon Collider Collaboration 4 developed a conceptual layout of a multi-Te V muon collider, including all stages from muon production to the storage of high-energy muons in a circular collider, and launched an ambitious R & D programme. Figure 1 gives an impression of the complexity of their proposed accelerator network to achieve multi-TeV muon collisions. In the context of this work, the idea of a neutrino factory was developed further and high lighted in a comprehensive paper by Geer. 5 The coincidence between the evidence of oscillations of solar 6 and atmospheric 7•8 Ve V and the possible oscillation of accelerator-produced as claimed by the LSND experimentµ. vµ. 9 on the one hand, and the intriguing potential of the neutrino factory for the study of neutrino

95 16 GeV/c 1.5 x 1022 (\ protons Proton Accelerator I year � ) ...__.... Pion Production Target � and Capture Solenoid

Pion Decay Channel

D D D D Muon Ionization D D D Cooling Channel D D D D D c D D D Stopped Muon 1021 MeV/c 1.5x "--� 100 Physics muons / year muons �

Muon Accelerators 100 MeV 2 TeV ->

TeV/c 2 � muons

Muon Collider Higgs, t t WW, 2 2 TeV x I , ... j

Figure Schematic layout of the US design of a 2 TeV muon collider. 1: 2 x

96 oscillations on the other hand, initiated a European study group on the neutrino factory within the context of an ECFA-sponsored 'Prospective Study of Muon Storage Rings in Europe' 10 led by Autin, Blonde! and Ellis. Much of the material presented below arose from the presentations and discussions in this study group. The concept of a muon collider lends itself naturally to a three-step scenario, with increasing technical (and financial) complexity: the neutrino factory at the low-energy end; the Higgs factory at intermediate energy; and the genuine muon collider at the high-energy end (the last two options are discussed by Janot 11 at this conference). While the concepts appear feasible in principle, it is recognized that formidable technical difficulties have yet to be overcome and the practical feasibility has not yet been demonstrated. The concept, if ever realized, lies far in the future. Specifically for the neutrino factory, which is considered the easiest part, it is estimated that from the technical point of view, a functioning machine is ten years away (let alone political and financial considerations).

2 The physics case

The physics case for the neutrino factory is largely - but not solely - driven by neutrino oscillations.

What do we know and what are the open questions? 2. 1 It is largely accepted that we observe neutrino oscillations with the following characteristics:

Ve 2 1 2 Solar disappear with Cl.m 2 10-4 to 10-5 eV , or with Cl.m 2 � 10- 0 eV , with either • 2 � small- or large-mixing sin 2812. i i 2 2 Atmospheric vµ. disappear with Cl.m 3 10-3 eV and large-mixing sin 2823, but do • � not oscillate into Ve . � Therefore, at least all three known neutrino families participate actively in the oscillations. • This suggests, as a minimum, a neutrino-mixing matrix analogous to the CKM matrix in the quark sector. There would be a minimum of six independent variables: three angles 11 2, 913, 823, a CF-violating phase 5 as well as two mass-squared differences·Cl.m and 1 . . 2 Cl.m 3· y In the likely limit Cl.m 3 Cl.m 2, the solar and the atmospheric neutrino oscillations can • � » effectively be described by the two-family oscillation formalism. Taking into account all � y three families, however, the solar-neutrino oscillation is governed by 1112, Cl.m 2 and 813, while the atmospheric-neutrino oscillation is governed by 823, Cl.m 3 and 813 . The two oscillation phenomena are linked through the angle 813. y � If the LSND claim of a transition Ve Vµ. with Cl.m2 1 eV2 is correct, a fourth neutrino • ---> � is needed which must be sterile since it is not seen in Z decay.

The open questions are, apart from the question of Dirac- or Majorana-type neutrinos, and accepting that the phenomenon of neutrino oscillation is real,

is there a sterile neutrino? • what are the values of the neutrino masses? • what are the values of the mixing angles? • is there CP violation in the neutrino-mixing matrix? •

97 2.2 What will we learn within the next fiveyears? From ongoing solar-neutrino experiments and from SuperKamiokande, but even more so from an impressive number of new experiments coming into operation (BOREXINO, ICARUS-600, I216(?), KamLAND, K2K, MiniBooNe, MINOS and SNO), we can expect the following infor mation within the next five years:

Neutrino oscillations will be ultimately confirmed, andsterile neutrinos will be ruled out • (MINOS, SNO, and SuperKamiokande together with K2K: from the ratio of the flavour blind neutral-current rate to the charged-current rate; direct check from MiniBooNE and 1216(?) of the LSND claim). A decision will be made as to which which of the various sets of possible solar-neutrino • oscillation parameters is correct (all solar-neutrino experiments, including BOREXINO and KamLAND in their capacity as long-baseline reactor neutrino experiments).

Of course, the above is the conservative view. The LSND claim may actually be confirmed, with important consequences forcosmology. Or the Homestake solar-neutrino experiment, which claims a larger deficitof solar neutrinos than all other solar-neutrino experiments and hasa large weight when the energy dependence of the deficitis exploited, may revise its results. In any such case, things would become even more interesting and would underline rather than undermine the physics casefor the neutrino factory.

2.3 What will be the questions in ten years frnm now? Adopting for the sake of the argument the conservative view, we may expect the main open questions in ten years from now to be:

what is the value of B1a ? • is there CP violation in the neutrino-mixing matrix? • As a consequence, the operating parameters of the neutrino factory should not be driven by today's questions but rather by the likely questions which will be asked in ten years from now. As will be argued later, the neutrino factory is the best machine known today to address both questions.

3 Neutrinos from muon decay

The decay of the muon is a well-known and precisely calculable process, there are only leptons as involved:

The neutrino energy spectrum in µ+ decay is shown in Fig. 2 for unpolarized muons, together with the monoenergetic line of vµ in 71'+ decay. While the fiµ peaks at the kinematic limit of 53 MeV, the spectrum peaks at 2/3 of it. Ve The neutrinos from muons decaying in the straight sections are Lorentz-boosted in the fonrnrd direction towards the detector which is located at some distance L. The neutrino flux in the forward direction is proportional to E�/ L2, and half of the neutrinos are contained within a cone angle of l/1 mµ/ Eµ . As the detector will be located at a distance of several 100 km and = even more, it will be much smaller than the transverse size of the beam. Taking into account the neutrino cross-section which rises linearly with energy, the event rate in the detector scales with M x E�/ where Mis the useful target mass of the detector. L2,

98 µ++ Dµ

, , le++ue+Uµ , I , ' , ' , , ' <1l , I \ , \ , Ql \ I 0.. \ t5!/) D ' \ 0 e I \ c I , \ I . \ I • E I ::i . I Ql • I . , z • , I I I I I I I , , , , 0 10 20 30 40 50 Neutrino energy (MeV)

Figure 2: Neutrino energy spectra in "+ and µ+ decay.

The energy spectrum of the neutrino flux in the distant detector is the same as shown in Fig. 2, but with the energy scale multiplied by the boost factor Eµ/(53 MeV). The momentum of the circulating muons is in principle a free parameter. Thanks to the dependence of the neutrino rates in the far detector, one always wins with higher momentumE! of the circulating muons, at any neutrino energy. Therefore, there is a strong incentive to run the neutrino factory at the highest possible muon momentum. In Table 1 a comparison is made between a conventional beam from and K decay, and v rr a beam from the neutrino factory. The comparison is made specificallyµ for the proposed NGS beam 12 from CERN to the Laboratorio Nazionale del Gran Sasso (hereafter called Gran Sasso) 732 km away, with a neutrino factory sending a beam over the same distance.

Table Comparison of a conventional neutrino beam with the beam from a neutrino factory. I: Conventional Neutrino factory Parents 71"+, K+ or ,K µ or µ+ 7l" beam iie = 1 : 1 v V : Backgroundµ �23 of µ�13 of Vµ none D1., v0 ii1, beam v,,t ii,, : Ve = 1: 1 Background �63 of �0.53 of iie none v , Variation of average energy limitedµ free within factor of �3 Uncertainty of energy spectrum ±103 <13 v Uncertainty of radial spectrum ±103 <13 v Uncertainty of absolute flux ±10% <13 v flux per year at 732 km 3 x 10' (optim. NGS) 3 x 10" v 2 per cm ) ( 4.5 x 1019 400 Ge V pot) (1021 injected 50 GeV µ) (vµ

99 Table Neutrino beam composition for different muon-polarization. 2: +1 -1 p = p = µ+ only Dµ Ve + Uµ µ - Ve + Uµ only uµ

4 Muon-polarization issues

Because of the well-defined helicities of the decay particles in muon decay, there is a strong correlation between the direction of emission and the muon-polarization vector. The Ve (De) plays a special role insofar as its helicity is opposite to that of the other two particles: the positron electron and This means, in the decay + ---+ for example, that the ( ) iiµ (uµ)· µ e+ueDµ Ve are always emitted in the opposite direction to the muon-polarization vector, so that they can never reach the far detector other than with negligible energy. Table 2 summarizes the situation for both muon polarities and both polarization states. Also, since the Ve (De) is always emitted together with a second particle in the same direction, it never attains the maximum energy permitted by the kinematic limit. This opens the way in principle to look near the kinematic limit for the appearance of u0 (De), which could then only

be the result of an oscillation of u1, (iiµ) into Ve (De). While this strategy is possible already with unpolarized muons, it improves considerably with muon-polarization. Therefore, having a large degree of polarization is potentially of great interest, 13 as it could circumvent our current inability to build multi-kiloton detectors able to distinguish the interactions of Ve an·d De, which is of interest in the search for CP violation. At any rate, whatever the polarization of the circulating muons is, it must be known precisely. Fortunately, this measurement is in principle straightforward through the measurement of the energy of the decay electrons at a given angle with respect to the muon momentum. A possible scheme to measure the muon-polarization, together with the muon momentum and the energy spread of the muon beam, has been worked out by Blonde! 14 and is included in the report of Janot 11 at this conference.

5 Physics reach of the neutrino factory

5.1 Neutrino-oscillation physics Ideally, one would like to study all transitions listed in Table 3 together with the transitions of the respective antiparticles.

Ta ble 3: Neutrino flavour transitions.

Ve beam u1, beam Vr beam

Appearance Ve-+ v,t v/L -+ Ve -+ Ve l/T Appearance Ve Vr Vµ Vr -+ Vµ -t -t l/1 Disappearance Ve -+ Vx Vµ -t llx Vr -t Vx

The neutrino factory adds significantly to the possibilities through the novel feature of a

well-defined Ve beam. A Vr beam is not within reach , however - neglecting possible T-violation

effects - all channels accessible with a ur beam are also accessible with the Ve and v1, beams of the neutrino factory.

JOO We note that in principle three types of detector are called for: one for tau detection, one for electron and one for muon detection, always including the separation of the charge of the respective lepton, which separates the interactions of neutrinos from those of antineutrinos. In practice, tau detection poses stringent requirements on the detector. One established way is to track the decay kink of the charged tau through precise track-coordinate measurement in emulsions. However, unlike the case of the conventional beam from. 7f+and K+ decay, there is vµ an equal-strength component in the beam which worsens significantly the background arising De from charmed particle decays. Table gives approximate numbers for the comparison between the situation at the neutrino factory 4and the 'optimized' NGS beam 12 from CERN to Gran

Sasso. The oscillation is assumed to take place with sin2 and t::..m2 = ( vµ -> Vr W 3 e V2 .) One year of operation with a 1 kt target and GeV = protons 1 on target3 xat 10 the- NGS is compared to one year of operation with a kt4.5 target x 1019 and 400 injected GeV muons at a distance of km from the neutrino factory. In1 the latter case,1021 the background50 from charm production is 732estimated 15 according to � Ncharm 3 x 10-5 Ncc(vµ) + 3 x 10-4 Ncc(De)

Table 4: detection with the OPERA emulsion detector in the optimized NGS and at the neutrino factory. Vr

CC events CC events Background events vµ Vr NGS Neutrino factory 2450 21.4 0.07 245 000 2140 44

While the Vr-appearance experiment would still give a large number of events above back ground, it will likely no longer be of interest when the neutrino factory comes into operation. The other detector type concerns identifying and measuring final-state electrons and muons. This is considerably easier for muons than for electrons, yet detectors might be contemplated which can do a good job in either case. However, we may again expect that at the time the neutrino factory comes into operation, the standard measurements such as spectral distortions due to disappearance of or or abnormal ratios of 'short' to 'long' events in calorimetric Ve vµ , detectors, will no longer be the focus of interest. What is likely to be the focusof interest is a measurement of the appearance of 'wrong-sign muons' which would arise from transitions and would constitute a sensitive measurement Ve -> vµ of the mixing angle Figure highlights the sensitivity which one might achieve with the neutrino factory, for 11different13 . momenta3 of the circulating muons. 16 The gain from choosing a large muon momentum is apparent. This physics case is discussed in more detail by Gavela 17 at this conference. The ultimate quest is for CP violation in the neutrino mixing matrix, forwhich the neutrino factory is the only known contender. CP violation would be signifiedin the asymmetry

which is most easily exploited in the appearance of µ- and µ+ in beams of and respectively. Ve De, Experimentally this seems simple; however, there are two obstacles. The first is that the ratio of neutrino to antineutrino cross-sections must be known with sufficient precision, say at the 0.13 level. This requires special cross-section measurements which can be done in a so-called 'near' detector, some m away from the centre of the neutrino factory's straight section. 250 The other problem is that our planet is not CP symmetric and produces, through the MSW effect, an apparent CP asymmetry which must be calculated and subtracted from the measured 18

IOI -2 10

-3 ··· ... 10 Eµ · · .. = ... 50 GeV . . ···· · .... ··. .... ··· .... -4 10

-1 10 Sin2 3 Figure 3: Sensitivity for the mixing angle at the Neutrmo291 Factory, for 1020 useful muon decays, detector x a distance of km, and for different muon 9momenta.13 Detection efficiency4 anJ a IJackground estimate arc included. 732

effect. The interplay between the intrinsic CF violation and the MSW effect poses an interesting challenge, which is yet to be explored in some detail. First indications arc that measurements at different distances together with measurements at different neutrino energies provide sufficient information so as to be able to disentangle the contributions to the measured effect. Again the neutrino factory is vital because by its very construction, it can serve two detectors at different distances simultaneously, and also hasquite some capacity in its neutrino intensity so as to serve detectors at very large distances. A recent calculation of the expected CP-violating a.symmetry has been made by Donini ct al. and is shown in Fig. 4. t9 This figure shows the significanceof the difference in the appearance of 11+ and µ- in a 10 kt detector as a function of the detector distance, with and without the MSW effect from traversing matter. The significanceis given for 2 x 1020 useful i.e. in the proper straight section) decaying ( muons. While this figure does not yet reflect a realistic detector configuration, it calls for the utmost intensity of stored muons when it comes to measuring CP-violating asymmetries. Measuring CF-violating asymmetries is bound to be difficult if not impossible within a

scheme of three neutrino families. For the transition <--> the CP-violating asymmetry is proportional to 6.mr which makes only the so-calledl/e 'large1/1, ,-mixi ng-angle' lvIS\N solution of 2 , the solar-neutrino deficit the only solution which va le CP-violating asymmetries. We must rely on Nature to have chosenenables the experimentally right set of oscillation obser b parameters. However, should there be a fourth, sterile, neutrino, large CF-violating asymmetries may occur due to new phases in the neutrino-mixing matrix. need the neutrino factory exploration. which would for their

107 � vacuum 12.5

8� 10 < 0 0 "' 75

-;;_ matter :!. ', In //

2000 Detector4000 distanc e (km)6000 8000

Figure Significance in standard deviations of a CP-violating asymmetry, as a function of the distance of the 4: 2 detector, calculated ll.mi2 1 eV , ll.rn�3 = eV2 , sin2 2912 1113 1/ 3 = 45° , = x • 3 x 3 = = 13°, 2 for 10- and 10- 0.5, b = 90°.

5. 2 Other interesting physics opportunities

With the neutrino factory, other interesting physics opportunities besides neutrino oscillations arise.

High-intensity, low-energy muons: • - searches for rare muon decays (e.g. e1 at the level of 14 µ, _, 10- ). High-precision neutrino physics: • in 'near' detector 4 8 CC events per kg for 1021 injected GeV muons; x 10 vµ. 50 precision measurements of neutrino-nucleon cross-sections; precision measurements of nucleon structure functions in unpolarized and polarized targets, for various nuclei; measurement of sin2 Bw to in neutrino-electron scattering. ±0.0001 Prolifictagged charm production. •

6 Accelerator issues

The current ideas about the accelerator network associated with the neutrino factory have been shaped at the recent NuFact'99 workshop held in Lyon (France). Whereas before the workshop three different designs existed, a unified design has now been agreed upon, the only remaining difference being the proton driver. 20 Even there the beam power on target (which is the relevant quantity for the production of charged 11") has been agreed upon: 4 M\V, for experts quite a daring but not unrealistic figure. This figure is important as it limits the production of muons to per year, which is now the baseline figure. However, it is recognized that the physics 1021 (in particular CP violation) in all likelihood wi ll ask for more, so per year is retained as an 1022 option to be kept in mind. Proton Booster Tairgel

Proton Recirculating &"-._/Linae Li nae ----. Far Detector 1

Near Detector

Far Detector",,.,,.' 2

Figure 5: Schematic layout of the neutrino factory complex.

The currently envisaged schematic layout of the neutrino factory complex is shown in Fig. 5. The protons deliver 4 MW onto a high-Z target. The target technology is very difficult, as very strong pressure waves have to be sustained. The target is embedded in a strong solenoidal field of some 20 T. The accepted forward-going are phase-rotated, i.e. their momentum spread 7r is reduced at the cost of bunch lengthening. After a drift space where the decay, the decay 7r muons are 'cooled'. with a relatively moderate compression of the phase space by a factor of 75. Since muons are unstable, all further actions on them must be highly efficient so as to accelerate them as fast as possible to their final momentum. The cooled muons enter first a recirculating linac, where they are repeatedly accelerated in a straight section, and returned in recirculating arcs. Each arc is optimized for a certain momentum. For physics reasons, a momentum range up to 50 GeV is desired, where 50 GeV is a limit imposed by accelerator-related rather than physics considerations. Finally the muons enter the muon storage ring. It has two straight sections, each pointing to 'far' detectors at different distances say 730 and km, implying inclination angles of 3° ( 5000 and 23°, respectively. The arcs are as short as possible. The two straight sections and the arcs are arranged in a near-planar geometry. At the end of one straight section, a 'near' detector would be located. While the overall design is considered feasible by the experts, it is recognized that a lot of R & D work must be successfully accomplished before a serious engineering design could be made. Particularly critical areas are the target zone and the muon cooling.

7 Summary

Taking everything together, the physics potential of the neutrino factory looks very good. The specific features of the muon beams originating from muon decay make possible experiments which are not feasible with conventional neutrino beams from and K decay. Outstanding 7r issues are transitions of and CP-violation studies. -+ Yet Nature must be v.kind toVµ have chosen the right set of oscillation parameters to make CP violation experimentally accessible. Ongoing as well as new experiments will tell within the next five years what the situation is, possibly making the case for the neutrino factory irresistibly strong.

HM. Technically, the neutrino factory could be in operation within ten years. However. in the real world of scarce resources, political and financial considerations are more likely to determine the time schedule than technical factors.

Acknowledgement

I should like to thank J. Tran Thanh Van for his invitation to this meeting which wasconducted in the usual wonderful atmosphere. My sincere thanks also go to the CERN Desktop Publishing Service for their competent and efficient help.

References

1. G.I. Budker, Proc. 7th Int. Conf. on High Energy Accelerators (Yerevan, 1969), p. 33; AIP Conf. Proc. 352 (1996) 4. 2. A.N. Skrinsky, Proc. Int. Seminar on Prospects of High Energy Physics (Morges, 1971); AIP Conf. Proc. 352 (1996) 6. 3. D.G. Koshkarev, CERN internal report CERN/ISR-DI/74-62 (1974). 4. C.M. Ankenbrandt et al., Status of muon collider research and development and future plans, report BNL-65-623 (Fermilab-PUB-98/179, LBNL-41935). 5. S. Geer, Phys. Rev. D57 (1998) 6989; and erratum. 6. J.N. Bahcall, P.I. Krastev and A.Yu. Smirnow, Where do we stand with solar neutrino oscillations, preprint hep-ph/9807216. 7. D. Casper, Results from Super-Kamiokande, these proceedings. 8. D. Michael, Results from MACRO, these proceedings. 9. G. Mills, Results from LSND, these proceedings. 10. B. Autin, A. Blonde! and J. Ellis (eds.), Prospective study of muon storage rings at CERN, Report CERN 99-02 (1999). 11. P. Janot, Physics with muon colliders, these proceedings. 12. G. Acquistapace et al ., The CERN neutrino beam to Gran Sasso (NGS), Report CERN 98-02 and INFN/ AE-98/05; Addendum CERN SL-99-034 DI and INFN/ AE-99/05. 13. D. Cline, contribution to the NuFact'99 workshop, 5-9 July 1999, Lyon (France). 14. A. Blonde!, in 'Prospective Study of Muon Storage Rings at CERN', Report CERN 99-02 (1999), eds. B. Autin, A. Blonde! and J. Ellis, p. 51 ff. 15. P. Strolin, contribution to the NuFact'99 workshop, 5-9 July 1999, Lyon (France). 16. A. Cervera Villanueva, F. Dydak and J.J. Gomez Cadenas, paper in preparation. 17. M.B. Gavela, Three-Family A.nalysis at the Neutrino Factory, these proceedings. 18. L. Wolfenstein, Phys. Rev. Dl 7 (1978) 2369; ibid. D20 (1979) 2634; S.P. Mikheyev and AYu. Smirnow, Sov. J. Nucl. Phys. 42 (1986) 913. 19. A. Donini, M.B. Gavela, P. Hernandez and S. Rigolin, paper in preparation. 20. E. Keil et al., contribution to the NuFact'99 workshop, 5--9 July 1999, Lyon (France); E. Keil, Seminar given at CERN on 15 July 1999; See also http://wwwslap .cern.ch/-keil/pub/talks/lyon.ps .

10'\