Table Of ContentCNGS neutrino beam systematics for θ
13
5
0 A. Ferraria∗ A. Guglielmi b P. R. Sala c
0
2 aCERN, 1211 Geneve 23, Switzerland.
n bIstituto Nazionale di Fisica Nucleare and Dept. of Physics, via Marzolo 8, 35131 Padova,Italy.
a
J
cIstituto Nazionale di Fisica Nucleare, via Celoria 16, 20133 Milano, Italy.
1
3
Energy spectra, intensityand composition of theCERNto Gran Sasso CNGS neutrinobeam for ν →ν and
µ τ
1 νµ → νe oscillation searches are presented. The associated beam systematics, which is the major ingredient for
v theνµ →νe search sensitivity,are obtained from thestudy of theprevious CERN WANF.
3
8
2
1
0 1. Introduction pact on sin22θ13 sensitivity is expected from the
5 systematics of the νe/νµ ratio, and in particular
Overthe nextfiveyearsthe presentgeneration
0 from its normalization error.
/ of oscillation experiments at accelerators with The CNGSbeam-linedesignwasaccomplished
h
long-baseline ν beams, K2K at KEK [1], MI-
p µ on the basis of the previous WANF νµ beam ex-
NOS at NUMI beam of FNAL [2] and ICARUS
- perienceatCERNSPS[7]forCHORUSandNO-
p and OPERA [3] at CNGS CERN to Gran Sasso
MAD experiments [8] which allowedfor a power-
e
neutrino beams [4] is expected to confirm the
h fulstudyoftheneutrinobeamprovidingastrong
ν → ν transitions observed in the atmospheric
: µ τ benchmark for conventional neutrino beams and
Xiv νofaancdcumraeacysurife|s∆inm222θ2|3>an1d0|−∆3me22V32|.witKh2inK1a0n%d in particular for the CNGS. In this sense the
23 CNGS beam systematics on ν /ν ratio can be
e µ
MINOS are looking for neutrino disappearance,
r derivedandpredictedfromthepreviousmeasure-
a by measuring the ν survival probability as a
µ ments and studies performed with the WANF.
function of neutrino energy while ICARUS and
OPERAwillsearchforevidenceofν interactions
τ 2. The WANF: a case of study
in conventional ν beams.
µ
Even not explicitly optimized for νµ → νe In the WANF νµ facility 450 GeV/c protons
searches these experiments can explore sin22θ13 were extracted by the SPS and sent into a seg-
oscillation parameter beyond the CHOOZ limit mented targetcomposedby 11 Be rods, 10 cm of
[5]. In particular ICARUS and OPERA [3] can length and 3 mm diameter each. The produced
reach a combined sensitivity on sin22θ13 (CP vi- positive (negative) mesons, essentially π and K,
olation and matter effects not accounted for) a werefocused(defocused)by twopulsedmagnetic
factor ∼ 5 better than CHOOZ in the allowed lenses (horn and reflector) into a 290 m vacuum
parameterregionofatmosphericneutrinooscilla- decay tunnel were neutrinos were produced. The
tions for five years exposure to the CNGS beam resulting ν beaminNOMAD at840mfromthe
µ
atnominalintensity[6]. Theνµ →νe oscillations Be target,was characterizedby anenergy∼24.3
willbesearchedforasνechargecurrenteventsex- GeV and a contamination of ∼ 6.8%, ∼ 1% and
cess over the νe contamination of the beam: the 0.3% of νµ, νe and νe respectively. Four main
key issue will be the knowledge of the neutrino processesassourcesofneutrinoswererecognized:
beam composition and spectrum. A strong im-
• 450 GeV/c proton interactions in the Be
∗OnleaveofabsencefromINFNSez. diMilano. targetwhich produce π±, K±, K0,... origi-
1
2
• nparitminagry96p%rotoofntshenoνnµiflnutexrainctNinOgMorAmDi;ssing nts / 8 GeV11x05 0010002 n m CC Data / MC11..112 n m CC
the target which produce in the horn and ve500 0.9
E
reflector walls, windows and beam-dump, 0 0 50 100 150 200 0.8 0 50 100 150 200
Visible energy (GeV) Visible energy (GeV)
mesons only weakly or not focused by the
• “obpfearcohmma-ropmpttenidcespua(t∼rritni1co5lse%”sagonefndνekνraa+otenνdseii)nn; tthheetdaercgaeyt vents / 8 GeV24000000 n- m CC Data / MC011...1824 n- m CC
E 0.6
and dump yielding ∼6% of ν ; 0
e 0 50 100 150 200 0 50 100 150 200
Visible energy (GeV) Visible energy (GeV)
• padffaefreotccitcuilnsegesdr∼ecio1nm0te%praoocnfteiνnoµtnsνaµanld+onννgeet.bhuetb∼ea4m0%-linoef ents / 16 GeV240000 n- eCC Data / MC01..1255 n- eCC
1.5 1.5 Ev 0 0
0 50 100 150 0 50 100 150
1 1 Visible energy (GeV) Visible energy (GeV)
0.5 0.5 Figure 2. Neutrino energy spectra (left) for the
data (points with statistical error bars) and the
0 0
0 100 200 300 0 100 200 300
p (GeV/c) p (GeV/c) Monte Carlo (histogram) for νµ CC, νµ CC and
1.5 1.5 ν CCinteractionsandtheircorrespondingratios
e
1 1 (right) in NOMAD [13].
0.5 0.5
0 0
0 100 200 300 0 100 200 300
p (GeV/c) p (GeV/c)
relative to the π+ is essential to calculate the ν
1.5 1.5 e
initial content in the beam.
1 1
DirectmeasurementsofπandKproductioninBe
0.5 0.5 by 400 GeV/c protons were performed by NA20
0 0 Collaboration [9] and by the SPY Collaboration
0 100 200 300 0 100 200 300
p (GeV/c) p (GeV/c)
1.4
V
1.2 Ge800
0.18 nts / 8 700 n eCC
0.6 Eve600
0.4 500
0.2
400
0
0 50 100 150 200 250 300 350
p (GeV/c) 300
Figure 1. FLUKA to SPY, NA20 meson produc-
tioncorrectionsasafunctionofthemomentump. 200
Thecurvesaretheresultsoffittingtheweightsat 100
fixedp (showedwiththeir uncertaintybars)with
0
0 20 40 60 80 100 120 140 160 180 200
combination of polynomial functions [13]. Visible energy (GeV)
In order to predict energy spectra, intensity Figure 3. Energy spectrum for the measured
andcompositionoftheνbeamasoundknowledge νe CC interactions (points with statistical error
ofmesonproductioninthe targetisrequired. In- bars) and the Monte Carlo (histogram) in NO-
deed the accurate description of the K+ and K0 MAD [13].
3
Figure 4. Schematic layout of the future CNGS neutrino beam line.
4
[10]at450GeV/c. AccuracyofMonteCarlogen- 3. The CNGS neutrino beam
eratorsofhadronicinteractionsforthemesonpro-
In the CNGS neutrino beam facility (Fig. 4)
ductioncouldlimit the sensitivitytoneutrinoos-
the primaryprotons at400GeV/c of momentum
cillation searches. The best agreement between
withnominalintensityof4.5·1019 pot/year(pro-
predictions and data was found with FLUKA
ton beam shared operations), will be extracted
standalone code [11] which reproduces the mea-
fromthe CERNSPSandsenttoacarbontarget,
sured π and K yields at the level of ∼ 20% in
13 graphite rods of 10 cm length, 5 and 4 mm of
the momentum region 30÷100 GeV/c which is
diameter, the first eight spaced by 9 cm. Simi-
expected to contribute most to ν flux [12]. Fur-
larly to previous WANF neutrino beam a mag-
thermore an accurate description of the primary
netic horn and a reflector will allow the focusing
proton beam spot, focusing system as well as of
(defocusing) of positive (negative)charged secon-
the materials inserted in the beam line from the
daries into a ∼ 1 km long decay tunnel where an
target to the dump (hadronic reinteraction pro-
intense ν neutrino beam is produced. A large
cesses) is also mandatory especially for the neu- µ
graphite and iron dump will absorb the residual
trino beam minor components.
hadrons at the end on the beam-line. A partic-
A complete analysis of this ν beam was per-
ular attention was devoted to have the full con-
formed by a beam line simulation from the Be
tainment of the proton beam spot in the target
target up to the NOMAD detector including the
and to reduce as well as possible the quantity of
FLUKAgeneratorwithfurthercorrectionstothe
material in the beam line in order to maximize
meson production based on the residual differ-
the neutrino flux in the Gran Sasso site.
encesbetweenthepredictedandmeasuredmeson
A full detailed simulation of the CNGS beam
yield in Be. These reweighting functions were
line was performed within FLUKA framework.
calculated as a function of the meson momen-
At the nominal proton beam intensity 4.5 ·
tum by integrating the particle production over
1019 pot/year (proton beam shared operations)
10mradofangularacceptanceandincludingalso
roughly 2900 ν CC/kt/year are expected at the
the transport efficiency as a function of produc- µ
Gran Sasso site. The increase the proton beam
tionanglealongthebeam-line. TheK0 reweight-
ing was obtained from the corresponding for K± intensity hasbeen studied, resultinginapossible
upgrade bigger than 40% [14].
usinga”quark-countingmodel”with15%ofun-
The muon neutrino flux will be characterized
certainty [13].
by an averageenergy of 17.4 GeV and ∼0.6%ν
The agreement with the measured ν interactions e
to ν contamination for E < 40 GeV (Fig. 5).
in NOMAD was at the few percent level (fig. 2) µ ν
Theν andν componentarebelow2%and0.2%
with a systematic normalization error of ∼ 7% µ e
respectively. The ν intrinsic level in the beam
onν CC/p.o.t. essentiallydeterminedby the ac- τ
curaµcy on the π+ and K+ production in the Be will be below 10−6 allowing for clean νµ → ντ
appearance experiments with both ICARUS and
target (3.4%) and on the protonbeam spot posi-
OPERA detectors. Due to the 732 Km of base-
tion on the Be target. The corresponding energy
line the contribution to neutrino beam from the
dependent error rangedbetween 2 and 5% in the
K±,K0isreducedbyafactor1.5÷2withrespect
2 < E < 100 GeV neutrino energy. Due to cor-
ν
to the WANF, while those from high energy pro-
relations between the sources of the ν and ν
µ e
ton interactions downstream of the graphite tar-
fluxes, the normalization uncertainties on ν /ν
e µ
getandfrompromptcharmedparticlesandkaons
ratio was smaller the the uncertainties on the in-
decayinthetargetanddumparenegligible. The
dividual ν and ν fluxes: 4.2% [13]. The cor-
µ e ν component will be mainly produced in the µ+
responding energy-dependent uncertainty ranged e
decayinsteadfromK+ andK0 asin the WANF.
from 4 to 6 %.
5
o.t. o.t.
p. p.
70 70
1 1
V / V /
Ge 103 Ge
2 5
/ m / e 10
n n
102
1
10
0 25 50 75 100 0 25 50 75 100
Neutrino energy (GeV) Neutrino energy (GeV)
Figure 5. Muon and electron neutrino spectra with parent particles at the Gran Sasso site.
4. CNGS beam systematics by pion decay and ν from the subsequent muon
e
decay,alargecancellationoccursintheν /ν er-
e µ
The previous scheme used for the WANF neu-
ror evaluation. Therefore the normalization un-
trino beam calculation based on FLUKA and
certaintyonν /ν isexpectedtobeonly∼3.1%.
e µ
SPY, NA20 hadroproduction data in Be is well
The corresponding bin-to-bin energy dependent
suitable for the CNGS neutrino neutrino beam.
error is estimated to be 2.5÷4% in the relevant
The Be-C scaling due to the different target ma-
part of the spectrum for ν , ν and ν /ν ratio.
µ e e µ
terialisexpectedtobealmostindependentonthe
secondarymesontype, to depend only weakly on
the transverse momentum of the meson, so con-
tributing lessthan2%tothe uncertaintyonneu- ainty00.1.28 ainty00.1.28
trino flux and less than 0.5% to νe/νµ ratio. ncert00..1146 ncert00..1146
Thesystematicsofneutrinoflux atGranSasso U0.12 U0.12
can be evaluated from the WANF by properly 0.1 0.1
0.08 0.08
rescaling the contribution of each meson to neu-
0.06 0.06
trino flux according to the different production 0.04 0.04
processes. A total normalization error of ∼ 4% 0.02 0.02
0 0
canbepredictedforbothν andν where∼3.8% 0 25 50 75 100 0 25 50 75 100
µ e E (GeV) E (GeV)
ofuncertainty is the secondaryproduction atthe
targetand∼1.1%errorisduetoparticlereinter- ainty00.1.28
actions along the beam-line. Owing to the tar- ncert00..1146
get configuration and the 732 km base-line, the U
0.12
position of proton beam on the target section, 0.1
0.08
the beam-line optics such as horn and reflector
0.06
currents and the amount of material included in 0.04
thesimulationarenotcriticalbutcontributeonly 0.02
0
0.8% to normalization error of ν and ν flux. 0 20 40 60 80 100
µ e E (GeV)
Dueto correlationsbetweenthe sourcesofthe ν
µ
andνefluxes,wherebothneutrinosareessentially Figure 6. Energy dependent uncertainty for νµ
produced from π+ of different momenta, the νµ and νe neutrino flux and for their ratio νe/νµ.
6
5. Conclusions 8, and references therein
TheNOMADColl.,Nucl.Instr.andMeth.A
The CERN to Gran Sasso CNGS neutrino
404 (1998) 96.
beam project for ν →ν oscillation search with
µ τ 9. H.W. Atherton et al., CERN-80-07,1980.
ICARUS and OPERA detectors largely benefits
10. G.Ambrosinietal.,Eur.Phys.J.C10(1999)
oftheWANFexperiencewithCHORUSandNO-
605.
MAD experiments.
11. A. Fasso’ et al., in A. Kling, F. Barao, M.
TheneutrinofluxcanbepredictedatGranSasso
Nakagawa, L. Tavora, P. Vaz, eds., Proceed-
within a systematic uncertainty ∼5%, the ν /ν
e µ ings of the Monte Carlo 2000 Conference
ratio will be known with ∼ 3.1% normalization
(Springer-Verlag,Berlin, 2001) 955.
error and 3÷4% energy dependent error, better
12. G. Collazuol et al., Nucl. Instr. and Meth. A
than quoted by ICARUS and OPERA, opening
449 (2000) 609.
the possibility to search for ν → ν oscillation
µ e 13. P. Astier et al., Nucl. Instr. and Meth. A 515
exploring sin22θ13 beyond the CHOOZ limit. (2003) 800.
However the sensitivity on both ν → ν and
µ τ A. Guglielmi and G. Collazuol, INFN/AE-
ν →ν oscillationchannelsarecompletely dom-
µ e 03/05 (2003).
inatedbythestatistics. In5yearsofCNGSstan-
14. M. Benedikt et al., CERN-AB-2004-022
dard operations, 4.5 · 1019 pot/year, 12 ν CC
τ OP/RF.
events can be recognized in ICARUS and simi-
larly in OPERA if ∆m2 = 2.6·10−3 eV2. In
23
the same time an excess of only 21 ν → ν CC
µ e
events over 120 ν CC from beam contamination
e
and ν → ν is expected in ICARUS (2.35 kton
µ τ
fiducial mass) for E <20 GeV if sin22θ13 ∼0.04
(θ13 ∼60). The foreseenincreaseof protonbeam
intensity is mandatory.
REFERENCES
1. M.H. Ahn et al.,Phys. Rev. Lett. 90 (2003)
021802.
2. The Fermilab NuMI Group, FNAL NuMI-
346,1998;E.Ablesetal.,FNALP-875,1995;
Nucl. Instrum. and Meth. A503 (2001) 122.
3. The ICARUS Coll., Nucl. Instrum. and
Meth. A461 (2001) 324; CERN-SPSC/2002-
27, SPSC-P-323,(2002).
The OPERA Coll., CERN-SPSC-P-318,
LNGS-P25-00; M. Komatsu and al.,
hep-ph/0210043(2002).
4. G. Acquistapace et al., CERN-98-02,
INFN/AE-98/05 (1998); CERN-SL/99-
034(DI), INFN/AE-99/05 Addendum.
5. M.Apollonioetal.,Eur.Phys.J.C27(2003),
331.
6. M. Komatsuet al., J. Phys.G 29 (2003)443.
7. L. Casagrande et al., CERN 96-06,1996.
8. TheCHORUSColl.,Phys.Lett.B497(2001)
