FarDetector inKoreafortheJ-PARCNeutrinoBeam 1 Neutrino Oscillations With A Next Generation Liquid Ar- gon TPC Detector in Kamioka or Korea Along The J-PARC Neutrino Beam A. Meregaglia,1 and A. Rubbia1 (1) Institut fu¨r Teilchenphysik, ETHZ, CH-8093 Zu¨rich, Switzerland Abstract The “baseline setup” for a possible, beyond T2K, next generation long base- lineexperimentalongtheJ-PARCneutrinobeamproducedatTokai,assumestwo 8 verylargedeep-undergroundWaterCerenkovimagingdetectorsofabout300kton 0 fiducial each,locatedone inKoreaandthe other inKamioka but atthe sameoff- 0 axis angle. In this paper, we consider the physics performance of a similar setup 2 but with a single and smaller, far detector, possibly at shallow depth, composed n of a 100 kton next generation liquid Argon Time Projection Chamber. The po- a tential location of the detector could be in the Kamioka area (L 295 km) or J on the Eastern Korean coast (L 1025 km), depending on the∼results of the T2K experiment. In Korea the o∼ff-axis angle could be either 2.5o 3o as in 5 SuperKamiokande, or 1o as to offer pseudo-wide-band beam conditi∼ons. 2 ∼ 1. Introduction ] The well-establishedobservationsofsolarandatmosphericneutrinos,inpartic- h ular from Superkamiokande (SK) [1], SNO [2] and KamLAND [3] and the recent p negative result from MiniBOONE[4], are very strong indicators of the validity of - p the 3 3 PMNS [5] mixing matrix U (να = Uαiνi) to describe all the observed e neutri×no flavor conversion phenomena. In order to complete this picture, all the h elements (magnitude and phase)of the mixing matrix must be determined. That [ iwnhcliuchdetsodthaeyUthe3ereeleims eonntly(|aUne3u|2pp=ersinb2ouθ1n3d.inTthhee sbteasntdcaorndstpraarianmt ceotemriezsatfiroonm),tfhoer 1 v CHOOZ [6] reactor experiment, corresponding to sin22θ13 . 0.14 (90%C.L.) at 5 the mass squared difference ∆m2 2.5 10−3 eV2. 3 Longbaselineneutrinoexp|erimaetnmt|s≃havet×hetasktocompletethepresentknowl- 0 edge on the mixing parameters, possibly including the complex phase. A non 4 vanishing Ue3 would open the possibility of CP/T violation in the leptonic sec- . tor, as a d|irec|t consequence of non-trivial complex phases in the 3 3 mixing 1 matrix. The determination of the missing elements is possible via th×e study of 0 ν ν transitions ata baseline L andenergyE with the choice E/L ∆m2 . 08 Aµt→theesame time, the sign of the parameter ∆m2atm is unknown and th∼is affeacttms, : in some region of the parameter space which depends on the value of sin22θ13, v our ability to unambiguously detect CP-violation, as we will illustrate it for our Xi setup later on. On the other hand, the sign of ∆m2 could be determined by atm (1) precision measurement of the ν ν oscillation probability as a function r µ → e a of energy at sufficiently long baselines (i.e. L & 300 km), (2) by a comparison of oscillations involving neutrinos and antineutrinos at sufficiently long baselines, (3) by a comparison of probabilities at similar energies with two detectors, one located at a shorter and one at a longer baseline, or (4) by some combination of all the above methods (1)-(3). Hence, for some region of the sin22θ13 angle, the determinationofCP-violationandtheneutrinomasshierarchymustbeaddressed (cid:13)c2007byUniversalAcademyPress,Inc. 2 J-PARC CERN SpS design upgrade ultimate CNGS + 1 2 [7] [19] [7] [13] [20] [20] Proton energy E 30 GeV 40 GeV 400 GeV p ppp( 1013) 33 67 >67 4.8 14 4.8 15 × T (s) 3.64 2 <2 6 6 6 6 c Efficiency 1.0 1.0 1.0 0.55 0.83 0.8 0.8 Running (d/y) 130 130 130 220 220 240 280 N / yr ( 1019) 100 380 700 7.6 33 12 43.3 pot × ≃ Beam power (MW) 0.6 1.6 4 0.5 1.5 0.5 1.6 E N 4 11.5 28 3 13.2 4.7 17.3 p pot × ( 1022 GeVpot/yr) × · Relative increase 3 7 4 1.5 6 × × × × × Timescale >2009 2014? >2014? >2008 >2016 ? Table 1 Design parameters for the various beams at J-PARC. Comparison with the dedicated CNGS intensity and assumed upgrades of the CERN SpS complex (see text). The beam power corresponds to the instantaneous power on the neutrino target while the product Ep×Npot corresponds to the total amount of energy de- posited on the target per year, which is more relevant to calculate neutrino event rates. simultaneously in order to avoid the mass hierarchy degeneracy problem. The purpose of this workshop was to investigate the possible options with an upgraded J-PARC neutrino beam after the present T2K experiment [7], in par- ticular by placing a second detector in Korea. The center of the T2K neutrino beam will go through undergroundbeneath SK, and will automatically reach the sealeveleastoftheKoreanshore. Therefore,placingadetectorinanappropriate locationinKoreawillprobeneutrinooscillationsatabaselineof1000to 1200km away from the source. The neutrino beam spectrum in Korea will depend on the off-axis angle and on the exact geographical location chosen, because of the non-cylindrical shape of the decay tunnel in the neutrino beam line. When the upper side of the beam ◦ ◦ at 2 to 3 off-axis angle is observed at SK, the lower side of the same beam at ◦ ◦ 0.5 to 3.0 off-axisanglecanbe observedinKorea. See Ref.[8]for beamspectra computed in some reference locations and for details. ThemeasurementsinT2Kmightindicatethattheθ13 angleisinaregionwhere the simultaneous determination of neutrino mass hierarchy and the CP violating phase becomes possible. In Ref. [9], the possibility of using two next generation 300ktonfiducialidenticalWaterCerenkovdetectorsplacedatdifferentbaseline ∼distancesbutatthesameoff-axisanglewasexplored. Theauthorsconcludedthat this setup would have the ability to resolve the mass hierarchy by comparing the event yields at the two baselines, one being more sensitive to matter effects than 3 the other. In the present paper, we consider an alternative setup composed of a single 100ktonliquidArgonTPClocatedsomewhereontheeasterncoastofKoreawith off-axis angles ranging from 1o to 2.5o. We explore the physics reach of this experiment in terms of sensit∼ivity to discover θ13, to determine the CP-violation phase δ and to resolve the mass hierarchy. 2. The liquid Argon Time Projection Chamber The liquid ArgonTime ProjectionChamber (LAr TPC) (See Ref. [10] and ref- erences therein) is a powerful detector for uniform and high accuracy imaging of massive active volumes. It is based on the fact that in highly pure Argon, ion- ization tracks can be drifted over distances of the order of meters. Imaging is provided by position-segmented electrodes at the end of the drift path, continu- ously recording the signals induced. T0 is provided by the prompt scintillation light. In this paper, we assume a detector following the GLACIER concept [11] with mass order of 100 kton. The pros and cons of the LAr TPC, in particular in comparison to the Water Cerenkov Imaging technique, can be summarized as follow: Pros: The liquid Argon TPC imaging should offer optimal conditions to • reconstruct with very high efficiency the electron appearance signal in the energy region of interest in the GeV range, while considerably suppressing the NC background consisting of misidentified π0’s. MC studies show that an efficiency above 90% for signal can be achieved while suppressing NC background to the permil level [12]. This MC result was shown to be true over a wide range of neutrino energy, typ. between 0 and 5 GeV. If verified experimentally, this implies that the intrinsic ν flux will be the dominant e background in a liquid Argon TPC coupled to a superbeam [13]. The sys- tematic erroronthis flux (.5%)will be determining the finalsensitivity of the experiment. Pros: thephysicsperformanceperunitmassoftheLArTPCisexpectedto • be superiorto that ofthe WC detector; hence, the LAr TPC detectorcould be smaller than the WC detector to achieve the same physics performance; the 100 ktondetector consideredhere is approximatelytwice the size of the Superkamiokande detector. In addition, a LAr TPC should allow operation at shallow depth. The constraints on the excavation and the related siting issues of the detector should hence be reduced in the case of a LAr TPC. Foraquantitativediscussiononthepossibleoperationatshallowdepth, see Ref. [14]. Pros: the imaging properties and the good energy resolution of the LAr • TPC would allow to consider all events around the GeV region and above, while the WC technology is essentially limited to quasi-elastic (QE) events and backgroundconsiderations limit the beam energy to lie below the GeV range. Hence,broaderbandbeams,asforexampleobtainedatJ-PARCwith smaller off-axis angles than SK, e.g. 1o off-axis, covering more features of the oscillation probability (e.g. first maxima, first minima, second maxima, etc.) can be contemplated. Cons: the community has less experience with the LAr TPC technology • than the WC; the largest detector ever operated, the ICARUS T300, has a modular design which is not easily extrapolated to the relevant masses. Significant R&D and improvements in the design are therefore required in ordertoreachascalabledesignwhichcouldofferapathfora100ktonmass 4 facilityinacosteffectiveway. ForanoverviewandresultsofonesuchR&D program,see Refs. [15, 16, 17, 18]. Cons: the procurement and underground handling of large amounts of liq- • uid Argon is more difficult than that for water, however, safe, surface or near-surface,storage of very large amounts of cryogen(with volumes larger than the ones considered here) has been achieved by the petrochemical in- dustry; liquid Argon is a natural by-product of air liquefaction which has large industrial and commercial applications and can be in principle pro- duced nearby any chosen location. 3. The J-PARC neutrino beam The measurements considered by future long baseline experiments will require to accumulate sufficiently large statistics in the far detectors to study with high precision the neutrino oscillation phenomena. The physics performance of these facilitieswillthereforedependontheabilityoftheinvolvedacceleratorcomplexes and of the neutrino beam infrastructures to offer stable and long-term fault-less operations. InTable1wesummarizethedesignparametersfortheJ-PARCbeam under the three conditions “baseline”, “upgraded” and “ultimate”: The “baseline” 0.6 MW T2K beam [7]: The J-PARC neutrino beam • is under constructionfor the T2Kexperimentandis planned to beginoper- ation in 2009 at low intensity. The final goal for the T2K experiment is to reachanintegratedintensity of5 1021 pots, or equivalently a beampower of 0.6 MW during 5 years. × ∼ The“upgraded”1.66MWbeam[19]: thereisaplantofurtherupgrade • the accelerator complex to potentially provide an increased beam power of 1.66MWtotheneutrinotarget. Thisupgradeshouldinprinciplenotrequire majormodificationsinthe beamlineinfrastructurewhichhasbeendesigned up to 2 MW. An“ultimate”4MWbeam[7]: Thephysicsreachofamegaton(Mton)- • classwaterCerenkovdetectors coupledto a beam powerof4MWwas origi- nallyconsideredinRef.[7]. Althoughsuchapowermightrequiresignificant upgradesoftheJ-PARCfacility,itwaschosenasdefaultvalueinthe“base- line setup” for the two Water Cerenkov detectors configuration in Korea and Kamioka [9]. For ease of comparison, we also assume this value, unless otherwise stated. Finally for reference, Table 1 includes the intensity expected at the CERN NGS as well as figures corresponding to assumed upgrades of the CERN accelerator complex[13,20]. TheupgradedCNGSperformancewillbediscussedinSection8.. 4. Expected event rates in Korea and physics analysis We carry out an analysis following our previous work outlined in Ref. [13]. For the fit procedure, we assume at this stage identical running periods for each horn polarity, i.e. assume equal neutrino and antineutrinos runs. In an actual experiment,theplainexplorationofsin22θ13 wouldsuggestmoreneutrinos,while theCP-violationsearchrequiresacomparisonbetweenneutrinosandantineutrino runs. Assuming that the neutrino flux at the far location will be precisely known by extrapolationof the near data measuredby detectors located at280m [7] and possibly at 2 km [12], we assumed a systematic error on the ν +ν¯ flux at the e e far locationof 5%. The simulated data andbackgroundis fitted in energybins of 100 MeV each. More refined treatments of the systematic errors, including also instrumental effects, will be included at a later stage. 5 Assuming an integrated intensity of 35 1021 pots (or 5 years of running at 4 MW) for each horn polarity mode (label×led neutrino and antineutrino runs for a total of 10 years), the number of events expected neglecting flavor oscillations are reported in Table 2 for three locations: (1) 295 km, OA 2.5o corresponding to the Kamioka region; (2) 1025 km, OA 1.0o corresponding to most east region of Korea; (3) 1025 km, OA 2.5o. In the Table, the interactions of neutrinos and antineutrinos are separated, however, for the analysis we sum them, as we do not consider the possibility to discriminate between the two on anevent-by-event basis. Thehigherratesareclearlyexpectedat295km,whilethe1/L2dependence reduces the flux in Korea by an order of magnitude. A significant fraction of the loss compared to Kamioka (OA 2.5o) can indeed be recovered by locating the Korean detector at OA 1.0o. In this case, the rate in Korea is approximately half that at Kamioka. In addition, given the different kinematics in 3-body kaon decays,the beam backgroundratio ν /ν is morefavorableatsmaller OAangles. e µ Based on these arguments, one would conclude that an OA 1.0o in Korea is favored compared to the OA 2.5o. However, it was pointed out by the authors of Ref. [9] that the same OA angle at Kamioka and Korea would allow to cancel certain systematic errors associated to the beam. At this stage we have not attempted to quantify this further. neutrino run antineutrino run Location ν CC ν CC (ν +ν ) / ν CC ν CC (ν +ν ) / µ e e e µ e e e (ν CC) (ν CC) (ν +ν ) (ν CC) (ν CC) (ν +ν ) µ e µ µ µ e µ µ J-PARC - 40 GeV/c protons - T2K optics - 4MW 295 km 2.5 deg 205000 3619 1.9 % 27562 1225 2.7 % (0-5 GeV) (5970) (416) (60404) (1136) 1025 km 1 deg 81650 716 0.9 % 9737 176 1.1 % ∼ (0-5 GeV) (3249) (60) (24415) (212) 1025 km 2.5 deg 16980 300 1.9 % 2283 101 2.7 % (0-5 GeV) (495) (34) (5003) (94) Table 2 Number of events calculated for 35×1021 p.o.t. (7×1021 pot per year × 5 years) for each polarity mode anda detectorof 100 kton. Acut of 5 GeV hasbeen set on neutrino energy. 5. Sensitivity on θ13 We reportinTable3theexpectednumberofoscillatedeventsincludingmatter effects for a normal neutrino hierarchy (ρ=2.8 g/cm3) for different values of the CP-violationphaseδ andsin2(2θ13)=0.002for the three same detectorlocations of Table 2. The parameters used for the oscillation are the following: ∆m2 32 6 neutrino run ν CC ν CC µ e Location sin2 (2θ13) + + = 0.002 ν CC ν¯ CC µ e no osc. δ=0 90o 270o 180o beam J-PARC - 40 GeV/c protons - T2K optics - 4MW 295 km 2.5 deg Matter (n.h.) 210970 274 39 393 158 4035 (0-5 GeV) √B =64 1025 km 1 deg Matter (n.h.) 85900 226 138 389 300 776 ∼ (0-5 GeV) √B =28 1025 km 2.5 deg Matter (n.h.) 17475 94 60 126 92 334 (0-5 GeV) √B =18 Table3 Numberofoscillatedeventscalculatedforadetectorof100ktonand35×1021 pots with horns on neutrino polarity. The parameters used for the oscillation are the following: ∆m232 = 2.5×10−3 eV2, ∆m212 =7×10−5 eV2, tg2 (θ12)= 0.45, sin2 (θ23)= 0.5, sin2 (2θ13)= 0.002, ρ=2.8 g/cm3. = 2.5 10−3 eV2, ∆m212 = 7 10−5 eV2, tg2 (θ12)= 0.45, sin2 (θ23)= 0.5, and sin2(2×θ13)=0.002. × Thenumberofoscillatedeventsdependsasexpectedonthechosenvalueofthe δ-phase. For the normal hierarchy and neutrinos, the smallest number of events is observed for δ around 90o. Relative to the intrinsic beam ν background, the e numberofoscillatedeventsissignificant(seetableforthe√B),evenforthesmall value of sin2(2θ13)=0.002. These results translate in the θ13 3σ C.L. sensitivity shown in Figure 1. The configurations 295 km, OA 2.5o and 1025 km, OA 1.0o yield similar sensitivities, while 1025km, OA2.5o is slightlyworse. The role ofthe antineutrino runfor the region 0 < δ < 150o is also illustrated with the dashed curve which corresponds to a neutrino run only. 6. CP-violation discovery For our definition of CP-violation sensitivity and the sensitivity fitting pro- cedure please refer to Ref. [13]. In short, the CP-violation can be said to be discovered if the CP-conserving values, δ =0 and δ =180o, can be excluded at a given C.L. The reach for discovering CP-violation is computed choosing a “true” value for δ (=0) as input at different true values of sin22θ13 in the (sin22θ13,δ)- plane, and f6or each point of the plane calculating the corresponding event rates expected in the experiment. This data is then fitted with the two CP-conserving valuesδ =0andδ =180o,leavingallotherparametersfree(includingsin22θ13 !). 7 Fig. 1 θ13 sensitivity at 3σ C.L. for ourthree locations. The dashed line corresponds toa neutrino run only for the295 km,OA 2.5o configuration. The opposite mass hierarchy is also fitted and the minimum of all cases is taken as final χ2. Leaving all unknown parameters free and letting vary the known ones within their experimentalerrorsin the fit, the CP-violation3σ C.L.discoverysensitivity is shown in Figure 2 for the three geographical configurations 295 km OA 2.5o, 1025 km OA 1o and OA 2.5o, assuming an integrated intensity of 35 1021 pots foreachhornpolaritymode(5yrsneutrinoand5yrsantineutrinorun×sforatotal of 10 years). For the shortest baseline (295 km), the mass hierarchy degeneracy affects the sensitivity at large sin22θ13 & 10−2. Given the limited sensitivity to matter effects at this baseline, the data could be fitted with a conservingvalue of δ and the wrong (opposite to true) sign of ∆m2 . 32 This effect can be qualitatively understood by noting that the neutrino beam is optimised to cover the first maximum of oscillation which corresponds, at the selected baseline, to about 600 MeV neutrinos. At this energy the oscillation bi-probabilities for neutrinos and antineutrinos with δ = 90o and ∆m2 > 0, essentiallycoincide withthatwithδ =0 andδ =180o for ∆m2 <0,thereforethe data can be very well fitted with a conserving value of δ and the opposite mass hierarchy. The mass hierarchy degeneracy problem can be resolved by choosing a longer baseline and/or by extending the energy range of the beam spectrum, like for example is the case when we reduce the OA angle. In the Korean configuration, the situation is resolved as can be seen from Figure 2 (right): the CP-violation discovery region for the OA 1 and OA 2.5 configurations are therefore favored for large values of sin22θ13, but it also emerges that the small off-axis angle configuration is favoured by the higher statistics, and for small values of θ13 the sensitivity is better than that obtained in Korea. To summarize, the results are plotted in terms of δ coverage in Figure 3. If 8 Fig. 2 CP-violation discovery at 3σ C.L. assuming an integrated intensity of 35×1021 pots for each horn polarity mode (5 yrs neutrino and 5 yrs antineutrino runsforatotalof 10years),for the(left) 295 kmwith (blueline) andwithout (red line)aprioriknowledgeofthemasshierarchyand(right)1025kmOA1o (blueline) and OA2.5o (red line). the Korean scenario is chosen, the OA 1o angle guarantees a better sensitivity. In the Kamioka scenario, the CP-coverage decreases for values of sin22θ13 & 0.01. Itis a strikingcoincidence that this value approximatelycorrespondsto the expected sensitivity of the T2K experiment! Hence, in case a signal is observed with sin22θ13 & 0.01 one should aim at the largest possible δ coverage, and a longer baseline should be chosen to avoidneutrino mass hierarchydegeneracy. In this case, the authors of Ref. [21] claim that one should pay attention to other systematic errors. On the contrary, if no signal is measured and sin2(2θ13) . 0.01, the priority is to reach the best CP-sensitivity for the smallest possible value of the mixing parameter θ13 and the shortest baseline could be favored. 7. Mass hierarchy determination For CP-violation discovery, the best setup depends on the value of sin22θ13. Thesituationformasshierarchydeterminationiscompletelydifferent: thelonger baseline represents the best solution. At the same OA 2.5o, the Korean baseline gives animprovementin the sensitivity of a factor 5 with respect to the Kamioka baseline (See Figure 4). The higher rate, i.e. the smaller OA angle, improves the mass hierarchy determination: going from OA 2.5o to OA 1o improves the sensitivity by a factor of 2. 8. Other options at J-PARC and comparison to the CERN case The main focus of the CERN program is the Large Hadron Collider (LHC) however CERN is engaged in long baseline neutrino physics with the CNGS project [22] and supports T2K as a “recognized” experiment. The CNGS beam has recently begun operation and first events have been collected in the OPERA detector [23]. The current optimization provides limited sensitivity to the ν µ νe reaction and OPERA should ultimately reach a sensitivity sin22θ13 . 0.→06 (90%C.L.) in 5 years of running with the nominal 4.5 1019 pot/yr [24]. The ICARUST600[10],stilltobe commissioned,willdetect×toofew containedCNGS events to competitively study electron appearance. Ideas to improve the ν ν sensitivity at the CNGS have been discussed in µ e → 9 Fig. 3 Fraction of δ coverage for CP-violation sensitivity at 3σ C.L. for our three configurations, corresponding to Figure 2. the past [25, 26]. Recently, in Ref. [13] we discussed the physics potential of an intensity upgraded and energy re-optimized CNGS neutrino beam coupled to a 100 kton liquid Argon TPC located at an appropriately chosen off-axis position, andshowedthatimprovementsinθ13sensitivity,searchforCP-violationandmass hierarchy determination were potentially possible. The discussion relied on the observationthatwhereasJ-PARCprovidesarapidcyclewithhighintensityproton bunches at 40 GeV, the CERN proton complex has fewer protons and a slower cycle but ca∼n accelerate up to 400 GeV. Hence, the resulting target beam powers are–onpaper–comparable(SeeTable1). Inparticular,itwasnotedthatfuture upgradesofthe CERNLHC injectionchain(to beenvisagedinthe contextofthe luminosity upgrades) could provide increased proton intensities in the SPS. This option labelled “CNGS+” in Table 1 accordingly envisioned 3.3 1020 pots/yr. The same idea was subsequently and independently re-analy×zed assuming a smaller detector of 20 kton located at an angle OA 0.8o at a baseline of 730 km (MODULAr [20]). Inthis case,twopossibleupgradesforCNGSbeamlabelledas “CNGS1” and “CNGS2” yielding 1.2 1020 pot/yr and 4.33 1020 pot/yr were considered. × × A CERN accelerator division report [27] subsequently indicated that with an upgradeof the SPSRF andnew injectors, itwouldindeed potentially be possible to accelerate 2.4 1020 pot/yr. This means that the CNGS+ exposure would correspondto a ru×nof 7 yearsinstead of5 yearsand the CNGS2 beam to 9 years instead of the assumed 5 years. Yet intensity limitations will be coming from the designoftheequipmentinthecurrentCNGSfacilityandfromradiationandwaste issues. The desired intensities would therefore require a major re-assessment or a complete reconstruction of the CERN neutrino beam infrastructure (we how- ever point out that the low energy beams considered here would accommodate a significantly reduced decay tunnel compared to the one of the CNGS). In order to compare these options with the possible upgrades at J-PARC, we 10 Fig. 4 Mass hierarchy determination at 3σ C.L. for several detectorconfigurations. focus onthe sin22θ13 sensitivity. The obtainedconclusionscanbe readilyextrap- olated to the CP-violation and mass hierarchy determination. Figure5showstheexpectedθ13sensitivityatthe3σC.L.fora20ktonLArTPC detector located at Kamioka after 5 years of neutrino run with (a) the upgraded beam power of 1.6 MW (b) the ultimate beam power of 4 MW. The expected sensitivity with the existing SK detector and 1.6 MW is also shown. Finally, the sensitivity of a 100 kton detector at the CNGS+ computed assuming 5 yeas of neutrino run and 5 years of antineutrino run is plotted. A 20 kton LAr TPC at Kamiokaisaneffectivewaytoimprovetheθ13-sensitivityoftheT2Kexperiment. Fig.6comparesthesensitivityofa20ktonLArTPCatKamiokatotheModuLAr expectation [20]. As argued above, assuming similar target masses, the upgraded T2K and up- graded CERN beams theoretically provide comparable sensitivities, however, we stress that the current CNGS beam line infrastructure has not been designed to exceed 7.8 1019 pot/yr, so further upgrades and/or reconstructions must be envisaged in×order to cope with the potential increase of intensity from the SPS accelerator complex. 9. An“ultimate”two-detectors configurationatJ-PARCandthesyn- ergies with proton decay searches So far we have considered one far detector, either in the Kamioka region or in Korea. In this last section, we discuss as “ultimate” configuration the one which uses of two very large detectors, possibly of different technologies, located at different baselines. In order to cope with the harder beam at the angle OA 1.0o we consider a 100 kton LAr TPC at 1025 km in Korea. We have argued in the previous para- graphs that the smaller off-axis angles give best performance for the detector in Korea. For the closer detector at 295 km OA 2.5o we consider either a 300 kton Water Cerenkov detector or a 100 kton LAr TPC. For the WC detector we used