Miniboone Oscillation Results with Complete Dataset

MiniBooNE Oscillation Results with Complete Dataset

Adrien Hourlier on behalf of the MiniBooNE Collaboration 2020/07/02

preprint available at arxiv:2006.16883 Comparing MiniBooNE and LSND

LSND (1993-1998) MiniBooNE (2002-2019) 0.8 GeV proton beam Decay At Rest neutrino flux 8 GeV proton beam Decay In Flight beam

LANSCE proton beam water target beamstop L ~ 30 m Advances in High Energy Physics !

E ~ 30 MeV Interaction Track Cherenkov Candidate + + � → � �� Muon

+ $ CCQE − � �¯�� $ +n→ p+$

Electron

e CCQE − e +n→ p+e Scintillation Neutral pion

NC!∘ Time ∘ Cerenkov Neutron +N→ +N+! capture on H

F"#$%& ': (Color online) MiniBooNE particle reconstruction [(]. From top to bottom, a muon neutrino charged-current quasielastic (CCQE) Different systematics. Same L/E baseline.interaction, an electron neutrino CCQE interaction, and a neutral current, neutral pion production (NC) ) interaction. *esecondandthe third columns show the characteristics of tracks and Cherenkov rings [+], and the last column shows the event∘ displays of candidate events. Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 ! 2

For the ] ] (] ] ) oscillation study, the follow- running, there was little neutrino cross section data in the ing three particle reconstruction algorithms were the most )00 MeV to few GeV energy regime. In response, MiniBooNE important: single" → Cherenkov# " → rings# from amuonand developed a highly successful campaign of cross section an electron and the two-ring electromagnetic shower topol- measurements, some of which are described here. *ese ogy from aneutralpiondecaytotwogammas.(1) Figure(2)' results are interesting by themselves and also can be used as shows the di,erent characteristics of these three signals, direct inputs to the BSM analyses, as described later in this including(3) examples of typical events in the detector [(]. paper. *ereconstructionalgorithmscanalsoreconstructmore MiniBooNE’s beam is among the .rst high-statistics, high complicated topologies important for constraining back- purity 1uxes in the energy range from )00 to )200 MeV. grounds and for cross section studies discussed below. *e *e observation of the resulting events in a large, isotropic charged-current single charged pion (CC) ) interaction detector with coverage is unique. Within this detector it is reconstruction algorithm [-] .ttwoCherenkovringsfrom+ relatively easy to achieve uniform angular acceptance. Also, .nal state particles, a charged lepton, and a! positive pion, the active veto4! makes it possible to measure NC interactions to .nd their kinematics. *e charged-current single neutral e,ectively. Insensitivity of hadronic details worked in pos- pion (CC) ) interaction reconstruction algorithm [/] .t itively. *ehadronmultiplicityo3en causes confusions for a charged lepton∘ and a neutral pion (which consists of tracker detectors. Although the MiniBooNE detector cannot two electromagnetic! showers, that is, the algorithm .ts for measure multiple hadron tracks, it measures total energy three Cherenkov rings). Another algorithm identi.es and of low energy hadrons (such as protons below Cherenkov reconstructs the neutral current elastic (NCE) interaction threshold from CCQE interactions) in calorimetric way, and, [)0], where the total kinetic energy of .nal state nucleons is as a result, the details of .nal state interactions (FSIs), such found using scintillation light. as rescattering, absorption, and charge exchange, do not Along with reconstruction of the light topology in the strongly a,ect reconstruction of kinematics. detector, event identi.cation also relies upon “subevents.” Perhaps most importantly to the overall impact of the *ese are bursts of light separated in time which indicate data, the MiniBooNE collaboration provided the cross sec- asequenceofdecay.Forexample,amuonwhichstopsand tion data in a form that is most useful to theorists. Tradi- then emits a decay (“Michel”) electron will produce two tionally, cross section data have been presented either as a subevents, one from the initial muon and the one from the function of neutrino energy ( ])or(-momentum transfer Michel electron. ( ). *is presentation is problematic in the MiniBooNE energy2 region, because of the importance# of nuclear e,ects: 3. MiniBooNE Cross Section Results Fermi$ motion smears the kinematics, binding energy shi3s the energy spectrum, nucleon correlations a,ect both energy All searches for BSM physics rely on a precise understand- dependence and normalization of cross sections, and pions ing of SM interactions. However, when MiniBooNE began may be created, absorbed, and charge-exchanged within The MiniBooNE Detector

• ⌀12.2 m sphere, ⌀10m ﬁducial volume • 800 tons of mineral oil, 450 tons ﬁducial mass • 2 optically isolated volumes • 1280 inner PMTs, 240 veto PMTs • Very well understood detector • 3% change of the energy scale over 17 years of running • Measurements of cross sections for >90% of the neutrino and anti-neutrino processes

(Neutrino mode)

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 3 Data taking timeline 2nd data set ν: 6.38x1020 POT

1st data set ν: 6.46x1020 POT 1st data set ν¯: 11.27x1020 POT Beam dump dark matter search

3rd data set ν: 5.91x1020 POT • We have added another ~6x1020 POT to the neutrino dataset since the previous data release. • The detector was turned oﬀ at the end of summer 2019, mothballed and waiting for future use… • Almost 17 years of running, or as much as 5 army ants worth of protons! Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 4 Data taking timeline 2nd data set ν: 6.38x1020 POT

1st data set ν: 6.46x1020 POT 1st data set ν¯: 11.27x1020 POT Beam dump dark matter search

graduated from high school! 3rd data set ν: 5.91x1020 POT • We have added another ~6x1020 POT to the neutrino dataset since the previous data release. • The detector was turned oﬀ at the end of summer 2019, mothballed and waiting for future use… • Almost 17 years of running, or as much as 5 army ants worth of protons! Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 4 Data taking timeline 2nd data set ν: 6.38x1020 POT

1st data set ν: 6.46x1020 POT 1st data set ν¯: 11.27x1020 POT Beam dump dark matter search

+ + J/ψ→ 7 K →π +invis. 10− invis.

favored • First dedicated search for direct detection of αµ accelerator-produced dark matter in a proton 10−8

= 0.5) BaBar beamline D α

, MB Elastic N • Beam aimed oﬀ of the target χ MB Electron

+ Timing D 98, 112004 Phys. Rev. • ν ﬂux reduced by ~50 MB Full N Direct = 3m −9 + Timing

V 10 Detection • Demonstrated the power of neutrino detectors Nucleon

(m XENON10/100- 4 -e ) χ

in searching for accelerator-produced dark V /m

matter χ −10 E137 (m 10 D α 2 ε NA64 Y =

10−11 Relic Density

LSND

10−3 10−2 10−1 2 mχ (GeV/c )

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 5 Excellent long term detector stability over 3 run periods Michel electron spectrum π0 mass 1.2 1.2 4000 4000 st 20072007st data data 20 20 data 2007 2007 data 1 1νe 6.46νe 6.46×10×10 POT POT 1.0 1.0 nd 2017nd data 20 20 2 22017νe 6.38 νdatae 6.38×10×10 POT POT 800 data 2017 2017 data 800 3rd 3νrd 5.91ν 5.91×1020×10 POT20 POT 0.8 0.8 20192019e data edata data 2019 2019 data 30003000 0.6 0.6 events/MeV events/MeV events/2 MeV 0.4 0.4 600 600 events/2 MeV 2000

Excess Events/MeV 0.2 Excess Events/MeV 0.2 2000 0.0 0.0 400 400 1000 −0.2−0.2 0.2 0.20.4 0.40.6 0.610000.8 0.81.0 1.01.2 1.21.4 1.41.6 1.6 MiniBooNEQE QE preliminary MiniBooNE preliminary 0 E E (GeV) (GeV) 20 ν 18.75ν 40 x 1020 POT 60 200 18.75 x 1020 POT 200 E [MeV] FIG. 7:FIG. The 7: plot The shows plot showsthe total the event total excesses event excesses in neutrino in neutrino mode for mode the for first, the second, first, second, and third and third 0 runningrunning periods. periods. Error bars Error include bars include only statistical only statistical uncertainties. uncertainties.20 40 60 100180 120160 140140 160120 180100 1.4 Ratio 2019/2007 2 0 0 2 12 12 E [MeV] ] [MeV/c mass invariantinvariant mass [MeV/c ] GordonGordon coecients, coecients, and the and probability the probability of1.2 pion of escape pion2019/2017 escape from the fromCnucleusisestimated the Cnucleusisestimated π π 12 12 to beto 62.5%. be 62.5%. The Theradiative radiative branching branching fraction fraction is 0.60% is 0.60% for C for andC 0.68% and 0.68% for H2 forafter H2 after 1 integratonintegraton over all over the all invariant the invariant mass range, massWe range, where can where the single theuse single gamma gammatwo production production standard branching branching candles to calibrate the energy scale over the different data sets • 1.40.8 Ratio 1.4 1.4 Ratio ratio increasesratio increases below below the pion the production pion production threshold. threshold.2019/2007 With these With values, these values, the ratio the of ratio single of single 2019/2007 2019/2007 Ratio 0 0 Scale0.6 up the energy response to match the original 2007 data release: gammagamma events events to NC to⇡ NCevents,⇡ events,R,• canR be, can1.2 estimated be estimated to2019/2017 be to be 1.2 2019/2017 2019/2017 1.2 20 40 60 E [MeV] R =0.R151=0.1510.00680.00681.5+0•1..5225+01 2015-2017.5220.00600.00601.5/0.6251.5/ =0.625 0. 0091neutrino = 0..0091. data => 2% energy scaling 1 1 ⇥ ⇥ ⇥ FIG.⇥ 1: The⇥ Michel⇥ ⇥ electron⇥ energy distribution for the first, second, and third running periods

Note thatNote single that single gamma gamma events events are assumedin are neutrino• assumed0.8 to2017-2019 come mode. to entirely come The events entirely from are from normalizedradiative neutrino radiative decay. to the decay. ﬁrst The running Thedata period. => The bottom 3% plot energy shows scaling 0.8 0.8 total uncertaintytotal uncertainty on this on ratio this is ratio 14.0%ratios is 14.0% (15.6%) of the (15.6%) third in neutrino running in neutrino period (antineutrino) to(antineutrino) the ﬁrst mode. and second mode. This running This periods. 0.6 0.6 nd 0.6 6 estimateestimate of R =0 of .R0091Adrien=0.00910.0013 0Hourlier agrees.0013 agrees fairly well fairly — with wellThe theoretical with XXIX theoretical calculations International calculations of the of single the Conference single on Neutrino Physics and Astrophysics — July 2 2020 ± ± 20 40 60 100 150 150 100 gammagamma event rateevent [29]. rate [29]. E [MeV] ] 2 [MeV/c mass invariant 0 π π0 invariant mass [MeV/c2]

FIG. 1: The Michel electron energy distribution6 for the first, second, and third running periods FIG. 3: neutrino The in NC ⇡0 periods mass distribution running third and for the second, first, first, second, the for and third running distribution mass periods0 ⇡ NC in The neutrino 3: FIG. 14 14 in neutrino mode. The events are normalized to the first running period. The bottom plot shows mode. the of The events ratios shows are normailzed plot bottom to The the first period. running running period. first the The to bottom plot normailzed are shows ratios events The of themode. ratios of the third running period to the first and second running periods. third running period to the first and second periods. running running periods. second and first the to period running third

6 8 8 Neutrino energy and 3ν prediction

Data (stat err.) +/- νe from µ 7 +/- νe from K 0 Events/MeV νe from K 6 π0 misid • Excess of data events with respect to our Δ → Nγ background prediction 5 dirt other • We report an excess of 560.6 ± 119.6 electron- Constr. Syst. Error 4 like events (neutrino mode) MiniBooNE preliminary 3 18.75 x 1020 POT 2 Neutrino mode • Signiﬁcance : 4.7 σ in neutrino mode only

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 QE Eν (GeV)

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 7 νe-like excess stable across 3 runs

1.2 st 20 1 νe 6.46×10 POT 1.0 nd 20 2 νe 6.38×10 POT rd 20 3 νe 5.91×10 POT 0.8 Comparing the data-prediction excess for three MiniBooNE preliminary • 0.6 data taking periods in neutrino mode 18.75 x 1020 POT Comparable statistics between the three data 0.4 Neutrino mode • releases Excess Events/MeV 0.2 • The observed excess remains well 0.0 compatible between the three data sets

−0.2 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 QE Eν (GeV)

FIG. 7: The plot shows the total event excesses in neutrino mode for the ﬁrst, second, and third running periods. Error bars include only statistical uncertainties.

Gordon coecients, and the probability of pion escape from the 12Cnucleusisestimated nd Adrien Hourlier — The XXIX International Conference 12on Neutrino Physics and Astrophysics — July 2 2020 8 to be 62.5%. The radiative branching fraction is 0.60% for C and 0.68% for H2 after integraton over all the invariant mass range, where the single gamma production branching ratio increases below the pion production threshold. With these values, the ratio of single gamma events to NC ⇡0 events, R, can be estimated to be

R =0.151 0.0068 1.5+0.522 0.0060 1.5/0.625 = 0.0091. ⇥ ⇥ ⇥ ⇥

Note that single gamma events are assumed to come entirely from radiative decay. The total uncertainty on this ratio is 14.0% (15.6%) in neutrino (antineutrino) mode. This estimate of R =0.0091 0.0013 agrees fairly well with theoretical calculations of the single ± gamma event rate [29].

14 Constraining the backgrounds

+/- νDatae from (stat µ err.) ν from µ+/-+/- 7 νee from K +/-0 νe from K νe from K0 Events/MeV Events/MeV from K πνe0 misid 6 π0 misid N Δ →→ Nγγ dirt 5 dirt other Constr. Syst. Syst. Error Error 4 MiniBooNE preliminary 3 18.75 x 1020 POT

2 Neutrino mode

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 EQE (GeV) Eν (GeV)

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 9 Constraining the backgrounds

2 Neutrino mode

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 EQE (GeV) Eν (GeV)

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 9 Constraining the backgrounds

Data (stat err.)+/- π0 MisID νe from µ ν from µ+/-+/- 7 νee from K constrained from in situ +/-0 νe fromfrom K K 0 νe 0 Events/MeV Events/MeV from K measurement of NC π πνe0 misid 6 π0 misid rate N νe from μ decay Δ →→ Nγγ dirt is constrained by in situ 5 dirt Δ-> Nγ resonance other νμ CCQE measurement Constr. Syst. Syst. Error Error constrained from in situ 4 measured NC π0 rate and MiniBooNE preliminary theoretical prediction 3 18.75 x 1020 POT νe from K decay constrained from in situ 2 Neutrino mode Dirt high energy events + SciBooNE high energy νμ constrained from in situ 1 dirt data sample event rate 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 EQE (GeV) Eν (GeV)

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 9 4 Phys. Rev. D 98, 112004 New : Dirt constraint with timing 120 Data Example RWM Trace νμ ⇡0 Decay-in-flight due to 0.3 100 ⇡0 short life time # CCQE Events Beam Trace Value (arb) Thin Target ⌫ ⇡± 80 Decay-in-flight after leaving target earance ⇡± ⌫ app single0.2 60 ν e a w the to ho 40 to leads the 0 0.1 ⇡ Decay-in-flight due to dirt illustrates 0 ⇡ short life time kground20 the in escaping Beam bac t panel er. Thick w Target as 0 Absorbed or decay- 0.5 0.51 0.52righ0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6 ⇡± photonsTime sincesho first pulse (µs) at-rest reduced neu- Dirt events: teraction ⇡± ) • in the k, trino flux tributing the trac ➡ beam-related neutrino interactions in the rocks whileFIG. 4. Zoomed-inof example of the BNB pulse microstruc- Figure 4-1: External interactions concontributing as turebac askground measured by theto RWM.the The dataνe pointsapp comeearance fromed surrounding the detector one a neutrino neutrino-modeto ⌫µ charged-current interactions in the Mini- FIG. 6. (Top) Production of dark matter and neutrinos signal.➡ timeThe shiftleft duepanel to extraillustrates flight path beforehoww particlesa neutrinov enterolume, interactionBooNE detectorelectromagneticin duringthe 2015–2016.dirt Theleads exampleobservto RWMa tracesinglewhenterms the beam hits a thin target. (Bottom) The production ho leadis plottedan by the readout value of the trace.an inof dark matter and suppression of neutrino generation when teractions the beam hits a thick target. photonthecon detectorverting in theinanalysis fiducial volume,can whileinto the right panel illustrates how fiducial ducing defined • No0 cut on the event timing within the beam spilltank Target Decay Pipe Beam Dump MiniBooNE Detector a π decaying near the illustrateswall of the tank can lead to one of theprophotons† isescaping the A. Beam o↵-target BNB simulation erting V v N (RF cavity structureExternal of 52.81 MHz)analysis the p cut 0 panel of con Be efore ⇡ fiducial volume4-1: undetected,the before converting into an electromagneticbAir Steel shoEarthwer. BooNEG4Beam was updated to include materials in • Event timing showsleft no significantin allexcess of off-bunch data The the beam line that would have changed the neutrino- w efore 50 m 0. 4m 487 m ➡ dirt constrained to better than 5σ b R mode flux (4.9)⌫ by less than a percent but are important Figure The erting the for the o↵-target beam configuration. Figure 7 shows v detectorFIG. 5. The production of dark matter in o↵-target run- a schematic of the beam-line geometry around the tar- e ning [20]. con near radius nd get, pointing out the materials that were added. An alu- Adriensignal. Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysicsat — July 2 2020 10, have to traverseyingas it undetected,enters the active detectoractiv before producing an observedatrac0f k,minum window at the end of the horn and a steel end cap the all, > with a small gap of air betweenis the end of the beam pipe photon0 deca w l and the steel beam dumpl were also added. Except for the erated would be (de)focused. For the restal of this paper, al π volume ters w windows and the−w end cap, the other materials that were en this mode of running will be denoted as− o↵-target, since added are hollowto around the beam center, and do not add defineda with respect to the detector wall, at radiuswithR0. The cut is defined−to in terms − it detector ts the beryllium target and horn wered not removed from the to thear primaryd meson production during on-target run- as en beam line. war ning.w The starting beamwith parameters for the o↵-target fiducial the ev for for erse to The decay pipe andR beam dump are buried in crushedR simulations were chosen by in situ measurements from of Evis and Rback−to−vwall to reject events with aggregate. There is a metal end cap at the downstream two multiwire planes, about one meter apart and about tra reject and four meters upstream of the target. ect end of the decayand pipe which prevents aggregate from en- boundary to to tering the, pipe. The beam dump consists of 104 inches The dark matter model does not have a charged- e resp l is cm, v wal of steelv followed by 36 inches of concrete and another thecurrent interaction component in its simplest form re- ha − bE to sulting in the assumption that the CCQE signature in with −to a1 26 inches of steel in the beam=100 direction. A detailed study of the neutrino flux comingf from the BNB in on- MiniBooNE (see Sec. V)comes does not have a dark matter back a0 close signal component. The CCQE distribution was used to R R < a a E , aand0b − Rtarget mode seen in the MiniBooNE> a , detector using the(4.9) definedback−to−wall 0b 1b vis< GEANT4forward[39]−to simulation−andwall package0BooNEG4Beamf can check the simulated o↵-target flux O↵. The nominal and o↵-target beam parameters and geometry produced 60% − all be found in Ref. [40]. On-target running consisted of kgrounds Evis −w neutrino, and antineutrinoo modes.ccurring The simulationst. were less CCQE events than measured, as shown in Fig. 8. −to updated to study the ots↵-target beam configuration and Inbac August of 2015 a remote-controlled robotic vehicle of are described below. righ was employed to survey the region between the target back cm/MeV, en reconstructedthan R ev 4-1 tified the where a0b =347.3 cm, a1b =0.595 cm/MeV,=0.595and a0f =100rejectcm, andFig.Rforward−to−wall is 1b in mis-iden that smaller a to as of This cm, safely 4-2. defined in an analogous fashion to reject evenfashionts occurring close to the boundaryrequiringwith , is olume, by γγ Fig. , =347.3 v tribution m of E e products escaping the fiduciala0b volume, as in Fig. 4-1 right.con analogous fiducial large rejected panels where an the a are ypothesis, t in k h righ As discussed in Sec. 3.5.1, a large contribution3.5.1,of mis-idenThosetified backgrounds comes reconstructed (4.10) defined escaping Sec. and of in ents. left from ν -induced NC π0 evenductsts. Those are rejected0 evby requiring that the reconstructedterms µ π wo-photon-tractop in pro NC t the 0 discussed the in π mass obtained under the Astwo-photon-track hypothesis, mγγ , is safelyfunctionsmaller than -induced under 2 , 0 νµ E e middle the true π mass, as illustrated in the top left andillustratedright panelsquadraticof Fig. 4-2.+ aThis2 from obtained as a e the a1E in mis- 0 mass 0 mass, using + requirement is applied usingπa quadraticπ function in terms of reconstructeda0 Ee, against applied 2 < er based true is m γγ illustratedow the t < as p are 0 cuts 2 2 4.3. applied, rejection 0 < mrequiremenγγ < a0 + a1Ee + a2Ee , Tab. (4.10) Those in are the en cuts giv enhance kgrounds. are bac a2 tification where a0, a1, and a2 are given in Tab. 4.3. h further , and iden -induced , a1 whic νµ Two more particle identification cuts area0 applied, as illustrated in the middle particle 4-2, other where more Fig. and bottom panels Fig. 4-2, which further enhanceo the rejection prejectower against mis- Tw panels also 126 0 0 and identified NC π and also reject other νµ-inducedbottombackgrounds.π Those cuts are based and NC tified iden 126 4 Phys. Rev. D 98, 112004 New : Dirt constraint with timing 120 Data Example RWM Trace νμ ⇡0 Decay-in-flight due to 0.3 100 ⇡0 short life time # CCQE Events Beam Trace Value (arb) Thin Target ⌫ ⇡± 80 Decay-in-flight after leaving target earance ⇡± ⌫ app single0.2 60 ν e a w the to ho 40 to leads the 0 0.1 ⇡ Decay-in-flight due to dirt illustrates 0 ⇡ short life time kground20 the in escaping Beam bac t panel er. Thick w Target as 0 Absorbed or decay- 0.5 0.51 0.52righ0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6 ⇡± photonsTime sincesho first pulse (µs) at-rest reduced neu- Dirt events: teractionthe ⇡± ) • in the k, trino flux ➡ beam-related neutrino interactions in the rockstributing of Data (stat err.) trac whileFIG.150 4. Zoomed-inMiniBooNE example preliminary of the BNB pulseBest Fit microstruc- Figure 4-1: External interactions contributing as background to the ν app+/- earance con ture as measured by the RWM.20 The datae pointsν from µ come fromed 40% of 18.75 x10 POT e Events one +/- from K surrounding the detector neutrino neutrino-mode ⌫µ charged-current interactionsνe in the Mini- FIG. 6. (Top) Production of dark matter and neutrinos Neutrino mode 0 a to νe from K olume, BooNE detector during 2015–2016. The example0 RWM trace when the beam hits a thin target. (Bottom) The production signal.➡ The left panel illustrates howwa neutrinov interactionelectromagneticin the dirt leadsπ misidobservto a singleterms time shift due to extra flight path before particles enter is plotted by the readout value of the trace.Δ → Nγ of dark matter and suppression of neutrino generation when ho lead 100an an in teractions νe dirt the beam hits a thick target. photonthecon detectorverting in theinanalysis fiducial volume,can whileinto the right panel illustratesother how fiducial ducing defined • No0 cut on the event timing within the beam spilltank Target Decay Pipe Beam Dump MiniBooNE Detector illustrates † is A. Beam o↵-target BNB simulation a π decaying near the wall of the tank can lead to one50 of theprophotons escaping the erting V v N (RF cavity structureExternal of 52.81 MHz)analysis the p cut 0 panel of con Be efore ⇡ fiducial volume4-1: undetected,the before converting into an electromagneticbAir Steel shoEarthwer. BooNEG4Beam was updated to include materials in • Event timing showsleft no significantin allexcess of off-bunch data The the beam line that would have changed the neutrino- w efore 50 m . 4m 487 m b 0 0 mode flux (4.9)⌫ by less than a percent but are important ➡ dirt constrainedThe to better than 5σ 0 R5 10 15 Figure erting the Bunch Time (ns) for the o↵-target beam configuration. Figure 7 shows v detectorFIG. 5. The production of dark matter in o↵-target run- a schematic of the beam-line geometry around the tar- e ning [20]. con near radius nd get, pointing out the materials that were added. An alu- Adriensignal. Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysicsat — July 2 2020 10, have to traverseyingas it undetected,enters the active detectoractiv before producing an observedatrac0f k,minum window at the end of the horn and a steel end cap the all, > with a small gap of air betweenis the end of the beam pipe photon0 deca w l and the steel beam dumpl were also added. Except for the erated would be (de)focused. For the restal of this paper, al π volume ters w windows and the−w end cap, the other materials that were en this mode of running will be denoted as− o↵-target, since added are hollowto around the beam center, and do not add defineda with respect to the detector wall, at radiuswithR0. The cut is defined−to in terms − it detector ts the beryllium target and horn wered not removed from the to thear primaryd meson production during on-target run- as en beam line. war ning.w The starting beamwith parameters for the o↵-target fiducial the ev for for erse to The decay pipe andR beam dump are buried in crushedR simulations were chosen by in situ measurements from of Evis and Rback−to−vwall to reject events with aggregate. There is a metal end cap at the downstream two multiwire planes, about one meter apart and about tra reject and four meters upstream of the target. ect end of the decayand pipe which prevents aggregate from en- boundary to to tering the, pipe. The beam dump consists of 104 inches The dark matter model does not have a charged- e resp l is cm, v wal of steelv followed by 36 inches of concrete and another thecurrent interaction component in its simplest form re- ha − bE to sulting in the assumption that the CCQE signature in with −to a1 26 inches of steel in the beam=100 direction. A detailed study of the neutrino flux comingf from the BNB in on- MiniBooNE (see Sec. V)comes does not have a dark matter back a0 close signal component. The CCQE distribution was used to R R < a a E , aand0b − Rtarget mode seen in the MiniBooNE> a , detector using the(4.9) definedback−to−wall 0b 1b vis< GEANT4forward[39]−to simulation−andwall package0BooNEG4Beamf can check the simulated o↵-target flux O↵. The nominal and o↵-target beam parameters and geometry produced 60% − all be found in Ref. [40]. On-target running consisted of kgrounds Evis −w neutrino, and antineutrinoo modes.ccurring The simulationst. were less CCQE events than measured, as shown in Fig. 8. −to updated to study the ots↵-target beam configuration and Inbac August of 2015 a remote-controlled robotic vehicle of are described below. righ was employed to survey the region between the target back cm/MeV, en reconstructedthan R ev 4-1 tified the where a0b =347.3 cm, a1b =0.595 cm/MeV,=0.595and a0f =100rejectcm, andFig.Rforward−to−wall is 1b in mis-iden that smaller a to as of This cm, safely 4-2. defined in an analogous fashion to reject evenfashionts occurring close to the boundaryrequiringwith , is olume, by γγ Fig. , =347.3 v tribution m of E e products escaping the fiduciala0b volume, as in Fig. 4-1 right.con analogous fiducial large rejected panels where an the a are ypothesis, t in k h righ As discussed in Sec. 3.5.1, a large contribution3.5.1,of mis-idenThosetified backgrounds comes reconstructed (4.10) defined escaping Sec. and of in ents. left from ν -induced NC π0 evenductsts. Those are rejected0 evby requiring that the reconstructedterms µ π wo-photon-tractop in pro NC t the 0 discussed the in π mass obtained under the Astwo-photon-track hypothesis, mγγ , is safelyfunctionsmaller than -induced under 2 , 0 νµ E e middle the true π mass, as illustrated in the top left andillustratedright panelsquadraticof Fig. 4-2.+ aThis2 from obtained as a e the a1E in mis- 0 mass 0 mass, using + requirement is applied usingπa quadraticπ function in terms of reconstructeda0 Ee, against applied 2 < er based true is m γγ illustratedow the t < as p are 0 cuts 2 2 4.3. applied, rejection 0 < mrequiremenγγ < a0 + a1Ee + a2Ee , Tab. (4.10) Those in are the en cuts giv enhance kgrounds. are bac a2 tification where a0, a1, and a2 are given in Tab. 4.3. h further , and iden -induced , a1 whic νµ Two more particle identification cuts area0 applied, as illustrated in the middle particle 4-2, other where more Fig. and bottom panels Fig. 4-2, which further enhanceo the rejection prejectower against mis- Tw panels also 126 0 0 and identified NC π and also reject other νµ-inducedbottombackgrounds.π Those cuts are based and NC tified iden 126 4 Phys. Rev. D 98, 112004 New : Dirt constraint with timing 120 Data Example RWM Trace νμ ⇡0 Decay-in-flight due to 0.3 100 ⇡0 short life time # CCQE Events Beam Trace Value (arb) Thin Target ⌫ ⇡± 80 Decay-in-flight after leaving target earance ⇡± ⌫ app single0.2 60 ν e a w the to ho 40 to leads the 0 0.1 ⇡ Decay-in-flight due to dirt illustrates 0 ⇡ short life time kground20 the in escaping Beam bac t panel er. Thick w Target as 0 Absorbed or decay- 0.5 0.51 0.52righ0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6 ⇡± photonsTime sincesho first pulse (µs) at-rest reduced neu- Dirt events: teractionthe ⇡± ) • in the k, trino flux ➡ beam-related neutrino interactions in the rockstributing of Data (stat err.) trac whileFIG.150 4. Zoomed-inMiniBooNE example preliminary of the BNB pulseBest Fit microstruc- Figure 4-1: External interactions contributing as background to the ν app+/- earance con ture as measured by the RWM.20 The datae pointsν from µ come fromed 40% of 18.75 x10 POT e Events one +/- from K surrounding the detector neutrino neutrino-mode ⌫µ charged-current interactionsνe in the Mini- FIG. 6. (Top) Production of dark matter and neutrinos Neutrino mode 0 a to νe from K olume, BooNE detector during 2015–2016. The example0 RWM trace when the beam hits a thin target. (Bottom) The production signal.➡ The left panel illustrates howwa neutrinov interactionelectromagneticin the dirt leadsπ misidobservto a singleterms time shift due to extra flight path before particles enter is plotted by the readout value of the trace.Δ → Nγ of dark matter and suppression of neutrino generation when ho lead 100an an in teractions νe dirt the beam hits a thick target. photonthecon detectorverting in theinanalysis fiducial volume,can whileinto the right panel illustratesother how fiducial ducing defined • No0 cut on the event timing within the beam spilltank Target Decay Pipe Beam Dump MiniBooNE Detector illustrates † is A. Beam o↵-target BNB simulation a π decaying near the wall of the tank can lead to one50 of theprophotons escaping the erting V v N (RF cavity structureExternal of 52.81 MHz)analysis the p cut 0 panel of con Be efore ⇡ fiducial volume4-1: undetected,the before converting into an electromagneticbAir Steel shoEarthwer. BooNEG4Beam was updated to include materials in • Event timing showsleft no significantin allexcess of off-bunch data The the beam line that would have changed the neutrino- w efore 50 m . 4m 487 m b 0 0 mode flux (4.9)⌫ by less than a percent but are important ➡ dirt constrainedThe to better than 5σ 0 R5 10 15 Figure erting the Bunch Time (ns) for the o↵-target beam configuration. Figure 7 shows v detectorFIG. 5. The production of dark matter in o↵-target run- a schematic of the beam-line geometry around the tar- e ning [20]. con near radius nd get, pointing out the materials that were added. An alu- Adriensignal. Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysicsat — July 2 2020 10, have to traverseyingas it undetected,enters the active detectoractiv before producing an observedatrac0f k,minum window at the end of the horn and a steel end cap the all, > with a small gap of air betweenis the end of the beam pipe photon0 deca w l and the steel beam dumpl were also added. Except for the erated would be (de)focused. For the restal of this paper, al π volume ters w windows and the−w end cap, the other materials that were en this mode of running will be denoted as− o↵-target, since added are hollowto around the beam center, and do not add defineda with respect to the detector wall, at radiuswithR0. The cut is defined−to in terms − it detector ts the beryllium target and horn wered not removed from the to thear primaryd meson production during on-target run- as en beam line. war ning.w The starting beamwith parameters for the o↵-target fiducial the ev for for erse to The decay pipe andR beam dump are buried in crushedR simulations were chosen by in situ measurements from of Evis and Rback−to−vwall to reject events with aggregate. There is a metal end cap at the downstream two multiwire planes, about one meter apart and about tra reject and four meters upstream of the target. ect end of the decayand pipe which prevents aggregate from en- boundary to to tering the, pipe. The beam dump consists of 104 inches The dark matter model does not have a charged- e resp l is cm, v wal of steelv followed by 36 inches of concrete and another thecurrent interaction component in its simplest form re- ha − bE to sulting in the assumption that the CCQE signature in with −to a1 26 inches of steel in the beam=100 direction. A detailed study of the neutrino flux comingf from the BNB in on- MiniBooNE (see Sec. V)comes does not have a dark matter back a0 close signal component. The CCQE distribution was used to R R < a a E , aand0b − Rtarget mode seen in the MiniBooNE> a , detector using the(4.9) definedback−to−wall 0b 1b vis< GEANT4forward[39]−to simulation−andwall package0BooNEG4Beamf can check the simulated o↵-target flux O↵. The nominal and o↵-target beam parameters and geometry produced 60% − all be found in Ref. [40]. On-target running consisted of kgrounds Evis −w neutrino, and antineutrinoo modes.ccurring The simulationst. were less CCQE events than measured, as shown in Fig. 8. −to updated to study the ots↵-target beam configuration and Inbac August of 2015 a remote-controlled robotic vehicle of are described below. righ was employed to survey the region between the target back cm/MeV, en reconstructedthan R ev 4-1 tified the where a0b =347.3 cm, a1b =0.595 cm/MeV,=0.595and a0f =100rejectcm, andFig.Rforward−to−wall is 1b in mis-iden that smaller a to as of This cm, safely 4-2. defined in an analogous fashion to reject evenfashionts occurring close to the boundaryrequiringwith , is olume, by γγ Fig. , =347.3 v tribution m of E e products escaping the fiduciala0b volume, as in Fig. 4-1 right.con analogous fiducial large rejected panels where an the a are ypothesis, t in k h righ As discussed in Sec. 3.5.1, a large contribution3.5.1,of mis-idenThosetified backgrounds comes reconstructed (4.10) defined escaping Sec. and of in ents. left from ν -induced NC π0 evenductsts. Those are rejected0 evby requiring that the reconstructedterms µ π wo-photon-tractop in pro NC t the 0 discussed the in π mass obtained under the Astwo-photon-track hypothesis, mγγ , is safelyfunctionsmaller than -induced under 2 , 0 νµ E e middle the true π mass, as illustrated in the top left andillustratedright panelsquadraticof Fig. 4-2.+ aThis2 from obtained as a e the a1E in mis- 0 mass 0 mass, using + requirement is applied usingπa quadraticπ function in terms of reconstructeda0 Ee, against applied 2 < er based true is m γγ illustratedow the t < as p are 0 cuts 2 2 4.3. applied, rejection 0 < mrequiremenγγ < a0 + a1Ee + a2Ee , Tab. (4.10) Those in are the en cuts giv enhance kgrounds. are bac a2 tification where a0, a1, and a2 are given in Tab. 4.3. h further , and iden -induced , a1 whic νµ Two more particle identification cuts area0 applied, as illustrated in the middle particle 4-2, other where more Fig. and bottom panels Fig. 4-2, which further enhanceo the rejection prejectower against mis- Tw panels also 126 0 0 and identified NC π and also reject other νµ-inducedbottombackgrounds.π Those cuts are based and NC tified iden 126 0 Data (stat err.) 150 Best Fit New : Radial distribution disfavors dirt and π background+/- νe from µ

Events +/- νe from K 0 νe from K π0 misid 800 Δ → Nγ Other dirt • Improved100 statistics allow fordirt more other N 0 700 Δ→ γ π misid

Events distributions to be investigated Intrinsic νe Best Fit 50 600 Data • Radial distribution shows that the excess is MiniBooNE preliminary spread evenly within the volume of the 0 500 20 0 5 10 15 18.75 x 10 POT Bunch Time (ns) FIG. 16:detector The bunch timing for data events in neutrino mode compared to the expected background. 400 Neutrino mode 0 Almost• An all ofexcess the excess data of events π occur, background as expected, within thewould ﬁrst 8 ns ofhave the bunch timing. 300 This datapeaked sample uses near events collected the withedge the new ﬁber timing system and represents about 40% of the(higher entire neutrino probability mode sample. of missing one of the γ) 200

TABLE• Similar II: The result approach of log-likelihood shape-onlyto SNO’s fits to the CC/NC radial distribution, constraint assuming only 100 statistical• Second errors, where best di↵erent candidate processes are normalized : NC toγ explain background the observed event excess. 0 The two-neutrino hypothesis fits the radial distribution best with a 2 =6.6/9ndf,whiletheNC⇡0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 3 (R/500cm) hypothesis has a much worse fitExcess with a 2 = 20 .shape8/9ndf. tests Hypothesis 2/9ndf 2 to Best Fit

2⌫ Oscillations 6.6 0 NC ⇡0 Background 20.8 14.2 NC Background 7.4 0.8

Intrinsic ⌫e Background 16.1 9.5 External Background 57.5 50.9 0 2 Example of a π mis-ID Other Background 8.6 2.0

ground estimate has been confirmed by measuring the absolute time of signalStatistics events relative only χ Figure 4-1: External interactions contributing as background to the ν appearance nd Adrien Hourlier — The XXIX International Conferencee on Neutrino Physics toand the protonAstrophysics beam microstructure — July (52.81 2 2020 MHz extraction frequency), which corresponds to11 signal. The left panel illustrates how a neutrino interaction in the dirt leads to a single buckets of beam approximately every 18.9 ns. Fig. 16 shows that the event excess peaks in photon converting in the analysis fiducial volume, while the right panel illustrates how the 8 ns window associated with beam bunch time, as expected from neutrino events in the a π0 decaying near the wall of the tank can lead to one of the photons escaping the detector, and is inconsistent with external neutrino events or beam-o↵ events, which would fiducial volume undetected, before converting into an electromagnetic shower. be approximately flat in time. Also, the observed background level outside of the beam agrees well with the predicted background estimate. The timing reconstruction performed have to traverse as it enters the active detector before producing an observed track, 16 defined with respect to the detector wall, at radius R0. The cut is defined in terms of Evis and Rback−to−wall to reject events with

R < a a E , and R > a , (4.9) back−to−wall 0b − 1b vis forward−to−wall 0f

where a0b =347.3 cm, a1b =0.595 cm/MeV, and a0f =100 cm, and Rforward−to−wall is defined in an analogous fashion to reject events occurring close to the boundary with products escaping the fiducial volume, as in Fig. 4-1 right. As discussed in Sec. 3.5.1, a large contribution of mis-identified backgrounds comes

0 from νµ-induced NC π events. Those are rejected by requiring that the reconstructed 0 π mass obtained under the two-photon-track hypothesis, mγγ , is safely smaller than the true π0 mass, as illustrated in the top left and right panels of Fig. 4-2. This requirement is applied using a quadratic function in terms of reconstructed Ee,

2 2 0 < mγγ < a0 + a1Ee + a2Ee , (4.10)

where a0, a1, and a2 are given in Tab. 4.3. Two more particle identiﬁcation cuts are applied, as illustrated in the middle and bottom panels Fig. 4-2, which further enhance the rejection power against mis-

0 identiﬁed NC π and also reject other νµ-induced backgrounds. Those cuts are based

126 0 Data (stat err.) 150 Best Fit New : Radial distribution disfavors dirt and π background+/- νe from µ

Events +/- νe from K 0 νe from K π0 misid 800 Δ → Nγ Other dirt • Improved100 statistics allow fordirt more other N 0 700 Δ→ γ π misid

2⌫ Oscillations 6.6 0 NC ⇡0 Background 20.8 14.2 NC Background 7.4 0.8

Intrinsic ⌫e Background 16.1 9.5 External Background 57.5 50.9 0 2 Example of a π mis-ID Other Background 8.6 2.0

R < a a E , and R > a , (4.9) back−to−wall 0b − 1b vis forward−to−wall 0f

2 2 0 < mγγ < a0 + a1Ee + a2Ee , (4.10)

0 identiﬁed NC π and also reject other νµ-induced backgrounds. Those cuts are based

126 New : Radial distribution disfavors dirt and π0 background

Radius [cm] Radius [cm] Radius [cm] 0 200 400 0 200 400 0 200 400

160 Best Fit 160 dirt 160 π0 misid Data - BkgMC Data - BkgMC Data - BkgMC 0 140 Oscillation 140 Dirt 140 π

120 120 120

Events [MC Rescaled] 100 Events [MC Rescaled] 100 Events [MC Rescaled] 100

80 80 80

60 60 60

40 40 40

20 20 20

0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 (R/500cm)3 (R/500cm)3 (R/500cm)3

• Scaling dirt and π0 backgrounds to fit the excess • The "best fit" corresponds to a 2ν oscillation with a sterile neutrino, and fit the observed excess better than a single background

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 12 New : Radial distribution disfavors dirt and π0 background

800 Other dirt

N 0 700 Δ→ γ π misid Events Intrinsic νe Best Fit

600 Data MiniBooNE preliminary • Investigate if the excess is caused by 500 18.75 x 1020 POT events spilling in or spilling out of the 400 Neutrino mode fiducial volume by varying the fiducial cut 300 • Excess remains significant when rejecting outer layers of the detector : excess within 200 the detector volume 100

0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 (R/500cm)3 Fiducial cut Excess signiﬁcance R<500 cm 560.6 119.64.7 ± R<400 cm 472.6 81.75.8 ± R<300 cm 208.8 40.35.2

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 ± Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 13 NC Δ → Nγ resonance • The production of NC Δ → Nγ is highly correlated to the measurement of NC π0 • Same probability of a NC interaction, the difference in final state is the relative rate of resonant production. ➡ Our predicted single γ/π0 ratio is ~0.9%, which takes into account pion absorption in the nucleus, higher mass resonances, coherent scattering, and non-resonant processes • Apply the same correction and fractional uncertainties to NC Δ → Nγ as NC π0 • Additional uncertainty to account for final state interactions (FSI) • The single gamma estimate agrees with theory

Phys. Lett. B 740 (2015) 16-22

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 14 New : Constraints from the beam dump run 14 2. Electron Reduction in neutrino production 20 Data • The signal distribution for this fit was defined as the ν from µ+/- 18 e • No changeevents to neutral that pass meson⌫-e cuts production with cos ✓e >and0.99. proton The from K+/- Events/GeV νe Events/MeV fit was a binned extended maximum-likelihood fit in 0 bremsstrahlung to first order 16 e νe from K three dimensions, E , cos ✓e, and bunch time, with a ➡ Can directly test modelsvis that predict the oscillation π0 misid single nuisance parameter to control the overall nor- 14 excessmalization does not of predictedscale as neutrino neutrino events. scales The region of Δ→ Nγ (e.g. vector0.9 < cos portal,✓e < 0 .inelastic99 was the dark control matter, region to …) constrain 12 dirt background events. Because of the well-defined control other D 98, 112004 Phys. Rev. 10 region, data from neutrino and antineutrino modes were • Expectednot : used 35.5±7.4 to constrain excess the prediction. events Two in [200,1250]⌫-e events were MeV 8 for a POT-scalingmeasured in excess o↵-target mode. After constraining the ⌫-e background the predicted number of events is 2.4 1.5. In 6 • Measuredthe : signal 6 events, region, zero8.8 eventsbackgrounds were measured expected with± a con- ➡ -2.8 excessstrained events prediction of 0.4 events. Statistical error dom- 4 inates the total error in the constrained prediction. No ➡ Explanation that scale only by POT instead of 2 dark matter candidate events were measured. neutrinoSystematic production uncertainties are ruled were out not at included 4.6σ in the fit 0 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 3 as the predicted number of background events has a sta- QE E (GeV) See Jordantistical et relativeal. Phys. uncertainty Rev. Lett. much 122,081801 greater than the (2019) pre- ν • dicted systematics, especially when considering some of QE for constraintsthe systematic on diff uncertaintieserent models are constrained by the con- FIG. 19. The E⌫ distribution for events that pass the ⌫e trolled region. The normalization parameter is fixed dur- oscillation cuts. Data comes from o↵-target mode. ing fitting so the data/fake data and null predictions are Adrien Hourlier —the The same XXIX for International the number Conference of events in on the Neutrino control Physics region. and Astrophysics — July 2nd 2020 15 When generating the fake data for the electron analysis, given in Ref. [29]. each bin is assumed to be independent with an underlying The simulation loop begins by determining the maxi- Poisson distribution with a mean equal to the predicted mum probability in the angular momentum distribution plus dark matter distribution. of each production channel, as it is not known analyt- ically [29]. This maximum is used in an acceptance- rejection algorithm to sample the angular momentum 3. Neutrino oscillation events in o↵-target mode distribution of each channel when generating dark matter trajectories. The total number of dark matter parti- As previously stated about 40% of the events that pass cles expected from each production channel is then calcu- ⌫-e cuts also pass neutrino oscillation cuts [23]. Figure 19 lated, and the output events are split between these chan- QE nels according to their fraction of the total dark matter shows the E⌫ distribution [Eq. (5) is used with the results from the electron track fit and EB = 0] for o↵-target production rate. running. Simulation predicted 8.8 events assuming there For the case of pseudoscalar meson decays, meson 4- are no oscillations. Six events were measured. All but momenta and positions are generated in the MiniBooNE one of the observed events were above 475 MeV. Impli- beam line by sampling an event list generated by the cations of this data are discussed in Sec. VII B. BooNEG4Beam simulations; see Sec. III. For the case of proton bremsstrahlung, the dark matter is simulated to occur at the front of the beam dump. VI. CONFIDENCE LEVEL LIMITS ON LIGHT The simulation attempts a given dark matter scat- DARK MATTER THEORY tering event for each dark matter trajectory from the previous step found to intersect with the MiniBooNE A fixed-target dark matter Monte Carlo simulator, Bd- detector. Possible interactions are elastic-nucleon (0⇡), NMC, was used to simulate the energy and position dis- elastic-electron, and inelastic-nucleon producing a single 0 0 tributions of the expected dark matter scattering signal pion (1⇡ if a ⇡ is produced, and 1⇡± if a ⇡± is pro- in the MiniBooNE detector [29]. There are a number of duced). The neutrino detector simulation, discussed in production channels in fixed-target experiments, though Sec. IV A, was used to simulate the response of the detec- often one will dominate for a given set of dark matter tor. This simulation used neutrino events generated by model parameters. For MiniBooNE, the decay of two NUANCE and contained the nuclear model and all final- pseudoscalar mesons, the ⇡0 and the ⌘ were considered, state interactions. We define the weight of each neutrino as well as production through proton bremsstrahlung simulated event as the ratio N (!) /N⌫ (!), where N (!) 2 plus vector mixing up to mV =1GeVc . The param- is the number of true interactions as a function of energy eter values and equations used in the simulation were transfer ! that generated the simulated event. N (!)

QE Eν (GeV) Excess interpretation in a sterile neutrino hypothesis

Data (stat err.) +/- νe from µ 7 +/- νe from K 0 νe from K Events/MeV 0 6 π misid Δ → Nγ dirt 5 other Constr. Syst. Error Best Fit 4 MiniBooNE preliminary 3 ν data : 18.75 x 1020 POT Combined (ν + ν¯) ﬁt 2

1 MiniBooNE preliminary Combined ﬁt 0 (ν + ν¯) 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 QE Eν (GeV)

• Combined (ν + ν¯) ﬁt • LSND and MiniBooNE points follow the same best ﬁt 2ν oscillation interpretation

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 16 1.2 st 20 1 νe 6.46×10 POT 1.0 nd 20 2 νe 6.38×10 POT 3rd ν 5.91×1020 POT 0.8 e

0.6 0.4

Excess Events/MeV 0.2 Preferred regions in sterile0.0 neutrino hypothesis 2 ) 2 10 • −0.2Neutrino mode excess 4.7σ, 68% CL 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 (eV 2 90% CL • Neutrino+Anti-neutrino modes excess : 4.8σ EQE (GeV) m ν Δ 95% CL 99% CL 2.5 3σ CL ν : 18.75×1020 POT 10 e : 11.27 1020 POT 4σ CL 2.0 νe × KARMEN2 90% CL (0.807,0.043 eV2) best fit 1.5 2 OPERA (0.01, 0.35 eV ) 90% CL MiniBooNE preliminary 1.0

Excess Events/MeV Combined ﬁt MiniBooNE preliminary (ν + ν¯) ν : 18.75 1020 POT 1 Combined (ν + ν¯) ﬁt 0.5 ν¯ : 11.27 1020 POT ν : 18.75 1020 POT ν¯ : 11.27 1020 POT 0.0

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 3.0 QE Eν (GeV) 10−1 LSND 90% CL Neutrino + Anti-Neutrino Mode LSND 99% CL (Δm2, sin2 2θ) = (0.043 eV2, 0.807) −2 2 10 χ /ndf = 21.7/15.5 (prob = 12.3%) 10−3 10−2 10−1 1 sin22θ

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 17 New : cos(θ) VS Evis Distributions for νe samples Data Prediction Data-Prediction

θ 1 60 θ 1 60 θ 1 20 cos cos cos 50 50 15 0.5 0.5 0.5 40 40 10

0 30 0 30 0 5 0 20 20 −0.5 MiniBooNE preliminary −0.5 MiniBooNE preliminary −0.5 MiniBooNE preliminary −5 Data (18.75 x1020) 10 Backgrounds 10 Data - Backgrounds Neutrino mode Neutrino mode Neutrino mode −10 1 0 1 0 1 − 500 1000 − 500 1000 − 500 1000 Evis [MeV] Evis [MeV] Evis [MeV] Best Fit

θ 1 6 cos 5 • Compare data to backgrounds in the (Evis,cosθ) space 0.5 • We believe that these distributions will guide theorists to 4 explain our data 0 3 These data will soon be provided in a data release with 2 • −0.5 MiniBooNE preliminary systematic errors Best Fit prediction 1 Neutrino mode 1 0 − 500 1000 Evis [MeV] Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 18 New : cos(θ) VS Evis Distributions for νe samples Data

θ 1 60 cos 50 0.5 Easier to visualize the excess in 1D plots of cosθ for 40 slices of Evis 0 30

20 −0.5 Data 10

1 0 − 500 1000 Evis [MeV] Prediction

θ 1 60 cos 50 0.5 40

0 30

20 −0.5 Background 10

1 0 − 500 1000 Evis [MeV]

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 19 New : cos(θ) VS Evis Distributions for νe samples Data

θ 1 60 cos 50 0.5 Easier to visualize the excess in 1D plots of cosθ for 40 slices of Evis 0 30

20 −0.5 Data 10

1 0 − 500 1000 Evis [MeV] Prediction

θ 1 60 cos 50 0.5 40

0 30

20 −0.5 Background 10

1 0 − 500 1000 Evis [MeV]

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 19 New : cos(θ) VS Evis Distributions for νe samples Data

θ 1 60 cos 50 0.5 Easier to visualize the excess in 1D plots of cosθ for 40 slices of Evis 0 30

20 −0.5 Data 60 10

−1 0 500 1000 40 Evis [MeV] Prediction 20

θ 1 60 cos 50 0 0.5 −1 −0.5 0 0.5 1 40 cos θ

0 30

20 −0.5 Background 10

1 0 − 500 1000 Evis [MeV]

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 19 New : cos(θ) VS Evis Distributions for νe samples Data

θ 1 60 cos 50 0.5 Easier to visualize the excess in 1D plots of cosθ for 40 slices of Evis 0 30

20 −0.5 Data 10

1 0 − 500 1000 Evis [MeV] Prediction

θ 1 60 cos 50 0.5 40

0 30

20 −0.5 Background 10

1 0 − 500 1000 Evis [MeV]

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 19 New : cos(θ) VS Evis Distributions for νe samples

125 MeV 175 MeV 225 MeV 275 MeV 325 MeV 375 MeV 425 MeV 475 MeV 525 MeV 575 MeV 625 MeV 675 MeV 725 MeV 775 MeV 825 MeV 875 MeV 925 MeV 975 MeV 1025 MeV 1075 MeV 1125 MeV 1175 MeV 1225 MeV Data 6060 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 Backgrounds

4040 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 Events

2020 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1 cos θ • The excess at low energy occurs across a wide range of cos θ

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 20 Summary

• MiniBooNE presented a full analysis of 17 years of data taking • The event excess has remained stable in shape and magnitude over the different data releases • We now have a 4.7σ significance in neutrino mode only, and a 4.8 σ significance in a combined (ν + ν¯) analysis for our nominal cut of R<500 cm. • Several explanations for the excess are disfavored • NC π0 • Dirt event • Dark matter run ruled out non-neutrino-related beam backgrounds • In the spirit of responding to requests for more information, we are now in a position to provide additional information on timing, visible energy VS angle of lepton scatter, and radius. • Further studies under way to better understand the excess, including investigating Meson Exchange Currents, stay tuned!

preprint available at arxiv:2006.16883

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 21 Thank you! Constraining Δ->Nγ background from NCπ0 background

• NC π0 produced by : - Δ production in 12C (52.2%) - Δ production in H2 (15.1%) - Coherent scattering on 12C (12.5%) - Coherent scattering H2(3.1%) - higher mass resonance (12.9%) - non resonant BG (4.2%)

• 2/3 of Δ decay in π0 • 62.5% π0 can escape the nucleus • Δ->Nγ BR = 0.60% for 12C • Δ->Nγ BR = 0.68% for H2

NΔ→Nγ = (0.151 × 0.0068 × 1.5) + (0.522 × 0.0060 × 1.5)/0.625 = 0.0091 NNCπ0

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 23 Theoretical Estimates for NC-γ production Agree well with MiniBooNE Estimates

Phys. Lett. B 740 (2015) 16-22

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 24 Theoretical Estimates for NC-γ production Agree well with MiniBooNE Estimates

Phys. Lett. B 719(2013) 409-414

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 25 Excess for cosθ VS Evis plot

125 MeV 175 MeV 225 MeV 275 MeV 325 MeV 375 MeV 425 MeV 475 MeV 525 MeV 575 MeV 625 MeV 675 MeV 725 MeV 775 MeV 825 MeV 875 MeV 925 MeV 975 MeV 1025 MeV 1075 MeV 1125 MeV 1175 MeV 1225 MeV

3030 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30

2020 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 (data-backgrounds)

1010 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 Events

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

-10−10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10 −10

-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1-1 0 1 cos θ

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 26 Neutrino only ﬁt 2 )

2 10 (eV 2

Data (stat err.) m

+/- Δ νe from µ 68% 7 +/- νe from K Data - expected background 0 2 90% νe from K Events/MeV 0 6 π misid 95% Δ → Nγ 99% dirt 10 5 other Excess Events/MeV 1.5 Best Fit 3σ Constr. Syst. Error 4 Best Fit σ 4 5σ 1 6σ 3 MiniBooNE preliminary MiniBooNE preliminary data : 18.75 x 1020 POT data : 18.75 x 1020 POT ν 0.5 ν 2 ν-only ﬁt ν-only ﬁt χ2-based contours, 1 1 no fake data 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 3.0 QE QE Eν (GeV) Eν (GeV)

ν-only fit 10−1 ν data : 18.75 x 1020 POT Only a small difference with respect to the MiniBooNE preliminary • (ν + ν¯) (sin22θ, Δm2) = (0.8426,0.039 eV2) combined fit BF 2 2 2 • nu-only : (sin 2θ, Δm ) = (0.843, 0.039 eV ) 10−2 10−3 10−2 10−1 1 2 2 2 2 • (ν + ν¯) : (sin 2θ, Δm ) = (0.807, 0.043 eV ) sin 2θ

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 27 VOLUME 89, NUMBER 1PHYSICALREVIEWLETTERS1JULY 2002

3 cosu , and ͑R͞RAV ͒ , derived from Monte Carlo calcu- lationsØ generated assuming no flavor transformation and the standard 8B spectral shape [6]. Background event pdfs are included in the analysis with fixed amplitudes determined by the background calibration. The extended maximum likelihood method used in the signal decompo- 161.9 126.4 sition yields 1967.7260.9 CC events, 263.6225.6 ES events, 149.5 and 576.5248.9 NC events [7], where only statistical uncertainties are given. Systematic uncertainties on fluxes derived by repeating the signal decomposition with per- turbed pdfs (constrained by calibration data) are shown in Table II. Normalized to the integrated rates above the kinetic en- Radial distribution in SNO 8 ergy threshold of Teff $ 5 MeV, the flux of B neutrinos measured with each reaction in SNO, assuming the stan- Phys. Rev. Lett. 89.011301 (2002) dard spectrum shape [6] is (all fluxes are presented in units of 106 cm22 s21) fSNO ෇ 1.7610.06͑stat͒10.09͑syst͒ , • NC and CCCC contributions20.05 have 20.09 SNO 10.24 10.12 differencef R-dependencies,ES ෇ 2.3920.23͑ statone͒2 can0.12 ͑syst͒ , use the radial distribution to estimate fSNO ෇ 5.0910.44͑stat͒10.46͑syst͒ . the relative NCcontribution2 0.43of each 20.43 Electronprocess. neutrino cross sections are used to calculate all •fluxes.SNO Theis only CC one and of ES the results many reported examples here are consistent withof experiment the earlier SNO using results radial [2] position for Teff to$ 6.75 MeV. The excessconstrain of the components NC flux over of thetheir CC data and ES fluxes implies neutrino flavor transformations. A simple change of variables resolves the data directly into electron ͑fe͒ and nonelectron ͑fmt͒ components [9], 10.05 10.09 fe ෇ 1.7620.05͑stat͒20.09͑syst͒ , 10.45 10.48 fmt ෇ 3.4120.45͑stat͒20.45͑syst͒ , Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 28 assuming the standard 8B shape. Combining the statistical and systematic uncertainties in quadrature, fmt 10.66 is 3.4120.64, which is 5.3s above zero, providing strong evidence for flavor transformation consistent with neutrino oscillations [10,11]. Adding the Super- Kamiokande ES measurement of the 8B flux [12] SK 10.08 fES ෇ 2.32 6 0.03͑stat͒20.07͑syst͒ as an additional 10.65 constraint, we find fmt ෇ 3.4520.62, which is 5.5s above zero. Figure 3 shows the flux of nonelectron flavor active neutrinos vs the flux of electron neutrinos deduced from the SNO data. The three bands represent the one standard deviation measurements of the CC, ES, and NC rates. The error ellipses represent the 68%, 95%, and 99% joint probability contours for f and f . FIG. 2 (color). (a) Distribution of cosu for R # 550 cm. e mt (b) Distribution of the volume weightedØ radial variable Removing the constraint that the solar neutrino energy 3 ͑R͞RAV ͒ . (c) Kinetic energy for R # 550 cm. Also shown spectrum is undistorted, the signal decomposition is re- 3 are the Monte Carlo predictions for CC, ES, and NC 1 bkgd peated using only the cosu and ͑R͞RAV ͒ information. neutron events scaled to the fit results, and the calculated The total flux of active 8BØ neutrinos measured with the spectrum of Cherenkov background (bkgd) events. The dashed NC reaction is lines represent the summed components, and the bands show SNO 11.57 10.55 61s uncertainties. All distributions are for events with fNC ෇ 6.4221.57͑stat͒20.58͑syst͒ , T $ 5 MeV. eff which is in agreement with the shape constrained value flux, the data are resolved into contributions from CC, above and with the standard solar model (SSM) prediction 8 11.01 ES, and NC events above threshold using pdfs in Teff , [13] for B, fSSM ෇ 5.0520.81. 011301-4 011301-4 8

3 3 3 ×10 ×10 ×10 5 (a) 14 Data Events Events 10 +/- ν e from µ +/- Events 4 νµ CCQE 12 ν e from K 0 ν e from K 0 data π misid (b) 10 Δ→ Nγ prediction 3 8 dirt 8 other

6 2

6 4 1 2

0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 ln(L /L ) ln(L /L ) 4 π0 e π0 e

3 3 ×10 ×10

1.6 1.2 2 Events Events 1.4 1 1.2 (c) (d) 0.8 1 0 0.5 1 1.5 2 2.5 3 fit ratio 0.8 0.6 1.5 QE Eν [GeV] 0.6 Data/MC 0.4

1 0.4

0.2 0.5 0.2

0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 0 ln(L /L ) ln(L /L ) 0 0.5 1 1.5 2 2.5 3 π0 e π0 e EQE[GeV] ν FIG. 10: Comparisons between the data and simulation for FIG. 8: An absolute comparison of the reconstructed neutrino the electron-pion likelihood distribution after successive cuts energy distribution for CCQE events between the neutrino are applied: (a) no PID cut, (b) electron-muon likelihood data (12.84 1020 POT) and the simulation (top). Also shown cut, (c) electron-muon plus electron-pion likelihood cuts, and is the ratio⇥ between the data and Monte Carlo simulation (d) electron-muon plus electron pion likelihood cuts plus a (bottom). gamma-gamma mass cut. The vertical lines in the ﬁgures PID variables : μ/e logshow likelihood the range of energy-dependent cut values.

3 3 ×10 ×10 3 3 (a) ×10 ×10 Data (a) Events 10 Events 9 +/- Data 7 ν e from µ 5 Events Events +/- from +/- ν e from K 8 ν e µ +/- 0 from K 8 ν e from K ν e Other 245.606 dirt 106.55 0 0 π misid 4 (b) 7 ν e from K 0 N π misid (b) 0 Δ→ γ Δ→ Nγ 255.785 π misid 719.89 dirt 6 Δ→ Nγ 6 2500 other 3 Events dirt 5 4 other Intrinsic νe 990.48 Best Fit 478.45

4 4 2 Data 2881.003 3 2000 2

arxiv:1805.12028[hep-ex] 2 1 2 MiniBooNE preliminary 1 1 0 0 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 ln(L /L ) ln(L /L ) 0 0 20 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 35018.75400 x 10 POT 2 2 1500 mγ γ (MeV/c ) mγ γ (MeV/c )

400 Neutrino mode

Events 350 Events 250 Events 600 1000 Events 600 300 200 (c) (d) 500 500 250 (c) (d)

150 400 400 200 500 300 300 150 100

100 200 200 50 50 0 100 0 1000.2 0.4 0.6 0.8 1 0 0 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 electron/muon log likelihood ln(L /L ) ln(L /L ) 0 0 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 2 2 mγ γ (MeV/c ) mγ γ (MeV/c ) FIG. 9: Comparisons between the data and simulation for FIG. 11: Comparisons between the data and simulation for the electron-muon likelihood distribution after successive cuts the gamma-gamma mass distribution after successive cuts are applied: (a) no PID cut, (b) electron-muon likelihood are applied: (a) no PID cut, (b) electron-muon likelihood cut, (c) electron-muon plus electron-pion likelihood cuts, and cut, (c) electron-muon plus electron-pion likelihood cuts, and (d) electron-muon plus electron pion likelihood cuts plus a (d) electron-muon plus electron pion likelihood cuts plus a gamma-gamma mass cut. The vertical lines in the ﬁgures gamma-gamma mass cut. The vertical lines in the ﬁgures show the range of energy-dependent cut values. show the range of energy-dependent cut values.

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 29 PID variables : π/e log8 likelihood

3 3 3 ×10 ×10 ×10 5 (a) 14 Data Events Events 10 +/- ν e from µ +/- Events 4 νµ CCQE 12 ν e from K 0 ν e from K Other 245.60 dirt 106.55 0 1800 data π misid (b) 10 Δ→ Nγ 0 prediction 3 Δ→ Nγ 255.78 π misid 719.89 8 dirt

8 Events other 1600 Intrinsic νe 990.48 Best Fit 478.45 6 2 1400 Data 2881.00 6 4 1 arxiv:1805.12028[hep-ex] 2 1200 MiniBooNE preliminary 0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 ln(L /L ) ln(L /L ) 4 0 e 0 e π π 1000 18.75 x 1020 POT

3 3 ×10 ×10 Neutrino mode 1.6 1.2 800 2 Events Events 1.4 1 600 1.2 (c) (d) 0.8 1 0 0.5 1 1.5 2 2.5 3 400 fit ratio 0.8 0.6 1.5 QE Eν [GeV] 0.6 Data/MC 0.4 200

1 0.4

0.2 0.2 0 0.5 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 electron/pion log likelihood 0 ln(L /L ) ln(L /L ) 0 0.5 1 1.5 2 2.5 3 π0 e π0 e EQE[GeV] ν FIG. 10: Comparisons between the data and simulation for FIG. 8: An absolute comparison of the reconstructed neutrino the electron-pion likelihood distribution after successive cuts energy distribution for CCQE events between the neutrino are applied: (a) no PID cut, (b) electron-muon likelihood data (12.84 1020 POT) and the simulation (top). Also shown cut, (c) electron-muon plus electron-pion likelihood cuts, and is the ratio⇥ between the data and Monte Carlo simulation (d) electron-muon plus electron pion likelihood cuts plus a (bottom). gamma-gamma mass cut. The vertical lines in the ﬁgures show the range of energy-dependent cut values.

nd 3 3 30 ×10 ×10 Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2 2020 3 3 (a) ×10 ×10 Data (a) Events 10 Events 9 +/- Data 7 ν e from µ 5 Events Events +/- from +/- ν e from K 8 ν e µ +/- 0 from K 8 ν e from K ν e 6 0 0 π misid 4 (b) 7 ν e from K π0 misid (b) Δ→ Nγ 5 dirt 6 Δ→ Nγ 6 other 3 dirt 5 other 4

4 4 2 3 3 2 2 1 2 1 1 0 0 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 ln(L /L ) ln(L /L ) 0 0 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 2 2 mγ γ (MeV/c ) mγ γ (MeV/c )

400

Events 350 Events 250 Events 600 Events 600

300 200 (c) (d) 500 500 250 (c) (d)

150 400 400 200

300 300 150 100

100 200 200 50 50 100 100

0 0 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 ln(L /L ) ln(L /L ) 0 0 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 2 2 mγ γ (MeV/c ) mγ γ (MeV/c ) FIG. 9: Comparisons between the data and simulation for FIG. 11: Comparisons between the data and simulation for the electron-muon likelihood distribution after successive cuts the gamma-gamma mass distribution after successive cuts are applied: (a) no PID cut, (b) electron-muon likelihood are applied: (a) no PID cut, (b) electron-muon likelihood cut, (c) electron-muon plus electron-pion likelihood cuts, and cut, (c) electron-muon plus electron-pion likelihood cuts, and (d) electron-muon plus electron pion likelihood cuts plus a (d) electron-muon plus electron pion likelihood cuts plus a gamma-gamma mass cut. The vertical lines in the ﬁgures gamma-gamma mass cut. The vertical lines in the ﬁgures show the range of energy-dependent cut values. show the range of energy-dependent cut values. 8

3 3 3 ×10 ×10 ×10 5 (a) 14 Data Events Events 10 +/- ν e from µ +/- Events 4 νµ CCQE 12 ν e from K 0 ν e from K 0 data π misid (b) 10 Δ→ Nγ prediction 3 8 dirt 8 other

6 2

6 4 1 2

0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 ln(L /L ) ln(L /L ) 4 π0 e π0 e

3 3 ×10 ×10

1.6 1.2 2 Events Events 1.4 1 1.2 (c) (d) 0.8 1 0 0.5 1 1.5 2 2.5 3 fit ratio 0.8 0.6 1.5 QE Eν [GeV] 0.6 Data/MC 0.4

1 0.4

0.2 0.5 0.2

0 0 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 −0.4 −0.35 −0.3 −0.25 −0.2 −0.15 −0.1 −0.05 0 0.05 0 ln(L /L ) ln(L /L ) 0 0.5 1 1.5 2 2.5 3 π0 e π0 e EQE[GeV] ν FIG. 10: Comparisons between the data and simulation for FIG. 8: An absolute comparison of the reconstructed neutrino the electron-pion likelihood distribution after successive cuts energy distribution for CCQE events between the neutrino are applied: (a) no PID cut, (b) electron-muon likelihood data (12.84 1020 POT) and the simulation (top). Also shown cut, (c) electron-muon plus electron-pion likelihood cuts, and is the ratio⇥ between the data and Monte Carlo simulation (d) electron-muon plus electron pion likelihood cuts plus a (bottom). gamma-gamma mass cut. The vertical lines in the ﬁgures show the range of energy-dependent cut values. PID variables : mγγ distribution 3 3 ×10 ×10 3 3 (a) ×10 ×10 Data (a) Events 10 Events 9 +/- Data 7 ν e from µ 5 Events Events +/- from +/- ν e from K 8 ν e µ +/- 0 from K 8 ν e from K ν e 6 0 0 π misid 4 (b) 7 ν e from K Other dirt 0 N π misid (b) Δ→ γ 5 800 dirt 6 Δ→ Nγ 0 6 Δ→ Nγ π misid other 3 dirt

5 other 4 Events Intrinsic νe Best Fit 4 4 700 2 3 3 Data 2 2 1 2 600 arxiv:1805.12028[hep-ex] 1 MiniBooNE preliminary 1 0 0 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 ln(L /L ) ln(L /L ) 0 0 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 500 20 m (MeV/c2) m (MeV/c2) 18.75 x 10 POT γ γ γ γ 400 400 Neutrino mode Events 350 Events 250 Events 600 Events 600 300 300 200 (c) (d) 500 500 250 (c) (d)

150 400 400 200 200

300 300 150 100 100 100 200 200 50 50 100 100 0 0 0 0 10 20 30 40 50 60 70 80 90 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 −0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 2 ln(L /L ) ln(L /L ) 0 0 µ e µ e 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Pi0 Mass [MeV/c ] 2 2 mγ γ (MeV/c ) mγ γ (MeV/c ) FIG. 9: Comparisons between the data and simulation for FIG. 11: Comparisons between the data and simulation for the electron-muon likelihood distribution after successive cuts the gamma-gamma mass distribution after successive cuts are applied: (a) no PID cut, (b) electron-muon likelihood are applied: (a) no PID cut, (b) electron-muon likelihood cut, (c) electron-muon plus electron-pion likelihood cuts, and cut, (c) electron-muon plus electron-pion likelihood cuts, and (d) electron-muon plus electron pion likelihood cuts plus a (d) electron-muon plus electron pion likelihood cuts plus a gamma-gamma mass cut. The vertical lines in the ﬁgures gamma-gamma mass cut. The vertical lines in the ﬁgures show the range of energy-dependent cut values. show the range of energy-dependent cut values.

Adrien Hourlier — The XXIX International Conference on Neutrino Physics and Astrophysics — July 2nd 2020 31