sign in sign up
<<
Home , SciBooNE

Unitarity and the three flavour neutrino mixing matrix.

Stephen Parke1 and Mark Ross-Lonergan2 1Theoretical Physics Department, Fermi National Accelerator Laboratory, P.O.Box 500, Batavia, IL 60510, USA 2IPPP, Department of Physics, Durham University, Durham DH1 3LE, UK Unitarity is a fundamental property of any theory required to ensure we work in a theoretically consistent framework. In comparison with the quark sector, experimental tests of unitarity for the 3x3 neutrino mixing matrix are considerably weaker. It must be remembered that the vast majority of our information on the neutrino mixing angles originates from νe and νµ disappearance experi- ments, with the assumption of unitarity being invoked to constrain the remaining elements. New physics can invalidate this assumption for the 3x3 subset and thus modify our precision measure- ments. We perform a reanalysis to see how global knowledge is altered when one refits oscillation results without assuming unitarity, and present 3σ ranges for allowed UPMNS elements consistent with all observed phenomena. We calculate the bounds on the closure of the six neutrino unitarity triangles, with the closure of the νeνµ triangle being constrained to be ≤ 0.03, while the remaining triangles are significantly less constrained to be ≤ 0.1 - 0.2. Similarly for the row and column nor- malization, we find their deviation from unity is constrained to be ≤ 0.2 - 0.4, for four out of six such normalisations, while for the νµ and νe row normalisation the deviations are constrained to be ≤ 0.07, all at the 3σ CL. We emphasise that there is significant room for new low energy physics, especially in the ντ sector which very few current experiments constrain directly.

2 With the knowledge of sin 2θ13 now almost at the 5% elements only, the νe normalisation is the sole condition level, and interplay between the long baseline accelerator that can be reasonably constrained without any further ν ν appearance data [1, 2] and short baseline reactor assumptions as to the origin of the non-unitarity [12]. µ → e νe νe disappearance [3–5] data, combined with prior In the quark sector, the analogous situation involv- → knowledge of θ23 from νµ νµ disappearance data [6–8], ing the Cabibbo-Kobayashi-Maskawa (CKM) matrix has → suggesting tentative global hints at δCP 3π/2, there is been subject to intense verification as many distinct ex- ≈ much merit to statements that we are now in the preci- periments have access to probes of all of the VCKM el- sion measurement era of neutrino physics. Our knowl- ements individually. Current data shows that the as- edge of the distinct Pontecorvo-Maki-Nakagawa-Sakata sumption of unitarity for the 3x3 CKM matrix is valid in (PMNS) neutrino mixing matrix elements comes from the quark sector to a high precision, with the strongest 2 2 2 the plethora of successful experiments that have run since normalisation constraint being Vud + Vus + Vub = the first strong evidence for neutrino oscillations, inter- 0.9999 0.0006 and the weakest| still| being| | significant| | ±2 2 2 preted as νµ ντ oscillations, was discovered by Super- at V + V + V = 1.044 0.06 [13]. Unlike the → ub cb tb Kamiokande in 1998 [9]. However, one must always re- quark| sector,| | however,| | experimental| ± tests of unitarity are member that our knowledge of the matrix elements is considerably weaker in the 3x3 UPMNS neutrino mixing predominately in the νe and νµ sectors, and comes pri- matrix. It remains an initial theoretical assumption in- marily from high statistics νe disappearance and νµ dis- herent in many analyses [14–16], but is the basis for the appearance experiments, with the concept of unitarity validity of the 3ν paradigm. being invoked to disseminate this information onto the This non-unitarity can arise naturally in a large va- remaining elements. With more statistics, the long base- riety of theories. A generic feature of many Beyond line νµ νe and νµ νe appearance experiments such the Standard Model scenarios is the inclusion of one → → as T2K [10] and NOνA [11] will aid in νµ sector precision or more new massive fermionic singlets, uncharged un- measurements. der the Standard Model (SM) gauge group, SU(3)C arXiv:1508.05095v1 [hep-ph] 20 Aug 2015 SU(2) U(1) . If these new sterile states mix with the× Unitarity of a mixing matrix is a necessary condition L × Y for a theoretically consistent description of the under- SM neutrinos then the true mixing matrix is enlarged lying physics, as non-unitarity directly corresponds to from the 3x3 UPMNS matrix to a nxn matrix, 3x3 a violation of probability in the calculated amplitudes. UPMNS In the neutrino sector unitarity can be directly veri- Ue1 Ue2 Ue3 Uen fied by precise measurement of each of the mixing ele- ··· Uµ1 Uµ2 Uµ3 Uµn † 1  z }| { ···  ments to confirm the unitarity condition: U U = = Extended Uτ1 Uτ2 Uτ3 Uτn † UPMNS = ··· . (1) UU . In this there are 12 dependant conditions, six  . . . . .   ......  of which we will refer to as normalisations (sum of the   Us 1 Us 2 Us 3 Us n squares of each row or column, e.g the νe normalisation  n n n ··· n  2 2 2   Ue1 + Ue2 + Ue3 = 1) and six conditions that mea- sure| | the| degree| with| | which each unitarity triangle closes These so-called sterile neutrinos have been a major ∗ ∗ ∗ (e.g the νeνµ triangle: Ue1Uµ1 + Ue2Uµ2 + Ue3Uµ3 = 0). discussion point for both the theoretical and experimen- Currently, from direct measurements of the individual tal communities for decades. If they have masses at or 2

Experiment Measured quantity with unitarity Without unitarity Normalisation

Reactor SBL 2 2 2 2 2 2 2 2 22 4|Ue3| 1 − |Ue3| = sin 2θ13 4|Ue3| Ue1| + |Ue2| |Ue1| + |Ue2| + |Ue3| (νe → νe)

Reactor LBL 2 2 2 4 2 2 2 2 22 4|Ue1| |Ue2| = sin 2θ12 cos θ13 4|Ue1| |Ue2| |Ue1| + |Ue2| + |Ue3| (νe → νe)

SNO (φCC /φNC 2 2 2 2 2 2 2 |U | = cos θ sin θ |Ue2| |Ue2| + |Uµ2| + |Uτ2| Ratio) e2 13 12

2 2 SK/T2K/MINOS 4|Uµ3| 1 − |Uµ3| = 2 2 2 2 2 22 2 2 2 2  4|Uµ3| Uµ1| + |Uµ2| |Uµ1| + |Uµ2| + |Uµ3| (νµ → νµ) 4 cos θ13 sin θ23 1 − cos θ13 sin θ23

T2K/MINOS 2 2 2 2 ∗ ∗ ∗ ∗ ∗ ∗ 2 4|Ue3| |Uµ3| = sin 2θ13 sin θ23 −4 Re{Ue3Uµ3 (Ue1Uµ1 + Ue2Uµ2)} |Ue1Uµ1 + Ue2Uµ2 + Ue3Uµ3| (νµ → νe)

SK/OPERA 2 2 2 4 ∗ ∗ ∗ ∗ ∗ ∗ 2 4|Uµ3| |Uτ3| = sin 2θ23 cos θ13 −4 Re{Uτ3Uµ3 (Uτ1Uµ1 + Uτ2Uµ2)} |Uµ1Uτ1 + Uµ2Uτ2 + Uµ3Uτ3| (νµ → ντ )

TABLE I: Example experiments and the leading order functions of UPMNS matrix elements they measure, in both the unitary and non-unitary case. The third column shows the normalisation that can be bound if the experimental measurements of the fluxes and backgrounds are known to a high enough degree. near the GUT scale, then the see-saw mechanism can tio BR(µ eγ). The strongest limits currently arise provide sufficiently small masses to the SM neutrinos from the MEG→ experiment, BR(µ eγ) < 5.7 10−13 [17], but a priori these new states can sit at practically at the 90% CL [21] which translates→ to a bound× on ∗ ∗ ∗ −5 any mass as there is no known symmetry to dictate Ue1Uµ1 + Ue2Uµ2 + Ue3Uµ3 < 10 [19] at the 90% CL. a scale. Although this extended nxn mixing matrix, | If, however, the new physics| that provides the non- should nature choose it, will indeed be unitary to unitarity enters at a much lower energy scale, as sev- preserve probability, the same is not true for any given eral current experimental hints suggest with anomalous mxm subset, with m < n. This is the canonical model results from LSND [22], MiniBooNE [23], the Gallium of how new physics, introduced at any scale, breaks anomaly [24, 25] and the Reactor anomaly [26], then observed unitarity in the neutrino sector. many of the most constraining experiments that bound the MUV scheme are not directly applicable. For physics entering at these lower scales, one must focus on direct An extensive body of work in the literature exists measurements of the individual mixing elements. To this on non-unitarity in the neutrino sector, most of which end neutrino oscillations are the most important exper- has been analysed with the rigorous model-independent imental probe we have access to. The most convincing approach of the Minimal Unitarity Violation (MUV) means of verification of unitarity in the neutrino sector scheme [18]. In this approach the new physics enters would be analogous to the quark sector, via direct and high above the energies involved in oscillation experi- independent measurement of all the UPMNS elements, ments, and the three neutrino Standard Model becomes a to overconstrain the parameter space and confirm that low-energy effective theory in which the unitarity of the the 12 unitarity constraints hold to within experimen- 3x3 mixing matrix is not assumed. It has been shown tal precision. However, we do not currently have access recently [19] that the current status under this scheme to enough experiments, at distinct mass differences, in is highly constrained by experiment, the weakest unitar- the νµ and ντ sectors to bound all of the elements to a 2 2 2 ity constraint is Uτ1 + Uτ2 + Uτ3 = 0.9947 1.0 sufficient degree to verify all 12 conditions. | | | | | | → at the 90 % CL, producing practically immeasurable de- The situation for progress in direct measurement is viations to the mixing angles with current experimen- therefore not promising. Thus we must look for alter- tal uncertainty. Some of the most stringent bounds in native ways to constrain the UPMNS elements. One can the MUV scheme come from rare lepton decays such as perform indirect searches of unitarity by searching for µ eγ. This is due to the fact that without the unitar- mixing elements outside those of the 3ν mixing regime. → ity condition in the 3x3 mixing matrix, the exact (in the These class of searches do not measure the 3x3 mixing massless ν limit) cancellation provided by the Glashow- elements per say, but rather by looking for additional Iliopoulos-Maiani mechanism [20] in the SM no longer states one can constrain the violations they would in- holds. In conjunction with the fact each pair of active duce in the 3x3 subset. One proceeds by noting all null neutrinos and charged leptons, e.g. (νe, eL), make up results in oscillations at mass differences distinct to those SU(2)L electroweak doublets, this results in the off di- of the 3ν paradigm. We do not wish to perform a global † agonal terms of UPMNSUPMNS being hugely constrained fit for new physics as this has been well covered in the by the current experimental limits on the branching ra- literature [27, 28], instead we focus on what unresolved 3 physics can do to our current precision, hence we do i = 1, 2, 3. Added assumptions on the exact origin of the not include any positive signals such as LSND or the non-unitarity in the 3x3 submatrix can lead to further MiniBooNE anomaly. Such an sterile driven approach correlations between the elements. In particular, if the requires additional assumptions on the exact origin of non-unitarity does indeed come from an enlarged mixing the non-unitarity, thus losing some model-independence. matrix then one must invoke Cauchy-Schwartz inequali- However, as an extended UPMNS matrix encompasses ties along with the unitarity constraints of the true ex- many beyond the Standard Model scenarios, it is natural tended mixing matrix to place six geometric constraints to include this in our analysis. To proceed one must then on the mixing elements [18], consider what scale the new physics enters at, however, as we do not focus on the origin of such non-unitarity we 3 2 3 3 ∗ 2 2 choose to marginalise over the new scale(s) assuming the UαiUβi 1 Uαi 1 Uβi , possibility they enter in at an oscillating scale, with at ≤ − | | ! − | | ! i=1 i=1 i=1 least ∆m2 10−2 eV2. Below this scale, states degen- X X X for α, β = (e, µ, τ), α = β, erate| with| SM ≥ neutrinos requires a much more detailed 6 τ 2 τ τ analysis. ∗ 2 2 A non-unitary mixing matrix can be parameterised as UαiUαj 1 Uαi 1 Uαj , ≤ − | | ! − | | ! a 3x3 matrix hosting 9 complex non-unitary elements, 5 α=e α=e α=e X X X phases of which can be removed by rephasing the lepton for i, j = (1, 2, 3), i = j. (3) 6 fields, leaving 13 parameters: 9 real positive numbers and 4 phases. The oscillation probability for a neutrino (anti- These geometric constraints enable precision measure- neutrino) of initial flavour α and energy Eν to transition ments in a single sector to be passed subsequently to 1 to a neutrino (anti-neutrino) of flavour β after a distance all elements of the mixing matrix . L with such a non-unitary mixing matrix is given by To perform the analysis, for each experiment consid- ered2 we take the observed amplitude of the ν ν (or 2 α → β (–) (–) ∗ να νβ) oscillation after normalisation, Aα;β, along- P να νβ = U Uαi (2) → 2 → βi side its published error σα;β and construct a χ = i=1 NU 2 2   X (P Aα;β) /σ for the associated non-unitary am- α;β − α:β ∗ ∗ 2 2 L NU 4 Re(U U U U ) sin ∆m plitudes as taken from equation (2), Pα;β , along with − βi βj αi αj ji 4E i

9 U U U | e1| | e2| | e3| w/o Unitarity (All data) 6 with Unitarity 2 (All data) χ

∆ 3 w/o Unitarity (No normalisation or sterile data) 0 9 U U U | µ1| | µ2| | µ3| 6 2 χ ∆ 3

0 9 U U U | τ1| | τ2| | τ3| 6 2 χ ∆ 3

0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 U U U | α1| | α2| | α3|

FIG. 1: Marginalised 1-D ∆χ2 for each of the magnitudes of the 3x3 neutrino mixing matrix elements, without (red solid) and with (black dashed) the assumption of unitarity, using all data. The x-axis is the magnitude of each individual matrix element, and the y-axis is the associated ∆χ2 after marginalisation over all parameters other than the one in question. Also shown is the fit in which no normalisation or sterile search data is used (blue dotted), to highlight their importance to the fit. The blue and red curves coincide closely with the unitary case in |Ue2| and |Ue3|. This analysis was preformed for the normal hierarchy, the inverse hierarchy providing the same qualitative result. no new interactions, such as an additional U(1)0 which good agreement with published results in contemporary can lead to strongly modified matter effects, are active global fits such as ν-fit [14]. at oscillation energies. Upon minimization the best fit points agree in both unitary and non-unitary fits. To compare how the pre- Also in Fig. (1), for the purpose of emphasising how cision varies we consider the frequentist 3σ ranges of the important normalisation and sterile searches are to the one-dimensional ∆χ2 projections without unitarity as- precision of the 3x3 UPMNS elements, without the as- sumed (with unitarity), where we marginalise over all sumption of unitarity, we performed an alternative, more parameters except the one in question, we obtain conservative fit, in which no normalisation or 3+N ster- ile search data is included. This fit only highlights the w/o Unitarity stark difference between the quark and neutrino sectors. U (with Unitarity) = | |3σ Whereas the quark sector can independently measure all 0.76 0.85 0.50 0.60 0.13 0.16 VCKM elements, some within 0.1% and all within 10% (0.79→→0.85) (0.50→→0.59) (0.14→→0.16) [13], in the neutrino sector we only have access to pre- 0.21 0.54 0.42 0.70 0.61 0.79 cision knowledge of the νe row in a completely model- (0.22→→0.52) (0.43→→0.70) (0.62→→0.79) . (4) independent manner. This is as expected, as the ν sec-   e   tor has access to high statistics experiments at both the 0.18 0.58 0.38 0.72 0.40 0.78  (0.24→→0.54) (0.47→→0.72) (0.60→→0.77)  solar and atmospheric mass scale, as well as a third inde-     pendent experiment in the solar flux measurements, due The ranges for the individual elements, assuming uni- to the MSW effect. In the ν sector, U 2 is known µ | µ3| tarity (bracketed numbers in above expression), are in to be non-zero from νµ disappearance experiments but 5

Normalisations constraints. 9 3σ Rows α =e We must stress that even if the 3σ ranges of the α =µ α =τ UPMNS elements agree closely with the unitarity case, 6 Columns this does not equate to the neutrino mixing matrix i =1 being unitary. In the unitary case the correlations are

2 i =2 much stronger and choosing an exact value for any one χ i =3 ∆ 2σ the mixing elements drastically reduces the uncertainty on the remaining elements. To better understand the 3 µ e level at which we know unitarity is conserved or not, we plot the resultant ranges for the normalisation in Fig 1σ (2). We see that the νe and νµ normalisation deviations from unity are relatively well constrained ( 0.06 and 0 -2 -1 ≤ 10 10 0.5 0.07 at 3σ CL respectively), primarily by reactor fluxes 1 ( U 2 + U 2 + U 2 ) or 1 ( U 2 + U 2 + U 2 ) and a combination of precision measurements of the rate − | α1| | α2| | α3| − | ei| | µi| | τi| Rows Columns and spectra of upward going muon-like events observed at Super-Kamiokande [53] and the multitude of long 2 and short baseline accelerator ν ν disappearance FIG. 2: 1-D ∆χ for deviation of both UPMNS row (solid) and µ → µ column (dashed) normalisations, fitted with all spectral and experiments. We note the νµ normalisation deviation normalisation data, when considering new physics that enters from unity is constrained slightly ( 1%) better than 2 −2 2 ≈ above |∆m | ≥ 10 eV . the νe normalisation. This is due to the large theoretical error, 5%, on total flux from reactors assumed [56]. The remaining normalisation deviations from unity are all constrained to be 0.2 - 0.4 at 3σ CL. as Uµ1 and Uµ2 only appear in the degenerate com- . | | |2 | 2 bination Uµ1 + Uµ2 , they cannot be distinguished individually.| This| | degeneracy| is very weakly broken by For the case of the six neutrino unitarity triangles, we the νµ νe appearance experiment T2K [1], and will be present the allowed ranges for their closures in Fig. (3). improved→ upon taking of more data and with future high For the three row triangles the bounds originate from a statistics NOνA [11] results. The addition of this nor- combination of the corresponding geometric constraints malisation and sterile data in the 3ν unitarity case does along with appearance data in the respective channel. not change anything in the fit. From here on we will The column triangles, however, are bound by the geomet- discuss only the main results, as calculated including all ric constraints only, and as the column normalisations are normalisation and sterile search data. proportionally less known, so too are the column unitar- The addition of this sterile search and normalisation ity triangles. Only one triangle does not contain a ντ data improves the situation significantly. If we define element, the νeνµ triangle, and hence it is the only tri- the shift in range of allowed values as the ratio of the angle in which it is excluded to be open by more than difference in 3σ ranges without and with unitarity, to 0.03 at the 3σ CL, compared to between 0.1 - 0.2 at the that derived with unitarity, the increase in parameter 3σ CL for the remaining triangles. This hierarchical sit- uation will not improve unless precise measurements can space for Uei , i = 2, 3 and Uµi , i = 1, 2, 3 are all 10% (4%,| 8%,| 8%, 7% and 4%| respectively),| with U ≤ be made in the ντ sector. | e1| taking the majority of the discrepancy in the νe sector, with an increase of allowed range of 68%, primarily If one wishes to proceed with measurements of unitar- due to the weaker bounds from KamLAND compared ity, without the assumption of an extended UPMNS ma- 2 to the SBL reactors, and that Ue1 forms the bulk of trix and its subsequent Cauchy-Schwartz bounds, then U 2 + U 2 + U 2. The entire| | ν sector, however, prospects for improvement are essentially limited to mea- | e1| | e2| | e3| τ may contain substantial discrepancies from unitarity suring the νe normalisation. Improvement of all νe ele- with shifts in allowed regions of 37%, 46% and 104% ments is possible, especially if the new generation reac- respectively. We have little or no current mechanisms tor experiments, JUNO [57] and RENO50 [58], proceed to directly measure any ντ elements and we have not as planned. See discussion by X. Qian et al. [12] for yet observed any oscillation amplitude peaks, even the a detailed discussion of the possible improvements. Sig- recent 5σ discovery of νµ ντ at OPERA [49] only nificant improvement in the νµ sector would require the st → sees the tail end of the 1 oscillation maximum and the measurement of νµ disappearance at the solar mass scale, observation of 5 events on a background of 0.25 0.05 well beyond what is currently technologically feasible. is not significant spectrally and can be equally be± fit by Improvements in the indirect 3+N sterile measure- a flat normalisation discrepancy. The precision we do ments are much more promising, the Short have is driven by the fact large deviations here cause Baseline Neutrino (SBN) [59] program consisting of the violations of unitarity too large in the νe and νµ sectors, SBND, MicroBooNE and ICARUS experiments on the passed through by the geometric Cauchy-Schwartz Booster beam, will be capable of probing a wide range 6

Unitarity Triangle Closures for precision high statistics measurements, both of 9 3σ Rows which are extremely difficult tasks. Perhaps crucially α,β =e,µ for ντ measurements, Hyper-Kamiokande [63] will be α,β =e,τ incredibly sensitive to atmospherically averaged steriles, α,β =µ,τ 2 Columns 0.1 eV , and will significantly improve the current 6 ≥ 2 2 2 i,j =1,2 bounds on Uτ1 + Uτ2 + Uτ3 in this regime, to | | | 2 | | 2 | 2

2 i,j =1,3 approximately 1 Uτ1 + Uτ2 + Uτ3 0.07 at the χ i,j =2,3 − | | | | | | ≤ ∆ 2σ 99% CL [64], which would bring it closer inline with the other sectors. 3 eµ In this paper we have emphasised the fact that 1σ current experimental bounds on unitarity within the 3ν 0 -1 10-2 10 0.5 paradigm allows for considerable violation, and without the unitarity assumption, the precision on the individual U U ∗ +U U ∗ +U U ∗ or U U ∗ +U U ∗ +U U ∗ | α1 β1 α2 β2 α3 β3 | | ei ej µi µj τi τj | Rows Columns UPMNS elements can vary significantly (up to 104% in the case of U ). However, we find no evidence for non- | τ3| FIG. 3: 1-D ∆χ2 for the absolute value of the closure of the unitarity. The prospects of directly measuring all the 12 three row (solid) and three column (dashed) unitarity tri- unitarity constraints with high precision are poor, and angles, fitted with all spectral and normalisation data, when even when one allows for additional model-dependant considering new physics that enters above |∆m2| ≥ 10−2 eV2. sterile searches we can only constrain the amount of There is one unique unitarity triangle, the νeνµ triangle, in non-unitarity to be . 0.2 - 0.4, for four out of six of that it does not contain any ντ elements and hence is con- the row and columns normalisations, with the νµ and νe strained to be unitary at a level half an order of magnitude normalisation deviations from unity constrained to be better than the others. By comparison to Fig. 2 one can 0.07, all at the 3σ CL, see Fig. 2. Similarly, five out of≤ clearly see the Cauchy-Schwartz constraints are satisfied. six of the unitarity triangles are only constrained to be . 0.1 - 0.2, with opening of the remaining νeνµ triangle being constrained to be 0.03, again at the 3σ CL, see of parameter space for 3+N models, increasing both the Fig. 3. One must be careful≤ when assessing the current appearance and disappearance bounds. Subsequently, experimental regime with the addition of new physics we the long baseline program DUNE [60] will also be are currently insensitive to, as without the assumption of able to significantly extend the constrained region of unitarity there is much room for new effects, especially νµ νe appearance to lower mass differences, leading → in the ντ sector where currently significant information to increased constraints on the νeνµ unitarity triangle comes from the unitarity assumption and not direct in this regime. An understanding of the neutrino flux measurements. and cross sectional uncertainties are crucial for unitarity measurements. Possible future experiments such as a fully fledged [61] or the nuStorm Acknowledgements: We thank Belen Gavela and facility [62], with the uncertainty on their fluxes of the Silvia Pascoli for fruitful discussions. MRL would like order 1%, will be able to constrain the νµ normalisation to thank the Theoretical Physics department at Fermi- and νeνµ triangle far beyond what is currently obtain- lab for hosting him during much of this work. SP thanks able. However, no one experiment can probe all scales the Kavli Institute for Theoretical Physics in UC Santa and complementarity is vital to definitively make a Barbara for their hospitalities, where part of this work statement about unitarity from new low-energy physics, was done. The authors acknowledge partial support from especially as there is little means to directly measure the the European Union FP7 ITN INVISIBLES (Marie Curie ντ sector. Improvement in ντ appearance requires new Actions, PITN-GA-2011-289442). This research was sup- experiments with both an intense, well known beam of ported in part by the National Science Foundation un- high enough energy νµ or νe to kinematically produce der Grant No. NSF PHY11-25915. Fermilab is operated charged taus, as well as a detector technology capable by the Fermi Research Alliance under contract no. DE- of efficiently identifying them to a degree necessary AC02-07CH11359 with the U.S. Department of Energy.

[1] K. Abe et al. (T2K), Phys.Rev. D91, 072010 (2015), [3] F. P. An et al. (Daya Bay), Phys. Rev. Lett. 108, 171803 arXiv:1502.01550 [hep-ex] . (2012), arXiv:1203.1669 [hep-ex] . [2] P. Adamson et al. (MINOS Collaboration), [4] J. K. Ahn et al. (RENO), Phys. Rev. Lett. 108, 191802 Phys.Rev.Lett. 110, 171801 (2013), arXiv:1301.4581 (2012), arXiv:1204.0626 [hep-ex] . [hep-ex] . [5] Y. Abe et al. (Double Chooz), Phys. Rev. D86, 052008 7

(2012), arXiv:1207.6632 [hep-ex] . vitz, and J. Spitz, Adv. High Energy Phys. 2013, 163897 [6] R. Wendell et al. (Super-Kamiokande), Phys. Rev. D81, (2013), arXiv:1207.4765 [hep-ex] . 092004 (2010), arXiv:1002.3471 [hep-ex] . [29] Y. Declais et al., Nucl. Phys. B434, 503 (1995). [7] A. B. Sousa (MINOS+, MINOS), Proceedings, 26th In- [30] I. E. Stockdale et al. (CCFR), Z. Phys. C27, 53 (1985). ternational Conference on Neutrino Physics and Astro- [31] D. Naples et al. (CCFR/NuTeV), Phys. Rev. D59, physics (Neutrino 2014), AIP Conf. Proc. 1666, 110004 031101 (1999), arXiv:hep-ex/9809023 [hep-ex] . (2015), arXiv:1502.07715 [hep-ex] . [32] A. Romosan et al. (CCFR/NuTeV), Phys. Rev. Lett. 78, [8] K. Abe et al. (T2K), Phys. Rev. Lett. 112, 181801 (2014), 2912 (1997), arXiv:hep-ex/9611013 [hep-ex] . arXiv:1403.1532 [hep-ex] . [33] K. S. McFarland et al., Phys. Rev. Lett. 75, 3993 (1995), [9] Y. Fukuda et al. (Super-Kamiokande Collaboration), arXiv:hep-ex/9506007 [hep-ex] . Phys.Rev.Lett. 81, 1562 (1998), arXiv:hep-ex/9807003 [34] F. Dydak et al. (CDHSW), Phys. Lett. B134, 281 (1984). [hep-ex] . [35] E. Eskut et al. (CHORUS), Nucl. Phys. B793, 326 [10] K. Abe et al. (T2K), Phys. Rev. Lett. 112, 061802 (2014), (2008), arXiv:0710.3361 [hep-ex] . arXiv:1311.4750 [hep-ex] . [36] M. Apollonio et al. (CHOOZ), Phys. Lett. B466, 415 [11] D. S. Ayres et al. (NOvA), (2004), arXiv:hep-ex/0503053 (1999), arXiv:hep-ex/9907037 [hep-ex] . [hep-ex] . [37] F. P. An et al. (Daya Bay), Chin. Phys. C37, 011001 [12] X. Qian, C. Zhang, M. Diwan, and P. Vogel, (2013), (2013), arXiv:1210.6327 [hep-ex] . arXiv:1308.5700 [hep-ex] . [38] F. P. An et al. (Daya Bay), Phys. Rev. Lett. 113, 141802 [13] K. A. Olive et al. (Particle Data Group), Chin. Phys. (2014), arXiv:1407.7259 [hep-ex] . C38, 090001 (2014). [39] M. Antonello et al. (ICARUS), Eur.Phys.J. C73, 2599 [14] M. Gonzalez-Garcia, M. Maltoni, and T. Schwetz, (2013), arXiv:1307.4699 [hep-ex] . (2014), arXiv:1409.5439 [hep-ph] . [40] B. Armbruster et al. (KARMEN), Phys. Rev. D65, [15] G. L. Fogli, E. Lisi, A. Marrone, D. Montanino, 112001 (2002), arXiv:hep-ex/0203021 [hep-ex] . A. Palazzo, and A. M. Rotunno, Phys. Rev. D86, 013012 [41] S. Abe et al. (KamLAND), Phys. Rev. Lett. 100, 221803 (2012), arXiv:1205.5254 [hep-ph] . (2008), arXiv:0801.4589 [hep-ex] . [16] D. V. Forero, M. Tortola, and J. W. F. Valle, Physical [42] A. Gando et al. (KamLAND), Phys. Rev. D83, 052002 Review D 90 (2014) 093006, Phys. Rev. D90, 093006 (2011), arXiv:1009.4771 [hep-ex] . (2014), arXiv:1405.7540 [hep-ph] . [43] P. Adamson et al. (MINOS), Phys. Rev. Lett. 110, [17] P. Minkowski, Phys.Lett. B67, 421 (1977). 251801 (2013), arXiv:1304.6335 [hep-ex] . [18] S. Antusch, C. Biggio, E. Fernandez-Martinez, [44] P. Adamson et al. (MINOS), Phys. Rev. D81, 052004 M. Gavela, and J. Lopez-Pavon, JHEP 0610, 084 (2010), arXiv:1001.0336 [hep-ex] . (2006), arXiv:hep-ph/0607020 [hep-ph] . [45] P. Astier et al. (NOMAD), Phys. Lett. B570, 19 (2003), [19] S. Antusch and O. Fischer, (2014), arXiv:1407.6607 [hep- arXiv:hep-ex/0306037 [hep-ex] . ph] . [46] P. Astier et al. (NOMAD), Nucl. Phys. B611, 3 (2001), [20] S. L. Glashow, J. Iliopoulos, and L. Maiani, Phys. Rev. arXiv:hep-ex/0106102 [hep-ex] . D 2, 1285 (1970). [47] R. Patterson, “First oscillation resultsfrom NOvA,” [21] J. Adam et al. (MEG Collaboration), Phys.Rev.Lett. (2015), Joint Experimental-Theoretical Seminar, Fermi- 110, 201801 (2013), arXiv:1303.0754 [hep-ex] . lab, August 6th, 2015. [22] A. Aguilar-Arevalo et al. (LSND Collaboration), [48] S. Avvakumov et al. (NuTeV), Phys. Rev. Lett. 89, Phys.Rev. D64, 112007 (2001), arXiv:hep-ex/0104049 011804 (2002), arXiv:hep-ex/0203018 [hep-ex] . [hep-ex] . [49] N. Agafonova et al. (OPERA), (2015), arXiv:1507.01417 [23] A. Aguilar-Arevalo et al. (MiniBooNE Collaboration), [hep-ex] . Phys.Rev.Lett. 110, 161801 (2013), arXiv:1207.4809 [50] N. Agafonova et al. (OPERA), JHEP 06, 069 (2015), [hep-ex] . arXiv:1503.01876 [hep-ex] . [24] W. Hampel, G. Heusser, J. Kiko, T. Kirsten, M. Lauben- [51] B. Aharmim et al. (SNO), Phys. Rev. C88, 025501 stein, E. Pernicka, W. Rau, U. Rnn, C. Schlosser, M. Wj- (2013), arXiv:1109.0763 [nucl-ex] . cik, R. v. Ammon, K. Ebert, T. Fritsch, D. Heidt, [52] G. Cheng et al. (SciBooNE, MiniBooNE), Phys. Rev. E. Henrich, L. Stieglitz, F. Weirich, M. Balata, F. Hart- D86, 052009 (2012), arXiv:1208.0322 . mann, M. Sann, E. Bellotti, C. Cattadori, O. Cre- [53] K. Abe et al. (Super-Kamiokande), Phys. Rev. D91, monesi, N. Ferrari, E. Fiorini, L. Zanotti, M. Altmann, 052019 (2015), arXiv:1410.2008 [hep-ex] . F. v. Feilitzsch, R. Mbauer, G. Berthomieu, E. Schatz- [54] T. Toshito (Super-Kamiokande), in 36th Rencontres de man, I. Carmi, I. Dostrovsky, C. Bacci, P. Belli, R. Bern- Moriond on Electroweak Interactions and Unified The- abei, S. d’Angelo, L. Paoluzi, A. Bevilacqua, M. Cri- ories Les Arcs, France, March 10-17, 2001 (2001) bier, L. Gosset, J. Rich, M. Spiro, C. Tao, D. Vignaud, arXiv:hep-ex/0105023 [hep-ex] . J. Boger, R. Hahn, J. Rowley, R. Stoenner, and J. We- [55] Y. Koshio, Proceedings, 26th International Conference on neser, Physics Letters B 420, 114 (1998). Neutrino Physics and Astrophysics (Neutrino 2014), AIP [25] J. Abdurashitov et al. (SAGE Collaboration), Phys.Rev. Conf. Proc. 1666, 090001 (2015). C59, 2246 (1999), arXiv:hep-ph/9803418 [hep-ph] . [56] A. Hayes, J. Friar, G. Garvey, G. Jungman, and [26] G. Mention, M. Fechner, T. Lasserre, T. Mueller, G. Jonkmans, Phys.Rev.Lett. 112, 202501 (2014), D. Lhuillier, et al., Phys.Rev. D83, 073006 (2011), arXiv:1309.4146 [nucl-th] . arXiv:1101.2755 [hep-ex] . [57] Y.-F. Li, J. Cao, Y. Wang, and L. Zhan, Phys.Rev. D88, [27] J. Kopp, P. A. N. Machado, M. Maltoni, and T. Schwetz, 013008 (2013), arXiv:1303.6733 [hep-ex] . JHEP 1305, 050 (2013), arXiv:1303.3011 [hep-ph] . [58] S.-B. Kim, (2013), Talk at International Workshop on [28] J. M. Conrad, C. M. Ignarra, G. Karagiorgi, M. H. Shae- RENO-50 toward Neutrino Mass Hierarchy, Seoul, South 8

Korea. (2014), arXiv:1402.5250 [hep-ex] . [59] M. Antonello et al. (LAr1-ND, ICARUS-WA104, Micro- [63] K. Abe et al., (2011), arXiv:1109.3262 [hep-ex] . BooNE), (2015), arXiv:1503.01520 [physics.ins-det] . [64] A. Himmel, “HK Sensitivity to Sterile Neutrinos and [60] C. Adams et al. (LBNE), (2013), arXiv:1307.7335 [hep- Lorentz Violation,” (2015), Talk at 6th Open Meeting ex] . for the HyperK Project, Kavli IPMU, The University of [61] S. Choubey et al. (IDS-NF), (2011), arXiv:1112.2853 Tokyo. [hep-ex] . [62] D. Adey et al. (nuSTORM), Phys. Rev. D89, 071301