Nonstandard interactions in propagation at the Deep Underground Neutrino Experiment
HJE
Nonstandard interactions in propagation at the Deep Underground Neutrino Experiment
Pilar Coloma 0 1
0 P. O. Box 500, Batavia, IL 60510 , U.S.A
1 Theoretical Physics Department, Fermi National Accelerator Laboratory
We study the sensitivity of current and future longbaseline neutrino oscillation experiments to the e ects of dimension six operators a ecting neutrino propagation through Earth, commonly referred to as NonStandard Interactions (NSI). All relevant parameters entering the oscillation probabilities (standard and nonstandard) are considered at once, in order to take into account possible cancellations and degeneracies between them. We nd that the Deep Underground Neutrino Experiment will signi cantly improve over current constraints for most NSI parameters. Most notably, it will be able to rule out the socalled LMAdark solution, still compatible with current oscillation data, and will be sensitive to o diagonal NSI parameters at the level of "
Beyond Standard Model; Neutrino Physics; CP violation

Underground
O(0:05
0:5). We also identify
two degeneracies among standard and nonstandard parameters, which could be partially
resolved by combining T2HK and DUNE data.
The formalism of NSI in propagation
Simulation details
Sampling of the parameter space
Experimental setups
5
Conclusions
A Implementation of prior constraints
1 Introduction
2
3
4
3.1
3.2
4.1
4.2
4.3
Results
Expected sensitivities for the DUNE experiment
Degeneracies
Comparison to other facilities and to prior experimental constraints
(LcL ~ )( ~yLL) ;
where LL stands for the lepton doublet, ~ = i 2 ,
being the SM Higgs doublet, and
is the scale of New Physics (NP) up to which the e ective theory is valid to. In eq. (1.1),
cd=5 is a coe cient which depends on the high energy theory responsible for the e ective
operator at low energies. Interestingly enough, the Weinberg operator is the only SM
gauge invariant d = 5 operator which can be constructed within the SM particle content.
Furthermore, it beautifully explains the smallness of neutrino masses with respect to the
rest of fermions in the SM through the suppression with a scale of NP at high energies.
When working in an e ective theory approach, however, an in nite tower of operators
would in principle be expected to take place. The e ective Lagrangian at low energies
would be expressed as:
L
e = LSM +
cd=5
O
d=5 +
cd=6
take place via d = 6 fourfermion e ective operators,1 in a similar fashion as in the case of
Fermi's theory of weak interactions. Fourfermion operators involving neutrino elds can
be divided in two main categories:
1. Operators a ecting chargedcurrent neutrino interactions. These include, for
instance, operators in the form (l
PL
)(q
P q0), where l stands for a charged lep
ton, P stands for one of the chirality projectors PR;L
5),
and
are lepton
2. Operators a ecting neutralcurrent neutrino interactions. These are operators in the
form (
PL
)(f
P f ). In this case, f stands for any SM fermion.
Operators belonging to the rst type will a ect neutrino production and detection
processes. For this type of NSI, near detectors exposed to a very intense neutrino beam would
be desired, in combination with a near detector, in order to collect a large enough event
sample [4]. Systematic uncertainties would play an important role in this case, since for
neutrino beams produced from pion decay the
ux cannot be computed precisely.2 For
recent studies on the potential of neutrino oscillation experiments to study NSI a ecting
neutrino production and detection, see e.g., refs. [7{12].
For operators a ecting neutralcurrent neutrino interactions the situation is very
different since these can take place coherently, leading to an enhanced e ect. Therefore,
longbaseline neutrino oscillation experiments, with L
O(500
1000) km, could potentially
place very strong constraints on NSI a ecting neutrino propagation. Moreover, unlike
atmospheric neutrino oscillation experiments [13{16], at longbaseline beam experiments the
beam is wellmeasured at a near detector, keeping systematic uncertainties under control.
Future longbaseline facilities, combined with a dedicated shortbaseline program [17{19]
to determine neutrino cross sections precisely, expect to be able to bring systematic
uncertainties down to the percent level. Therefore, they o er the ideal benchmark to constrain
NSI in propagation. This will be the focus of the present work.
As a benchmark setup, we consider the proposed Deep Underground Neutrino
Experiment [20] (DUNE) and determine the bounds that it will be able to put on NSI a
ecting neutrino propagation through matter. For comparison, we will also show the
sensitivity reach for the current generation of longbaseline neutrino oscillation experiments,
1In principle, the largest e ects from NSI are expected to come from d = 6 operators since they appear
at low order in the expansion. However, this is might not be always the case [2]. The situation might be
2A di erent situation would take place at beams produced from muon decay, such as Neutrino Factories
or the more recently proposed nuSTORM facility. In this case, the ux uncertainties are expected to remain
{ 2 {
i.e., T2K [21] and NOvA [22]. Finally, we will also compare its reach to a proposed future
neutrino oscillation experiment with much larger statistics but a much shorter baseline, to
illustrate the importance of the longbaseline over the size of the event sample collected.
As an example, we will consider the reach of the T2HK experiment [23].
The impact of NSI in propagation at longbaseline experiments has been studied
extensively in the literature, see refs. [24{32] for an incomplete list, or see refs. [33, 34] for recent
reviews on the topic. In particular, the reach of the LBNE experiment (very similar to the
DUNE setup considered in this work) was studied in ref. [35]. However, this study was
performed under the assumption of a vanishing 13, and only one nonstandard parameter
was switched on at a time. In the current work, we will follow the same approach as in
ref. [32]: all NSI parameters are included at once in the simulations, in order to explore
possible correlations and degeneracies among them. As we will see, this will reveal two
important degeneracies, potentially harmful for standard oscillation analyses.
The recent determination of 13 also has important consequences for the sensitivity to
NSI parameters. On one hand, the large value of 13 makes it possible for the interference
terms between standard and nonstandard contributions to the oscillation amplitudes to
become relevant (see, e.g., ref. [36] for a recent discussion). In addition, the value of 13
has now been determined to an extremely good accuracy by reactor experiments [37{39],
while the current generation of longbaseline facilities expects to signi cantly improve the
precision on the atmospheric parameters in the upcoming years [40]. At the verge of the
precision Era in neutrino experiments, it thus seems appropriate to reevaluate the
sensitivity of current and future longbaseline experiments to NSI parameters and, in particular,
of the DUNE proposal.
The paper is structured as follows. In section 2 we introduce the NSI formalism;
section 3 describes the simulation procedure and the more technical details of the experimental
setups under study; section 4 summarizes our results, and we present our conclusions in
section 5. Finally, appendix A contains some more technical details regarding the
implementation of previous constraints on the oscillation parameters in our simulations.
2
The formalism of NSI in propagation
NSI a ecting neutrino propagation (from here on, we will refer to them simply as NSI)
take place through the following fourfermion e ective operators:
LNSI =
p
contribution of a given operator with coe cient "fP , but only on their sum over avours
and chirality. The e ects of these operators appear in the neutrino evolution equation, in
{ 3 {
HJEP03(216)
the avour basis,3 as:
0
B
i
d
(2.2)
where
ij =
(1=ne) Pf;P nf
mi2j =2E, U is the lepton avor mixing matrix, A
fP , with nf the f type fermion number density and GF the Fermi coupling
p
2 2GF ne and "
constant. The three diagonal entries of the modi ed matter potential in eq. (2.2) are real
parameters, while the o diagonal parameters are generally complex.
.
the Hamiltonian. The three complex NSI parameters "e ; "e and "
will be parametrized
Due to the requirement of SM gauge invariance, in principle any operators responsible
of neutrino NSI would be generated simultaneously with analogous operators involving
charged leptons [2, 42{44]. Thus, the tight experimental constraints on charged lepton
avor violating processes can be automatically applied to operators giving NSI, rendering
the e ects unobservable at neutrino experiments. However, there are ways in which the
charged lepton constraints can be avoided, e.g., if the NSI are generated through operators
involving the Higgs, or from interactions with a new light gauge boson, see e.g., refs. [2, 42,
43, 45]. At this point, however, model dependence comes into play. In the present work,
we will explore how much the current bounds can be improved from a direct measurement
at neutrino oscillation experiments, without necessarily assuming the viability of a model
which can lead to large observable e ects.
Direct constraints on NSI can be derived either from4 scattering processes [43, 48{
50] or from neutrino oscillation data [51{54]. Currently, the strongest bounds for NSI in
propagation come from the global t to neutrino oscillation data in ref. [54]. At the 90% CL,
most constraints on the e ective " parameters are around
this is "~ee, for which only O(1) can be derived from current data. An important conclusion
derived from the global ts performed in refs. [51{54] is the presence of strong degeneracies
in the data. In presence of NSI in propagation, global analyses of neutrino oscillation data
are fully compatible with two solutions:
the LMA solution: the standard Large Mixing Angle (LMA) solution corresponds to
mixing angles fully compatible with the results obtained from a global t to neutrino
oscillation data in absence of NSI. The results are fully compatible with the hypothesis
3If production or detection NSI were present, though, the e ective production and detection
avour
eigenstates would not coincide with the standard
avour ones [41]. However, for simplicity we will consider
in this work that no signi cant NSI a ecting production or detection are present.
4Stronger limits can be derived from monojet and multilepton constraints at colliders [46, 47]. However,
these bounds are somewhat modeldependent and, in particular, fade away for models where the NSI come
from interactions via a new light mediator.
{ 4 {
of no NSI. There is a slight preference for a nonzero value of "~ee in the t, which arises
from the nonobservation of the upturn in the solar neutrino transition probability.
the LMAdark solution: this solution is obtained for "~ee
3. In this case, all the
oscillation parameters remain essentially unchanged, except for 12 which now lies
in the higher octant [51]. It should be stressed that this solution is fully compatible
with all current oscillation data, and there is no signi cant tension in the t.
In this work, we will consider that both solutions are equally viable, and will be considered
literature. Perturbative expansions of the relevant oscillation probabilities to this work can
be found, for instance, in ref. [25, 32, 55]. The main impact of NSI on the probabilities can
be summarized as follows:
The major impact on the
! e and
! e oscillation probabilities is expected
to come from the " e and " e parameters, as well as from "~ee. The dependence with
" e and " e appears at the same order in the perturbative expansion, and therefore
nontrivial correlations are expected to take place between them. The dependence
with the CPviolating phases ( ,
e and
e) is also expected to be nontrivial.
On the other hand, the disappearance channels
!
and
!
are mainly
a ected by the presence of "~
and " . The dependence of the oscillation probability
on these parameters will be brie y discussed in section 4.2.
Before nalizing this section it should be mentioned that, in the event of sizable NSI
e ects in propagation, the currently measured values of the oscillation parameters may be
a ected. In our simulations, we leave the atmospheric parameters free within their current
experimental priors, and all parameters (standard and nonstandard) will be tted
simultaneously. However, some comments are in order. Firstly, the measured value of 13 observed
at the Daya Bay experiment is not expected to be signi cantly a ected, due to the short
baseline and low neutrino energies involved. It can thus be considered as precise input for
the longbaseline analyses. A di erent situation may take place for the atmospheric mixing
angle 23, though, since its determination comes mainly from atmospheric and longbaseline
experiments, where NSI could be sizable. Nevertheless, in refs. [53, 54] it was found that the
determination of the atmospheric parameters is not signi cantly a ected by the addition of
a generalized matter potential. Finally, longbaseline experiments are not very sensitive to
the solar parameters, and in this case they have to rely in previous measurements. We will
consider the input values and priors at 1 from ref. [54], where the allowed con dence
regions were obtained under the assumption of a generalized matter potential with NSI e ects.
{ 5 {
3.1
Sampling of the parameter space
In our simulations, all relevant standard and nonstandard parameters are marginalized
over. This amounts to a total of fourteen parameters: six standard oscillation parameters
(the three mixing angles, the CPviolating phase and the two mass splittings), ve moduli
for the nonstandard parameters ("~ee; "~ ; j" ej; j" ej and j" j) and three nonstandard
CPviolating phases ( e
;
e and
). In order to sample all parameters e ciently, a
Monte Carlo Markov Chain (MCMC) algorithm is used. The Monte Carlo Utility Based
Experiment Simulator (MonteCUBES) C library [
58
] has been used to incorporate MCMC
sampling into the General Long Baseline Experiment Simulator (GLoBES) [59, 60]. For
the implementation of the NSI probabilities in matter, we use the nonStandard Interaction
Event Generator Engine (nSIEGE), distributed along with the MonteCUBES package.
Parameter estimation through MCMC methods is based on Bayesian inference. The
aim is to determine the probability distribution function of the di erent model parameters
given some data set d, i.e., the posterior probability P ( j d):
P = P ( j d) = L(d j )P ( )
:
P (d)
(3.1)
where L(d j
) is the likelihood, i.e., the probability of observing the data set d given a
certain set of values for the parameters
, and P (d) is the total probability of measuring
the data set d and can be regarded as a normalization constant. The prior P ( ) is the
probability that the parameters assume the value
regardless of the data d, that is, our
prior knowledge of the parameters. For the standard parameters, the assumed priors are
taken to be gaussian, and in agreement with the current experimental uncertainties (see
the pro les shown for the NSI with up quarks in gure 6 in ref. [54], rescaled accordingly
as "
3 "u , see ref. [54] for details.
At least 50 MCMC chains have been used in all our simulations, and the number
of distinct samples after combination always exceeds 106. The convergence of the whole
sample improves as R ! 1, with R being the ratio between the variance in the complete
sample and the variance for each chain. We have checked that, for most of the parameters
the convergence of the whole sample is much better than R
1 = 5
10 3, and in all cases
is better than 10 2. More technical details related to the sampling of the parameter space
can be found in appendix A.
3.2
Experimental setups
In this work we have considered several facilities among the current and future generation
of neutrino oscillation experiments:
DUNE. We consider a 40 kton
ducial liquid argon detector placed at 1300 km from the
source, onaxis with respect to the beam direction. The neutrino beam con guration
considered in this work corresponds to the 80 GeV con guration from ref. [62], with
{ 6 {
HJEP03(216)
a beam power of 1.08 MW. The detector performance has been simulated following
ref. [62], with migration matrices for neutral current backgrounds from ref. [63].
Three years of running time are assumed in both neutrino and antineutrino modes.
Systematic uncertainties of 2% and 5% are assumed for the signal and background
rates, respectively.
NOvA. The NOvA experiment has a baseline of 810 km, and the detector is exposed to an
o axis (0:8 ) neutrino beam produced from 120 GeV protons at Fermilab. The
implementation of the NOvA experiment follows refs. [22, 64]. The ducial mass of the
detector is 14 kton, and 6:0
1020 protons on target (PoT)/year are assumed. Again,
a running time of 3 years in both neutrino and antineutrino modes is considered.
T2K+NOvA. In this case, the expected results for the T2K experiment after 30
PoT in neutrino mode5 are added to the NO A results. The SuperKamiokaNDE
detector is placed o axis (2:5 ) with respect to the beam direction at L = 295 km, and
has a
ducial mass of 22.5 kton. The neutrino uxes have been taken from ref. [65].
The signal and background rejection e ciencies have been set to match the event
rates and sensitivities from ref. [21] for the same exposure, and rescaled up to the
larger statistics considered here. Given the much larger uncertainties in antineutrino
mode, only neutrino data is considered for T2K.
T2HK. The T2HK experiment is a proposed upgrade for the T2K experiment, with a
much larger detector (560 kton
ducial mass) located at the same o axis angle and
at the same distance as for the T2K experiment [23]. In this case, the signal and
background rejection e ciencies have been taken as in ref. [66]. The number of events
as well as the physics performance is consistent with the values reported in tables VIII
and IX in ref. [67]. These correspond to 3(7) years of data taking in (anti)neutrino
mode with a beam power of 750 MW. Systematic uncertainties of 5% and 10% are
assumed for the signal and background rates, respectively.
For all the setups simulated in this work, systematic uncertainties are taken to be correlated
among all contributions to the signal and background event rates, but uncorrelated between
di erent oscillation channels. In principle, a more detailed systematics implementation
should be performed, taking into account the possible impact of a near detector, correlations
between systematics a ecting di erent channels, etc. However, a careful implementation
of systematic errors would add a large number of nuisance parameters to the problem,
which would have to be marginalized over during the simulations. This would considerably
complicate the problem, and is beyond the scope of the present work.
For reference, the total expected event rates for the four experiments considered in this
work are summarized in table 1. The true values assumed for the oscillation parameters
are in good agreement with the best t values from ref. [
56
]: 12 = 33:5 , sin2 2 13 = 0:085,
23 = 42 ,
=
study the sensitivities of neutrino oscillation experiments to the NSI parameters, their true
5This corresponds to roughly ve times the PoT accumulated by the beginning of 2015 [21].
{ 7 {
! e
! e
!
(unosc.)
DUNE
NO A
T2K
1136/287
111/232
82/28
95/23
12/17
{/{
T2HK
considered in this work. The rates for the appearance channels are provided for the oscillation
parameters assumed in our simulations (under the assumption of no NSI), while for the disappearance
channels we provide the number of unoscillated events. Signal and background rejection e ciencies
have been taken into account in all cases.
values are set to zero in all cases. The matter density is xed to the value given by the
Preliminary Reference Earth Model [68]. We have checked that allowing it to vary within
a 2% range does not signi cantly a ect our nal sensitivities to NSI parameters, while it
slowed down the simulations.
4
Results
This section summarizes the results obtained for the expected sensitivities to NSI in
propagation for the setups considered in this work. We will rst summarize the expected results
for the DUNE experiment in more detail in section 4.1; a discussion of the
degeneracies found among standard and nonstandard parameters will be performed in section 4.2;
nally, a comparison to the expected results from T2K, NOvA and from the T2HK
experiment will then be performed in section 4.3.
Our results will be presented in terms of credible intervals, or credible regions, which
are obtained as follows. The total sample of points collected during the MCMC is projected
onto a particular plane in the parameter space. After projection, the regions containing a
given percentage (68%, 90% and 95%, in this work) of the distinct samples are identi ed.
4.1
Expected sensitivities for the DUNE experiment
The DUNE sensitivities to NSI parameters are summarized in gure 1. The gure shows
one and twodimensional projections of the MCMC results onto several planes. The
parameters used in the projections are indicated in the left and low edge of the collection
of panels. In the onedimensional distributions, the vertical band indicates the credible
interval at 68% level, while the dashed line shows the value which maximizes the posterior
probability. In the twodimensional projections, the red, green and blue lines show the 68%,
90% and 95% credible regions. In our simulations, all standard and nonstandard
parameters are left free and marginalized over. Similar projections for the standard oscillation
parameters can be found in appendix A, see gure 7.
{ 8 {
The vertical green bands indicate the credible intervals at 68%.
Several features can be observed from gure 1. Most notably, two important
degeneracies appear in the sensitivities: the rst a ects the determination of "~ , while the second
degeneracy is observed in the "~ee
" e plane. We will discuss these degeneracies in more
detail in section 4.2. A second important conclusion that can be derived from
gure 1 is
that DUNE will already be able to explore the LMA dark solution at more than 90% CL.
This can be observed in the leftmost column in
gure 1, where the range of values of "~ee
compatible with the LMAdark solution are disfavoured at more than 90%. We will return
to this point again in section 4.2.
When considering operators which are not diagonal in
avor space, it is important
to bear in mind that they may be accompanied by new sources of CPviolation. The
presence of such new phases may considerably a ect our sensitivity to the moduli of the
{ 9 {
Φ 50
100
planes. Dashed green lines indicate the results when no prior constraints are included on the NSI
parameters, while solid blue lines indicate the results after imposing prior constraints on the NSI
parameters. For reference, the vertical lines indicate prior constraints (at 90% CL, 1 d.o.f.) as
extracted from ref. [54].
NSI parameters, due to destructive and constructive interference e ects. For this reason,
we show in
gure 2 the twodimensional projections for the expected credible regions but
in this time after projecting the MCMC results on the j" j
planes. As can be seen
from the gure, the e ect is rather large for the three operators considered, and the bounds
are modi ed by a factor of between two and three in all cases. The dependence with the
CPphases is also di erent depending on the parameter under study.
The case where the dependence of the sensitivity with the CP phase is most notable
is the case of " . In this case, the sensitivity for values of
close to
=2 can be up
to a factor of three worse than the sensitivity around CPconserving values. While in the
former case the sensitivity would not be able to improve over current constraints, in the
latter case DUNE would be able to improve over current constraints by a factor of two.
The dependence with can be well understood from the leading order expansion of the disappearance channel [25, 32, 55]:
P
= P std
Ref" g (AL) sin ( 31L) + O("2) ;
(4.1)
p
where A
2 2GF ne stands for the standard matter potential, ij = ( mi2j =2E), and P std
is the oscillation probability in absence of NSI. Additional terms, which depend on both the
real and imaginary parts of " , enter the probability at second order in the perturbative
expansion, and provide some sensitivity in the regions with
=2. At second order,
the probability P
also depends on "~ , and will be further discussed in section 4.2.
The situation is a bit more convoluted for " e and " e due to their combined e ect on
the appearance oscillation probabilities, see for instance ref. [55]. In the case of " e, we nd
that DUNE will improve over current constraints regardless of the value of its associated
CPphase. The sensitivity changes by a factor of 2 depending on the value of
e, and
uctuates between 0.05 and 0.1. The results for " e also show a sizable dependence with
the value of
e. However, in this case the prior constraints play a very relevant role, as can
be seen from the comparison between the dashed green and solid blue lines in the panel for
" e in gure 2. Whereas before imposing prior constraints on the NSI parameters negative
values of e are perfectly allowed in the t, once the prior constraints on NSI are imposed
this is no longer the case. This has important consequences in the analysis, and implies
that DUNE will be sensitive to values of " e down to 0.05 for values of e
=2. The
reason for this is as follows. As it was shown in
gure 1, DUNE is insensitive to large
values of "~ee and j" ej as long as their moduli lie along the two lines identi ed in gure 1
(see the projected allowed regions in the "~ee
" e plane). For negative values of "~ee, the
degeneracy condition can only be satis ed for values of e
=2, as we will discuss
in more detail in section 4.2. However, prior constraints on NSI rule out a large fraction
of the parameter space for "~ee 2 ( 2; 0). Therefore, once these are included in the t,
the degeneracy condition can no longer be satis ed, which is translated into an increased
sensitivity at DUNE for " e, at the level of 0.05 for e .
Finally, it is important to keep in mind that the new CPviolating phases could have
an impact on standard CPviolating searches, see for instance ref. [32] for a study in the
context of Neutrino Factories, or ref. [69] for a pseudoanalytical study at DUNE. This will
be further discussed in section 4.2.
4.2
Degeneracies
When studying the sensitivity of DUNE to NSI, we have identi ed two important
degeneracies between both standard and nonstandard parameters. The rst one has been
previously reported in the literature (see, e.g., refs. [32, 55, 70, 71]), and takes place between
the parameters "~
and
23
23
=4. This degeneracy can be understood analytically
at the level of the oscillation probabilities. As already mentioned in section 2, the
sensitivity to the "~
parameter comes from the
and
disappearance channels. A perturbative
oscillation probability on
23, "
where A stands for the standard matter potential,
ij = ( mi2j =2E), and P std is the
oscillation probability in absence of NSI. Note the di erent combination of oscillatory
L
L
50 DUNE+T2HK
NSI w priors
NSI no priors
NSI w priors
Θ45
40
combination with T2HK data. Three cases are shown for DUNE: the standard case when no NSI
are allowed in the t, a case where marginalization is performed over NSI parameters within previous
constraints, and a case where no previous constraints are assumed over NSI during the t. The
combination with T2HK data is only shown in the case where prior NSI constraints are imposed
in the t. Right: same results, projected in the 23
"~
plane. The dot indicates the true input
values considered.
phases in the terms in eq. (4.2). The second term in principle should be subleading with
respect to the rst term, since it depends quadratically on a combination of
23 (
0:05, in
our case) and ", as opposed to the rst term which is linear. However, for energies matching
the oscillation peak, the rst term will be strongly suppressed with the oscillatory phase.
Due to the simultaneous dependence of P
on
23 and "~ , a degeneracy appears
in this plane. In fact, while in the standard scenario the DUNE experiment is able to
successfully resolve the octant of 23 (see gure 8 in appendix A), when NSI are marginalized
over in the t this is no longer the case, and the fake solution in the higher octant reappears.
This is explicitly shown in gure 3. The left panel shows the results projected onto the
23
plane for three di erent scenarios: when no NSI are considered in the analysis (solid
yellow), when NSI are marginalized over within current priors (dashed green) and when
NSI are marginalized over with no priors on the NSI parameters (dotted blue). As it can
be seen from the
gure, the higher octant solution is not allowed by the data when NSI
are not included in the t, but reappears if they are marginalized over (see also gures 7
and 8 in appendix A). The reason is that there is a strong degeneracy between "~
and 23,
explicitly shown in the right panel. In the case where no prior uncertainties are assumed for
the NSI parameters (dotted blue line), two additional solutions appear around 23 = 45 .
However, these take place for values of "~
in tension with current constraints, and are
therefore partially removed when the prior on the "
parameter is imposed (dashed green
lines). Finally, we nd that when T2HK is added to the DUNE data the degeneracy is
almost completely solved, as it is shown by the dotdashed gray contours.
1.0
0.8
È 0.6
e
Τ
¶
È
DUNE  no priors
DUNE  w priors
T2HK+DUNE  w priors
2
1
0
Ž
¶
ee
L
°
H
e
Τ
Φ
100
50
0
50
100
150
1
2
3
3
2
1
1
2
3
0
Ž
¶
ee
j" ej plane for DUNE and for DUNE+T2HK, as indicated in
the legend. For DUNE we also show the resulting region when no prior uncertainties are imposed
on NSI during the t. In all cases, the contours enclose the 90% credible regions.
The second degeneracy we found in this study takes place between the CP violating
phase , and the NSI parameters "~ee and " e (including its CP phase). In this case, due
to the large values of "~ee involved, perturbation theory cannot be used to understand the
interplay of parameters. The degeneracy is explicitly shown in gure 4, for DUNE and for
DUNE+T2HK, in the planes "~ee
j" ej (left panel) and "~ee
e (right panel). As can be
seen from this
gure, there is a nontrivial dependence with the CPviolating phase
which is responsible of this degeneracy: while for small values of "~ee all values of
e
equally probable, as the value of "~ee increases only certain values of
e are possible (namely,
a negative phase for "~ee < 0, while only positive phases are allowed if "~ee > 0). This also
illustrates why in gure 2 the sensitivity to " e improves so dramatically in the region where
e < 0. Again in this case, when T2HK is added to the DUNE data the degeneracy is again
partially solved, although not completely, as can be seen from the solid contours in gure 4.
The fact that this degeneracy depends on the value of
e suggests that it might have
a relevant impact on CPviolation searches. This is shown explicitly
gure 5, where the
oscillation probabilities are shown for the
!
e and
!
oscillation channels at
L = 1300 km as a function of the neutrino energy, for three di erent cases. The solid
blue lines show the probabilities in the standard case, with true values of the oscillation
parameters matching the best t values from ref. [
56
] and
=
90 . The dashed red line,
on the other hand, shows the probabilities for "~ee =
2 and " e = 0:45,
e =
150 , where the rest of the NSI parameters are taken to be zero and the standard
ones are unchanged with respect to the standard scenario. Finally, the dotted green line
shows the probabilities for "~ee = 1, " e = 0:25,
=
90 . The three
probabilities are identical, as can be seen from the gure, which could eventually lead to
a misinterpretation of the data and a wrong determination of the value of . To the best
HJEP03(216)
2
4
6
8
10
E HGeVL
!
HJEP03(216)
lation channels, under the assumption of standard oscillations only, and two di erent set of NSI
parameters. Set (a) corresponds to "~ee =
set (b) assumes "~ee = 1, j" ej = 0:25,
2 and j" ej = 0:45,
e =
of our knowledge, this degeneracy has not been studied previously in the literature.6 A
detailed study would be needed to address its impact on CP violation searches at DUNE.
This remains beyond the scope of this work and is left for future studies.
4.3
Comparison to other facilities and to prior experimental constraints
It is interesting to compare the DUNE sensitivities to current constraints as well as to
other oscillation experiments currently in operation (such as T2K and/or NOvA) or being
planned for the future (such as T2HK). Our results from this comparison are presented in
gure 6, where the colored bands indicate the credible intervals found at 90% found for each
of the NSI parameters, either for the experiments alone or in combination with one another.
Results are presented for the moduli of the di erent NSI parameters, after marginalization
over the remaining oscillation parameters and the CPphases. The results are compared to
the constraints from previous experiments (see table 1 or gure 6 in ref. [54]), indicated by
the dashed vertical lines. We have found that the combination of T2K and NOvA is not
sensitive to NSI below the current constraints derived in ref. [54], due to the presence of
strong degeneracies among di erent oscillation parameters, and therefore their results are
not shown in this gure.
The most important feature in gure 6 can be seen in the uppermost panel, where the
sensitivity to "~ee is shown and compared to the currently allowed regions by global ts to
neutrino oscillation data. As can be seen from this panel, under the assumption of no
relevant NSI e ects in the oscillation probability, both DUNE and T2HK will be able to probe
the LMAdark solution. The possibility of ruling out the LMAdark solution with
long6The degeneracy in the "~ee
" e plane shows similar features to the degeneracy studied in refs. [70{72].
Both degeneracies might be related but there are important di erences. While the degeneracy studied in
refs. [70{72] appeared in the disappearance probabilities, our degeneracy takes place in the appearance
channels instead and involves the new CPphases. Furthermore, the relation between " e and "~ee is also
di erent: while in our case the degeneracy imposes a linear relation between the two parameters, in refs. [70{
72] the degeneracy took place along a parabola. This indicates that a possible way to break this degeneracy
could be through combination with atmospheric neutrino data.
baseline experiments was already pointed out previously in the literature. For instance, in
ref. [45] it was found that NOvA could rule out this solution at approximately 85% CL.
We
nd, however, that the NOvA experiment on its own (or in combination with T2K)
will not be able to rule out the LMAdark solution. Due to the strong degeneracy between
"~ee and " e (see section 4.2), it is always possible to reconcile the t and the simulated
data by assuming simultaneously large values for "~ee and " e. This degeneracy is partially
solved when prior constraints are imposed on " e; however, we nd that a small region of
the parameter space around "~ee
3 and j" ej
0:45 still provides a good t to the data.
Conversely, DUNE and/or T2HK will be sensitive enough to the presence of NSI in order
to rule out the LMAdark solution on their own. The rejection power is then increased if
prior constraints on NSI parameters are included, as expected (dark bands in gure 6).
According to our results, the DUNE experiment will also be able to improve current
constraints on " e and " e by a factor of between 2 and 5, and at least by a factor of two
with respect to the results expected at T2HK alone, as can be seen from the comparison
of the light colored bands. In the case of " , the sensitivity when no prior is imposed goes
above the current experimental constraint, indicating that the sensitivity to this parameter
is somewhat limited. However, as it was shown in gure 2, the sensitivity to this parameter
depends strongly on the value of its CPviolating phase, and DUNE is expected to improve
over the current limit as long as
6
=
=2, see gure 2.
Finally, it is worth pointing out that, on its own, DUNE will not be able to improve
over current constraints for "~ , for the set of true oscillation parameters assumed in this
work. In this case, combination with T2HK would be essential. As can be seen from
the second panel in
gure 6, before combination none of the two experiments is able to
improve over current experimental constraints, although they favour di erent regions in
the parameter space. Thus, after combination, the sensitivity to "~
is notably improved,
yielding a slightly better result than the ones from current limits.
5
Conclusions
Neutrino physics is entering the precision Era. After the discovery of the third mixing angle
in the leptonic mixing matrix, and in view of the precision measurements performed by the
reactor experiments (most notably, Daya Bay) and longbaseline experiments (MINOS,
T2K and, in the near future, NO A), it appears timely to reevaluate the sensitivity of
current and future oscillation experiments to possible NonStandard neutrino Interactions
(NSI). We have focused on the impact of NSI on neutrinos in propagation through matter,
something for which the planned Deep Underground Neutrino Experiment (DUNE) is very
well suited for, due to its relatively high energies and very long baseline. Given the current
experimental and theoretical e ort to keep systematic uncertainties below the 2%5% level,
it o ers a very wellsuited environment to conduct New Physics searches.
In this work, a Monte Carlo Markov Chain (MCMC) has been used to explore the
multidimensional parameter space surrounding the global minimum of the 2. The total number
of parameters which are allowed to vary in the t is fourteen: six standard oscillation
parameters, ve moduli for the nonstandard parameters, and three new CPviolating phases.
Prior experimental constraints, completely modelindependent, have been implemented in
Credible Intervals at 90%
rak
d
T2HK+DUNE
A
A
M
M
L
L
4
3
2
1
2
3
Credible Intervals at 90%
0.6
0.4
0.2
¶ΜΜ
T2HK+DUNE
¶ΜΤ
and after combining their respective data sets. Darker (Lighter) bands show the results when priors
constraints on NSI parameters are (not) included in the
t. The vertical gray areas bounded by
the dashed lines indicate the allowed regions at 90% CL (taken from the SNODATA lines for f=u
in ref. [54]).
DUNE with no priors on NSI
DUNE with priors Current constraint
"~ee
"~
j" ej
j" ej
j
" j
cients accompanying the NSI fourfermion operators a ecting neutrino propagation in matter. The
rede nition "~
"
"
has been used, see section 2 for details. For comparison, the last
column shows the current constraints at 90% CL extracted from a global t to neutrino oscillation
data (taken from the SNODATA lines for f=u in ref. [54]).
our simulations, see section 3.1 and appendix A for details. By including all (standard and
nonstandard) parameters at once in the simulation, we derive conservative and completely
modelindependent limits on each of the coe cients accompanying the new operators
entering the e ective operator expansion. At the same time, we fully take into account possible
degeneracies among di erent parameters entering the oscillation probabilities.
We have identi ed two potentially important degeneracies among standard and
nonstandard parameters. The rst one takes place in the disappearance channels between
23 and "~ , and could be potentially harmful for the octant sensitivity of the DUNE
experiment. While in the standard case we nd that the DUNE experiment is able to reject
the higher octant solution, this is no longer the case if the "~
parameter is marginalized
over during the t. The second degeneracy takes place between "~ee, " e,
e and
in the
appearance channels. The interplay between the di erent parameters in this case is
nontrivial and it involves one of the nonstandard CPviolating phases,
e. This degeneracy
could potentially pose a challenge for standard CPviolating searches and a more careful
study will be left for future work.
One of the most relevant results shown in the present study is that the DUNE
experiment will be able to probe the socalled LMAdark solution. The LMAdark solution, which
is fully compatible with current oscillation data [54], favors a large nonstandard matter
potential driven by "~ee
3 and a solar mixing angle in the second octant, 12 >
=4. We
nd that, for the true oscillation parameters assumed in this work, the credible regions at
90% do not include the LMAdark region, see gures 4 and 6.
We nd that DUNE will be able to improve over current constraints on " e by at least
a factor of ve, and on " e by at least a 20%. The sensitivity to " e shows a signi cant
(and nontrivial) dependence with the value of its associated CPphase and, in particular, is
HJEP03(216)
signi cantly a ected by the current prior on "~ee (see gures 4 and 2). Regarding " , DUNE
will be able to improve over current constraints as long as
=2, see gure 2. Finally,
we nd that DUNE will not be able to improve over current constraints on "~ , for the set of
6
=
true oscillation parameters assumed in this work. For convenience, the expected sensitivity
of DUNE to NSI parameters is summarized in table 2, where the credible intervals are
given at 90%.
Finally, we have also compared the expected reach for the DUNE experiment to that of
the current generation of longbaseline experiments and to the future T2HK proposal. We
found that the combination of T2K and NOvA will not be sensitive enough to the presence
of NSI in order to improve over current constraints from oscillation data. The T2HK
experiment on its own will not be able to improve over current constraints either for most
parameters, with the exception of " e. Interestingly enough, we nd that the combination of T2HK
and DUNE is able to partially resolve the degeneracies discussed in section 4.2. In
particular, the combination of DUNE and T2HK would yield a strong improvement in the
determination of "~
and solve almost completely the degeneracy between "~
and 23, see gure 3.
Note added: the preprint version of ref. [73] was made available online two days before
the present manuscript. In ref. [73], a very similar analysis was performed for nonstandard
interactions in propagation at DUNE.
Acknowledgments
I am especially grateful to Enrique FernandezMartinez for support regarding the use of the
MonteCUBES software as well as for useful discussions and comments on the manuscript.
I would like to thank Jacobo LopezPavon and Stephen Parke for useful comments on
the manuscript, and Alexander Friedland, Andre de Gouvea and Thomas Schwetz for
useful discussions. I would also like to thank David Vanegas Forero for his help in writing
the T2K
les with the 2013
uxes, and Michele Maltoni for useful communications
regarding the prior constraints on NSI coming from current oscillation data. I acknowledge
partial support by the European Union through the ITN INVISIBLES (Marie Curie
Actions, PITNGA2011289442 INVISIBLES). Fermilab is operated by the Fermi Research
Alliance under contract DEAC0207CH11359 with the U.S. Department of Energy.
A
Implementation of prior constraints
In order to restrict the region sampled by the MCMC to the physical region of interest,
priors have been implemented for all parameters (standard and nonstandard) in our
simulations, with the only exception of the standard CPviolating phase , since current hints
only have a limited statistical signi cance at the 1
2
CL (see, however, refs. [
56, 74
]
for recent discussions on this topic). Since the measurements on 13 and 23 do not come
from a direct measurement of the angles themselves, these priors have been implemented
according to the quantities that are directly measured at oscillation experiments. For 13
this amount to imposing a gaussian prior on sin2 2 13. In the case of 23, however, the
situation is a bit more complicated. The most precise determination of 23 comes from
the observation of
disappearance at longbaseline experiments, which measure an
\effective" mixing angle sin2 2
, see e.g., refs. [75, 76]. Given the large value of 13, the
Prior (at 68%)
0.02
3%
3%
agreement with the current uncertainties from ref. [
56
], except for sin2 2 23 for which the prior has
been relaxed by a factor of two.
correspondence
$ 23 no longer takes place. Instead, the following relation holds:
sin
= sin 23 cos 13 :
(A.1)
HJEP03(216)
Therefore, a gaussian prior a ecting 23 has been implemented on this e ective angle
instead, since this is the quantity which is actually constrained by longbaseline experiments.
The DUNE experiment will provide the most precise determination of this parameter,
though. Therefore, in this case only a mild prior has been imposed, relaxing the current
constraints by a factor of two, in order to ease convergence of the simulations. Finally,
for the solar mixing angle we have implemented a gaussian prior on sin2 2 12 since, in
practice, this is the only quantity that can be determined from current oscillation data.
Table 3 summarizes the priors implemented for the standard oscillation parameters, which
are assumed to be gaussian.
For the NSI parameters, we have implemented nongaussian priors, extracted from the
results for SNODATA lines from
gure 6 in ref. [54], for f=u. These have been rescaled
according to the relation "
= 3:051"u . We have considered that both the LMA and
LMAdark solutions are equally allowed by the data.
Finally, a typical problem usually encountered when a multidimensional parameter
space is explored using a MCMC has to do with the existence of multiple minima. If the
2 between di erent minima is large enough, the MCMC will generally tend to sample only
one of them, leaving the rest unexplored. This is specially relevant in neutrino oscillations,
where degeneracies are expected to arise between di erent parameters, even in absence of
NSI [78{81]. This problem is dealt with in our simulations by the use of \degeneracy steps",
chosen speci cally to make sure that all possible degeneracies are explored by the MCMC.
For example, since a nonmaximal value of 23 has been considered in our simulations, an
obvious choice in this case is to add a larger step in the 23 direction so as to guarantee that
the octant degeneracy is appropriately sampled. Additional steps in the " directions have
also been set up in order to guarantee that all possible degenerate solutions are found in the
simulations (for instance, in order to guarantee that the LMAdark solution is appropriately
sampled, we have added a step in the "~ee direction with
"~ee = 4).
Figure 7 shows explicitly that the octant degeneracies are well sampled in our
simulations. This gure shows the same type of one and twodimensional projections of the
MCMC results as in gure 1, for the standard oscillation parameters,7 assuming no priors
over the NSI parameters. As it can be clearly seen from this gure, the octant degeneracy
7Longbaseline experiments are not sensitive to the solar parameters and therefore their measurement
is not expected to improve over the assumed priors. For this reason we only show the projections for
13; 23;
m321 and . Nevertheless, solar parameters are always left free during marginalization, within the
assumed priors listed in table 3.
23 0.8
2.48
δ
2
2
for the standard oscillation parameters, after marginalizing over all NSI parameters. No prior
constraints on NSI parameters are have been imposed. The red, green and blue lines indicate the
credible regions at 68%, 90% and 95%. The vertical green bands indicate the credible intervals at 68%.
in the 23 axis has been properly sampled by our MCMC, and three well separated regions
are obtained. For comparison, gure 7 shows the same projections when no NSI are
allowed in the t (i.e., only standard parameters are allowed in the t). In this case, the
octant degeneracies disappear, in agreement with the results in previous literature (see,
e.g., refs. [35, 82]).
Finally, it should be mentioned that the T2HK experiment [23] is not sensitive to the
neutrino mass ordering at high con dence level for all possible values of the CPviolating
phase
and all values of the atmospheric mixing angle. Therefore, degeneracies in the
m231 direction are expected to take place, and should be explored as well.
Nevertheless, the determination of the mass ordering might come instead from a combination of
di erent facilities [83{89], from atmospheric data at HK [23], or from the combination of
T2K+NO A at some level, if the current hint for
=2 persists in the future.
Therefore, we will adopt an optimistic approach in this paper and assume that the neutrino mass
ordering is determined by the time these experiments nish taking data. Normal ordering
has been assumed in all our simulations.
0.85
δ
This article is distributed under the terms of the Creative Commons
Attribution License (CCBY 4.0), which permits any use, distribution and reproduction in
any medium, provided the original author(s) and source are credited.
[INSPIRE].
JHEP 09 (2009) 038 [arXiv:0906.1461] [INSPIRE].
[4] R. Alonso et al., Summary report of MINSIS workshop in Madrid, arXiv:1009.0476
[5] nuSTORM collaboration, D. Adey et al., nuSTORM  neutrinos from STORed Muons:
proposal to the Fermilab PAC, arXiv:1308.6822 [INSPIRE].
[6] ISS Physics Working Group collaboration, A. Bandyopadhyay, Physics at a future
neutrino factory and superbeam facility, Rept. Prog. Phys. 72 (2009) 106201
[arXiv:1305.4350] [INSPIRE].
[arXiv:1310.5917] [INSPIRE].
[9] I. Girardi, D. Meloni and S.T. Petcov, The Daya Bay and T2K results on sin2 2 13 and
nonstandard neutrino interactions, Nucl. Phys. B 886 (2014) 31 [arXiv:1405.0416]
reactor neutrino experiments, J. Phys. G 42 (2015) 065003 [arXiv:1411.5330] [INSPIRE].
[11] S.K. Agarwalla, P. Bagchi, D.V. Forero and M. Tortola, Probing nonstandard interactions
at Daya Bay, JHEP 07 (2015) 060 [arXiv:1412.1064] [INSPIRE].
[12] M. Blennow, S. Choubey, T. Ohlsson and S.K. Raut, Exploring source and detector
nonstandard neutrino interactions at ESS SB, JHEP 09 (2015) 096 [arXiv:1507.02868]
[13] S. Choubey, A. Ghosh, T. Ohlsson and D. Tiwari, Neutrino physics with nonstandard
interactions at INO, JHEP 12 (2015) 126 [arXiv:1507.02211] [INSPIRE].
[14] S. Choubey and T. Ohlsson, Bounds on nonstandard neutrino interactions using PINGU,
Phys. Lett. B 739 (2014) 357 [arXiv:1410.0410] [INSPIRE].
[15] T. Ohlsson, H. Zhang and S. Zhou, E ects of nonstandard neutrino interactions at PINGU,
Phys. Rev. D 88 (2013) 013001 [arXiv:1303.6130] [INSPIRE].
[16] I. Mocioiu and W. Wright, Nonstandard neutrino interactions in the mutau sector, Nucl.
Phys. B 893 (2015) 376 [arXiv:1410.6193] [INSPIRE].
[17] CAPTAIN collaboration, H. Berns et al., The CAPTAIN detector and physics program, in
Community summer study 2013: Snowmass on the Mississippi (CSS2013), Minneapolis MN
U.S.A. July 29{August 6 2013 [arXiv:1309.1740] [INSPIRE].
[18] MINER A collaboration, L. Fields, CCQE results from MINER A, AIP Conf. Proc. 1663
(2015) 080006 [INSPIRE].
[19] ArgoNeuT and MicroBooNE collaborations, A.M. Szelc, Recent results from ArgoNeuT
and status of MicroBooNE, Nucl. Part. Phys. Proc. 265266 (2015) 208 [INSPIRE].
[20] DUNE collaboration, R. Acciarri et al., LongBaseline Neutrino Facility (LBNF) and Deep
Underground Neutrino Experiment (DUNE) conceptual design report volume 2: the physics
program for DUNE at LBNF, arXiv:1512.06148 [INSPIRE].
[21] T2K collaboration, K. Abe et al., Measurements of neutrino oscillation in appearance and
disappearance channels by the T2K experiment with 6:6
D 91 (2015) 072010 [arXiv:1502.01550] [INSPIRE].
1020 protons on target, Phys. Rev.
arXiv:1209.0716 [INSPIRE].
longbaseline neutrino oscillation experiment using a JPARC neutrino beam and
[INSPIRE].
[INSPIRE].
[INSPIRE].
[24] P. Huber, T. Schwetz and J.W.F. Valle, Confusing nonstandard neutrino interactions with
oscillations at a neutrino factory, Phys. Rev. D 66 (2002) 013006 [hepph/0202048]
interactions, Phys. Rev. D 78 (2008) 053007 [arXiv:0804.2261] [INSPIRE].
[31] D. Meloni, T. Ohlsson, W. Winter and H. Zhang, Nonstandard interactions versus
nonunitary lepton
avor mixing at a neutrino factory, JHEP 04 (2010) 041
[arXiv:0912.2735] [INSPIRE].
[INSPIRE].
044201 [arXiv:1209.2710] [INSPIRE].
[32] P. Coloma, A. Donini, J. LopezPavon and H. Minakata, Nonstandard interactions at a
neutrino factory: correlations and CPviolation, JHEP 08 (2011) 036 [arXiv:1105.5936]
[34] O.G. Miranda and H. Nunokawa, Non standard neutrino interactions: current status and
future prospects, New J. Phys. 17 (2015) 095002 [arXiv:1505.06254] [INSPIRE].
[35] P. Huber and J. Kopp, Two experiments for the price of one? The role of the second
oscillation maximum in long baseline neutrino experiments, JHEP 03 (2011) 013 [Erratum
ibid. 05 (2011) 024] [arXiv:1010.3706] [INSPIRE].
[36] A. Friedland and I.M. Shoemaker, Searching for novel neutrino interactions at NO A and
beyond in light of large 13, arXiv:1207.6642 [INSPIRE].
[37] Daya Bay collaboration, F.P. An et al., Observation of electronantineutrino disappearance
at Daya Bay, Phys. Rev. Lett. 108 (2012) 171803 [arXiv:1203.1669] [INSPIRE].
[arXiv:1204.0626] [INSPIRE].
[39] Double CHOOZ collaboration, Y. Abe et al., Indication for the disappearance of reactor
electron antineutrinos in the Double CHOOZ experiment, Phys. Rev. Lett. 108 (2012) 131801
[41] P. Langacker and D. London, Lepton number violation and massless nonorthogonal
HJEP03(216)
neutrinos, Phys. Rev. D 38 (1988) 907 [INSPIRE].
[42] Z. Berezhiani and A. Rossi, Limits on the nonstandard interactions of neutrinos from e+e
colliders, Phys. Lett. B 535 (2002) 207 [hepph/0111137] [INSPIRE].
[43] S. Davidson, C. PenaGaray, N. Rius and A. Santamaria, Present and future bounds on
nonstandard neutrino interactions, JHEP 03 (2003) 011 [hepph/0302093] [INSPIRE].
[44] S. Antusch, J.P. Baumann and E. FernandezMartinez, Nonstandard neutrino interactions
with matter from physics beyond the standard model, Nucl. Phys. B 810 (2009) 369
[arXiv:0807.1003] [INSPIRE].
[45] Y. Farzan, A model for large nonstandard interactions of neutrinos leading to the LMAdark
solution, Phys. Lett. B 748 (2015) 311 [arXiv:1505.06906] [INSPIRE].
[46] A. Friedland, M.L. Graesser, I.M. Shoemaker and L. Vecchi, Probing nonstandard standard
model backgrounds with LHC monojets, Phys. Lett. B 714 (2012) 267 [arXiv:1111.5331]
[47] D.B. Franzosi, M.T. Frandsen and I.M. Shoemaker, New or
missing energy?
Discriminating dark matter from neutrino interactions at the LHC, arXiv:1507.07574
[48] J. Barranco, O.G. Miranda, C.A. Moura and J.W.F. Valle, Constraining nonstandard
interactions in ee or ee scattering, Phys. Rev. D 73 (2006) 113001 [hepph/0512195]
[49] C. Biggio, M. Blennow and E. FernandezMartinez, Loop bounds on nonstandard neutrino
interactions, JHEP 03 (2009) 139 [arXiv:0902.0607] [INSPIRE].
[50] C. Biggio, M. Blennow and E. FernandezMartinez, General bounds on nonstandard
neutrino interactions, JHEP 08 (2009) 090 [arXiv:0907.0097] [INSPIRE].
[51] O.G. Miranda, M.A. Tortola and J.W.F. Valle, Are solar neutrino oscillations robust?, JHEP
10 (2006) 008 [hepph/0406280] [INSPIRE].
[52] F.J. Escrihuela, O.G. Miranda, M.A. Tortola and J.W.F. Valle, Constraining nonstandard
neutrinoquark interactions with solar, reactor and accelerator data, Phys. Rev. D 80 (2009)
105009 [Erratum ibid. D 80 (2009) 129908] [arXiv:0907.2630] [INSPIRE].
[53] M.C. GonzalezGarcia, M. Maltoni and J. Salvado, Testing matter e ects in propagation of
atmospheric and longbaseline neutrinos, JHEP 05 (2011) 075 [arXiv:1103.4365] [INSPIRE].
[54] M.C. GonzalezGarcia and M. Maltoni, Determination of matter potential from global
analysis of neutrino oscillation data, JHEP 09 (2013) 152 [arXiv:1307.3092] [INSPIRE].
[55] T. Kikuchi, H. Minakata and S. Uchinami, Perturbation theory of neutrino oscillation with
nonstandard neutrino interactions, JHEP 03 (2009) 114 [arXiv:0809.3312] [INSPIRE].
MonteCUBES, Comput. Phys. Commun. 181 (2010) 227 [arXiv:0903.3985] [INSPIRE].
HJEP03(216)
experiments with GLoBES (General Long Baseline Experiment Simulator), Comput. Phys.
Commun. 167 (2005) 195 [hepph/0407333] [INSPIRE].
of neutrino oscillation experiments with GLoBES 3:0: General Long Baseline Experiment
Simulator, Comput. Phys. Commun. 177 (2007) 432 [hepph/0701187] [INSPIRE].
[61] SuperKamiokande collaboration, G. Mitsuka et al., Study of nonstandard neutrino
interactions with atmospheric neutrino data in SuperKamiokande I and II, Phys. Rev. D 84
(2011) 113008 [arXiv:1109.1889] [INSPIRE].
[62] LBNF letter of intent, submitted to the Fermilab PAC P1062, U.S.A. December 2014.
[63] LBNE collaboration, T. Akiri et al., The 2010 interim report of the LongBaseline Neutrino
Experiment collaboration physics working groups, arXiv:1110.6249 [INSPIRE].
[64] M. Blennow, P. Coloma, A. Donini and E. FernandezMartinez, Gain fractions of future
neutrino oscillation facilities over T2K and NO A, JHEP 07 (2013) 159 [arXiv:1303.0003]
[INSPIRE].
[INSPIRE].
[65] T2K collaboration, K. Abe et al., T2K neutrino
ux prediction, Phys. Rev. D 87 (2013)
012001 [Addendum ibid. D 87 (2013) 019902] [arXiv:1211.0469] [INSPIRE].
[66] P. Coloma, P. Huber, J. Kopp and W. Winter, Systematic uncertainties in longbaseline
neutrino oscillations for large 13, Phys. Rev. D 87 (2013) 033004 [arXiv:1209.5973]
[67] K. Abe et al., Letter of intent: the HyperKamiokande experiment  detector design and
physics potential, arXiv:1109.3262 [INSPIRE].
Interiors 25 (1981) 297 [INSPIRE].
[68] A.M. Dziewonski and D.L. Anderson, Preliminary reference earth model, Phys. Earth Planet.
[69] M. Masud, A. Chatterjee and P. Mehta, Probing CPviolation signal at DUNE in presence of
nonstandard neutrino interactions, arXiv:1510.08261 [INSPIRE].
[70] A. Friedland, C. Lunardini and M. Maltoni, Atmospheric neutrinos as probes of
neutrinomatter interactions, Phys. Rev. D 70 (2004) 111301 [hepph/0408264] [INSPIRE].
[71] A. Friedland and C. Lunardini, Two modes of searching for new neutrino interactions at
MINOS, Phys. Rev. D 74 (2006) 033012 [hepph/0606101] [INSPIRE].
[72] A. Friedland and C. Lunardini, A test of tau neutrino interactions with atmospheric
neutrinos and K2K, Phys. Rev. D 72 (2005) 053009 [hepph/0506143] [INSPIRE].
[73] A. de Gouv^ea and K.J. Kelly, Nonstandard neutrino interactions at DUNE,
arXiv:1511.05562 [INSPIRE].
1350093 [arXiv:1209.5658] [INSPIRE].
[arXiv:1406.2551] [INSPIRE].
JHEP 10 (2001) 001 [hepph/0108085] [INSPIRE].
superbeams using liquid argon detectors, JHEP 03 (2014) 087 [arXiv:1304.3251] [INSPIRE].
[arXiv:1306.2500] [INSPIRE].
experiments: an `adequate' con guration for LBNO, JHEP 03 (2014) 094
[arXiv:1308.5979] [INSPIRE].
PINGU and Daya Bay II, JHEP 09 (2013) 089 [arXiv:1306.3988] [INSPIRE].
T2K, NO A and reactor experiments, JHEP 04 (2013) 009 [arXiv:1212.1305] [INSPIRE].
[1] S. Weinberg , Baryon and lepton nonconserving processes , Phys. Rev. Lett . 43 ( 1979 ) 1566 [2] M.B. Gavela , D. Hernandez , T. Ota and W. Winter , Large gauge invariant nonstandard neutrino interactions , Phys. Rev. D 79 ( 2009 ) 013007 [arXiv: 0809 .3451] [INSPIRE].
[3] M.B. Gavela , T. Hambye , D. Hernandez and P. Hernandez , Minimal avour seesaw models, [7] A.N. Khan , D.W. McKay and F. Tahir , Sensitivity of mediumbaseline reactor neutrino masshierarchy experiments to nonstandard interactions , Phys. Rev. D 88 ( 2013 ) 113006 [8] T. Ohlsson , H. Zhang and S. Zhou , Nonstandard interaction e ects on neutrino parameters at mediumbaseline reactor antineutrino experiments , Phys. Lett. B 728 ( 2014 ) 148 [22] NO A collaboration , R.B. Patterson , The NO A experiment: status and outlook , [ 23 ] HyperKamiokande proto collaboration, K. Abe et al., Physics potential of a HyperKamiokande, Prog. Theor. Exp. Phys . 2015 ( 2015 ) 053C02 [arXiv: 1502 .05199] [25] J. Kopp , M. Lindner , T. Ota and J. Sato , Nonstandard neutrino interactions in reactor and superbeam experiments , Phys. Rev. D 77 ( 2008 ) 013007 [arXiv: 0708 .0152] [INSPIRE].
[26] J. Kopp , P.A.N. Machado and S.J. Parke , Interpretation of MINOS data in terms of nonstandard neutrino interactions , Phys. Rev. D 82 ( 2010 ) 113002 [arXiv: 1009 .0014] [27] J. Kopp , M. Lindner and T. Ota , Discovery reach for nonstandard interactions in a neutrino factory , Phys. Rev. D 76 ( 2007 ) 013001 [ hep ph/0702269] [INSPIRE].
[28] M. Blennow , T. Ohlsson and J. Skrotzki , E ects of nonstandard interactions in the MINOS experiment , Phys. Lett. B 660 ( 2008 ) 522 [ hep ph/0702059] [INSPIRE].
[29] M. Blennow , D. Meloni , T. Ohlsson , F. Terranova and M. Westerberg , Nonstandard interactions using the OPERA experiment , Eur. Phys. J. C 56 ( 2008 ) 529 [30] J. Kopp , T. Ota and W. Winter , Neutrino factory optimization for nonstandard [33] T. Ohlsson, Status of nonstandard neutrino interactions , Rept. Prog. Phys . 76 ( 2013 ) disappearance in the RENO experiment , Phys. Rev. Lett . 108 ( 2012 ) 191802 experiment , Prog. Theor. Exp. Phys . 2015 ( 2015 ) 043C01 [arXiv: 1409 .7469] [INSPIRE].
[56] M.C. GonzalezGarcia , M. Maltoni and T. Schwetz , Updated t to three neutrino mixing: status of leptonic CPviolation , JHEP 11 ( 2014 ) 052 [arXiv: 1409 .5439] [INSPIRE].
[57] A. Friedland , C. Lunardini and C. PenaGaray , Solar neutrinos as probes of neutrino matter interactions , Phys. Lett. B 594 ( 2004 ) 347 [ hep ph/0402266] [INSPIRE].
[58] M. Blennow and E. FernandezMartinez , Neutrino oscillation parameter sampling with [59] P. Huber , M. Lindner and W. Winter , Simulation of longbaseline neutrino oscillation [60] P. Huber , J. Kopp , M. Lindner , M. Rolinec and W. Winter , New features in the simulation [74] J. Elevant and T. Schwetz , On the determination of the leptonic CP phase , JHEP 09 ( 2015 ) [75] S.K. Raut , E ect of nonzero 13 on the measurement of 23, Mod . Phys. Lett. A 28 ( 2013 ) [76] P. Coloma , H. Minakata and S.J. Parke , Interplay between appearance and disappearance channels for precision measurements of 23 and , Phys. Rev. D 90 ( 2014 ) 093003 [77] Daya Bay collaboration , L. Zhan, Recent results from Daya Bay , PoS(NEUTEL2015) 017 [78] J. BurguetCastell , M.B. Gavela , J.J. GomezCadenas , P. Hernandez and O. Mena , On the measurement of leptonic CPviolation , Nucl. Phys. B 608 ( 2001 ) 301 [hep ph/0103258]