Constraints on nonstandard intermediate boson exchange models from neutrinoelectron scattering
Abstract
Constraints on couplings of several beyondStandardModelphysics scenarios, mediated by massive intermediate particles including (1) an extra Zprime, (2) a new light spin1 boson, and (3) a charged Higgs boson, are placed via the neutrinoelectron scattering channel to test the Standard Model at a low energymomentum transfer regime. Data on and scattering from the TEXONO and LSND, respectively, are used. Upper bounds to coupling constants of the flavorconserving and flavorviolating new light spin1 boson and the charged Higgs boson with respect to different mediator masses are determined. The relevant parameter spaces are extended by allowing light mediators. New lower mass limits for extra Zprime gauge boson models are also placed.
Corresponding Author: ] muhammed.deniz@deu.edu.tr
I Introduction
Recent discovery of nonzero neutrino mass and mixing undoubtedly implies new physics beyond the Standard Model (BSM). Nevertheless, the origin of the masses of neutrinos and absolute mass scale remain unknown. The seesaw mechanisms, Rparityviolating supersymmetry (SUSY), TeV scale loop mechanisms, extra dimensions, and string theory are the most popular proposals attempting to answer these questions and explain the origin of neutrino mass king2015 (). However, in the underlying new physics BSM, it is mostly expected that the structure of electroweak charged and neutral currents of the Standard Model (SM) would also change. Such changes in the neutrino sector lead to nonstandard interactions (NSI) of neutrinos. In many works on NSI, new interactions are generally mediated by new particles, which are assumed to be heavier than the electroweak scale. Hence, these are carried out in the form of effective fourfermion interaction at low energy. Furthermore, it is also possible mediated new particles can have relatively low masses.
Neutrino interactions, being pure leptonic processes, are one of the most appropriate mechanisms to test the electroweak theory of the SM jpanmanwjmarciano (); ohlsson2013 (); miranda2015 (); jerler125 (). Therefore, the elastic scattering can be used to search for BSM new physics scenarios mediated by a massive intermediate boson such as the extra Zprime gauge boson (), new light spin1 boson (NLS1B), and charged Higgs boson (CHB), which are predicted by certain models to describe new interactions mediated by new particles in addition to the SM electroweak and gauge bosons.
In this paper, we report experimental constraints on the masses and coupling parameters for the exchange of the NLS1B and CHB as well as lower mass limits on from and elastic scattering.
Ii NeutrinoElectron Scattering and Data
ii.1 Standard Model
Since incoming neutrinos are of electrontype flavor, elastic scattering can occur via both charged current and neutral current interaction. Therefore, their interference, which is destructive, also contributes to the cross section. The SM differential cross section of elastic scattering can be expressed in the laboratory frame as mdeniz033004 (); sbilmis073011 (); jiunn011301 (); sbilmis033009 (); bkayser87 ()
(1)  
where is the Fermi coupling constant, is the kinetic energy of the recoil electron, is the incident neutrino energy, and the coefficients of and are given in Table 1 in terms of chiral coupling constants and , weak mixing angle , and vectoraxialvector coupling constants and , which are defined as and , respectively.
Coefficients  

ii.2 Input data
The analysis of the experimental data sets of TEXONO, as a sample of the antineutrino channel at low energy with three different detectors located at KSNL whose energy ranges are different, and LSND, as a sample of the neutrino channel at high energy, is reported. The published results of the differential cross section measurements are used for each sample. The results from three independent data sets of the TEXONO interaction are compared with those from the LSND interaction.
 TEXONO Experiment

Three experimental data sets taken with different detectors are used as follows:
 CsI(Tl)

29882/7369 kgdays of reactor on/off data: electroweak interaction cross section, , , weak mixing angle , and charge radius squared were measured with an effective mass of 187 kg CsI(Tl) crystal scintillator array at 3 8 MeV. The rootmeansquare (RMS) energy resolutions are 5.8%, 5.2%, and 4.0% at Cs, K, and Tl peaks, respectively. The residual reactor on reactor off event rate spectrum at 3 8 MeV shown in Fig. 16(b) of Ref. mdeniz072001 () is used for this analysis.
 HPGe

570.7/127.8 kgdays of reactor on/off data: New limits are set to the neutrino magnetic moment and axion with a target mass of 1.06 kg highpurity germanium detector hblihtwong () at 12 64 keV. The RMS energy resolution of HPGe is 880 keV at GaK shell xray energy soma2016 (). The residual reactor on reactor off event rate spectrum at 12 64 keV shown in Fig. 13 of Ref. hblihtwong () is used for this analysis.
 PCGe

124.2/70.3 kgdays of reactor on/off data: New limits are set to neutrino millicharge and low mass weakly interacting massive particle (WIMP) with a fiducial mass of 500 g point contact germanium (PCGe) detector jiunn011301 () at the 0.3 12 keV energy region. The RMS energy resolution of PCGe is 87 keV at GaK shell xray energy soma2016 (). The residual reactor on reactor off event rate spectrum at 0.3 12 keV shown in Fig. 4 of Ref. jiunn011301 () is used for this analysis.
 LSND Experiment

The Liquid Scintillator Detector at the Los Alamos Neutron Science Center was exposed to electron neutrinos produced at the proton beam stop with electron recoil energy of 18 50 MeV. The cross section for the elastic scattering reaction and weak mixing angle were measured. The energy resolution was determined from the shape of the electron energy spectrum and was found to be 6.6% at the 52.8 MeV end point. The observed and expected distribution of beamexcess events at 18 50 MeV published in Fig. 10 of Ref. auerbach2001 () are adopted in this analysis.
 LAMPF Experiment

A 15 ton finegrained tracking calorimeter surrounded by multiwire proportional chambers (MWPCs) was exposed to electron neutrinos from muon decay at rest with of MeV at the Los Alamos Meson Physics Facility, now renamed the Los Alamos Neutron Science Center. In this experiment, neutrinoelectron elastic scattering was observed and electroweak parameters were measured. From the agreement between the measured and SM expectation, limits on neutrino properties (such as neutrino flavor changing neutral currents and neutrino electromagnetic moments) and limits on the masses of new bosons [such as neutral tensor and pseudo(scalar) boson, charged Higgs boson, and a purely lefthanded charged (neutral) vector boson] were derived in Ref. allen93 ().
ii.3 Analysis methods
The expected event rate of can be calculated as
(2) 
where is the electron number density per kg of target mass, and is the neutrino spectrum. represents different interaction channels such as SM, NLS1B, CHB, etc.
The measurable differential cross section is denoted by and corresponds to a convolution of the detector energy resolution to the physical differential cross section . In practice, as far as BSM scenarios and experimental data studied in this work are concerned, the variations of with energy are gradual, so that the resolution smearing does not significantly alter the measured spectra in the region of interest. The difference between and is less than 0.1%. Accordingly, resolution effects can be neglected in this analysis. is expressed in units of and for the CsI(Tl) and Ge data sets, respectively.
The published neutrino spectra for auerbach2001 () are used to derive the SM differential cross sections for the LSND analysis. The number of measured physical and background events are taken from Fig. 10 and Table III of Ref. auerbach2001 (). The published total cross section measured values are used for normalization:
(3) 
The results on physics couplings from this analysis are expressed either as “bestfit statistical systematic uncertainties” at the 1 level, or in terms of limits at a 90% or 95% confidence level (C.L.). The statistical uncertainties are derived by the minimum method, defined as
(4) 
where is the measured rate; and and are the expected event rates for the SM and (with , NLS1B, CHB, etc.), respectively; and is the th bin statistical uncertainty published by the experiments. The published systematic uncertainties of the experiments contribute to shifts of the bestfit values in the physics couplings. The two contributions are added in quadrature to give rise to the combined uncertainties, from which the 90% or 95% C.L. limits can be derived using the prescription of Ref. Gary98 ().
Iii Intermediate Bosons Beyond the Standard Model
Some of the BSM involve exchanging of massive intermediate bosons such as the extra , NLS1B, and CHB in addition to the SM and gauge bosons. A Feynman diagram of neutrino and antineutrino scattering off electron for various NSI scenarios is illustrated in Fig. 1.
Some of the new physics BSMs have a mechanism giving mass to neutrinos such as lowenergy SUSY with Rparity breaking, an extra Higgs boson, unified SUSY models, etc. Indeed, any BSM physics model should reproduce current data and therefore should include massive neutrinos. In addition, there are some recent modeldependent BSM studies in the literature ferzan (); laha (). In this paper, we only study some specific models for new interactions with massive virtual bosons. In the following sections these BSM scenarios and their corresponding experimental constraints will be discussed in detail.
NSI can simply be considered as modifications of coupling constants with additional new terms in the chiral couplings of in general. Therefore, for the flavorconserving (FC) NSI cases, the new couplings can be expressed as
(5) 
Coefficients  

The and scattering differential cross sections can be written in terms of new couplings of FC and flavorviolating (FV) NSI of neutrinos given in Table 2. The differential cross section of BSM contributions can be obtained by using Eq. (1) together with the coefficients from Table 2 considering both FC NSI and FV NSI with or .
Model  Best fit  /dof  bounds  Projected (1%)  Current limit  

for (1 )  at 95% C.L.  bounds  [PDG 2016]  
(GeV)  at 95% C.L. (GeV)  at 95% C.L. (GeV)  
String  8.7/9  (ATLAS)  
Type  8.7/9  (ATLAS)  
8.7/9  (CMS)  
7.8/9  (RVUE)  
8.7/9  (ATLAS) 
iii.1 Extra gauge boson
Intermediate particles of electroweak interaction in addition to SM and gauge bosons, have engaged particle physicists’ attention for a long while since they are a common feature of many models aiming to define the nature of BSM. The gauge boson, the new gauge boson, was proposed as a theoretical particle resulting from the expansion of electroweak interactions in particle physics. Its name comes from the SM boson.
New massive U(1) gauge bosons emerge in grand unified and superstring theories such as SO(10) and London86 (), in theories of extra spacetime dimensions of the SM gauge bosons petriello (). In this study, we will not restrict ourselves to SM gauge bosons. In fact, we will consider a possible new vector boson predicted in many extensions of the SM called the gauge boson, which is a massive, electrically neutral and colorsinglet hypothetical particle of spin 1.
There are various physical models of BSM that suggest different bosons. The most popular of them are the stringtype model, leftright symmetric model, and the sequential Standard Model (SSM). The stringtype model, based on symmetries, contains the and , which means that the two states (i.e., and ) are included and can mix by some angle . The mixing of these two states is given by their linear combination as erler ().
The new coupling parameters of BSM are generally obtained by modifying the ordinary coupling constants of the SM. Therefore, the new cross sections for the interactions via the exchange of an extra gauge boson can be obtained by replacing the SM couplings appearing in Eq. (1) with the new modified couplings accordingly.
The new differential cross section of models for elastic scattering can be obtained by modifying the couplings with
(6) 
where , , and .
In this paper, three main models of the stringtype model barrancoZprime () have been investigated: the model where , the model where , and the model where .
One of the other popular models proposing a heavy neutral vector boson is the leftright symmetric model, which has breaking dynamical symmetry. The leftright symmetric model is based on , where and are associated to the lefthanded and righthanded weak isospins, respectively, and is associated to the charge , where and are the baryon and lepton number, respectively. The couplings are constructed in this model as
(7) 
where the parameters of and can be described as
(8) 
Finally, the SSM, , is defined as having the same couplings with quarks and leptons which are identical to those of the SM , and decays of only known fermions. This model serves as a useful reference case when comparing the researches with wellmotivated models erler (). The differential cross section for this model can be written as
(9)  
The differential cross sections for various extra models with the use of CsI(Tl) as a target at a specific value of are displayed in Fig. 2, where the SM contribution is superimposed for comparison. As it can be seen in the figure, the cross sections of different models demonstrate similar behavior with respect to the recoil energy of the electron. Working at the MeVenergy regime has many more advantages than working at low energy since the cross sections of the SM were measured more precisely with CsI(Tl) data. Therefore, more stringent limits are set to the mass of the extra gauge boson with the CsI(Tl) detector data set compared to those of Ge detector data sets.
By adopting a minimum analysis, the bestfit results and the lower bounds for the mass of the gauge boson obtained from the CsI(Tl) detector data set for each model are given in Table 3. The projected sensitivities and the present bounds from the LHC experiment are also given for comparison. It can be seen that the bounds from lowenergy neutrinoelectron scattering experiments are much less stringent than those of highenergy collider experiments, due to worse statistics and in general a larger background.
The realistic sensitivities of future reactor scattering experiments are discussed in Table VII of Ref. mdeniz072001 (). The main improvement is due to background suppression. The effects of a projected accuracy of 1% to the various bounds are also shown in Table 3. The aim of this extrapolation is to see how the mass bounds are related to the experimental accuracies. It may provide intuitive scaling for the future neutrino experiments. Moreover, the mixingparameterindependent sensitivities of CsI(Tl) and HPGe detector data at 95% C.L. for stringtype models are shown in Fig. 3. It can be seen from the figure that the model, where , can provide a more stringent limit.
()  ()  

TEXONO PCGe  LSND  TEXONO CsI(Tl)  LSND 
iii.2 New light spin1 boson
The exchange of new massive particles can be a possible origin of NSI of neutrinos, manifested as anomalies in the measurable total or differential cross sections. These massive particles, however, can be as light as in the order of a few MeV scale, which is the range of lowenergy experiments. The NLS1B is one of the examples of such kinds of particles. A spin1 particle could also be involved in explaining the NuTeV anomaly Celine2004 (). In addition to this, the NLS1B may also explain the muon anomalous magnetic moment value Gninenko2001 (). Moreover, spin1 bosons can couple to dark matter and the nonbaryonic matter of the Universe in the MeV scale region. They could be responsible for the annihilation that is seen as the unexplained 511 keV gamma emissions anomaly from the galactic bulge Hooper2007 (). Furthermore, the NLS1B particle, which is lighter than quarks, would explain the anomalous violation in the mixing of neutral Bmesons. Sechul2011 (). The effective Lagrangian for the NLS1B can be written as ChengWC2013 ()
(10)  
where are the chiral projectors and the labels correspond to lepton flavor or .
The scattering differential cross section for NLS1B exchange contributions can be obtained by modifying the chiral couplings as given in Table 2. The can be defined in terms of the coupling parameters , and the mass of as
(11) 
where or , and can be defined as
(12) 
We can alternatively define new couplings and similar to the SM chiral couplings of in the case of one of the couplings not being zero as
(13)  
where can be defined as
(14) 
()  ()  

TEXONO PCGe  LSND  TEXONO CsI(Tl)  LSND 
The differential cross sections as a function of the recoil energy with typical reactor spectra for the NLS1B at some specific coupling and mass parameters using the Ge detector as a target for both the FC and FV cases are displayed in Fig. 4 for illustration, where the SM contribution is superimposed. As can be seen in this figure, the cross section shows different behavior with respect to the recoil energy , which provides more advantages in the measurements of the couplings at low energy for the low mass values of . Because of the dependency in the cross section, working at lowenergy threshold provides better sensitivity in the coupling for small mass values of . The term in the denominator can be safely neglected for high values of . Therefore, the CsI(Tl) detector data are expected to provide more stringent limits since the cross section is measured at a good sensitivity in the 38 MeV range.
In particular, the NLS1B cross section has dependency, which is directly proportional to the sensitivity of the couplings of . When and become comparable, the term in the denominator cannot be neglected anymore. This term, however, causes us to lose the sensitivity in for low mass values of . When gets bigger, i.e., MeV, can be neglected and the sensitivity stays fixed at the minimum value as shown in Fig. 5. In this case, the and conditions are satisfied.
For the FC NLS1B interaction, the allowed regions at 90% C.L. in the parameter space of and with various 1, 2, 5, 25 MeV for TEXONO and 3, 4, 5, 25 MeV for LSND are illustrated in Figs. 5(a) and 5(c), respectively.
Similarly, for the FV NLS1B interaction, the allowed regions at 90% C.L. for the couplings of and with various 1, 2, 5, 25 MeV for TEXONO and 3, 4, 5, 25 MeV for LSND are illustrated in Figs. 5(b) and 5(d), respectively.
The global fitting for allowed regions of TEXONO and LSND for the couplings of and at 90% C.L. with various 1, 2, 5, 25 MeV are illustrated in Figs. 5(e) and 5(f) for FC and FV NSI, respectively.
By adopting a oneparameteratatime analysis in the minimum analysis, the bounds at 90% C.L. on the FC and FV NLS1B couplings for low and high mass values are given in Tables 4 and 5, and the upper limits at 90 % C.L. are illustrated in Fig. 6 with respect to mass parameter . As shown in Table 4, Table 5, and Fig. 6, the TEXONO PCGe and HPGe data provide better constraints in and parameter spaces compared to LSND for both FC and FV NLS1B in the case of . On the other hand, TEXONO CsI(Tl) gives better constraints in the and parameter spaces compared to LSND for both FC and FV NLS1B in the case of .
The 90% C.L. upper limits for the couplings of , and , vs mass parameter of for TEXONO and LSND are illustrated in Figs. 6(a) and 6(b) and Figs. 6(e) and 6(f), respectively. The 90% C.L. upper limits for the couplings of and , versus mass parameter of for TEXONO CsI(Tl) and LSND are illustrated in Figs. 6(c) and 6(d) and Figs. 6(g) and 6(h), respectively.
Similarly, the allowed regions at 90% C.L. in the parameter space of and with various for TEXONO CsI(Tl) and LSND are illustrated in Figs. 7(a) and 7(c), respectively, for the FC NLS1B. In the case of the FV NLS1B, the allowed regions at 90% C.L. in the parameter space of and with various for TEXONO CsI(Tl) and LSND are illustrated in Figs. 7(b) and 7(d), respectively.
The global fitting for allowed regions of TEXONO and LSND for the couplings of and at 90% C.L. with various 1, 2, 5, 25 MeV are illustrated in Figs. 7(e) and 7(f) for FC and FV NSI, respectively.
TEXONO  LSND  

1 MeV  4.99  8.03 
2 MeV  0.09  2.91 
2.2 MeV  0.02  1.22 
2.5 MeV  0.06  0.32 
2.9 MeV  0.02  2.88 
3 MeV  4.58  1.47 
4 MeV  8.45  0.46 
5 MeV  11.46  2.26 
6 MeV  14.27  17.67 
10 MeV  25.02  57.57 
iii.3 Charged Higgs boson
Leptons, quarks and gauge bosons acquire their mass through the Higgs mechanism Englert64 (), while neutrinos still remain massless in the SM. In order to introduce and explain the smallness of neutrino masses without requiring an extra righthanded neutrino, one of the simplest models among other mechanisms is the Higgs triplet model (HTM), through which neutrinos gain their mass Cheng80 (); Ong13 (). In HTM, apart from the neutral scalar Higgs boson (), there also appear singly charged () and doubly charged () ones, since Higgs triplets under the standard gauge group have two units of weak hypercharge.
There are many phenomenological studies at highenergy accelerator experiments such as LHC and Tevatron in the literature Fileviez08 (); Akeroyd11 (). However, in this study we also consider the lowenergy frontier with and elastic scattering, which are pure leptonic processes providing an elegant test to the SM of electroweak theory. The Feynman diagrams of and scattering via the exchange of CHB are displayed in Fig. 8.
In the HTM, the electroweak parameter is predicted at the tree level as , where and are the vacuum expectation values of the doublet Higgs field and triplet Higgs field, respectively. However, the experimental value of this parameter shinya12 () requires that be smaller than a few GeV, i.e., 3.5 GeV at %95 C.L., and hence . Taking these into account, the interaction Lagrangian for the coupling of the CHB to leptons can be written as
(15) 
where is the coupling constant; e, , or ; is the charge conjugation; and is the chiral projector.
The and scattering differential cross sections for CHB exchange contributions are found, respectively, to be
(16) 
and
(17) 
The differential cross section for CHB with relevant parameters for TEXONO CsI(Tl) is displayed in Fig. 9 with different mass parameters, where SM contribution is superimposed for comparison.
For the high mass value of CHB, the terms of or in the denominator can be neglected. Therefore, becomes a fitting parameter. From the best fit,
(18) 
is obtained at for TEXONO CsI(Tl) data. Similarly,
(19) 
is obtained at for LSND. They can be converted to their corresponding upper limit at 90% C.L. of
(20) 
for TEXONO CsI(Tl) and
(21) 
for LSND. TEXONO provides more stringent limits than those from LSND and a previous study given in Ref. coarasa96 () as for the LAMPF experiment allen93 () at 90% C.L., which was derived based on the measurement value of .
On the other hand, for the low mass value of CHB, the term in the denominator can be neglected. Therefore, only becomes a fitting parameter. From the best fit,
(22) 
is obtained at with its corresponding upper limit at 90% C.L. of
(23) 
for TEXONO.
Similarly, for LSND, from the best fit,
(24) 
is obtained at with its corresponding upper limit at 90% C.L. of
(25) 
for the low mass value of CHB.
The upper limit of coupling with respect to the CHB mass values of for TEXONO CsI(Tl) and LSND at 90% C.L. for low and high mass values are shown in Figs. 10(a) and 10(b), respectively. Upper bounds at 90% C.L. on the coupling of for and scattering for various mass values of are listed in Table 6. In this study, the parameter space is extended and consequently a new window is opened for the low mass CHB.
Iv SUMMARY and PROSPECTS
In summary, in this article, some of the BSM new physics scenarios including massive intermediate particles such as the NLS1B, , and CHB have been discussed and their potential to explain some of the anomalous effects that cannot be explained by SM has been addressed.
The experimental results of upper bounds for NSI using data from the analysis of the and elastic scattering interaction cross section measurements were placed in the framework of these BSM scenarios. The existing experimental sensitivities were improved, and the parameter space was extended by including the lowenergy regime.
Particularly, in the NLS1B study, a new research window has been opened for a low mass NLS1B in the lowenergy regime due to dependency in the cross section. For a low mass NLS1B, the coupling becomes directly proportional to ; therefore, working at low energy and low threshold becomes substantially important to see the effect of BSM. In this study we found that TEXONO gives better constraints in parameter space compared to the neutrinoelectron channel, i.e., the LSND and LAMPF experiments, for both the FC and FV NLS1B cases.
In the literature, many studies on high energies have targeted high mass values of CHB. However, in this study we also considered the lowenergy frontier with and elastic scattering, which are pure leptonic processes providing an elegant test of the electroweak theory of SM. We have found new limits on the CHB couplings with respect to mass covering the low mass CHB region.
In our study of , the current limits were not improved since the experimental uncertainties are big compared to the heavy expectation value of the mass. However, it is still interesting enough to look for the mass limits of at the lowenergy, lowmomentum regime. This study showed that if the experimental uncertainties were improved by 1%, the current existing limits could be reached via the neutrinoelectron scattering channel. By the help of the projection, it is possible to investigate the relationship between the mass bounds and experimental accuracies that may provide intuitive scaling for future neutrino experiments.
V Acknowledgments
This work is supported by Contract No. 114F374 under the Scientific and Technological Research Council of Turkey (TÜBİTAK); Contract No. 1042112M001038MY3 from the Ministry of Science and Technology, Taiwan; and Contract No. 2017ECP2 from the National Center of Theoretical Sciences, Taiwan.
References
 (1) S.F. King, J. Phys. G 42, 123001 (2015).
 (2) J. Panman, in Precision Tests of the Standard Electroweak Model, edited by P. Langacker (World Scientific, Singapore, 1995), pp. 504544; W.J. Marciano and Z. Parsa,, J. Phys. G 29, 2629 (2003).
 (3) T. Ohlsson, Reports on Progress in Physics 76(4), 044201 (2013).
 (4) O.G. Miranda and H. Nunokawa, New Journal of Physics 17(9), 095002 (2015).
 (5) J. Erler and P. Langacker, Phys. Lett. B 667, 125 (2008), and references therein.
 (6) M. Deniz et al., Phys. Rev. D 82, 033004 (2010); M. Deniz et al., Phys. Rev. D 95,033008 (2017).
 (7) S. Bilmis et al., Phys. Rev. D 85, 073011 (2012).
 (8) JiunnWei Chen, HsinChang Chi, HauBin Li, C. P. Liu, Lakhwinder Singh, Henry T. Wong, ChihLiang Wu, and ChihPan Wu, Phys. Rev. D 90, 011301(R) (2014).
 (9) S. Bilmis, I. Turan, T.M. Aliev, M. Deniz, L. Singh, and H.T. Wong, Phys. Rev. D 92, 033009 (2015).
 (10) B. Kayser, E. Fischbach, S.P. Rosen, and H. Spivack, Phys. Rev. D 20, 87 (1979).
 (11) M. Deniz et al., Phys. Rev. D 81, 072001 (2010).
 (12) H.B. Li et al., Phys. Rev. Lett. 90, 131802 (2003); H.T. Wong et al., Phys. Rev. D 75, 012001 (2007).
 (13) A.K. Soma et al., Nucl. Instr. and Meth. A 836, 67 (2016).
 (14) L.B. Auerbach et al, Phys. Rev. D 63, 112001 (2001).
 (15) R.C. Allen et al, Phys. Rev. D