Challenge to Anomalous Phenomena in Solar Neutrino

We suggest a would-be solution to the solar neutrino tension why solar neutrinos appear to mix differently from reactor antineutrinos, in theoretical respect. To do that, based on an extended theory with light sterile neutrinos added we derive a general transition probability of neutrinos born with one flavor tuning into a different flavor. Three new mass-squared differences are augmented in the extended theory: two $\Delta m^2_{\rm ABL}\lesssim{\cal O}(10^{-11})\,{\rm eV}^2$ optimized at astronomical-scale baseline (ABL) oscillation experiments and one $\Delta m^2_{\rm SBL}\sim{\cal O}(1)\,{\rm eV}^2$ optimized at reactor short-baseline (SBL) oscillation experiments. With a so-called composite matter effect that causes a neutrino flavor change via the effects of sinusoidal oscillation including the Mikheyev-Smirnov-Wolfenstein matter effect, we find that the value of $\Delta m^2$ measured from reactor antineutrino experiments can be fitted with that from the $^8$B solar neutrino experiments for roughly $\Delta m^2_1\lesssim10^{-13}\,{\rm eV}^2$ and $\Delta m^2_2\simeq10^{-11}\,{\rm eV}^2$. Future precise measurements of $^8$B and $pep$ solar neutrinos may confirm and improve the value of $\Delta{m}^2_2$.


I. INTRODUCTION
One of the great discoveries in particle physics is the experimental evidence of neutrino oscillations, implying that neutrinos are massive particles and that the three flavor neutrinos ν e , ν µ , ν τ are mixtures of neutrinos with definite masses ν i (with i = 1, 2, ...) which are identified based on the charged particles (electron, muon, tau) produced via weak interaction [1].
Despite the successful description of the properties of three known neutrino species through decades of experimentation, we are faced with a growing list of anomalous phenomenaexperimentally unexpected results that conflict with the three active neutrino oscillation standard framework (3νSF [1]): (i) the so-called "short-baseline (SBL) anomalies" [2][3][4][5] (including MiniBooNE data [6]), anomalous results measured in several experiments at distance less than 1 km, and (ii) the so-called "solar neutrino tension" [7][8][9] (see also Refs. [10,11]), a discrepancy between the oscillation parameter determined in solar neutrino experiments and the one measured in reactor neutrino oscillation experiments. Although the solar neutrino tension (below 2σ uncertainties) seems small, the discrepancy has been the long-standing tension in the neutrino physics of 3νSF [7][8][9]. In fact, it is well known that none of the 8 B measurements performed by Sudbury Neutrino Observatory (SNO) [10], Super-Kamiokande (SK) [8], and Borexino [11] has shown any evidence of the low energy spectrum turn-up expected in the standard Mikheyev-Smirnov-Wolfenstein (MSW)-the large mixing angle, LMA, solution for the value of mass-squared difference ∆m 2 favored by KamLAND [12][13][14].
Moreover, the recent observation of low-energy 8 B solar neutrino flux (as low as ∼ 3.5 MeV) by SK [8] has marked the raise of the solar neutrino tension. The anomalous phenomena (the SBL anomalies plus solar neutrino tension) could be due to some unknown physical phenomena that are sensitive to SBL oscillations and solar neutrino oscillations, or else unlikely statistical fluctuations in the current data, or experimental errors. If such anomalous phenomena point to potential problems with the 3νSF predictions, they may provide a new level of importance as possible routes to "new neutrino physics", an extended theory for phenomena unexplained by the 3νSF.
In theoretical respects, the most straightforward interpretation of those anomalous phenomena could be neutrino oscillations with new parameters. Problem is, in the 3νSF [1], there are only six oscillation parameters: two mass-squared differences ∆m 2 Sol and ∆m 2 Atm , three mixing angles θ 23 , θ 13 , θ 12 , and one Dirac CP phase δ CP . These parameters are not sufficient to fit the anomalous phenomena, indicating that the current theoretical understanding of neutrino oscillation may not be complete. If the anomalous phenomena are interpreted in terms of neutrino oscillations, it may suggest the presence of other types of light neutrinos called sterile neutrinos that do not have weak interactions 1 . An attempt to explain both the SBL anomalies and the ultra-high energy neutrino events at IceCube [15] has been proposed in a framework of neutrino oscillation [17].
The goal of this work is to study for a theoretical understanding of the solar neutrino tension why solar neutrinos at SNO, SK, and Borexino experiments appear to mix differently from reactor antineutrinos at KamLAND, despite that both neutrinos are sensitive to the same oscillation parameters in the 3νSF. To do that, first, based on an extended theory with light sterile neutrinos added [17] we derive a general transition probability between the massive neutrinos that a flavor eigenstate ν α becomes flavor eigenstate ν β with α, β = e, µ, τ , that can also have a potential for explaining both the anomalous phenomena and ultra-high energy neutrino events at IceCube, simultaneously 2 . Second, we re-examine the MSW matter effects in our theoretical framework and suggest a solution to the solar neutrino tension with a so-called composite matter effect that causes a neutrino flavor change with new oscillatory terms containing ∆m 2 ABL O(10 −11 ) eV 2 optimized at astronomical-scale baseline (ABL) ( L es = 149.6×10 6 km, earth-sun distance) oscillation experiments together with the MSW matter effect. An important point is that, contrary to the MSW effect [19,20] that causes a change in the flavor content of a neutrino but without sinusoidal oscillation, the so-called composite matter effect causes a neutrino flavor change via the effects of sinusoidal oscillation, as well as the MSW matter effect.
This work is organized as follows. In section II we provide an introduction to the model setup, masses and mixings. In section III we compute a general transition probability for three flavor neutrinos with their three light sterile neutrino pairs, subsequently, in section III-A we investigate possible mass orderings and show how additional oscillation parameters 1 The authors in Ref. [16] have extended the Pontecorvo-Maki-Nakagawa-Sakata (PMNS) mixing matrix to interpret the SBL anomalies as neutrino oscillations. While, instead of the way of extending the PMNS matrix, in order to introduce new oscillation parameters the authors in Ref . [17] have parameterized with unitary condition in a way that a diagonal form of 2 × 2 partitioned matrix holding the 3 × 3 PMNS mixing matrix is linearly multiplied by a 6 × 6 mixing matrix of active to sterile neutrinos. 2 However, in this paper, we will not study phenomenological interpretations of the SBL anomalies and astronomical neutrino data at IceCube. See Refs. [17,18]. could be constrained by cosmological data (the sum of active neutrino masses) and the effective neutrino mass in both β-decay and neutrinoless-double-beta (0νββ)-decay experiments, and then we interpret reactor anti-neutrinos at KamLAND in the new oscillation framework in section III-B. In section IV-A, we re-examine the MSW matter effects in the extended 3νSF, and we analytically study why the oscillation parameters determined in solar neutrino experiments are not in complete agreement with the measurements collected in other types of experiments in section IV-B. Conclusions are drawn in section V.

II. MASSES AND MIXINGS
In the basis of interaction eigenstates, ψ L ≡ (ν L S c R ) T , where active neutrinos are in the up-stairs and sterile neutrinos are in the down-stairs, the most general renormalizable Lagrangian for neutrinos reads in the charged lepton basis at low energies [17] where g is the SU(2) coupling constant, θ W is the Weinberg angle, ℓ = (e, µ, τ ), ν L = (ν e , ν µ , ν τ ), and S R = (S 1 , S 2 , ...S n ). The light neutral fermions S α do not take part in the standard weak interaction and thus are not excluded by LEP results, while the number of active neutrinos that are coupled with the W ± and Z bosons is N ν = 2.984 ± 0.008 [21].
After electroweak symmetry breaking, Eq. (1) describes 3 × n Majorana neutrinos. In the case of n = 3 sterile neutrinos, the 6 × 6 Majorana neutrino mass matrix is 3  [23] where m D and M R are Dirac and Majorana neutrino masses, see e.g. Ref. [18,22]. n L , that is, The phenomenology of Eq. (2) depends on the values of the matrices M L , M D , and M S , that is, the mass eigenvalues and mixings of M D , M S and M L . We impose unitary condition to W ν (such that satisfys the unitary condition W ν W † ν = W † ν W ν = I 6×6 ), which preserves norm and thus probability amplitude, and choose the 6 × 6 unitary neutrino transformation matrix as [17,18] where U L corresponds to the PMNS mixing matrix U PMNS , U R is an unknown unitary 3 × 3 matrix, V 1 = diag(1, 1, 1)/ √ 2, and V 2 = diag(e iχ 1 , e iχ 2 , e iχ 3 )/ √ 2 with χ i being arbitrary phases. The 6 × 6 unitary mixing matrix V ν forms a bridge between active and sterile neutrinos: In the mass eigenstates ν 1 , ν 2 , ν 3 , S c 1 , S c 2 , S c 3 basis the Hermitian matrix M ν M † ν can be diagonalized as a real and positive 6 × 6 mass-squared matrix by the unitary transformation where a real positive diagonal matrixM = U T R M D U L = diag(m 1 , m 2 , m 3 ) and two complex diagonal matricesM L = U T L M L U L andM S ≡ U T R M S U R are used. In Eq. (6), the parameter δ is defined 4 as where the mass-squared eigenvalues (real and positive) are given by and the mixing angles θ k between active and sterile neutrino is given by From Eq. (8) we define the mass splitting for k-th generation as Here, for simplicity, we take φ k = −π/4 and assume |(M S ) k |, m k ≫ |(M L ) k |, or equivalently, in Eq. (2). Then the mass-squared eigenvalues of k-th generation of Eq. (8) can be expressed in terms of θ k and ∆m 2 k as which in turn, together with Eq. (10), leads to The three active neutrino states emitted by weak interactions are described in terms of the mass eigenstates ν k , S k (k = 1, 2, 3) and the 3 × 3 PMNS mixing matrix U ≡ U PMNS as with the massive states 4 The real positive condition ofM = U T R M D U L derives the phase of V 2 to be equivalent to that of δ.
in which the field redefinitions ν k → e −iπ/4 ν k and S k → e i3π/4 S k have been used. Since the active neutrinos are massive and mixed, the weak eigenstates ν α (with flavor α = e, µ, τ ) produced in a weak gauge interaction are linear combinations of the mass eigenstates with definite masses. The charged gauge interaction in Eq. (1) for the neutrino flavor production and detection is written in the charged lepton basis as Thus in the mass eigenstate basis the PMNS leptonic mixing matrix [1] at low energies is visualized in the charged weak interaction, which is expressed in terms of three mixing angles, θ 12 , θ 13 , θ 23 , and three CP-odd phases (one δ CP for the Dirac neutrino and twoφ 1,2 for the Majorana neutrino) as where s ij ≡ sin θ ij , c ij ≡ cos θ ij and P ν is a diagonal phase matrix what is that particles are Majorana ones. Three-flavor oscillation parameters from global fit results at the best-fit values and (1σ) 3σ confidence intervals in Ref. [7] for normal mass ordering, NO, [inverted one, IO] in Table-I. Super-Kamiokande atmospheric data [7]. NO = normal neutrino mass ordering; IO = inverted mass ordering with ∆m 2 Sol ≡ m 2 for IO. On the other hand, as shown in Ref. [8], in the 3νSF there is a clear discrepancy for masssquared difference at 2σ C.L.: the value ∆m 2 KL preferred by KamLAND [14] is somewhat higher than the one ∆m 2 ⊙ from solar experiments [8,11,[24][25][26], while their mixing angles are the same, which is as follows at 2σ C.L. for θ 13 The oscillation parameter ∆m 2 determined in solar neutrino experiments is not in complete agreement with the measurements collected in other types of reactor and accelerator experiments: the so-called "solar neutrino tension". Let us first bring out a transition probability of new oscillations with the help of the neutrino mixing matrix Eq. (4). The transition probability between the massive neutrinos that a neutrino eigenstate ν a becomes an eigenstate ν b follows from the time evolution of mass eigenstates as where a, b = e, µ, τ, s 1 , s 2 , s 3 , L is the distance between the neutrino detector and the neutrino source, E is the neutrino energy, andM ν ≡ W T ν M ν W ν . We are interested in the flavor transition between the active neutrinos ν e , ν µ , ν τ . From Eq. (18) the flavor transition probability between the active neutrinos ν e,µ,τ can be generically expressed in terms of the oscillation parameters θ, ∆m 2 , L, E, and the mixing components of the 3 × 3 PMNS matrix U αi as where In the model the mixing parameters θ and ∆m 2 are determined by nature, so experiments should choose L and E to be sensitive to oscillations through a given ∆m 2 . As expected, in the limit of m s i → m ν i and θ i → 0 (i = 1, 2, 3), the new transition probability of Eq. (19) is recovered to the standard form of that for three neutrinos in vacuum [1].

A. Possible mass orderings
In the new oscillation framework of Eq. (19) there appear three additional oscillation parameter sets (∆m 2 i , θ i ) in addition to the six standard oscillation parameters. To resolve a tension between the mass-squared differences of solar and reactor neutrinos in Eq. (17), as well as to accommodate an eV sterile neutrino for a possible solution to the SBL anomalies [17] and ultra-high energy neutrino events at the IceCube detector [17,18,27], we assume together with | cos 2θ 1(2) | ≈ 1 and a sizable θ 3 from Eq. (10), where L ABL and L SBL stand for astronomical-scale baselines ( L es ) and short-baselines ( 1km), respectively. Hence Eq. (22) means the effects of the pseudo-Dirac neutrinos for the first and second generation characterized by ∆m 2 1(2) can be detected through ABL oscillation experiments without damaging the three-active neutrino oscillation experimental data, while that for the third generation characterized by ∆m 2 3 can be measured through SBL oscillation experiments [17]. Moreover, Eq. (22) indicates that the mass splittings ∆m 2 1(2) should be constrained by results in reactor and accelerator based neutrino experiments, such as the results of reactor experiments (optimized at oscillation lengths L ∼ km with ∆m 2 31 ∼ 2.5 × 10 −3 eV 2 and Eν e ∼ MeV) and solar neutrino oscillation experiments (optimized at L ∼ O(10 ∼ 100) km with ∆m 2 21 ∼ 7.5 × 10 −5 eV 2 and Eν e ∼ MeV), since they can modify the reactor angle θ 13 , atmospheric mixing angle θ 23 , and the LMA solution θ 12 . Whereas 5 the mass splitting ∆m 2 3 ∼ O(1) eV 2 can be constrained by SBL ( 1km) oscillation experiments (and possibly long baseline oscillation experiments) as shown in Ref. [17].
Considering the hierarchy of mass splittings Eq. (22), there are two possible neutrino mass spectra 6 by taking an order of eV mass splitting ∆m 2 3 = m 2 ν 3 − m 2 s 3 < 0 into account as in Ref. [17]: (i) the normal mass ordering m 2 , because of the observed hierarchy ∆m 2 Atm ≫ ∆m 2 21 and the requirement of a MSW resonance for solar neutrinos ∆m 2 21 > 0, see Eq. (46). Since oscillation experiments are insensitive to the absolute scale of neutrino masses, we study how the new mixing parameters (θ k , ∆m 2 k ) can be constrained through the sum of three active neutrinos m ν . Cosmology is mostly sensitive to the total energy density in neutrinos, directly proportional to the sum of the neutrino masses which can be expressed in terms of θ k and ∆m 2 k , see Eq. (23). Together with the known mass-squared differences (|∆m 2 31 | ∼ |∆m 2 32 |, ∆m 2 21 ) in Table-I of the standard oscillation form, the sum of the neutrino masses can be parameterized in terms of new 5 Recent cosmological data have a tendency toward disfavoring an excess of radiation beyond the three neutrinos and photons: considering the Big Bang Nucleosynthesis (BBN) and CMB limits to ∆N eff ν < 0.2 at 95% CL [28]. The SBL anomalies including MiniBooNE data may indicate the existence of eV-mass sterile neutrino, while present cosmological data coming from CMB+LSS, and BBN do not prefer extra fully thermalized sterile neutrino in the eV-mass range since they violate the hot dark matter limit on the neutrino mass. It can be realized by requiring sterile neutrinos do not or partially equilibrium at the BBN epoch when the initial lepton asymmetry is large [29,30]. Especially in Ref. [30] showed quantitatively the amount of thermalization as a function of neutrino parameters (mass splitting, mixing, and initial lepton asymmetry). See also Refs. [31]. 6 Here the possibilities of mass ordering via the sign of ∆m 2 1(2) in Eq. (12) are not considered.
where m ν k = ∆m 2 k / sin 2θ k (1 + sin 2θ k )/2. Cosmological and astrophysical measurements provide powerful constraints on the sum of neutrino masses complementary to those from accelerators and reactors. Bounds on the sum of the three active neutrino masses can be summarized as a lower limit for the sum of the neutrino masses 3 i=1 m ν i 0.06 eV (for NO) and 0.103 eV (for IO) could be provided by the neutrino oscillation measurements; upper limits 7 at 95% CL are given in Ref. [32].
The existence of massive neutrino at the eV scale can also be constrained by β-decay experiments [35] and by 0νββ-decay experiments [36]. The two types of mass ordering, discussed above, should be compatible with the existing constraints on the absolute scale of neutrino masses m j . The most sensitive experiments on the search of the effects of neutrino masses in β-decay use the tritium decay process 3 H → 3 He + e − +ν e . Non-zero neutrino masses distort the measurable spectraum of the emitted electron. The most stringent upper bounds on theν e mass, mν e , have been obtained from direct searches in the Mainz [37] and Troitsk [38] experiments at 95% CL: 2.05 eV , Troitsk [38] .
The upcoming KATRIN experiment [39] planned to reach sensitivity of mν e ∼ 0.20 eV will probe the region of the QD spectrum in the model. Note that the bounds in Eq. (26) coming 7 Massive neutrinos could leave distinct signatures on the CMB and large-scale structure (LSS) at different epochs of the Universe's evolution [34]. To a large extent, these signatures could be extracted from the available cosmological observations, from which the total neutrino mass could be constrained.  [39]. For i=1,2,3 m ν , we take the values from Eq. (26). And the best-fit values in Table-I  there appears only third generation in ββ0ν-decay rate. Using Eq. (25) one can easily see vanishing ββ0ν-decay rate, as shown in Ref. [17]. Hence if the ββ0ν-decay rate is measured in the near future the model would explicitly be excluded.
In order to display new physical effects, we investigate the influence of ∆m 2 k and sin 2θ k on the sum of active neutrino masses and the effective mass in β-decay. Plugging the experimental constraints of Table-I  and 0.103 eV (for IO) and upper limits 0.340 ∼ 0.715 eV at 95% CL [32]. In the plots we consider 10 −16 < ∆m 2 1(2) < 10 −5 eV 2 for Fig. 1 since they should be less than the measured ∆m 2 Sol , while for Fig. 2 only eV-mass scale of sterile neutrino since too heavy neutrino is conflict with cosmology ∆N eff ν < 0.2 at 95% CL [28,30].

B. Interpretation of reactor neutrino at KamLAND
As shown in Eq. (17), the KamLANDν e survival probability has reported a precise determination of ∆m 2 KL = ∆m 2 21 and θ KL = θ 12 at 99.998% C.L. [12,13]. It has confirmed the LMA solution which can theoretically be explained via the MSW solar matter effects in neutrino oscillations [19,20].
Nuclear fission reactors are a powerful source ofν e with energies around a few MeV. Thus, the expected oscillation length is O(10 ∼ 100) km, which is a reasonable distance relative to a reactor to place a detector and observeν e disappearance. Assuming CPT (charge-

IV. SOLAR NEUTRINO TENSION AND ITS POSSIBLE SOLUTION
The low-energy 8 B solar neutrinos (as low as ∼ 3.5 MeV) observed by SK [8], within the 3νSF, translate into the mass-squared difference ∆m 2 ⊙ somewhat lower than ∆m 2 KL measured by KamLAND but the same mixing angle θ ⊙ as θ KL measured by KamLAND, as shown in Eq. (17). Moreover, it is well known that none of the 8 B measurements performed by SNO [10], SK [8], and Borexino [11] shows any evidence of the low energy spectrum turn-up expected in the standard MSW-LMA solution for the value of ∆m 2 KL favored by KamLAND. Hence the clear discrepancy at less than 2σ [7][8][9] indicates that the current theoretical understanding of neutrino oscillation may not be complete. Some unknown physical phenomena that only affect solar neutrinos must be at play to explain such discrepancy.
In this regard, we investigate why solar neutrinos at SNO [25], SK [8] and Borexino [11] appear to mix differently from reactor antineutrinos at KamLAND [12,13] for theoretical understanding of the solar neutrino tension, by re-examining the MSW matter effects in our new framework of Eq. (1).

A. Solar matter effects in neutrino oscillations
We discuss, first, how neutrinos produced in the Sun propagate towards the surface of the Sun and then to a detector on Earth, and study the relevant transition probabilities.
To do this, we construct an effective Hamiltonian for matter effects in the Sun by assuming that neutrinos propagating in matter interact coherently with the particles in the medium.
where k = 1, 2, 3, α = e, µ, τ , and I 3 (0 3 ) stands for the 3 × 3 unit (null) matrix. Here the parameter A α is a measure of the importance of matter effect with the matter-induced effective potential; V α and V s = 0 are the potentials experienced by the active neutrinos and the sterile neutrinos, respectively, and E is the neutrino energy. For anti-neutrino the Hamiltonian can be obtained by the substitution V α → −V α and W ν → W * ν . ν e 's have charged-current (CC) interactions with electrons and neutral-current (NC) interactions with nucleons V e = √ 2G F (N e − N n /2), while ν µ 's and ν τ 's have only NC interactions V µ = V τ = √ 2G F (−N n /2) whose equivalence is available at tree level in the weak interactions [42], and any S α 's have no interactions, V s = 0, where G F is the Fermi constant, and N e (N n ) is the average electron (neutron) number per unit volume along the neutrino path.
Consider, for example, electron neutrinos (ν e s) generated in the Sun by nuclear reactions.
Assuming that there are no sterile neutrinos initially when the ν e s are produced by the weak interactions at the core of the Sun. The flavor state ν e propagates as a mass eigenstate ν mi in the medium and at any later time the eigenstate ν mi has an "active" and "sterile" where U L is a 3 × 3 transformation matrix of the form responsible for the mixing in matter for active neutrinos, which is a product of three matrices I k U ij ≡Ĩ k (ϕ m k ) U(θ m ij ) rotation in corresponding planes by the set angles with the phase matricesĨ 1 = diag(1, e iϕ m 1 , e −iϕ m 1 ),Ĩ 2 = diag(e iϕ m 2 , 1, e −iϕ m 2 ), andĨ 3 = diag(e iϕ m 3 , e −iϕ m 3 , 1), and U R is given by the mixing matrix in vacuum since between the sterile neutrino themselves have no weak interactions in matter.
The mixing parameters in U L and V ν of Eq. (30) appear in the final amplitudes of ν α → ν β with α, β = e, µ, τ , see Eq. (55), when projecting the flavor states onto propagation basis states at the neutrino production and detection. Since the mixing parameters only in U L (correspondingly instantaneous mass eigenstates) become functions of V α for active neutrino propagation in a medium with varying density, it is expected that an active neutrino propagates with both mixing angle between active neutrinos θ m ij and mixing angle between active and sterile neutrinos θ k . Hence an intermediate instantaneous eigenstates ξ m ≡ (ν mi , S mi ) T can be defined with a transformation relation in matter where a unitary mixing matrixW m transforms between the interaction eigenstates ψ f and the (instantaneous) mass eigenstates ξ m . In turn, another unitary mixing matrix X m ≡W † m W m in matter can be defined as the matrix that connects the eigenstates ξ m with the eigenstates n m : in which the mixing matrix X m in matter between the active and sterile neutrinos becomes the one in vacuum. The sterile neutrinos get generated as vacuum-like states via coherent oscillations during propagation.

Mixing in matter
With the help of the mixing of Eq. (32), we can seclude the active neutrinos from the sterile neutrinos. In the interaction basis ψ f = (ν α , S α ) T , then, the effective Hamiltonian H m of Eq. (29) in matter is modified to where irrelevant terms of the Hamiltonian proportional to the unit matrix are omitted.
Considering the matter effects in the Sun, which are of special relevance for solar neutrinos.
As ν e s produced at the core of the Sun move outward, A eµ will decrease as the density decreases, and the neutrinos will eventually go through the resonance regions, and proceed number density at the point of ν e production in the Sun [45]. Then the difference of the potentials for ν e and ν µ,τ , i.e. A eµ ≡ A e − A µ = 2E(V e − V µ ), due to the charged current scattering of ν e on electrons (ν e e → ν e e) [19], reads A eµ = 2 √ 2 E G F N e (R): After crossing the Sun, the make-up of the neutrino state existing the Sun will depend on the relative size of A eµ versus A r ij (at neutrino-state resonance point) whose parameters are determined by nature. Similar to Eq. Using with good accuracy, since the condition of A r 13 ≫ A eµ (R, E) is satisfied in all ranges of solar neutrino energies. Thus the corrections to the θ 23 and θ m 13 can be safely neglected in solar matter. Then each equation in Eq. (42) can be approximated with good accuracy as A r 13 ≃ ∆m 2 31 cos 2θ 13 and A r 12 ≃ ∆m 2 21 cos 2θ 12 / cos 2 θ 13 . Therefore, in solar matter a matter mixing angle θ m 12 is only effective, and the mixing matrix Eq. (31) becomes and the mass-squared eigenvalues in Eq. (41) become M 2 ν 1 ≃ m 2 ν 1 + A eµ cos 2 θ 13 cos 2 θ m 12 + ∆m 2 21 sin 2 (θ 12 − θ m 12 ) , M 2 ν 2 ≃ m 2 ν 1 + A eµ cos 2 θ 13 sin 2 θ m 12 + ∆m 2 21 cos 2 (θ 12 − θ m 12 ) , In vacuum limit i.e. θ m ij → θ ij with A eµ → 0, from Eq. (41) we clearly see that the masssquared eigenvalues M 2 ν i go to m 2 ν i . From Eqs. (40) and (41) For a medium with varying density the eigenstates ν mi are no longer eigenstates of the Hamiltonian of Eq. (35). Indeed, since U L is x dependent, the neutrino propagation equation is written as Considering Eq. (44) reduces the above neutrino propagation equation to the two-neutrino state problem in solar matter where the eigenstates are redefined as ν mi → e i(M 2 ν 1 +M 2 ν 2 )/4E ν mi with i = 1, 2 and If the density is slowly changing, on a distance scale of roughly the wavelength in matter, the off-diagonal term dθ m 12 /dx can be negligible. Here we assume the adiabatic condition |dθ m 12 /dx| ≪ |∆M 2 21 |/4E, then the states ν mi become the eigenstates of the Hamiltonian of Eq. (35).
Next, to find a mixing between the active and sterile neutrinos we perform a basis rotation Then the associated Hamiltonian for the active to sterile neutrinos in vacuum is given in the mass eigenstates ξ = (ν i , S i ) T by where X ν =W † ν W ν . In the propagation basis ξ m = (ν mi , S mi ) T of Eq. (33), the associated Hamiltonian in matter is given with the replacement of m 2 ν k by the effective mass M 2 ν k of Eq. (45) as where X ν is equivalent to X m in Eq.
Note that the eigenvalues in the eigenequation H as m |n m = H as mi |n m do not have matter effects and behave like those in vacuum.

Three-active-neutrino Oscillation Probabilities in Matter
We are interested in the flavor transition between the active neutrinos ν e , ν µ , ν τ . The transition probability in matter between the massive neutrinos that a neutrino eigenstate ν α becomes an eigenstate ν β follows from the time evolution of mass eigenstates as where α, β denote either e, µ, or τ , andM 2 ν = diag(M 2 ν 1 , M 2 ν 2 , M 2 ν 3 , m 2 s 1 , m 2 s 2 , m 2 s 3 ) in Eq. (54). We have assumed adiabatic case in Eq. (49) so that n m = (N mi , S mi ) T propagate from production to the surface of matter and the mass eigenstates in matter do not mix. From Eq. (55) the flavor transition probability between the active neutrinos ν e,µ,τ can be explicitly expressed in terms of the oscillation parameters θ, θ m , L, E, and ∆m 2 as where sin 2∆ kj and sin∆ kj are given by Eqs. (20) and (21), respectively, except for the replacement of ∆m 2 kj by ∆M 2 kj = M 2 ν k −M 2 ν j , and the 3 ×3 mixing components U αi ≡ U αi (L) in vacuum at neutrino detection point L and U αi ≡ U αi (x 0 ) in matter at neutrino production point x 0 . As expected, in the limit of θ m ij → θ ij (i = j = 1, 2, 3) with A eµ → 0, the transition probability in matter of Eq. (56) is recovered to the form of Eq. (19) in vacuum; moreover, with the limit of ∆m k → 0 and θ k → 0 it is recovered to the 3νSF, i.e. the standard form of that for three active neutrinos, in vacuum [1].  = 1, 2) and averaging out the associated oscillating phases in the propagation between the Sun and the Earth for given solar neutrino energies, the solar ν e transition at the exposed surface of the Earth can be described by where the sensitivity of θ 3 is negligible due to the tiny value of sin 4 θ 13 ≃ 5 × 10 −4 and θ m 13 ≃ θ 13 of Eq. (43). The above ν e transition probability is different from that of the conventional 3νSF, P m3νSF νe→νe ≃ cos 2 θ 13 cos 2 θ m 13 cos 2 θ m 12 cos 2 θ 12 +sin 2 θ m 12 sin 2 θ 12 +sin 2 θ 13 sin 2 θ m 13 [1], in that it contains oscillatory terms of ∆m 2 1 (2) . We refer to the matter effect that causes a neutrino flavor change via both the effects of sinusoidal oscillation and the MSW matter effect as a "composite matter effect". Contrary to the conventional MSW matter effect that causes a change of electron neutrino but without sinusoidal oscillating terms [19,20], the so-called composite matter effect causes an electron neutrino change via the effects of sinusoidal oscillation induced by the oscillatory terms of ∆m 2 1(2) , as well as the MSW matter effect, as shown in Eq. (57). In the ν e transition probability of Eq. (57) from the composite matter effect to vacuum oscillations, we will show that the value of ∆m 2 KL measured in KamLAND [12,13] can be compatible with that measured in SNO [25], SK [8], and Borexino [11] at higher energies (> 3 MeV) through new oscillation effects induced by the terms containing ∆m 2 1(2) : • If the matter parameter A eµ at ν e production point is much larger than the resonant density, i.e. A eµ (x 0 , E) ≫ A r 12 in Eq. (46), the matter mixing angle θ m 12 of Eq. (44) goes to π/2 for ∆m 21 > 0 and cos 2θ 12 > 0 (θ 12 < π/4). Then, the solar ν e transition can be described 9 as P m νe→νe ≃ cos 4 θ 13 sin 2 θ 12 cos 2 ∆m 2 2
for E > few MeV (such as 8 B and hep neutrinos). As expected, for the earth-sun distance L es being much smaller than the oscillation length, i.e. L es ≪ 4πE/∆m 2 2 , we can obtain a similar form to the 3νSF, P m3νSF νe→νe ≃ cos 4 θ 13 sin 2 θ 12 + sin 4 θ 13 . Consequently, for 8 B neutrinos with energies above a few MeV, the SK+SNO [46] and Borexino [11] data can explain the MSW-LMA solution to solar neutrino oscillations.
For L es ≫ 4πE/∆m 2 2 in Eq. (58), however, the oscillating phase averages out and the LMA solution cannot be explained.
On the other hand, as indicated in Refs. [8,25], the recent observation of low-energy 8 B solar neutrino flux (as low as ∼ 3.5 MeV) by SK [8] has marked the raise of the solar neutrino tension that is a discrepancy appearing in the 3νSF, see Eq. (17). This discrepancy could be due to the oscillating phase of Eq. (58) that only affects solar neutrinos. Interestingly enough, such discrepancy appearing in the 3νSF can be removed by the new oscillation effect without significantly modifying the MSW-LMA solution to solar neutrino oscillations. For oscillation lengths L osc 2 = 4πE/∆m 2 2 being optimized to L es , that is, a bound of ∆m 2 2 1.7×10 −11 eV 2 is roughly derived for energies above 1 MeV in order not to significantly modify the current LMA solution, as illustrated in Fig. 3. In the 3νSF, the mass-squared difference from the 8 B solar neutrinos at SK [8] is somewhat lower than that from the reactor neutrino at KamLAND [12,13] at less than 2σ, as shown in Eq. (17). On the other hand, in our new oscillation framework, the data from the 8 B solar neutrino experiments can be well fitted with that of the KamLAND ν e , as illustrated in Fig. 3: we find, numerically, that a solution to the solar neutrino tension can happen roughly at 0.9 × 10 −11 eV 2 < ∆m 2 2 1.7 × 10 −11 eV 2 , with the global fit of mixing parameters of 3νSF in Table-I. Note here that the above estimation is derived from the 3σ data (instead of the 2σ data) of Table-I. • If the matter parameter A eµ at ν e production is well below the resonant value A r 12 in Eq. (42), i.e. A r 12 ≫ A eµ (x 0 , E), the corresponding matter effects are negligible, leading to θ m 12 → θ 12 . For L es ≫ L osc 1 = 4πE/∆m 2 1 with 0.1 MeV E few MeV, the oscillating phase averages out. Then, it leads to P m νe→νe ≃ cos 4 θ 12 /2 + sin 4 θ 12 cos 2 ∆m 2 2 4E L which cannot explain the Borexino pp, 7 Be, and pep data [11]. In fact, since the oscillation length L osc 1 at the low energy range in Eq. (57) can be sensitive to the earth-sun distance L es , it should be at least one order of magnitude larger than L es in order for the pp, 7 Be, and pep data shown in the left plot of Fig. 3 to fit well: whose condition leads to a bound 10 of ∆m 2 1 |∆m 2 1 | 10 −13 eV 2 .
Then, the ν e survival probability at the exposed surface of the Earth is given by  Fig. 3, for energies less than 1 MeV.
• If the matter parameter A eµ at ν e production is only slightly below the resonant value, 10 This bound satisfies the Borexino pp data [11], as clearly shown in the right plot of Fig. 3, and which is stronger by one order of magnitude than a bound ∆m 2 1 < 1.8 × 10 −12 eV 2 at 3σ in Ref. [47].
All the data (especially, including the red and purple points) shown in Fig. 3 can be consistent with the rough constraints from Eqs. (60) and (62) ∆m 2 1 10 −13 eV 2 , ∆m 2 2 ≃ 10 −11 eV 2 , through the composite matter effects. This indicates that our new oscillation scenario can be a good candidate for explanation of a MSW-LMA solution to the solar neutrino tension.
Future precise measurements of 8 B and pep solar neutrinos may confirm and improve the value of ∆m 2 2 as a solution to the solar neutrino tension, including future measurements of hep solar neutrino which has not been detected yet. Moreover, future measurements of the carbon-nitrogen-oxygen (CNO) cycle neutrinos, one of two sets of nuclear fusion reactions [1], will give a full understanding of solar neutrinos at less than few MeV, as well as the nuclear fission processes inside the Sun.

V. CONCLUSION
This is the first theoretical study of a would-be solution to the so-called solar neutrino tension why solar neutrinos at SNO, SK, and Borexino experiments appear to mix differently from reactor antineutrinos at KamLAND. Three gauge-singlet neutrinos added to the standard model Lagrangian make the neutrinos massive, as required by experimental observations. A unitary condition is imposed to the 6 × 6 mixing matrix which connects the interaction eigenstates with the mass eigenstates. Then the extended theory with three light sterile neutrinos forms pseudo-Dirac pairs that augment three additional oscillation parameter sets (∆m 2 i , θ i ) besides the six oscillation parameters of the 3νSF (∆m 2 Sol , ∆m 2 Atm , θ 12 , θ 23 , θ 13 , δ CP ): two ∆m 2 ABL O(10 −11 ) eV 2 optimized at ABL ( L es = 149.6 × 10 6 km, earth-sun distance) oscillation experiments and one ∆m 2 SBL ∼ O(1) eV 2 optimized at reactor SBL oscillation experiments (with their corresponding mixing angles |θ 1(2) | ≈ 0 ≪ |θ 3 | ∼ O(1)). If the light sterile neutrinos exist and have particular masses, each of them should produce a unique feature that is detectable by its optimized experiment.
In the extended theory, we have derived a general transition probability between the mas-sive neutrinos (that a flavor eigenstate ν α becomes flavor eigenstate ν β with α, β = e, µ, τ ) that can have a potential for explaining the anomalous phenomena (the solar neutrino tension plus SBL anomalies) in terms of neutrino oscillations. Assuming no sterile neutrinos are initially generated when electron neutrinos are produced in the Sun by nuclear reactions. Then, we have re-examined the MSW matter effects in our theoretical framework and suggested a solution to the solar neutrino tension with a so-called composite matter effect that causes a neutrino flavor change with new oscillatory terms containing ∆m 2 ABL O(10 −11 ) eV 2 , so that |∆m 2 1 | ≪ |∆m 2 2 | O(10 −11 ) eV 2 ≪ |∆m 2 3 | ∼ O(1) eV 2 . We stress that, contrary to the conventional matter effect that causes a change in the flavor content of a neutrino but without sinusoidal oscillation, the composite matter effect causes a neutrino flavor change via the effects of sinusoidal oscillation induced by the oscillatory terms containing ∆m 2 ABL , as well as the MSW matter effect.
With the composite matter effect of our theoretical framework, we have shown that the values of ∆m 2 measured in reactor KamLAND, ∆m 2 KL = 7.49 +0.19 −0.18 × 10 −5 eV 2 [14], can be compatible with those measured in solar neutrino experiments (SNO, SK, and Borexino) at energies (> 3 MeV) for ∆m 2 1 10 −13 eV 2 and ∆m 2 2 ≃ 10 −11 eV 2 , as summarized in Fig. 3. This indicates that our new oscillation scenario can be a good candidate for explanation of a MSW-LMA solution to the solar neutrino tension. Future precise measurements of 8 B and pep solar neutrinos may confirm and improve the value of ∆m 2 2 as a solution to the solar neutrino tension, including future measurements of hep solar neutrino which has not been detected yet.