Dirac CP Phase in Neutrino Oscillations

Updated 14 January 2026

The Dirac CP phase is a fundamental parameter in the PMNS matrix that governs observable CP violation in neutrino oscillations.
It is probed through asymmetries in neutrino-antineutrino appearance channels in long-baseline and μDAR-based experiments, exploiting both sin δ and cos δ sensitivities.
Theoretical frameworks use sum rules and residual symmetries to tightly constrain δ, often predicting values near maximal CP violation.

The Dirac CP phase (commonly denoted δ₍CP₎ or simply δ) is a fundamental parameter in the lepton mixing matrix (PMNS matrix) that governs CP-violating effects in neutrino oscillations. As the only currently accessible source of leptonic CP violation in oscillatory phenomena, it encapsulates the complex structure of flavor in the neutrino sector, with profound implications for matter–antimatter asymmetry generation, flavor model building, and the ultimate structure of the lepton sector. The determination and theoretical understanding of δ₍CP₎ is central to ongoing and future experimental programs.

1. Parametrization and Theoretical Definition

The leptonic mixing matrix, or PMNS matrix $U$ , is typically given in the Particle Data Group (PDG) standard parameterization: $U = \begin{pmatrix} c_{12}c_{13} & s_{12}c_{13} & s_{13} e^{-i\delta} \ -s_{12}c_{23}-c_{12}s_{23}s_{13}e^{i\delta} & c_{12}c_{23}-s_{12}s_{23}s_{13}e^{i\delta} & s_{23}c_{13} \ s_{12}s_{23}-c_{12}c_{23}s_{13}e^{i\delta} & -c_{12}s_{23}-s_{12}c_{23}s_{13}e^{i\delta} & c_{23}c_{13} \end{pmatrix}$ where $s_{ij} = \sin \theta_{ij}$ , $c_{ij} = \cos \theta_{ij}$ , and $\delta$ is the Dirac CP phase. In the Majorana neutrino scenario, two additional Majorana phases appear, but these do not affect oscillations.

A mathematically rephasing-invariant, explicit formula for δ in an arbitrary basis is given by: $\delta = \arg\left[\frac{U_{e1} U_{e2} U_{\mu 3} U_{\tau 3}}{U_{e3} \det U}\right]$ This expression, derived via explicit rephasing to the standard form, isolates the Dirac phase from all unphysical phases and remains invariant under arbitrary left- and right-handed field redefinitions (Yang, 16 Dec 2025, Yang, 7 Jul 2025). All physical CP-violating effects in oscillation phenomena arise via δ entering the Jarlskog invariant: $J_{CP} = c_{12} c_{23} c_{13}^2 s_{12} s_{23} s_{13} \sin \delta$ with direct proportionality of observable CP asymmetries to $\sin \delta$ .

2. Phenomenological Consequences and Experimental Probes

Neutrino oscillation probabilities in vacuum and matter are sensitive to δ exclusively through terms containing the imaginary parts of PMNS matrix element products. For instance, the probability for α→β oscillation contains CP-odd terms of the form: $2 \sum_{i<j} \Im[U_{\alpha i}U_{\beta i}^* U_{\alpha j}^* U_{\beta j}] \sin \phi_{ij}$ where δ appears exclusively in the imaginary part, generating differences between neutrino and antineutrino appearance rates (CP asymmetries) (Ohlsson et al., 2013).

Experimental strategies for measuring δ exploit the CP-violating difference in appearance probabilities, optimized through the definition of key observables:

ΔAₐᵦ^m: the maximal swing of the CP asymmetry as δ varies, directly measuring experimental sensitivity.
ΔAₐᵦ^{CP}(δ): the difference in CP asymmetry between δ and δ=0, providing a direct handle on intrinsic CP-violation (Ohlsson et al., 2013).

Conventional long-baseline experiments (T2K, NOvA, DUNE, T2HK) access sin δ at the first oscillation maximum via ν_μ→ν_e appearance. The advent of non-traditional setups like μDAR-based configurations (e.g., TNT2K) introduces both sin δ and cos δ sensitivity, resolving degeneracies such as δ↔π–δ and obviating error blow-up at maximal CP-violation (Ge, 2017).

Table: Representative experimental outcomes as a function of technology and approach.

Proposal	Characteristic Sensitivity to δ	Key Outcome
T2K/NOvA	sin δ only; ~30°–40°	Two-fold degeneracy persists
TNT2K (μDAR)	sin δ and cos δ; ~10°–20°	Single-valued, robust δ
ESSnuSB/DUNE	High coverage, multi-mode	Excludes multiple scenarios

3. Theoretical Constraints, Model Predictions, and Sum Rules

Numerous theoretical frameworks give rise to specific predictions or constraints for the allowed range and structure of δ, often via sum rules, flavor symmetry, or texture-based approaches.

Residual Symmetry and Sum Rules: A single horizontal Z₂ symmetry in the neutrino sector yields the sum rule (Ge et al., 2011, Ge et al., 19 Nov 2025):

$4\,c_{23}s_{23}c_{12}s_{12}s_{13} \cos\delta = (s_{12}^2 - c_{12}^2 s_{13}^2)(c_{23}^2 - s_{23}^2)$

enforcing δ near ±90° (maximal CP violation) for experimentally allowed mixing angles.

Perturbed Mixing Scenarios: Corrections to leading-order bimaximal, tribimaximal, or democratic mixing matrices through single or double complex rotations generate analytic relations linking δ to measured θ₁₂, θ₁₃, θ₂₃, providing model-independent formulas for cos δ. This often implies tight bands around quasi-maximal CP-violation, but with qualitative distinctions depending whether perturbations are in the neutrino or charged-lepton sector (Garg, 2018, Ahn et al., 2022, Delgadillo et al., 2018).
Froggatt–Nielsen Textures: In FN models with det[M_ν]=0, scanning the parameter space restricted by oscillation data yields δ in two symmetric regions around ±π/2, quantitatively δ∈±(0.4–2.9) rad, with stronger localization at ±π/2 if θ₂₃ > 45°. This strictly excludes CP-conserving points δ=0,π (Kaneta et al., 2017).
Statistical/Random Matrix Approaches: Broken democracy with random perturbations to the democratic matrix naturally leads to bifurcated probability peaks near δ=±π/2, in sharp contrast to anarchy models with uniformly flat distributions (Ge et al., 2018). This demonstrates the "democratization" of CP-violating phases when patterns of residual symmetry are weakly broken.
Extended Frameworks (3+1, Sterile Neutrinos): Adding sterile states introduces multiple independent “Dirac-like” phases (I₁, I₂, I₃), with only I₁ reducing to δ in the 3×3 limit. Solutions consistent with global fits can favor normal hierarchy and sub-0.1 eV lightest mass, with nonzero CP phases in active-sterile mixing channels (Becerra-García et al., 2021).

4. Interpretations in Early Universe and Leptogenesis

The Dirac CP phase, while directly impacting oscillation phenomena, also has cosmological consequences:

Thermal Leptogenesis: In extensions of the type-I seesaw (notably ISS+LSS), δ₍CP₎ can be the dominant source of CP violation for baryogenesis via leptogenesis if precise conditions are met (linear seesaw present, resonance enhancement, moderate washout). Successful realization requires δ≈3π/2 and constraints from cLFV and collider processes, but parameter space compatible with data is available (Dolan et al., 2018, Karmakar et al., 2015).
Early Universe Lepton Asymmetries: The impact of δ on neutrino degeneracy parameters at BBN is suppressed except in cases of initial mu-tau asymmetry or substantial muonic backgrounds. The effect on Y_p can reach ~0.1% under favorable initial conditions, offering, in principle, a route to probe δ in the precision cosmological era (Gava et al., 2010).

5. Analytical Structure, Extraction, and Ambiguities

Extraction from Experimental Data: CP-violating effects appear generically as linear combinations a cos δ + b sin δ in appearance probabilities, making direct extraction possible with appropriate event statistics and precise control of systematics. Multi-energy/channel approaches accessing both terms (as via μDAR + accelerator spectra) are essential for unambiguous determination (Ohlsson et al., 2013, Ge, 2017).
Charged-Lepton Corrections and Basis Ambiguities: The relation between the observed δ and the "intrinsic" CP phase in the neutrino sector can be substantially shifted by small charged-lepton mixings. Perturbative expansions show corrections up to 30° or more, mandating their treatment in any global-fit or grand unification analysis (Yang, 7 Jul 2025). Full rephasing-invariant mapping between arbitrary mixing matrices and the PDG form allows systematic extraction of the physical Dirac and Majorana phases from any model or experimental fit (Yang, 16 Dec 2025).
Distinct Cases in Mass Ordering: Some theoretical analyses indicate that certain sum rules or approaches are only predictive for inverted ordering (e.g., via sum of oscillation probabilities), sometimes excluding normal hierarchy as a solution with a unique, nontrivial δ (Todorovic, 2021).

6. Outstanding Questions and Experimental Prospects

The precision measurement of the Dirac CP phase is a stated goal of present and next-generation facilities (DUNE, Hyper-K, ESSnuSB, T2HK, TNT2K, JUNO). Current fits suggest strong but non-maximal CP violation (δ close to 3π/2), with model frameworks providing either robust predictions for maximality, strong correlations with atmospheric angle or mass hierarchy, or broader bands depending on theoretical assumptions (Ge et al., 19 Nov 2025, Kaneta et al., 2017). Residual symmetries (e.g., Z₂ˢ, \overline{Z}_2^s) generate sharply peaked predictions and clear correlations between δ and θ₂₃, enabling near-future critical tests.

Table: Theoretical predictions and corresponding experimental targets.

Scenario	Predicted δ range / feature	Experimental implication
Residual Z₂ symmetry	δ ≃ ±90° (“maximal”)	DUNE/Hyper-K resolve maximality
FN + det M_ν=0	δ ∈ ±(0.4–2.9) rad	Exclude δ=0,π; measure octant
Broken democracy	δ peaked at ±π/2	Distribution near maximal values
Charged-lepton corr.	δ shifted, model-dependent	Sub-30° correction possible
3+1 w/μ–τ breaking	δ ∈ global NH 1σ window; active CP	Multiple nonzero Dirac phases

Ultimately, the determination of the Dirac CP phase will either reveal a key piece of the Standard Model flavor puzzle or necessitate departures from minimal frameworks if deviations from theoretical predictions are observed.

7. Special Cases: Novel CP Phases and Beyond-Standard Frameworks

Exact Dirac Equation Derivation: In the two-flavor, relativistic Dirac equation framework, a novel physical CP phase (the difference Θ_R–Θ_L) arises when right-handed flavor mixing is physical, accessible only if right-handed neutrino interactions can be probed (e.g., in left-right symmetric or BSM models) (Kimura et al., 2021).
Model-Independent Rephasing: All observable CP-violating phases—once reference basis ambiguities are eliminated—can be systematically written as combinations of the arguments of matrix elements and the determinant, as explicitly demonstrated for three generations and generalized to all fermionic mixing matrices (Yang, 16 Dec 2025).