Introduction
One of recent interests with respect to the strong interaction governed by the quantum chromodynamics (QCD) may be to clarify the phase structure on the plane spanned by the temperature and density [1]. Especially, at finite density and low temperature, some authors have pointed out [2 , 3 , 4 , 5 ] that there exists a possibility of a quark spin polarization due to a pseudovector-type quark-antiquark condensation by using the low energy effective model of QCD such as the Nambu-Jona-Lasinio (NJL) model [6 ] . Also, as was shown in the previous work, the spin polarization may also occur due to a tensor-type quark-antiquark condensate in the NJL model [7 ] and the magnetic moment density is evaluated in the hybrid star with a quark matter in the inner core of compact stars [8 ] .
By carrying out the Weinberg transformation, the pseudovector condensate due to the pseudovector interaction in the NJL model can be recast into the inhomogeneous condensate [9 ] . Under the inhomogeneous condensate, the excited modes, especially the Nambu-Goldstone (NG) modes [6 , 10 , 11 ] have been investigated by the Landau-Ginzburg-Wilson effective Lagrangian [12 , 13 ] .
In this paper, a possibility of the existence of the pseudovector-type quark-antiquark condensed phase in the quark matter is investigated and its possibility is indicated taking account of the vacuum effect that leads to the chiral symmetry breaking, which may be neglected in some previous investigations [4 , 5 ] . Further, the massless Nambu-Goldstone modes on the pseudovector condensate due to the pseudovector interaction in the NJL model, which leads to the quark spin polarization, are investigated directly.
We consider two condensates. One is a usual chiral condensate ⟨ ψ ˉ ψ ⟩ \langle\bar{\psi}\psi\rangle ⟨ ψ ˉ ψ ⟩ . Another is the pseudovector condensate, namely ⟨ ψ ˉ γ μ = 3 τ i = 3 ψ ⟩ = ⟨ ψ † Σ 3 τ 3 ψ ⟩ \langle\bar{\psi}\gamma^{\mu=3}\tau_{i=3}\psi\rangle= \langle\psi^\dagger\Sigma_3\tau_3\psi\rangle ⟨ ψ ˉ γ μ = 3 τ i = 3 ψ ⟩ = ⟨ ψ † Σ 3 τ 3 ψ ⟩ , where Σ 3 \Sigma_3 Σ 3 represents the third component of the spin matrix and γ μ \gamma^\mu γ μ and τ i \tau_i τ i represent the Dirac gamma matrices and the isospin matrices, respectively. Although the quark spin polarization, of course, does not appear in vacuum, the quark spin polarization may occur in a certain region of the QCD phase diagram such as the high density and low temperature quark matter. It will be shown that the pseudovector condensed phase may appear in a certain region with respect to the quark chemical potential. Therefore, in consideration of a possibility of existence of both the pseudovector condensate and the chiral condensate in a certain parameter region, the NG modes will be investigated in the case of the coexistence of the pseudovector condensate densate and the chiral condensates as well as in the case of the pseudovector condensate only. In addition to the above two condensates, we consider the tensor condensate, namely ⟨ ψ ˉ γ μ = 1 γ μ = 2 τ i = 3 ψ ⟩ = ⟨ ψ ˉ Σ 3 τ 3 ψ ⟩ \langle\bar{\psi}\gamma^{\mu=1}\gamma^{\mu=2}\tau_{i=3}\psi\rangle=\langle\bar{\psi}\Sigma_{3}\tau_{3}\psi\rangle ⟨ ψ ˉ γ μ = 1 γ μ = 2 τ i = 3 ψ ⟩ = ⟨ ψ ˉ Σ 3 τ 3 ψ ⟩ . This condensate does not coexist with the chiral condensate as was shown in the previous paper [14 ] . Thus, the NG modes will be also investigated on the tensor condensate only.
This paper is organized as follows: In the next section, the gap equations for both the chiral condensate and the pseudovector condensate are derived in the NJL model with the pseudovector interaction between quarks. In section 3, the numerical results for the gap equations are shown and the possibility of the existence of the psuedovector condensed phase at finite density is indicated. In section 4, the massless modes are investigated on the pseudovector condensate. Also, in section 5, the gap equations for the tensor-type quark-antiquark condensate, which may also leads to the quark spin polarization, is revisited following to the previous paper [14 ] in the NJL model with a tensor interaction between quarks. The detailed calculations to derive the massless modes are shown in Appendix § 5.1 . The last section is devoted to a summary and concluding remarks.
Gap equations with respect to the chiral condensate and the pseudovector condensate
Let us introduce the NJL model with vector and pseudovector interactions which retains the chiral symmetry [6 , 15 , 16 , 17 ] :
L N J L = ψ ˉ i γ μ ∂ μ ψ + G S [ ( ψ ˉ ψ ) 2 + ( ψ ˉ i γ 5 τ ψ ) 2 ] + G [ ( ψ ˉ γ μ τ ψ ) 2 + ( ψ ˉ γ 5 γ μ τ ψ ) 2 ] , (1) \mathcal{L}_{\mathrm{NJL}}=\bar{\psi}i\gamma^\mu\partial_\mu\psi+G_{S}\left[(\bar{\psi}\psi)^{2}+(\bar{\psi}i\gamma_{5}\boldsymbol{\tau}\psi)^{2}\right]+G\left[(\bar{\psi}\gamma^\mu\boldsymbol{\tau}\psi)^{2}+(\bar{\psi}\gamma_{5}\gamma^\mu\boldsymbol{\tau}\psi)^{2}\right], \tag*{(1)} L NJL = ψ ˉ i γ μ ∂ μ ψ + G S [ ( ψ ˉ ψ ) 2 + ( ψ ˉ i γ 5 τ ψ ) 2 ] + G [ ( ψ ˉ γ μ τ ψ ) 2 + ( ψ ˉ γ 5 γ μ τ ψ ) 2 ] , (1)
where τ \boldsymbol{\tau} τ represents the isospin matrices with respect to the flavor s u ( 2 ) su(2) s u ( 2 ) -symmetry. The generating functional Z Z Z is written as
Z = ∫ D ψ D ψ ˉ exp ( i ∫ d 4 x L N J L ) . (2) Z=\int\mathcal{D}\psi\mathcal{D}\bar{\psi}\exp\left(i\int d^{4}x\mathcal{L}_{\mathrm{NJL}}\right). \tag*{(2)} Z = ∫ D ψ D ψ ˉ exp ( i ∫ d 4 x L NJL ) . (2)
Here, we use the auxiliary field method and introduce the functional Gaussian integral leading to unit as
1 = ∫ D σ ′ D π ′ exp ( i ∫ d 4 x [ − 1 G S π ′ 2 ( σ ′ 2 + π ′ 2 ) ] ) , 1 = ∫ D ρ ′ μ D a ′ μ exp ( i ∫ d 4 x [ − 1 G ( ρ ′ μ ρ μ ′ + a ′ μ a μ ′ ) ] ) . (3) \begin{aligned}
1&=\int\mathcal{D}\sigma'\mathcal{D}\pi'\exp\left(i\int d^{4}x\left[-\frac{1}{G_{S}\vphantom{\pi'^2}}\left(\sigma'^{2}+\pi'^{2}\right)\right]\right),\\
1&=\int\mathcal{D}\rho'^{\mu}\mathcal{D}a'^{\mu}\exp\left(i\int d^{4}x\left[-\frac{1}{G}\left(\rho'^{\mu}\rho'_{\mu}+a'^{\mu}a'_{\mu}\right)\right]\right). \tag*{(3)}
\end{aligned} 1 1 = ∫ D σ ′ D π ′ exp ( i ∫ d 4 x [ − G S π ′2 1 ( σ ′ 2 + π ′ 2 ) ] ) , = ∫ D ρ ′ μ D a ′ μ exp ( i ∫ d 4 x [ − G 1 ( ρ ′ μ ρ μ ′ + a ′ μ a μ ′ ) ] ) . (3)
Inserting the above unit into the generating functional Z Z Z in Eq. (2) and replacing σ ′ \sigma' σ ′ , π ′ \pi' π ′ , ρ ′ μ \rho'^{\mu} ρ ′ μ and a ′ μ a'^{\mu} a ′ μ into σ \sigma σ , π \pi π , ρ μ \rho^{\mu} ρ μ and a μ a^{\mu} a μ as
σ ′ = σ + G S ( ψ ˉ ψ ) , π ′ = π + G S ( ψ ˉ i γ 5 τ ψ ) , ρ ′ μ = ρ μ + G ( ψ ˉ γ μ τ ψ ) , a ′ μ = a μ + G ( ψ ˉ γ 5 γ μ τ ψ ) , (4) \begin{aligned}
\sigma' &= \sigma+ G_{S}(\bar{\psi}\psi),\\
\pi' &= \pi+ G_{S}(\bar{\psi}i\gamma_{5}\tau\psi),\\
\rho'^{\mu} &= \rho^{\mu} + G(\bar{\psi}\gamma^{\mu}\tau\psi),\\
a'^{\mu} &= a^{\mu} + G(\bar{\psi}\gamma_{5}\gamma^{\mu}\tau\psi), \tag*{(4)}
\end{aligned} σ ′ π ′ ρ ′ μ a ′ μ = σ + G S ( ψ ˉ ψ ) , = π + G S ( ψ ˉ i γ 5 τ ψ ) , = ρ μ + G ( ψ ˉ γ μ τ ψ ) , = a μ + G ( ψ ˉ γ 5 γ μ τ ψ ) , (4)
the generating functional is rewritten as
Z = ∫ D ψ D ψ ˉ ∫ D σ D π D ρ μ D a μ exp ( i ∫ d 4 x L ~ ) , (5) Z = \int\mathcal{D}\psi\mathcal{D}\bar{\psi}\int\mathcal{D}\sigma\mathcal{D}\pi\mathcal{D}\rho^{\mu}\mathcal{D}a^{\mu}\exp\left(i\int d^{4}x\widetilde{\mathcal{L}}\right), \tag*{(5)} Z = ∫ D ψ D ψ ˉ ∫ D σ D π D ρ μ D a μ exp ( i ∫ d 4 x L ) , (5)
where
L ~ = ψ ˉ ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 ρ μ ⋅ γ μ τ − 2 a μ ⋅ γ 5 γ μ τ ) ψ − 1 G S ( σ 2 + π 2 ) − 1 G ( ρ μ ⋅ ρ μ + a μ ⋅ a μ ) . (6) \widetilde{\mathcal{L}} = \bar{\psi}(i\gamma^{\mu}\partial_{\mu}-2\sigma-2\pi\cdot i\gamma_{5}\tau-2\rho_{\mu}\cdot\gamma^{\mu}\tau-2a_{\mu}\cdot\gamma_{5}\gamma^{\mu}\tau)\psi-\frac{1}{G_{S}}(\sigma^{2}+\pi^{2})-\frac{1}{G}(\rho^{\mu}\cdot\rho_{\mu}+a^{\mu}\cdot a_{\mu}). \tag*{(6)} L = ψ ˉ ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 ρ μ ⋅ γ μ τ − 2 a μ ⋅ γ 5 γ μ τ ) ψ − G S 1 ( σ 2 + π 2 ) − G 1 ( ρ μ ⋅ ρ μ + a μ ⋅ a μ ) . (6)
Here, from δ L ~ / δ σ = 0 \delta\widetilde{\mathcal{L}}/\delta\sigma=0 δ L / δ σ = 0 and so on, we obtain
σ = − G S ( ψ ˉ ψ ) , π = − G S ( ψ ˉ i γ 5 τ ψ ) , ρ μ = − G ( ψ ˉ γ μ τ ψ ) , a μ = − G ( ψ ˉ γ 5 γ μ τ ψ ) . (7) \begin{aligned}
\sigma&= -G_{S}(\bar{\psi}\psi),\\
\pi&= -G_{S}(\bar{\psi}i\gamma_{5}\tau\psi),\\
\rho_{\mu} &= -G(\bar{\psi}\gamma_{\mu}\tau\psi),\\
a_{\mu} &= -G(\bar{\psi}\gamma_{5}\gamma_{\mu}\tau\psi). \tag*{(7)}
\end{aligned} σ π ρ μ a μ = − G S ( ψ ˉ ψ ) , = − G S ( ψ ˉ i γ 5 τ ψ ) , = − G ( ψ ˉ γ μ τ ψ ) , = − G ( ψ ˉ γ 5 γ μ τ ψ ) . (7)
Further, integrating ψ \psi ψ and ψ ˉ \bar{\psi} ψ ˉ , the generating functional can be expressed as
Z = ∫ D σ D π D ρ μ D a μ exp ( i Γ ) , Z = \int\mathcal{D}\sigma\mathcal{D}\pi\mathcal{D}\rho^{\mu}\mathcal{D}a^{\mu}\exp(i\Gamma), Z = ∫ D σ D π D ρ μ D a μ exp ( i Γ ) ,
Γ = ∫ d 4 x [ − 1 G S ( σ 2 + π 2 ) − 1 G ( ρ μ ⋅ ρ μ + a μ ⋅ a μ ) ] − i ln Det ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 ρ μ ⋅ γ μ τ − 2 a μ ⋅ γ 5 γ μ τ ) . (8) \Gamma= \int d^{4}x\left[-\frac{1}{G_{S}}(\sigma^{2}+\pi^{2})-\frac{1}{G}(\rho^{\mu}\cdot\rho_{\mu}+a^{\mu}\cdot a_{\mu})\right]-i\ln\operatorname{Det}\left(i\gamma^{\mu}\partial_{\mu}-2\sigma-2\pi\cdot i\gamma_{5}\tau-2\rho_{\mu}\cdot\gamma^{\mu}\tau-2a_{\mu}\cdot\gamma_{5}\gamma^{\mu}\tau\right). \tag*{(8)} Γ = ∫ d 4 x [ − G S 1 ( σ 2 + π 2 ) − G 1 ( ρ μ ⋅ ρ μ + a μ ⋅ a μ ) ] − i ln Det ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 ρ μ ⋅ γ μ τ − 2 a μ ⋅ γ 5 γ μ τ ) . (8)
It should be noted that the pseudovector mode a 3 a_{3} a 3 is written as
a 3 = − G ψ ˉ γ 5 γ 3 τ ψ = − G ψ † Σ 3 τ ψ . (9) \begin{aligned}
a_{3} &= -G\bar{\psi}\gamma_{5}\gamma_{3}\tau\psi\\
&= -G\psi^{\dagger}\Sigma_{3}\tau\psi. \tag*{(9)}
\end{aligned} a 3 = − G ψ ˉ γ 5 γ 3 τ ψ = − G ψ † Σ 3 τ ψ . (9)
Here, we use the Dirac representation for the Dirac gamma matrices, namely,
γ μ = ( γ 0 , γ ) , γ 0 = ( 1 0 0 − 1 ) , γ = ( 0 σ − σ 0 ) , γ 5 = ( 0 1 1 0 ) , \gamma^\mu=(\gamma^0,\boldsymbol{\gamma}),\qquad
\gamma^0=\begin{pmatrix}1&0\\0&-1\end{pmatrix},\qquad
\boldsymbol{\gamma}=\begin{pmatrix}0&\boldsymbol{\sigma}\\-\boldsymbol{\sigma}&0\end{pmatrix},\qquad
\gamma_5=\begin{pmatrix}0&1\\1&0\end{pmatrix}, γ μ = ( γ 0 , γ ) , γ 0 = ( 1 0 0 − 1 ) , γ = ( 0 − σ σ 0 ) , γ 5 = ( 0 1 1 0 ) ,
γ 0 γ 5 γ 3 ≡ Σ 3 = ( σ 3 0 0 σ 3 ) , (10) \gamma^0\gamma_5\gamma_3\equiv\Sigma_3=\begin{pmatrix}\sigma_3&0\\0&\sigma_3\end{pmatrix}, \tag*{(10)} γ 0 γ 5 γ 3 ≡ Σ 3 = ( σ 3 0 0 σ 3 ) , (10)
where σ \boldsymbol{\sigma} σ and σ 3 \sigma_3 σ 3 represent the Pauli spin matrices and its third component. Thus, Σ 3 \Sigma_3 Σ 3 is the third component of the spin matrices.
Hereafter, let us assume that there exist the chiral condensate σ 0 = ⟨ σ ⟩ \sigma_0=\langle\sigma\rangle σ 0 = ⟨ σ ⟩ and pseudovector condensate a 0 = ⟨ a μ = 3 i = 3 ⟩ a_0=\langle a_{\mu=3}^{i=3}\rangle a 0 = ⟨ a μ = 3 i = 3 ⟩ with μ = 3 \mu=3 μ = 3 and the third component of the isospin, namely
σ 0 ≡ ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩ , \sigma_0\equiv\langle\sigma\rangle=-G_S\langle\bar{\psi}\psi\rangle, σ 0 ≡ ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩ ,
a 0 ≡ ⟨ a μ = 3 i = 3 ⟩ = − G ⟨ ψ † Σ 3 τ 3 ψ ⟩ . (11) a_0\equiv\langle a_{\mu=3}^{i=3}\rangle=-G\langle\psi^\dagger\Sigma_3\tau_3\psi\rangle. \tag*{(11)} a 0 ≡ ⟨ a μ = 3 i = 3 ⟩ = − G ⟨ ψ † Σ 3 τ 3 ψ ⟩ . (11)
Thus, the pseudovector condensate ⟨ a μ = 3 i = 3 ⟩ \langle a_{\mu=3}^{i=3}\rangle ⟨ a μ = 3 i = 3 ⟩ can be regarded as the quark spin polarization. Under these quark-antiquark condensates, the effective potential V ( σ 0 , a 0 ) V(\sigma_0,a_0) V ( σ 0 , a 0 ) can be derived from the effective action Γ \Gamma Γ as
Γ ( σ = σ 0 , a 3 3 = a 0 ) = − V ( σ 0 , a 0 ) ∫ d 4 x , \Gamma(\sigma=\sigma_0,a_3^3=a_0)=-V(\sigma_0,a_0)\int d^4x, Γ ( σ = σ 0 , a 3 3 = a 0 ) = − V ( σ 0 , a 0 ) ∫ d 4 x ,
V ( σ 0 , a 0 ) = 1 G S σ 0 2 − 1 G a 0 2 − ∫ d 4 k i ( 2 π ) 4 ln det [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = 1 G S σ 0 2 − 1 G a 0 2 − ∫ d 4 k i ( 2 π ) 4 tr ln [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] . (12) \begin{aligned}
V(\sigma_0,a_0)=\frac{1}{G_S}\sigma_0^2-\frac{1}{G}a_0^2-\int\frac{d^4k}{i(2\pi)^4}\ln\det\left[\gamma^\mu k_\mu-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3\right] \\
&=\frac{1}{G_S}\sigma_0^2-\frac{1}{G}a_0^2-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\ln\left[\gamma^\mu k_\mu-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3\right]. \tag*{(12)}
\end{aligned} V ( σ 0 , a 0 ) = G S 1 σ 0 2 − G 1 a 0 2 − ∫ i ( 2 π ) 4 d 4 k ln det [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = G S 1 σ 0 2 − G 1 a 0 2 − ∫ i ( 2 π ) 4 d 4 k tr ln [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] . (12)
where trace is taken by the Dirac gamma matrices, isospin space and color space. Here, it should be noted that a i = 3 μ = 3 a μ = 3 i = 3 = − a 0 2 a_{i=3}^{\mu=3}a_{\mu=3}^{i=3}=-a_0^2 a i = 3 μ = 3 a μ = 3 i = 3 = − a 0 2 is used because, in this paper, the metric tensor is adopted as
g μ ν = diag. ( 1 , − 1 , − 1 , − 1 ) . (13) g^{\mu\nu}=\operatorname{diag.}(1,-1,-1,-1). \tag*{(13)} g μν = diag. ( 1 , − 1 , − 1 , − 1 ) . (13)
The condensates σ 0 \sigma_0 σ 0 and a 0 a_0 a 0 are determined by the following gap equations:
∂ V ∂ σ 0 = 2 G S σ 0 − ∫ d 4 k i ( 2 π ) 4 tr [ − 2 γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = 0 , \frac{\partial V}{\partial\sigma_0}=\frac{2}{G_S}\sigma_0-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\frac{-2}{\gamma^\mu k_\mu-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right]=0, ∂ σ 0 ∂ V = G S 2 σ 0 − ∫ i ( 2 π ) 4 d 4 k tr [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 − 2 ] = 0 ,
∂ V ∂ a 0 = − 2 G a 0 − ∫ d 4 k i ( 2 π ) 4 tr [ 1 γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ( − 2 γ 5 γ 3 τ 3 ) ] = 0. (14) \frac{\partial V}{\partial a_0}=-\frac{2}{G}a_0-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\frac{1}{\gamma^\mu k_\mu-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}(-2\gamma_5\gamma^3\tau_3)\right]=0. \tag*{(14)} ∂ a 0 ∂ V = − G 2 a 0 − ∫ i ( 2 π ) 4 d 4 k tr [ γ μ k μ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ( − 2 γ 5 γ 3 τ 3 ) ] = 0. (14)
which, after some tedious calculation for the gamma matrices, lead to
σ 0 [ 1 + 48 G S ∫ d 4 k i ( 2 π ) 4 k 2 − 4 σ 0 2 + 4 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) ] = 0 a 0 [ 1 − 48 G ∫ d 4 k i ( 2 π ) 4 k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) ] = 0 (15) \begin{aligned}
\sigma_0\left[1+48G_S\int\frac{d^4k}{i(2\pi)^4}\frac{k^2-4\sigma_0^2+4a_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\right]&=0\\
a_0\left[1-48G\int\frac{d^4k}{i(2\pi)^4}\frac{k^2+4\sigma_0^2-4a_0^2+2k_3^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\right]&=0 \tag*{(15)}
\end{aligned} σ 0 [ 1 + 48 G S ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 − 4 σ 0 2 + 4 a 0 2 ] a 0 [ 1 − 48 G ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 ] = 0 = 0 (15)
where k 2 = k 0 2 − k 2 k^2=k_0^2-\boldsymbol{k}^2 k 2 = k 0 2 − k 2 and k 3 k_3 k 3 means the third component of the space vector k \boldsymbol{k} k .
Numerical evaluations for the solutions of the gap equations
The NJL model is a cutoff theory. We determine the three-momentum cutoff Λ \Lambda Λ and the coupling constant G S G_S G S in order to reproduce the pion decay constant f π = 92 f_\pi=92 f π = 92 MeV and the quark mass M = 2 σ 0 = 313 M=2\sigma_0=313 M = 2 σ 0 = 313 MeV. In this paper, we take Λ = 643 \Lambda=643 Λ = 643 MeV and G S = 5.20 G_S=5.20 G S = 5.20 GeV− 2 ^{-2} − 2 . However, the interaction strength G G G for the pseudovector interaction cannot be determined. We here take G G G as a free parameter.
In this section only, the notations are changed for simplicity as follows:
G = − G p v , σ 0 = σ s 2 , a 0 = σ p v 2 . (16) G=-G_{pv},\qquad\sigma_0=\frac{\sigma_s}{2},\qquad a_0=\frac{\sigma_{pv}}{2}. \tag*{(16)} G = − G p v , σ 0 = 2 σ s , a 0 = 2 σ p v . (16)
Of course, σ s \sigma_s σ s corresponds to the quark mass M M M . Further, in order to consider the quark matter with finite quark chemical potential μ \mu μ , we add one term μ ψ † ψ \mu\psi^\dagger\psi μ ψ † ψ in the effective Lagrangian density (6) . Thus, we should replace k 0 k_0 k 0 into k 0 + μ k_0+\mu k 0 + μ in the gap equations. Then, the gap equations (15) are recast into
σ s [ 1 + 48 G S ∫ d 4 k i ( 2 π ) 4 ( k 0 + μ ) 2 − k 2 − σ s 2 + σ p v 2 ( ( k 0 + μ ) 2 − k 2 − σ s 2 − σ p v 2 ) 2 − 4 σ p v 2 ( k 3 2 + σ s 2 ) ] = 0 σ p v [ 1 + 48 G p v ∫ d 4 k i ( 2 π ) 4 ( k 0 + μ ) 2 − k 2 + σ s 2 − σ p v 2 + 2 k 3 2 ( ( k 0 + μ ) 2 − k 2 − σ s 2 − σ p v 2 ) 2 − 4 σ p v 2 ( k 3 2 + σ s 2 ) ] = 0 (17) \begin{aligned}
\sigma_s\left[1+48G_S\int\frac{d^4k}{i(2\pi)^4}\frac{(k_0+\mu)^2-\boldsymbol{k}^2-\sigma_s^2+\sigma_{pv}^2}{\left((k_0+\mu)^2-\boldsymbol{k}^2-\sigma_s^2-\sigma_{pv}^2\right)^2-4\sigma_{pv}^2(k_3^2+\sigma_s^2)}\right]&=0\\
\sigma_{pv}\left[1+48G_{pv}\int\frac{d^4k}{i(2\pi)^4}\frac{(k_0+\mu)^2-\boldsymbol{k}^2+\sigma_s^2-\sigma_{pv}^2+2k_3^2}{\left((k_0+\mu)^2-\boldsymbol{k}^2-\sigma_s^2-\sigma_{pv}^2\right)^2-4\sigma_{pv}^2(k_3^2+\sigma_s^2)}\right]&=0 \tag*{(17)}
\end{aligned} σ s [ 1 + 48 G S ∫ i ( 2 π ) 4 d 4 k ( ( k 0 + μ ) 2 − k 2 − σ s 2 − σ p v 2 ) 2 − 4 σ p v 2 ( k 3 2 + σ s 2 ) ( k 0 + μ ) 2 − k 2 − σ s 2 + σ p v 2 ] σ p v [ 1 + 48 G p v ∫ i ( 2 π ) 4 d 4 k ( ( k 0 + μ ) 2 − k 2 − σ s 2 − σ p v 2 ) 2 − 4 σ p v 2 ( k 3 2 + σ s 2 ) ( k 0 + μ ) 2 − k 2 + σ s 2 − σ p v 2 + 2 k 3 2 ] = 0 = 0 (17)
If there exist the nontrivial solutions σ s ≠ 0 \sigma_s\ne0 σ s = 0 and/or σ p v ≠ 0 \sigma_{pv}\ne0 σ p v = 0 of the above gap equations, we can rewrite the gap equations introducing the three-momentum cutoff Λ \Lambda Λ as
σ s = − 24 ( 2 π ) 2 G S ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 [ σ s − 2 σ s σ p v k 3 2 + σ s 2 ϵ k ( − ) ( θ ( μ − ϵ k ( − ) ) − θ ( μ + ϵ k ( − ) ) ) + σ s + 2 σ s σ p v k 3 2 + σ s 2 ϵ k ( + ) ( θ ( μ − ϵ k ( + ) ) − θ ( μ + ϵ k ( + ) ) ) ] , (18) \begin{aligned}
\sigma_s={}&-\frac{24}{(2\pi)^2}G_S\int_0^\Lambda dk_{12}\int_0^{\sqrt{\Lambda^2-k_{12}^2}}dk_3\,k_{12}\left[\frac{\sigma_s-\frac{2\sigma_s\sigma_{pv}}{\sqrt{k_3^2+\sigma_s^2}}}{\epsilon_k^{(-)}}\left(\theta(\mu-\epsilon_k^{(-)})-\theta(\mu+\epsilon_k^{(-)})\right)\right.\\
&\left.\qquad+\frac{\sigma_s+\frac{2\sigma_s\sigma_{pv}}{\sqrt{k_3^2+\sigma_s^2}}}{\epsilon_k^{(+)}}\left(\theta(\mu-\epsilon_k^{(+)})-\theta(\mu+\epsilon_k^{(+)})\right)\right], \tag*{(18)}
\end{aligned} σ s = − ( 2 π ) 2 24 G S ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 ϵ k ( − ) σ s − k 3 2 + σ s 2 2 σ s σ p v ( θ ( μ − ϵ k ( − ) ) − θ ( μ + ϵ k ( − ) ) ) + ϵ k ( + ) σ s + k 3 2 + σ s 2 2 σ s σ p v ( θ ( μ − ϵ k ( + ) ) − θ ( μ + ϵ k ( + ) ) ) , (18)
σ p v = − 24 ( 2 π ) 2 G p v ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 [ σ p v − k 3 2 + σ s 2 ϵ k ( − ) ( θ ( μ − ϵ k ( − ) ) − θ ( μ + ϵ k ( − ) ) ) + σ p v + k 3 2 + σ s 2 ϵ k ( + ) ( θ ( μ − ϵ k ( + ) ) − θ ( μ + ϵ k ( + ) ) ) ] , (19) \begin{aligned}
\sigma_{pv}={}&-\frac{24}{(2\pi)^2}G_{pv}\int_0^\Lambda dk_{12}\int_0^{\sqrt{\Lambda^2-k_{12}^2}}dk_3\,k_{12}\left[\frac{\sigma_{pv}-\sqrt{k_3^2+\sigma_s^2}}{\epsilon_k^{(-)}}\left(\theta(\mu-\epsilon_k^{(-)})-\theta(\mu+\epsilon_k^{(-)})\right)\right.\\
&\left.\qquad+\frac{\sigma_{pv}+\sqrt{k_3^2+\sigma_s^2}}{\epsilon_k^{(+)}}\left(\theta(\mu-\epsilon_k^{(+)})-\theta(\mu+\epsilon_k^{(+)})\right)\right], \tag*{(19)}
\end{aligned} σ p v = − ( 2 π ) 2 24 G p v ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 [ ϵ k ( − ) σ p v − k 3 2 + σ s 2 ( θ ( μ − ϵ k ( − ) ) − θ ( μ + ϵ k ( − ) ) ) + ϵ k ( + ) σ p v + k 3 2 + σ s 2 ( θ ( μ − ϵ k ( + ) ) − θ ( μ + ϵ k ( + ) ) ) ] , (19)
where
ϵ k ( ± ) = k 2 + σ s 2 + σ p v 2 ± 2 σ p v k 3 2 + σ s 2 , k 12 = k 1 2 + k 2 2 . (20) \epsilon_k^{(\pm)}=\sqrt{k^2+\sigma_s^2+\sigma_{pv}^2\pm2\sigma_{pv}\sqrt{k_3^2+\sigma_s^2}},\qquad k_{12}=\sqrt{k_1^2+k_2^2}. \tag*{(20)} ϵ k ( ± ) = k 2 + σ s 2 + σ p v 2 ± 2 σ p v k 3 2 + σ s 2 , k 12 = k 1 2 + k 2 2 . (20)
In Eqs. (18) and (19) , the terms with θ ( μ + ϵ k ( ± ) ) \theta(\mu+\epsilon_k^{(\pm)}) θ ( μ + ϵ k ( ± ) ) in the right-hand sides represent the vacuum effects which should not be omitted in this paper, while these effects were often neglected in the previous work.
Similarly, the effective potential (12) is calculated as follows:
V ( σ s , σ p v ) = σ s 2 4 G S + σ p v 2 4 G p v − 12 ( 2 π ) 2 ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 [ ( ϵ k ( − ) − μ ) θ ( ϵ k ( − ) − μ ) + ( ϵ k ( + ) − μ ) θ ( ϵ k ( + ) − μ ) ] . (21) \begin{aligned}
V(\sigma_s,\sigma_{pv})={}&\frac{\sigma_s^2}{4G_S}+\frac{\sigma_{pv}^2}{4G_{pv}}-\frac{12}{(2\pi)^2}\int_0^\Lambda dk_{12}\int_0^{\sqrt{\Lambda^2-k_{12}^2}}dk_3k_{12}\left[(\epsilon_k^{(-)}-\mu)\theta(\epsilon_k^{(-)}-\mu)\right.\\
&\left.\qquad+(\epsilon_k^{(+)}-\mu)\theta(\epsilon_k^{(+)}-\mu)\right]. \tag*{(21)}
\end{aligned} V ( σ s , σ p v ) = 4 G S σ s 2 + 4 G p v σ p v 2 − ( 2 π ) 2 12 ∫ 0 Λ d k 12 ∫ 0 Λ 2 − k 12 2 d k 3 k 12 [ ( ϵ k ( − ) − μ ) θ ( ϵ k ( − ) − μ ) + ( ϵ k ( + ) − μ ) θ ( ϵ k ( + ) − μ ) ] . (21)
Let us, first, consider the effective potential without the chiral condensate, namely σ s = 0 \sigma_s=0 σ s = 0 , in order to show the existence of the nontrivial solution σ p v ≠ 0 \sigma_{pv}\ne0 σ p v = 0 at finite quark chemical potential. Figure 1 shows the effective potential (right panels) and the pseudovector condensate σ p v \sigma_{pv} σ p v (left panels) at some values of the pseudovector interaction strength G p v G_{pv} G p v , respectively. In Figure 1 , the effective potential V ( σ p v ) V(\sigma_{pv}) V ( σ p v ) is normalized by V ( σ p v = 0 ) = 0 V(\sigma_{pv}=0)=0 V ( σ p v = 0 ) = 0 , namely V ( σ p v ) = V ( σ s = 0 , σ p v ) − V ( σ s = 0 , σ p v = 0 ) V(\sigma_{pv})=V(\sigma_s=0,\sigma_{pv})-V(\sigma_s=0,\sigma_{pv}=0) V ( σ p v ) = V ( σ s = 0 , σ p v ) − V ( σ s = 0 , σ p v = 0 ) . It is shown that, in the case of G p v = 1.1 G S G_{pv}=1.1G_S G p v = 1.1 G S , the effective potential has a minimum at σ p v = 0 \sigma_{pv}=0 σ p v = 0 in all the chemical potential μ \mu μ . However, in the case above G p v = 1.2 G S G_{pv}=1.2G_S G p v = 1.2 G S , the effective potential has a minimum at a point with σ p v ≠ 0 \sigma_{pv} \ne0 σ p v = 0 . Of course, as G p v G_{pv} G p v becomes larger, then σ p v \sigma_{pv} σ p v becomes larger. Thus, there is a possibility that the chiral symmetry is broken without the chiral condensate, σ s = 0 \sigma_s=0 σ s = 0 , for rather large pseudovector interaction strength. .
Figure 1. The effective potentials (right panels) and the pseudovector condensates σ p v \sigma_{pv} σ p v (left panels) are shown with the pseudovector interaction strength G p v = 1.1 G S G_{pv}=1.1G_S G p v = 1.1 G S , 1.2 G S 1.2G_S 1.2 G S and 1.3 G S 1.3G_S 1.3 G S , respectively.
Next, let us consider both the condensates σ s \sigma_s σ s and σ p v \sigma_{pv} σ p v simultaneously. In Figure 2 , the contour plots of the effective potential as a function of the chiral and the psuedovector condensates, σ s \sigma_s σ s and σ p v \sigma_{pv} σ p v , are depicted with different values of the quark chemical potential μ \mu μ in the case of G p v = 1.2 G S G_{pv}=1.2G_S G p v = 1.2 G S . At the quark chemical potential being 0, the absolute minimum point of the effective potential is given by the nonzero chiral condensate σ s \sigma_s σ s and σ p v = 0 \sigma_{pv}=0 σ p v = 0 . As the quark chemical potential becomes larger, the local minimum of the effective potential develops at σ p v ≠ 0 \sigma_{pv}\ne0 σ p v = 0 and σ s = 0 \sigma_s=0 σ s = 0 . However, the absolute minimum is still the point with σ s ≠ 0 \sigma_s\ne0 σ s = 0 and σ p v = 0 \sigma_{pv}=0 σ p v = 0 . On the other hand, around μ = 312 \mu=312 μ = 312 MeV, the local minimum with
Figure 2. The contour plots of the effective potential as a function of the chiral and the psuedovector condensates, σ s \sigma_s σ s and σ p v \sigma_{pv} σ p v , are depicted with the quark chemical potential μ = 0 \mu=0 μ = 0 , 300 MeV, 312 MeV, 320 MeV and 400 MeV, respectively, in the case of G p v = 1.2 G S G_{pv}=1.2G_S G p v = 1.2 G S .
σ s = 0 \sigma_s = 0 σ s = 0 and σ p v ≠ 0 \sigma_{pv} \ne0 σ p v = 0 is converted to the absolute minimum and vice versa. In the region with larger quark chemical potential, such as μ ≈ 400 \mu\approx400 μ ≈ 400 MeV, the local minimum with σ s ≠ 0 \sigma_s \ne0 σ s = 0 disappears. Thus, the first order phase transition is seen from the chiral condensed phase to the pseudovector condensed phase.
Figure 3 is the same as Figure 2 except for the pseudovector interaction strength, namely G p v = 1.3 G S G_{pv} = 1.3G_S G p v = 1.3 G S instead of G p v = 1.2 G S G_{pv} = 1.2G_S G p v = 1.2 G S . In this strong coupling case, the absolute minimum of the effective potential gives nonzero σ p v \sigma_{pv} σ p v and σ s = 0 \sigma_s = 0 σ s = 0 even at the quark chemical potential μ \mu μ being 0. On the other hand, the usual vacuum with σ s ≠ 0 \sigma_s \ne0 σ s = 0 and σ p v = 0 \sigma_{pv} = 0 σ p v = 0 corresponds to the local minimum of the effective potential. This situation is unphysical because, if this situation is realized, it results in the quark spin polarization due to the pseudovector condensate in vacuum. Thus, even if there is the pseudovector interaction between quarks, the interaction strength must be restricted.
Figure 3. The contour plots of the effective potential as a function of the chiral and the psuedovector condensates, σ s \sigma_s σ s and σ p v \sigma_{pv} σ p v , are depicted with the quark chemical potential μ = 0 \mu= 0 μ = 0 , 200 200 200 MeV, 300 300 300 MeV and 400 400 400 MeV, respectively, in the case of G p v = 1.3 G S G_{pv} = 1.3G_S G p v = 1.3 G S .
Massless Meson Modes
In this section, the massless Nambu-Goldstone modes on the chiral and/or pseudovector condensates are investigated. The meson propagator Δ α ( x , y ) \Delta_\alpha(x,y) Δ α ( x , y ) can be derived from the effective action Γ \Gamma Γ as
Δ α − 1 ( x , y ) = − δ 2 Γ δ α ( x ) δ α ( y ) (22) \Delta_\alpha^{-1}(x,y)=-\frac{\delta^2\Gamma}{\delta\alpha(x)\delta\alpha(y)} \tag*{(22)} Δ α − 1 ( x , y ) = − δ α ( x ) δ α ( y ) δ 2 Γ (22)
where α ( x ) = σ ( x ) \alpha(x)=\sigma(x) α ( x ) = σ ( x ) , π ( x ) \boldsymbol{\pi}(x) π ( x ) , ρ μ ( x ) \rho_\mu(x) ρ μ ( x ) or a μ ( x ) a_\mu(x) a μ ( x ) . In the momentum representation, the two-point vertex function Γ α ( p ) \Gamma_\alpha(p) Γ α ( p ) can be expressed as
Γ α ( p ) ( 2 π ) 4 δ 4 ( p + q ) = δ 2 Γ δ α ( p ) δ α ( q ) ∣ σ = σ 0 , a 3 3 = a 0 , π = ρ μ = a μ ≠ 3 = a μ = 3 i = 0 (23) \left.\Gamma_\alpha(p)(2\pi)^4\delta^4(p+q)=\frac{\delta^2\Gamma}{\delta\alpha(p)\delta\alpha(q)}\right|_{\sigma=\sigma_0,a_3^3=a_0,\pi=\rho_\mu=a_{\mu\ne3}=a^i_{\mu=3}=0} \tag*{(23)} Γ α ( p ) ( 2 π ) 4 δ 4 ( p + q ) = δ α ( p ) δ α ( q ) δ 2 Γ σ = σ 0 , a 3 3 = a 0 , π = ρ μ = a μ = 3 = a μ = 3 i = 0 (23)
where p p p and q q q represent the four-momenta. Therefore, for small p 2 p^2 p 2 , Γ α ( p ) \Gamma_\alpha(p) Γ α ( p ) behaves like
Γ α ( p ) = − Δ α − 1 ( p ) ∝ m α 2 − p 2 + ⋯ (24) \Gamma_\alpha(p)=-\Delta_\alpha^{-1}(p)\propto m_\alpha^2-p^2+\cdots \tag*{(24)} Γ α ( p ) = − Δ α − 1 ( p ) ∝ m α 2 − p 2 + ⋯ (24)
where m α m_\alpha m α means the meson mass.
Going back to §2, two-point vertex functions are obtained as
Γ σ ( p ) = − 2 G S − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 ) 1 k̸ + p̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] (25) \Gamma_\sigma(p)=-\frac{2}{G_S}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\times(-2)\frac{1}{\not{k}+\not{p}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right] \tag*{(25)} Γ σ ( p ) = − G S 2 − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 ) k + p − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] (25)
Γ π , i ( p ) = − 2 G S − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 i γ 5 τ i ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 i γ 5 τ i ) 1 k̸ + p̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] (26) \Gamma_{\pi,i}(p)=-\frac{2}{G_S}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2i\gamma_5\tau_i)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\times(-2i\gamma_5\tau_i)\frac{1}{\not{k}+\not{p}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right] \tag*{(26)} Γ π , i ( p ) = − G S 2 − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 i γ 5 τ i ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 i γ 5 τ i ) k + p − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] (26)
Γ ρ , i μ ν ( p ) = − 2 G T g μ ν − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 γ μ τ i ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 γ ν τ i ) 1 k̸ + p̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] (27) \Gamma_{\rho,i}^{\mu\nu}(p)=-\frac{2}{G^T}g^{\mu\nu}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2\gamma^\mu\tau_i)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\times(-2\gamma^\nu\tau_i)\frac{1}{\not{k}+\not{p}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right] \tag*{(27)} Γ ρ , i μν ( p ) = − G T 2 g μν − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 γ μ τ i ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 γ ν τ i ) k + p − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] (27)
Γ a , i μ ν ( p ) = − 2 G T g μ ν − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 γ 5 γ μ τ i ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 γ 5 γ ν τ i ) 1 k̸ + p̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] (28) \Gamma_{a,i}^{\mu\nu}(p)=-\frac{2}{G^T}g^{\mu\nu}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2\gamma_5\gamma^\mu\tau_i)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\times(-2\gamma_5\gamma^\nu\tau_i)\frac{1}{\not{k}+\not{p}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right] \tag*{(28)} Γ a , i μν ( p ) = − G T 2 g μν − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 γ 5 γ μ τ i ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 γ 5 γ ν τ i ) k + p − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] (28)
where k̸ = γ μ k μ \not{k}=\gamma^\mu k_\mu k = γ μ k μ and so on.
The case of σ s ( = 2 σ 0 ) ≠ 0 \sigma_s(=2\sigma_0)\ne0 σ s ( = 2 σ 0 ) = 0 and σ p v ( = 2 a 0 ) = 0 \sigma_{pv}(=2a_0)=0 σ p v ( = 2 a 0 ) = 0
The chiral symmetry breaking usually occurs due to the chiral condensate ⟨ ψ ˉ ψ ⟩ = − σ 0 / G S \langle\bar{\psi}\psi\rangle=-\sigma_0/G_S ⟨ ψ ˉ ψ ⟩ = − σ 0 / G S . In this subsection, we treat the case of σ 0 ( = σ s / 2 ) ≠ 0 \sigma_0(=\sigma_s/2)\ne0 σ 0 ( = σ s /2 ) = 0 and a 0 ( = σ p v / 2 ) = 0 a_0(=\sigma_{pv}/2)=0 a 0 ( = σ p v /2 ) = 0 . This case is nothing but the dynamical chiral symmetry breaking in the original NJL model. Therefore, as is well known, the NG boson is pion. Here, for the caution’s sake, let us review its treatment. From Eq. (15) , the gap equation for σ 0 ≠ 0 \sigma_0\ne0 σ 0 = 0 and a 0 = 0 a_0=0 a 0 = 0 is as follows:
1 + 48 G S ∫ d 4 k i ( 2 π ) 4 1 k 2 − 4 σ 0 2 = 0. (29) 1+48G_S\int\frac{d^4k}{i(2\pi)^4}\frac{1}{k^2-4\sigma_0^2}=0 . \tag*{(29)} 1 + 48 G S ∫ i ( 2 π ) 4 d 4 k k 2 − 4 σ 0 2 1 = 0. (29)
Then, the inverse propagator in Eq. (26) can be calculated as
Γ π , i ( p → 0 ) = − 2 G S − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 i γ 5 τ i ) 1 k̸ − 2 σ 0 ( − 2 i γ 5 τ i ) 1 k̸ − 2 σ 0 ] = − 2 G S + 4 ∫ d 4 k i ( 2 π ) 4 tr [ 1 − k̸ − 2 σ 0 ⋅ 1 k̸ − 2 σ 0 ] = − 2 G S − 96 ∫ d 4 k i ( 2 π ) 4 1 k 2 − 4 σ 0 2 = 0. (30) \begin{aligned}
\Gamma_{\pi,i}(p\to0)&=-\frac{2}{G_S}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2i\gamma_5\tau_i)\frac{1}{\not{k}-2\sigma_0}(-2i\gamma_5\tau_i)\frac{1}{\not{k}-2\sigma_0}\right]\\
&=-\frac{2}{G_S}+4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\frac{1}{-\not{k}-2\sigma_0}\cdot\frac{1}{\not{k}-2\sigma_0}\right]\\
&=-\frac{2}{G_S}-96\int\frac{d^4k}{i(2\pi)^4}\frac{1}{k^2-4\sigma_0^2}\\
&=0 . \tag*{(30)}
\end{aligned} Γ π , i ( p → 0 ) = − G S 2 − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 i γ 5 τ i ) k − 2 σ 0 1 ( − 2 i γ 5 τ i ) k − 2 σ 0 1 ] = − G S 2 + 4 ∫ i ( 2 π ) 4 d 4 k tr [ − k − 2 σ 0 1 ⋅ k − 2 σ 0 1 ] = − G S 2 − 96 ∫ i ( 2 π ) 4 d 4 k k 2 − 4 σ 0 2 1 = 0. (30)
where the gap equation (29) is used in the last line in (30). Thus, from Eq. (24) with p → 0 p\to0 p → 0 , all pion masses with any isospin component i i i are equal to zero. Therefore, the pions are the NG bosons.
The case of σ s ( = 2 σ 0 ) = 0 \sigma_s(=2\sigma_0)=0 σ s ( = 2 σ 0 ) = 0 and σ p v ( = 2 a 0 ) ≠ 0 \sigma_{pv}(=2a_0)\ne0 σ p v ( = 2 a 0 ) = 0
Next, let us consider the case with σ 0 = 0 \sigma_0=0 σ 0 = 0 and a 0 ≠ 0 a_0\ne0 a 0 = 0 . In this case, the gap equation in Eq. (15) can be expressed as
1 − 48 G ∫ d 4 k i ( 2 π ) 4 k 2 − 4 a 0 2 + 2 k 3 2 ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 = 0. (31) 1-48G\int\frac{d^4k}{i(2\pi)^4}\frac{k^2-4a_0^2+2k_3^2}{(k^2-4a_0^2)^2-16a_0^2k_3^2}=0 . \tag*{(31)} 1 − 48 G ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 k 2 − 4 a 0 2 + 2 k 3 2 = 0. (31)
From Eqs. (27) and (28) , the inverse propagator for rho meson with μ = 3 \mu=3 μ = 3 and i ≠ 3 i\ne3 i = 3 can be rewritten as
Γ ρ , i ≠ 3 33 ( p → 0 ) = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 ] . (32) \Gamma_{\rho,i\ne3}^{33}(p\to0)=\frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^3\tau_i\frac{1}{\not{k}-2a_0\gamma_5\gamma^3\tau_3}\gamma^3\tau_i\frac{1}{\not{k}-2a_0\gamma_5\gamma^3\tau_3}\right] . \tag*{(32)} Γ ρ , i = 3 33 ( p → 0 ) = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 ] . (32)
Similarly, from Eq. (28) , the inverse propagator for a meson with μ = 3 \mu= 3 μ = 3 and i ≠ 3 i \ne3 i = 3 can be rewritten as
Γ a , i ≠ 3 33 ( p → 0 ) = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ γ 5 γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 γ 5 γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 ] = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 γ 3 τ i 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 ] , (33) \begin{aligned}
\Gamma_{a,i\ne3}^{33}(p \to0) &= \frac{2}{G} - 4 \int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}} \operatorname{tr}\left[\gamma_{5}\gamma^{3}\tau_{i}\frac{1}{\not{k}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\gamma_{5}\gamma^{3}\tau_{i}\frac{1}{\not{k}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\right] \\
&= \frac{2}{G} - 4 \int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}} \operatorname{tr}\left[\gamma^{3}\tau_{i}\frac{1}{\not{k}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\gamma^{3}\tau_{i}\frac{1}{\not{k}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\right], \tag*{(33)}
\end{aligned} Γ a , i = 3 33 ( p → 0 ) = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ γ 5 γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 γ 5 γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 ] = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 γ 3 τ i k − 2 a 0 γ 5 γ 3 τ 3 1 ] , (33)
where we used the anticommutation relation { γ μ , γ 5 } = 0 \{\gamma^{\mu},\gamma_{5}\} = 0 { γ μ , γ 5 } = 0 and γ 5 2 = 1 \gamma_{5}^{2} = 1 γ 5 2 = 1 . Thus, the same expressions for the inverse propagators of rho meson with μ = 3 \mu= 3 μ = 3 and i ≠ 3 i \ne3 i = 3 and a a a meson with μ = 3 \mu= 3 μ = 3 and i ≠ 3 i \ne3 i = 3 are obtained. After some calculations described in Appendix A, we finally obtain Γ ρ , i ≠ 3 33 ( p → 0 ) \Gamma_{\rho,i\ne3}^{33}(p \to0) Γ ρ , i = 3 33 ( p → 0 ) and Γ a , i ≠ 3 33 ( p → 0 ) \Gamma_{a,i\ne3}^{33}(p \to0) Γ a , i = 3 33 ( p → 0 ) as
Γ ρ / a , i ≠ 3 33 ( p → 0 ) = 2 G − 96 ∫ d 4 k i ( 2 π ) 4 k 2 − 4 a 0 2 + 2 k 3 2 ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 . (34) \Gamma_{\rho/a,i\ne3}^{33}(p \to0) = \frac{2}{G} - 96 \int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}} \frac{k^{2} - 4a_{0}^{2} + 2k_{3}^{2}}{(k^{2} - 4a_{0}^{2})^{2} - 16a_{0}^{2}k_{3}^{2}} . \tag*{(34)} Γ ρ / a , i = 3 33 ( p → 0 ) = G 2 − 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 k 2 − 4 a 0 2 + 2 k 3 2 . (34)
By using the gap equation in Eq. (31) ,
Γ ρ / a , i ≠ 3 33 ( p → 0 ) = 0 (35) \Gamma_{\rho/a,i\ne3}^{33}(p \to0) = 0 \tag*{(35)} Γ ρ / a , i = 3 33 ( p → 0 ) = 0 (35)
is obtained. Because Γ ρ / a , i ≠ 3 33 ( p ) \Gamma_{\rho/a,i\ne3}^{33}(p) Γ ρ / a , i = 3 33 ( p ) corresponds to the inverse propagator p 2 − m ρ / a 2 p^{2} - m_{\rho/a}^{2} p 2 − m ρ / a 2 , thus ρ i = 1 o r 2 3 \rho_{i=1\ \mathrm{or}\ 2}^{3} ρ i = 1 or 2 3 and a i = 1 o r 2 3 a_{i=1\ \mathrm{or}\ 2}^{3} a i = 1 or 2 3 mesons have no masses. Therefore, these mesons are the massless bosons. We can check that the other mesons have non-zero masses.
The case of σ s ( = 2 σ 0 ) ≠ 0 \sigma_{s}(= 2\sigma_{0}) \ne0 σ s ( = 2 σ 0 ) = 0 and σ p v ( = 2 a 0 ) ≠ 0 \sigma_{pv}(= 2a_{0}) \ne0 σ p v ( = 2 a 0 ) = 0
In the case of σ 0 ≠ 0 \sigma_{0} \ne0 σ 0 = 0 and a 0 ≠ 0 a_{0} \ne0 a 0 = 0 , the gap equations are nothing but Eq. (15) :
1 + 48 G S ∫ d 4 k i ( 2 π ) 4 k 2 − 4 σ 0 2 + 4 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = 0 , 1 − 48 G ∫ d 4 k i ( 2 π ) 4 k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = 0. (36) \begin{aligned}
1 + 48G_{S}\int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}}\frac{k^{2} - 4\sigma_{0}^{2} + 4a_{0}^{2}}{(k^{2} - 4\sigma_{0}^{2} - 4a_{0}^{2})^{2} - 16a_{0}^{2}(k_{3}^{2} + 4\sigma_{0}^{2})} &= 0, \\
1 - 48G\int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}}\frac{k^{2} + 4\sigma_{0}^{2} - 4a_{0}^{2} + 2k_{3}^{2}}{(k^{2} - 4\sigma_{0}^{2} - 4a_{0}^{2})^{2} - 16a_{0}^{2}(k_{3}^{2} + 4\sigma_{0}^{2})} &= 0 . \tag*{(36)}
\end{aligned} 1 + 48 G S ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 − 4 σ 0 2 + 4 a 0 2 1 − 48 G ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 = 0 , = 0. (36)
First, let us consider the pion inverse propagator in Eq. (26) . Since τ 3 \tau_{3} τ 3 is already included in Γ π , i ( p ) \Gamma_{\pi,i}(p) Γ π , i ( p ) , let us consider Γ π , i = 3 ( p ) \Gamma_{\pi,i=3}(p) Γ π , i = 3 ( p ) . Then, we obtain
Γ π , i = 3 ( p → 0 ) = − 2 G S + 4 ∫ d 4 k i ( 2 π ) 4 tr [ γ 5 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 γ 5 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] , (37) \begin{aligned}
\Gamma_{\pi,i=3}(p \to0)
&= -\frac{2}{G_{S}} + 4 \int\frac{\mathrm{d}^{4}k}{i(2\pi)^{4}}\operatorname{tr}\left[\gamma_{5}\frac{1}{\not{k}-2\sigma_{0}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\gamma_{5}\frac{1}{\not{k}-2\sigma_{0}-2a_{0}\gamma_{5}\gamma^{3}\tau_{3}}\right], \tag*{(37)}
\end{aligned} Γ π , i = 3 ( p → 0 ) = − G S 2 + 4 ∫ i ( 2 π ) 4 d 4 k tr [ γ 5 k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 γ 5 k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] , (37)
where τ 3 2 = 1 \tau_3^2=1 τ 3 2 = 1 is used. The detailed calculation is described in Appendix A. After some calculations, we easily obtain Γ π , i = 3 ( p → 0 ) \Gamma_{\pi,i=3}(p\to0) Γ π , i = 3 ( p → 0 ) as
Γ π , i = 3 ( p → 0 ) = − 2 G S − 96 ∫ d 4 x i ( 2 π ) 4 k 2 − 4 σ 0 2 + 4 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 ( k 3 2 + 4 σ 0 2 ) = 0 , (38) \Gamma_{\pi,i=3}(p\to0)=-\frac{2}{G_S}-96\int\frac{d^4x}{i(2\pi)^4}\frac{k^2-4\sigma_0^2+4a_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0(k_3^2+4\sigma_0^2)}
=0, \tag*{(38)} Γ π , i = 3 ( p → 0 ) = − G S 2 − 96 ∫ i ( 2 π ) 4 d 4 x ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 ( k 3 2 + 4 σ 0 2 ) k 2 − 4 σ 0 2 + 4 a 0 2 = 0 , (38)
where we used the first gap equation in Eq. (36) . Thus, the third component of pion, π 3 \pi_3 π 3 , is the massless boson.
Secondly, let us consider the inverse propagator of a meson with μ = 3 \mu=3 μ = 3 and i ≠ 3 i\ne3 i = 3 in Eq. (28) . By using the anticommutation relations for the Dirac and the isospin matrices, we first obtain Γ a , i ≠ 3 33 ( p → 0 ) \Gamma^{33}_{a,i\ne3}(p\to0) Γ a , i = 3 33 ( p → 0 ) as
\Gamma^{33}_{a,i\ne3}(p\to0)=\frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^3\frac{1}{\slashed{k}+2\sigma_0+2a_0\gamma_5\gamma^3\tau_3}\gamma^3\frac{1}{\slashed{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right] \tag*{(39)}
After tedious calculation described in Appendix A, finally, we obtain the inverse propagator Γ a , i ≠ 3 33 ( p → 0 ) \Gamma^{33}_{a,i\ne3}(p\to0) Γ a , i = 3 33 ( p → 0 ) as
Γ a , i ≠ 3 33 ( p → 0 ) = 2 G − 96 ∫ d 4 k i ( 2 π ) 4 k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 ( k 3 2 + 4 σ 0 2 ) = 0 , (40) \Gamma^{33}_{a,i\ne3}(p\to0)=\frac{2}{G}-96\int\frac{d^4k}{i(2\pi)^4}\frac{k^2+4\sigma_0^2-4a_0^2+2k_3^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0(k_3^2+4\sigma_0^2)}
=0, \tag*{(40)} Γ a , i = 3 33 ( p → 0 ) = G 2 − 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 ( k 3 2 + 4 σ 0 2 ) k 2 + 4 σ 0 2 − 4 a 0 2 + 2 k 3 2 = 0 , (40)
where we used the second gap equation for a 0 a_0 a 0 in Eq. (36) . Thus, a a a mesons with μ = 3 \mu=3 μ = 3 and i = 1 , 2 i=1,2 i = 1 , 2 , a i = 1 o r 2 3 a^3_{i=1\ \mathrm{or}\ 2} a i = 1 or 2 3 , are the massless bosons adding to π 3 \pi_3 π 3 .
Now, we have three NG modes. As will be discussed in section 6, in the case with both the chiral condensate and the pseudovector condensate, five generators of the chiral s u V ( 2 ) × s u A ( 2 ) su_V(2)\times su_A(2) s u V ( 2 ) × s u A ( 2 ) symmetry should be broken. Thus, two NG modes are missing. Here, let us show that the pion mode, π 1 \pi_1 π 1 and the rho meson mode ρ 2 μ = 3 \rho^{\mu=3}_2 ρ 2 μ = 3 mix. Also, it will be shown that the pion mode, π 2 \pi_2 π 2 and the rho meson mode ρ 1 μ = 3 \rho^{\mu=3}_1 ρ 1 μ = 3 mix.
Here, the two-point vertex function in Eq. (23) is rewritten as
Γ α β ( p ) ( 2 π ) 4 δ 4 ( p + q ) = δ 2 Γ δ α ( p ) δ β ( q ) ∣ σ = σ 0 , a 3 3 = a 0 , π = ρ μ = a μ ≠ 3 = a μ = 3 i = 0 , (41) \Gamma_{\alpha\beta}(p)(2\pi)^4\delta^4(p+q)=\left.\frac{\delta^2\Gamma}{\delta\alpha(p)\delta\beta(q)}\right|_{\sigma=\sigma_0,a_3^3=a_0,\pi=\rho_\mu=a_{\mu\ne3}=a^i_{\mu=3}=0}, \tag*{(41)} Γ α β ( p ) ( 2 π ) 4 δ 4 ( p + q ) = δ α ( p ) δ β ( q ) δ 2 Γ σ = σ 0 , a 3 3 = a 0 , π = ρ μ = a μ = 3 = a μ = 3 i = 0 , (41)
where Δ α − 1 ( p ) = − Γ α α ( p ) \Delta^{-1}_{\alpha}(p)=-\Gamma_{\alpha\alpha}(p) Δ α − 1 ( p ) = − Γ α α ( p ) . Then, the two-point vertex functions for α \alpha α or β = π 1 , π 2 , ρ 1 μ = 3 , ρ 2 μ = 3 \beta=\pi_1,\pi_2,\rho^{\mu=3}_1,\rho^{\mu=3}_2 β = π 1 , π 2 , ρ 1 μ = 3 , ρ 2 μ = 3 are similarly calculated as
Γ π 1 π 1 ( p → 0 ) = Γ π 2 π 2 ( p → 0 ) = − 2 G S − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 i γ 5 τ 1 / 2 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 i γ 5 τ 1 / 2 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = − 2 G S − 96 ∫ d 4 k i ( 2 π ) 4 k 2 − 4 σ 0 2 − 4 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = 96 ∫ d 4 k i ( 2 π ) 4 8 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) . (42) \begin{aligned}
\Gamma_{\pi_1\pi_1}(p\to0)&=\Gamma_{\pi_2\pi_2}(p\to0)\\
&=-\frac{2}{G_S}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2i\gamma_5\tau_{1/2})\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right.\\
&\qquad\left.\times(-2i\gamma_5\tau_{1/2})\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right]\\
&=-\frac{2}{G_S}-96\int\frac{d^4k}{i(2\pi)^4}\frac{k^2-4\sigma_0^2-4a_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\\
&=96\int\frac{d^4k}{i(2\pi)^4}\frac{8a_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}. \tag*{(42)}
\end{aligned} Γ π 1 π 1 ( p → 0 ) = Γ π 2 π 2 ( p → 0 ) = − G S 2 − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 i γ 5 τ 1/2 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 i γ 5 τ 1/2 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] = − G S 2 − 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 − 4 σ 0 2 − 4 a 0 2 = 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) 8 a 0 2 . (42)
Here, the gap equation (36) is used to obtain the last equality. Similarly, we obtain Γ ρ 2 μ = 3 ρ 2 μ = 3 \Gamma_{\rho_2^{\mu=3}\rho_2^{\mu=3}} Γ ρ 2 μ = 3 ρ 2 μ = 3 as
Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) = 2 G − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 γ 3 τ 2 / 1 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 γ 3 τ 2 / 1 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = 2 G − 96 ∫ d 4 k i ( 2 π ) 4 k 2 + 2 k 3 2 − 4 σ 0 2 − 4 a 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = 96 ∫ d 4 k i ( 2 π ) 4 8 σ 0 2 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) . (43) \begin{aligned}
\Gamma_{\rho_2^{\mu=3}\rho_2^{\mu=3}}(p\to0)&=\Gamma_{\rho_1^{\mu=3}\rho_1^{\mu=3}}(p\to0)\\
&=\frac{2}{G}-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2\gamma^3\tau_{2/1})\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right.\\
&\qquad\left.\times(-2\gamma^3\tau_{2/1})\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right]\\
&=\frac{2}{G}-96\int\frac{d^4k}{i(2\pi)^4}\frac{k^2+2k_3^2-4\sigma_0^2-4a_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\\
&=96\int\frac{d^4k}{i(2\pi)^4}\frac{8\sigma_0^2}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}. \tag*{(43)}
\end{aligned} Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) = G 2 − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 γ 3 τ 2/1 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 γ 3 τ 2/1 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] = G 2 − 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) k 2 + 2 k 3 2 − 4 σ 0 2 − 4 a 0 2 = 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) 8 σ 0 2 . (43)
where the gap equation (36) is used to obtain the last equality. It should be noted that, for π 1 / 2 \pi_{1/2} π 1/2 and ρ 2 / 1 \rho_{2/1} ρ 2/1 modes, the off-diagonal elements of the two-point vertex functions exist. From (41) , the off-diagonal elements are calculated, for example α = π a \alpha=\pi_a α = π a and β = ρ b μ = 3 \beta=\rho_b^{\mu=3} β = ρ b μ = 3 , as
Γ π 1 ρ 2 μ = 3 ( p → 0 ) = − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 i γ 5 τ 1 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 γ 3 τ 2 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = 96 ∫ d 4 k i ( 2 π ) 4 − 8 σ 0 a 0 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = Γ ρ 2 μ = 3 π 1 ( p → 0 ) , (44) \begin{aligned}
\Gamma_{\pi_1\rho_2^{\mu=3}}(p\to0) &=-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2i\gamma_5\tau_1)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right.\\
&\qquad\left.\times(-2\gamma^3\tau_2)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right]\\
&=96\int\frac{d^4k}{i(2\pi)^4}\frac{-8\sigma_0a_0}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\\
&=\Gamma_{\rho_2^{\mu=3}\pi_1}(p\to0), \tag*{(44)}
\end{aligned} Γ π 1 ρ 2 μ = 3 ( p → 0 ) = − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 i γ 5 τ 1 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 γ 3 τ 2 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] = 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) − 8 σ 0 a 0 = Γ ρ 2 μ = 3 π 1 ( p → 0 ) , (44)
Γ π 2 ρ 1 μ = 3 ( p → 0 ) = − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( − 2 i γ 5 τ 2 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 × ( − 2 γ 3 τ 1 ) 1 k̸ − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 ] = 96 ∫ d 4 k i ( 2 π ) 4 8 σ 0 a 0 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) = Γ ρ 1 μ = 3 π 2 ( p → 0 ) . (45) \begin{aligned}
\Gamma_{\pi_2\rho_1^{\mu=3}}(p\to0) &=-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-2i\gamma_5\tau_2)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right.\\
&\qquad\left.\times(-2\gamma^3\tau_1)\frac{1}{\not{k}-2\sigma_0-2a_0\gamma_5\gamma^3\tau_3}\right]\\
&=96\int\frac{d^4k}{i(2\pi)^4}\frac{8\sigma_0a_0}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}\\
&=\Gamma_{\rho_1^{\mu=3}\pi_2}(p\to0). \tag*{(45)}
\end{aligned} Γ π 2 ρ 1 μ = 3 ( p → 0 ) = − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( − 2 i γ 5 τ 2 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 × ( − 2 γ 3 τ 1 ) k − 2 σ 0 − 2 a 0 γ 5 γ 3 τ 3 1 ] = 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) 8 σ 0 a 0 = Γ ρ 1 μ = 3 π 2 ( p → 0 ) . (45)
The other off-diagonal elements are equal to zero due to the property of the traceless:
Γ π 1 π 2 ( p → 0 ) = Γ π 2 π 1 ( p → 0 ) = Γ π 1 ρ 1 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 π 1 ( p → 0 ) = Γ ρ 2 μ = 3 π 2 ( p → 0 ) = Γ π 2 ρ 2 μ = 3 ( p → 0 ) = Γ ρ 2 μ = 3 ρ 1 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 ρ 2 μ = 3 ( p → 0 ) = 0. (46) \begin{aligned}
\Gamma_{\pi_1\pi_2}(p\to0)=\Gamma_{\pi_2\pi_1}(p\to0)=\Gamma_{\pi_1\rho_1^{\mu=3}}(p\to0)=\Gamma_{\rho_1^{\mu=3}\pi_1}(p\to0) \\
&=\Gamma_{\rho_2^{\mu=3}\pi_2}(p\to0)=\Gamma_{\pi_2\rho_2^{\mu=3}}(p\to0)=\Gamma_{\rho_2^{\mu=3}\rho_1^{\mu=3}}(p\to0)=\Gamma_{\rho_1^{\mu=3}\rho_2^{\mu=3}}(p\to0) \\
&=0. \tag*{(46)}
\end{aligned} Γ π 1 π 2 ( p → 0 ) = Γ π 2 π 1 ( p → 0 ) = Γ π 1 ρ 1 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 π 1 ( p → 0 ) = Γ ρ 2 μ = 3 π 2 ( p → 0 ) = Γ π 2 ρ 2 μ = 3 ( p → 0 ) = Γ ρ 2 μ = 3 ρ 1 μ = 3 ( p → 0 ) = Γ ρ 1 μ = 3 ρ 2 μ = 3 ( p → 0 ) = 0. (46)
Thus, the mass matrix derived from the two-point vertex functions is written as
( Γ π 1 π 1 ( p → 0 ) Γ π 1 ρ 2 μ = 3 ( p → 0 ) Γ π 1 π 2 ( p → 0 ) Γ π 1 ρ 1 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 π 1 ( p → 0 ) Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 π 2 ( p → 0 ) Γ ρ 2 μ = 3 ρ 1 μ = 3 ( p → 0 ) Γ π 2 π 1 ( p → 0 ) Γ π 2 ρ 2 μ = 3 ( p → 0 ) Γ π 2 π 2 ( p → 0 ) Γ π 2 ρ 1 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 π 1 ( p → 0 ) Γ ρ 1 μ = 3 ρ 2 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 π 2 ( p → 0 ) Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) ) = K ( 8 a 0 2 − 8 σ 0 a 0 0 0 − 8 σ 0 a 0 8 σ 0 2 0 0 0 0 8 a 0 2 8 σ 0 a 0 0 0 8 σ 0 a 0 8 σ 0 2 ) , (47) \begin{aligned}
&\begin{pmatrix}
\Gamma_{\pi_1\pi_1}(p\to0) & \Gamma_{\pi_1\rho_2^{\mu=3}}(p\to0) & \Gamma_{\pi_1\pi_2}(p\to0) & \Gamma_{\pi_1\rho_1^{\mu=3}}(p\to0)\\
\Gamma_{\rho_2^{\mu=3}\pi_1}(p\to0) & \Gamma_{\rho_2^{\mu=3}\rho_2^{\mu=3}}(p\to0) & \Gamma_{\rho_2^{\mu=3}\pi_2}(p\to0) & \Gamma_{\rho_2^{\mu=3}\rho_1^{\mu=3}}(p\to0)\\
\Gamma_{\pi_2\pi_1}(p\to0) & \Gamma_{\pi_2\rho_2^{\mu=3}}(p\to0) & \Gamma_{\pi_2\pi_2}(p\to0) & \Gamma_{\pi_2\rho_1^{\mu=3}}(p\to0)\\
\Gamma_{\rho_1^{\mu=3}\pi_1}(p\to0) & \Gamma_{\rho_1^{\mu=3}\rho_2^{\mu=3}}(p\to0) & \Gamma_{\rho_1^{\mu=3}\pi_2}(p\to0) & \Gamma_{\rho_1^{\mu=3}\rho_1^{\mu=3}}(p\to0)
\end{pmatrix}\\
&=K\begin{pmatrix}
8a_0^2 & -8\sigma_0a_0 & 0 & 0\\
-8\sigma_0a_0 & 8\sigma_0^2 & 0 & 0\\
0 & 0 & 8a_0^2 & 8\sigma_0a_0\\
0 & 0 & 8\sigma_0a_0 & 8\sigma_0^2
\end{pmatrix}, \tag*{(47)}
\end{aligned} Γ π 1 π 1 ( p → 0 ) Γ ρ 2 μ = 3 π 1 ( p → 0 ) Γ π 2 π 1 ( p → 0 ) Γ ρ 1 μ = 3 π 1 ( p → 0 ) Γ π 1 ρ 2 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) Γ π 2 ρ 2 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 ρ 2 μ = 3 ( p → 0 ) Γ π 1 π 2 ( p → 0 ) Γ ρ 2 μ = 3 π 2 ( p → 0 ) Γ π 2 π 2 ( p → 0 ) Γ ρ 1 μ = 3 π 2 ( p → 0 ) Γ π 1 ρ 1 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 ρ 1 μ = 3 ( p → 0 ) Γ π 2 ρ 1 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) = K 8 a 0 2 − 8 σ 0 a 0 0 0 − 8 σ 0 a 0 8 σ 0 2 0 0 0 0 8 a 0 2 8 σ 0 a 0 0 0 8 σ 0 a 0 8 σ 0 2 , (47)
where
K = 96 ∫ d 4 k i ( 2 π ) 4 1 ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) . (48) K=96\int\frac{d^4k}{i(2\pi)^4}\frac{1}{(k^2-4\sigma_0^2-4a_0^2)^2-16a_0^2(k_3^2+4\sigma_0^2)}. \tag*{(48)} K = 96 ∫ i ( 2 π ) 4 d 4 k ( k 2 − 4 σ 0 2 − 4 a 0 2 ) 2 − 16 a 0 2 ( k 3 2 + 4 σ 0 2 ) 1 . (48)
It is seen from Eq. (47) that the mode mixing occurs between π 1 \pi_1 π 1 and ρ 2 \rho_2 ρ 2 and/or π 2 \pi_2 π 2 and ρ 1 \rho_1 ρ 1 , respectively, because the mass matrix becomes to the block diagonal.
As fo the mixing between π 1 \pi_1 π 1 and ρ 2 \rho_2 ρ 2 , namely
M π 1 ρ 2 2 μ = 3 = ( Γ π 1 π 1 ( p → 0 ) Γ π 1 ρ 2 μ = 3 ( p → 0 ) Γ ρ 2 π 1 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) ) = K ( 8 a 0 2 − 8 σ 0 a 0 − 8 σ 0 a 0 8 σ 0 2 ) (49) M_{\pi_1\rho_2}^{2\mu=3} =
\begin{pmatrix}
\Gamma_{\pi_1\pi_1}(p\to0) & \Gamma_{\pi_1\rho_2}^{\mu=3}(p\to0) \\
\Gamma_{\rho_2\pi_1}^{\mu=3}(p\to0) & \Gamma_{\rho_2^{\mu=3}\rho_2^{\mu=3}}(p\to0)
\end{pmatrix}
=K
\begin{pmatrix}
8a_0^2 & -8\sigma_0a_0 \\
-8\sigma_0a_0 & 8\sigma_0^2
\end{pmatrix} \tag*{(49)} M π 1 ρ 2 2 μ = 3 = ( Γ π 1 π 1 ( p → 0 ) Γ ρ 2 π 1 μ = 3 ( p → 0 ) Γ π 1 ρ 2 μ = 3 ( p → 0 ) Γ ρ 2 μ = 3 ρ 2 μ = 3 ( p → 0 ) ) = K ( 8 a 0 2 − 8 σ 0 a 0 − 8 σ 0 a 0 8 σ 0 2 ) (49)
we can diagonalize the above mass matrix by using the unitary matrix U U U as
U = 1 σ 0 2 + a 0 2 ( σ 0 a 0 a 0 − σ 0 ) , U − 1 = U † = 1 σ 0 2 + a 0 2 ( σ 0 a 0 a 0 − σ 0 ) , U=\frac{1}{\sqrt{\sigma_0^2+a_0^2}}
\begin{pmatrix}
\sigma_0 & a_0 \\
a_0 & -\sigma_0
\end{pmatrix},
\qquad
U^{-1}=U^\dagger=\frac{1}{\sqrt{\sigma_0^2+a_0^2}}
\begin{pmatrix}
\sigma_0 & a_0 \\
a_0 & -\sigma_0
\end{pmatrix}, U = σ 0 2 + a 0 2 1 ( σ 0 a 0 a 0 − σ 0 ) , U − 1 = U † = σ 0 2 + a 0 2 1 ( σ 0 a 0 a 0 − σ 0 ) ,
U M π 1 ρ 2 2 μ = 3 U − 1 = ( 0 0 0 8 K ( σ 0 2 + a 0 2 ) ) . (50) UM_{\pi_1\rho_2}^{2\mu=3}U^{-1}=
\begin{pmatrix}
0 & 0 \\
0 & 8K(\sigma_0^2+a_0^2)
\end{pmatrix}. \tag*{(50)} U M π 1 ρ 2 2 μ = 3 U − 1 = ( 0 0 0 8 K ( σ 0 2 + a 0 2 ) ) . (50)
Thus, we conclude that the ( σ 0 π 1 + a 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 (\sigma_0\pi_1+a_0\rho_2^{\mu=3})/\sqrt{\sigma_0^2+a_0^2} ( σ 0 π 1 + a 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 is the massless NG mode and ( a 0 π 1 − σ 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 (a_0\pi_1-\sigma_0\rho_2^{\mu=3})/\sqrt{\sigma_0^2+a_0^2} ( a 0 π 1 − σ 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 is the massive mode with the mass 8 K ( σ 0 2 + a 0 2 ) \sqrt{8K(\sigma_0^2+a_0^2)} 8 K ( σ 0 2 + a 0 2 ) .
Secondly, as fo the mixing between π 2 \pi_2 π 2 and ρ 1 \rho_1 ρ 1 , namely
M π 2 ρ 1 2 μ = 3 = ( Γ π 2 π 2 ( p → 0 ) Γ π 2 ρ 1 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 π 2 ( p → 0 ) Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) ) = K ( 8 a 0 2 8 σ 0 a 0 8 σ 0 a 0 8 σ 0 2 ) , (51) M_{\pi_2\rho_1}^{2\mu=3}=
\begin{pmatrix}
\Gamma_{\pi_2\pi_2}(p\to0) & \Gamma_{\pi_2\rho_1}^{\mu=3}(p\to0) \\
\Gamma_{\rho_1^{\mu=3}\pi_2}(p\to0) & \Gamma_{\rho_1^{\mu=3}\rho_1^{\mu=3}}(p\to0)
\end{pmatrix}
=K
\begin{pmatrix}
8a_0^2 & 8\sigma_0a_0 \\
8\sigma_0a_0 & 8\sigma_0^2
\end{pmatrix}, \tag*{(51)} M π 2 ρ 1 2 μ = 3 = ( Γ π 2 π 2 ( p → 0 ) Γ ρ 1 μ = 3 π 2 ( p → 0 ) Γ π 2 ρ 1 μ = 3 ( p → 0 ) Γ ρ 1 μ = 3 ρ 1 μ = 3 ( p → 0 ) ) = K ( 8 a 0 2 8 σ 0 a 0 8 σ 0 a 0 8 σ 0 2 ) , (51)
the mass matrix can be diagonalized by using the unitary matrix U U U as
U = 1 σ 0 2 + a 0 2 ( a 0 σ 0 σ 0 − a 0 ) , U − 1 = U † = 1 σ 0 2 + a 0 2 ( a 0 σ 0 σ 0 − a 0 ) , U=\frac{1}{\sqrt{\sigma_0^2+a_0^2}}
\begin{pmatrix}
a_0 & \sigma_0 \\
\sigma_0 & -a_0
\end{pmatrix},
\qquad
U^{-1}=U^\dagger=\frac{1}{\sqrt{\sigma_0^2+a_0^2}}
\begin{pmatrix}
a_0 & \sigma_0 \\
\sigma_0 & -a_0
\end{pmatrix}, U = σ 0 2 + a 0 2 1 ( a 0 σ 0 σ 0 − a 0 ) , U − 1 = U † = σ 0 2 + a 0 2 1 ( a 0 σ 0 σ 0 − a 0 ) ,
U M π 2 ρ 1 2 μ = 3 U − 1 = ( 8 K ( σ 0 2 + a 0 2 ) 0 0 0 ) . (52) UM_{\pi_2\rho_1}^{2\mu=3}U^{-1}=
\begin{pmatrix}
8K(\sigma_0^2+a_0^2) & 0 \\
0 & 0
\end{pmatrix}. \tag*{(52)} U M π 2 ρ 1 2 μ = 3 U − 1 = ( 8 K ( σ 0 2 + a 0 2 ) 0 0 0 ) . (52)
Thus, we conclude that the ( σ 0 π 2 − a 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 (\sigma_0\pi_2-a_0\rho_1^{\mu=3})/\sqrt{\sigma_0^2+a_0^2} ( σ 0 π 2 − a 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 is the massless NG mode and ( a 0 π 2 + σ 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 (a_0\pi_2+\sigma_0\rho_1^{\mu=3})/\sqrt{\sigma_0^2+a_0^2} ( a 0 π 2 + σ 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 is the massive mode with the mass 8 K ( σ 0 2 + a 0 2 ) \sqrt{8K(\sigma_0^2+a_0^2)} 8 K ( σ 0 2 + a 0 2 ) .
Thus, we have five NG modes in the case of the chiral symmetry breaking due to both the chiral condensate σ 0 \sigma_0 σ 0 ( = ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩ ) (=\langle\sigma\rangle=-G_S\langle\bar{\psi}\psi\rangle) ( = ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩) and the pseudovector condensate a 0 a_0 a 0 ( = ⟨ a 3 μ = 3 ⟩ = − G A ⟨ ψ † Σ 3 ψ ⟩ ) (=\langle a_3^{\mu=3}\rangle=-G_A\langle\psi^\dagger\Sigma_3\psi\rangle) ( = ⟨ a 3 μ = 3 ⟩ = − G A ⟨ ψ † Σ 3 ψ ⟩) , namely
a 1 μ = 3 , a 2 μ = 3 , π 3 , σ 0 π 1 + a 0 ρ 2 μ = 3 σ 0 2 + a 0 2 , σ 0 π 2 − a 0 ρ 1 μ = 3 σ 0 2 + a 0 2 . (53) a_1^{\mu=3},\quad a_2^{\mu=3},\quad\pi_3,\quad
\frac{\sigma_0\pi_1+a_0\rho_2^{\mu=3}}{\sqrt{\sigma_0^2+a_0^2}},\quad
\frac{\sigma_0\pi_2-a_0\rho_1^{\mu=3}}{\sqrt{\sigma_0^2+a_0^2}}. \tag*{(53)} a 1 μ = 3 , a 2 μ = 3 , π 3 , σ 0 2 + a 0 2 σ 0 π 1 + a 0 ρ 2 μ = 3 , σ 0 2 + a 0 2 σ 0 π 2 − a 0 ρ 1 μ = 3 . (53)
Gap equation with respect to the tensor condensate and the massless modes on the tensor condensate
Gap equation for the tensor condensate
Let us introduce the NJL model with tensor interaction, instead of the pseudovector interaction, which retains the chiral symmetry[6 , 15 , 16 , 17 ] :
L N J L , T = ψ ˉ i γ μ ∂ μ ψ + G S [ ( ψ ˉ ψ ) 2 + ( ψ ˉ i γ 5 τ ψ ) 2 ] + G T [ ( ψ ˉ γ μ γ ν τ ψ ) 2 + ( ψ ˉ i γ 5 γ μ γ ν ψ ) 2 ] . (54) \mathcal{L}_{\mathrm{NJL,T}}=\bar{\psi}i\gamma^{\mu}\partial_{\mu}\psi+G_{S}\left[(\bar{\psi}\psi)^{2}+(\bar{\psi}i\gamma_{5}\tau\psi)^{2}\right]+G_{T}\left[(\bar{\psi}\gamma^{\mu}\gamma^{\nu}\tau\psi)^{2}+(\bar{\psi}i\gamma_{5}\gamma^{\mu}\gamma^{\nu}\psi)^{2}\right]. \tag*{(54)} L NJL , T = ψ ˉ i γ μ ∂ μ ψ + G S [ ( ψ ˉ ψ ) 2 + ( ψ ˉ i γ 5 τ ψ ) 2 ] + G T [ ( ψ ˉ γ μ γ ν τ ψ ) 2 + ( ψ ˉ i γ 5 γ μ γ ν ψ ) 2 ] . (54)
In a similar way developed in §2, we use the auxiliary field method and introduce the functional Gaussian integral leading to unit as
1 = ∫ D σ ′ D π ′ exp ( i ∫ d 4 x [ − 1 G S ( σ ′ 2 + π ′ 2 ) ] ) , 1=\int\mathcal{D}\sigma'\mathcal{D}\pi'\exp\left(i\int d^{4}x\left[-\frac{1}{G_{S}}\left(\sigma'^{2}+\pi'^{2}\right)\right]\right), 1 = ∫ D σ ′ D π ′ exp ( i ∫ d 4 x [ − G S 1 ( σ ′ 2 + π ′ 2 ) ] ) ,
1 = ∫ D t ′ μ ν D p ′ μ ν exp ( i ∫ d 4 x [ − 1 G T ( t ′ μ ν ⋅ t μ ν ′ + p ′ μ ν p μ ν ′ ) ] ) . (55) 1=\int\mathcal{D}t'^{\mu\nu}\mathcal{D}p'^{\mu\nu}\exp\left(i\int d^{4}x\left[-\frac{1}{G_{T}\!}\left(t'^{\mu\nu}\cdot t'_{\mu\nu}+p'^{\mu\nu}p'_{\mu\nu}\right)\right]\right). \tag*{(55)} 1 = ∫ D t ′ μν D p ′ μν exp ( i ∫ d 4 x [ − G T 1 ( t ′ μν ⋅ t μν ′ + p ′ μν p μν ′ ) ] ) . (55)
Inserting the above unit into the generating functional Z Z Z in Eq. (2) and replacing σ ′ \sigma' σ ′ , π ′ \pi' π ′ , t ′ μ ν t'^{\mu\nu} t ′ μν and p ′ μ ν p'^{\mu\nu} p ′ μν into σ \sigma σ , π \pi π , t μ ν t^{\mu\nu} t μν and p μ ν p^{\mu\nu} p μν as
σ ′ = σ + G S ( ψ ˉ ψ ) , \sigma'=\sigma+G_{S}(\bar{\psi}\psi), σ ′ = σ + G S ( ψ ˉ ψ ) ,
π ′ = π + G S ( ψ ˉ i γ 5 τ ψ ) , \pi'=\pi+G_{S}(\bar{\psi}i\gamma_{5}\tau\psi), π ′ = π + G S ( ψ ˉ i γ 5 τ ψ ) ,
t ′ μ ν = t μ ν + G T ( ψ ˉ γ μ γ ν τ ψ ) , t'^{\mu\nu}=t^{\mu\nu}+G_{T}(\bar{\psi}\gamma^{\mu}\gamma^{\nu}\tau\psi), t ′ μν = t μν + G T ( ψ ˉ γ μ γ ν τ ψ ) ,
p ′ μ ν = p μ ν + G T ( ψ ˉ i γ 5 γ μ γ ν ψ ) . (56) p'^{\mu\nu}=p^{\mu\nu}+G_{T}(\bar{\psi}i\gamma_{5}\gamma^{\mu}\gamma^{\nu}\psi). \tag*{(56)} p ′ μν = p μν + G T ( ψ ˉ i γ 5 γ μ γ ν ψ ) . (56)
the generating functional is rewritten as
Z = ∫ D ψ D ψ ˉ ∫ D σ D π D t μ ν D p μ ν exp ( i ∫ d 4 x L ~ ) . (57) Z=\int\mathcal{D}\psi\mathcal{D}\bar{\psi}\int\mathcal{D}\sigma\mathcal{D}\pi\mathcal{D}t^{\mu\nu}\mathcal{D}p^{\mu\nu}\exp\left(i\int d^{4}x\widetilde{\mathcal{L}}\right). \tag*{(57)} Z = ∫ D ψ D ψ ˉ ∫ D σ D π D t μν D p μν exp ( i ∫ d 4 x L ) . (57)
where
L ~ = ψ ˉ ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 t μ ν ⋅ γ μ γ ν τ − 2 p μ ν i γ 5 γ μ γ ν ) ψ − 1 G S ( σ 2 + π 2 ) − 1 G T ( t μ ν ⋅ t μ ν + p μ ν p μ ν ) . (58) \widetilde{\mathcal{L}}=\bar{\psi}\left(i\gamma^{\mu}\partial_{\mu}-2\sigma-2\pi\cdot i\gamma_{5}\tau-2t_{\mu\nu}\cdot\gamma^{\mu}\gamma^{\nu}\tau-2p_{\mu\nu}i\gamma_{5}\gamma^{\mu}\gamma^{\nu}\right)\psi-\frac{1}{G_{S}}\left(\sigma^{2}+\pi^{2}\right)-\frac{1}{G_{T}\!}\left(t^{\mu\nu}\cdot t_{\mu\nu}+p^{\mu\nu}p_{\mu\nu}\right). \tag*{(58)} L = ψ ˉ ( i γ μ ∂ μ − 2 σ − 2 π ⋅ i γ 5 τ − 2 t μν ⋅ γ μ γ ν τ − 2 p μν i γ 5 γ μ γ ν ) ψ − G S 1 ( σ 2 + π 2 ) − G T 1 ( t μν ⋅ t μν + p μν p μν ) . (58)
Here, from δ L ~ / δ σ = 0 \delta\widetilde{\mathcal{L}}/\delta\sigma=0 δ L / δ σ = 0 and so on, we obtain
σ = − G S ( ψ ˉ ψ ) , \sigma=-G_{S}(\bar{\psi}\psi), σ = − G S ( ψ ˉ ψ ) ,
π = − G S ( ψ ˉ i γ 5 τ ψ ) , \pi=-G_{S}(\bar{\psi}i\gamma_{5}\tau\psi), π = − G S ( ψ ˉ i γ 5 τ ψ ) ,
t μ ν = − G T ( ψ ˉ γ μ γ ν τ ψ ) , t_{\mu\nu}=-G_{T}(\bar{\psi}\gamma_{\mu}\gamma_{\nu}\tau\psi), t μν = − G T ( ψ ˉ γ μ γ ν τ ψ ) ,
p μ ν = − G T ( ψ ˉ i γ 5 γ μ γ ν ψ ) . (59) p_{\mu\nu}=-G_{T}(\bar{\psi}i\gamma_{5}\gamma_{\mu}\gamma_{\nu}\psi). \tag*{(59)} p μν = − G T ( ψ ˉ i γ 5 γ μ γ ν ψ ) . (59)
Further, integrating ψ \psi ψ and ψ ˉ \bar{\psi} ψ ˉ , the generating functional can be expressed as
\begin{aligned}
Z &= \int\mathcal{D}\sigma\mathcal{D}\pi\mathcal{D}t^{\mu\nu} \mathcal{D}p^{\mu\nu} \exp(i\Gamma),\\
\Gamma&= \int d^{4}x \left[-\frac{1}{G_{S}}(\sigma^{2}+\pi^{2})-\frac{1}{G_{T}(t^{\mu\nu}\cdot t_{\mu\nu}+p^{\mu\nu}p_{\mu\nu})\right.\\
&\qquad\left.-i\ln\operatorname{Det}\left(i\gamma^{\mu}\partial_{\mu}-2\sigma-2\boldsymbol{\pi}\cdot i\gamma_{5}\boldsymbol{\tau}-2t_{\mu\nu}\cdot\gamma^{\mu}\gamma^{\nu}\boldsymbol{\tau}-2p_{\mu\nu}i\gamma_{5}\gamma^{\mu}\gamma^{\nu}\right)\right].
\end{aligned}} \tag*{(60)}\end{aligned}
It should be noted that the tensor mode t 12 t^{12} t 12 is written as
t 12 = − G T ψ ˉ γ 1 γ 2 τ ψ = i G T ψ ˉ Σ 3 τ ψ . (61) \begin{aligned}
t^{12} &= -G_{T}\bar{\psi}\gamma^{1}\gamma^{2}\tau\psi\\
&= iG_{T}\bar{\psi}\Sigma_{3}\tau\psi. \tag*{(61)}
\end{aligned} t 12 = − G T ψ ˉ γ 1 γ 2 τ ψ = i G T ψ ˉ Σ 3 τ ψ . (61)
Here, we use the Dirac representation for the Dirac gamma matrices, namely,
γ 1 γ 2 = − i Σ 3 = − i ( σ 3 0 0 σ 3 ) . (62) \gamma^{1}\gamma^{2}=-i\Sigma_{3}=-i
\begin{pmatrix}
\sigma_{3} & 0\\
0 & \sigma_{3}
\end{pmatrix}. \tag*{(62)} γ 1 γ 2 = − i Σ 3 = − i ( σ 3 0 0 σ 3 ) . (62)
Hereafter, let us assume that there exist the chiral condensate σ 0 = ⟨ σ ⟩ \sigma_{0}=\langle\sigma\rangle σ 0 = ⟨ σ ⟩ and tensor condensate t 0 = ⟨ t i = 3 μ = 1 ν = 2 ⟩ ( = − ⟨ t i = 3 μ = 2 ν = 1 ⟩ ) t_{0}=\langle t^{\mu=1\ \nu=2}_{i=3}\rangle(=-\langle t^{\mu=2\ \nu=1}_{i=3}\rangle) t 0 = ⟨ t i = 3 μ = 1 ν = 2 ⟩ ( = − ⟨ t i = 3 μ = 2 ν = 1 ⟩) , namely
σ 0 ≡ ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩ , \sigma_{0}\equiv\langle\sigma\rangle=-G_{S}\langle\bar{\psi}\psi\rangle, σ 0 ≡ ⟨ σ ⟩ = − G S ⟨ ψ ˉ ψ ⟩ ,
t 0 ≡ ⟨ t μ = 1 μ = 2 i = 3 ⟩ = i G T ⟨ ψ ˉ Σ 3 τ 3 ψ ⟩ . (63) t_{0}\equiv\langle t^{i=3}_{\mu=1\mu=2}\rangle=iG_{T}\langle\bar{\psi}\Sigma_{3}\tau_{3}\psi\rangle. \tag*{(63)} t 0 ≡ ⟨ t μ = 1 μ = 2 i = 3 ⟩ = i G T ⟨ ψ ˉ Σ 3 τ 3 ψ ⟩ . (63)
Under these quark-antiquark condensates, the effective potential V ( σ 0 , t 0 ) V(\sigma_{0},t_{0}) V ( σ 0 , t 0 ) can be derived from the effective action Γ \Gamma Γ as
Γ ( σ = σ 0 , t 3 12 = t 0 ) = − V ( σ 0 , t 0 ) ∫ d 4 x , \Gamma(\sigma=\sigma_{0},t^{12}_{3}=t_{0})=-V(\sigma_{0},t_{0})\int d^{4}x , Γ ( σ = σ 0 , t 3 12 = t 0 ) = − V ( σ 0 , t 0 ) ∫ d 4 x ,
V ( σ 0 , t 0 ) = 1 G S σ 0 2 + 2 G T t 0 2 − ∫ d 4 k i ( 2 π ) 4 ln det [ γ μ k μ − 2 σ 0 − 4 t 0 γ 1 γ 2 τ 3 ] = 1 G S σ 0 2 + 2 G T t 0 2 − ∫ d 4 k i ( 2 π ) 4 tr ln [ γ μ k μ − 2 σ 0 − 4 t 0 γ 1 γ 2 τ 3 ] , (64) \begin{aligned}
V(\sigma_{0},t_{0})&=\frac{1}{G_{S}}\sigma_{0}^{2}+\frac{2}{G_{T}}t_{0}^{2}-\int\frac{d^{4}k}{i(2\pi)^{4}}\ln\det\left[\gamma^{\mu}k_{\mu}-2\sigma_{0}-4t_{0}\gamma^{1}\gamma^{2}\tau_{3}\right]\\
&=\frac{1}{G_{S}}\sigma_{0}^{2}+\frac{2}{G_{T}}t_{0}^{2}-\int\frac{d^{4}k}{i(2\pi)^{4}}\operatorname{tr}\ln\left[\gamma^{\mu}k_{\mu}-2\sigma_{0}-4t_{0}\gamma^{1}\gamma^{2}\tau_{3}\right], \tag*{(64)}
\end{aligned} V ( σ 0 , t 0 ) = G S 1 σ 0 2 + G T 2 t 0 2 − ∫ i ( 2 π ) 4 d 4 k ln det [ γ μ k μ − 2 σ 0 − 4 t 0 γ 1 γ 2 τ 3 ] = G S 1 σ 0 2 + G T 2 t 0 2 − ∫ i ( 2 π ) 4 d 4 k tr ln [ γ μ k μ − 2 σ 0 − 4 t 0 γ 1 γ 2 τ 3 ] , (64)
where trace is taken by the Dirac gamma matrices, isospin space and color space. Here, it should be noted that t i = 3 μ = 2 ν = 1 = − t i = 3 μ = 1 ν = 2 = − t 0 t^{\mu=2\ \nu=1}_{i=3}=-t^{\mu=1\ \nu=2}_{i=3}=-t_{0} t i = 3 μ = 2 ν = 1 = − t i = 3 μ = 1 ν = 2 = − t 0 is used.
Hereafter, we consider the tensor condensate t 0 t_{0} t 0 only because both the tensor condensate and the chiral condensate σ 0 \sigma_{0} σ 0 may not coexist according to our previous work.[14 ] Thus, we set σ 0 = 0 \sigma_0=0 σ 0 = 0 in Eq. (64) . The condensate t 0 t_0 t 0 is determined by the following gap equation:
∂ V ∂ t 0 = 4 G T t 0 − ∫ d 4 k i ( 2 π ) 4 tr [ 1 γ μ k μ − 4 t 0 γ 1 γ 2 τ 3 ( − 4 γ 1 γ 2 τ 3 ) ] = 0 , (65) \frac{\partial V}{\partial t_0}=\frac{4}{G_T}t_0-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\frac{1}{\gamma^\mu k_\mu-4t_0\gamma^1\gamma^2\tau_3}(-4\gamma^1\gamma^2\tau_3)\right]=0, \tag*{(65)} ∂ t 0 ∂ V = G T 4 t 0 − ∫ i ( 2 π ) 4 d 4 k tr [ γ μ k μ − 4 t 0 γ 1 γ 2 τ 3 1 ( − 4 γ 1 γ 2 τ 3 ) ] = 0 , (65)
which, after some tedious calculation for the gamma matrices, leads to
t 0 [ 1 − 32 G T ∫ d 4 k i ( 2 π ) 4 k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) ] = 0 , (66) t_0\left[1-32G_T\int\frac{d^4k}{i(2\pi)^4}\frac{k^2+16t_0^2+2(k_1^2+k_2^2)}{(k^2+16t_0^2)^2+64t_0^2(k_1^2+k_2^2)}\right]=0, \tag*{(66)} t 0 [ 1 − 32 G T ∫ i ( 2 π ) 4 d 4 k ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) ] = 0 , (66)
where k 2 = k 0 2 − k 2 k^2=k_0^2-\mathbf{k}^2 k 2 = k 0 2 − k 2 , and k 1 k_1 k 1 and k 2 k_2 k 2 mean the first and the second components of the momentum vector k \mathbf{k} k .
Massless modes on the tensor condensate t 0 t_0 t 0
Thus, two-point vertex functions for α = t μ ν \alpha=t^{\mu\nu} α = t μν and p μ ν p^{\mu\nu} p μν in Eq. (23) are obtained by using Γ \Gamma Γ in Eq. (60) as
Γ t , i μ ν , ρ σ ( p ) = − 4 G T ( g μ ρ g ν σ − g μ σ g ν ρ ) − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 4 γ μ γ ν τ i ) 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 × ( − 4 γ ρ γ σ τ i ) 1 k̸ + p̸ − 4 t 0 γ 1 γ 2 τ 3 ] , (67) \Gamma_{t,i}^{\mu\nu,\rho\sigma}(p)=-\frac{4}{G_T}(g^{\mu\rho}g^{\nu\sigma}-g^{\mu\sigma}g^{\nu\rho})-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-4\gamma^\mu\gamma^\nu\tau_i)\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\times(-4\gamma^\rho\gamma^\sigma\tau_i)\frac{1}{\not{k}+\not{p}-4t_0\gamma^1\gamma^2\tau_3}\right], \tag*{(67)} Γ t , i μν , ρ σ ( p ) = − G T 4 ( g μ ρ g ν σ − g μ σ g ν ρ ) − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 4 γ μ γ ν τ i ) k − 4 t 0 γ 1 γ 2 τ 3 1 × ( − 4 γ ρ γ σ τ i ) k + p − 4 t 0 γ 1 γ 2 τ 3 1 ] , (67)
Γ p , i μ ν , ρ σ ( p ) = − 4 G T ( g μ ρ g ν σ − g μ σ g σ ρ ) − ∫ d 4 k i ( 2 π ) 4 tr [ ( − 4 i γ 5 γ μ γ ν ) 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 × ( − 4 i γ 5 γ ρ γ σ ) 1 k̸ + p̸ − 4 t 0 γ 1 γ 2 τ 3 ] , (68) \Gamma_{p,i}^{\mu\nu,\rho\sigma}(p)=-\frac{4}{G_T}(g^{\mu\rho}g^{\nu\sigma}-g^{\mu\sigma}g^{\sigma\rho})-\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(-4i\gamma_5\gamma^\mu\gamma^\nu)\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\times(-4i\gamma_5\gamma^\rho\gamma^\sigma)\frac{1}{\not{k}+\not{p}-4t_0\gamma^1\gamma^2\tau_3}\right], \tag*{(68)} Γ p , i μν , ρ σ ( p ) = − G T 4 ( g μ ρ g ν σ − g μ σ g σ ρ ) − ∫ i ( 2 π ) 4 d 4 k tr [ ( − 4 i γ 5 γ μ γ ν ) k − 4 t 0 γ 1 γ 2 τ 3 1 × ( − 4 i γ 5 γ ρ γ σ ) k + p − 4 t 0 γ 1 γ 2 τ 3 1 ] , (68)
where we used the relations ∂ ( t α β t α β ) / ∂ t μ ν = ∂ ( t μ ν t μ ν + t ν μ t μ ν ) / ∂ t μ ν = 4 t μ ν \partial(t^{\alpha\beta}t_{\alpha\beta})/\partial t_{\mu\nu}=\partial(t^{\mu\nu}t_{\mu\nu}+t^{\nu\mu}t_{\mu\nu})/\partial t_{\mu\nu}=4t^{\mu\nu} ∂ ( t α β t α β ) / ∂ t μν = ∂ ( t μν t μν + t ν μ t μν ) / ∂ t μν = 4 t μν and ∂ t μ ν t α β / ∂ t ρ σ = g μ ρ g ν σ − g μ σ g ν ρ \partial t^{\mu\nu}t_{\alpha\beta}/\partial t_{\rho\sigma}=g^{\mu\rho}g^{\nu\sigma}-g^{\mu\sigma}g^{\nu\rho} ∂ t μν t α β / ∂ t ρ σ = g μ ρ g ν σ − g μ σ g ν ρ .
Here, we only consider the case with σ 0 = 0 \sigma_0=0 σ 0 = 0 and t 0 ≠ 0 t_0\ne0 t 0 = 0 . In this case, the gap equation in Eq. (66) can be expressed as
1 − 32 G T ∫ d 4 k i ( 2 π ) 4 k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) = 0. (69) 1-32G_T\int\frac{d^4k}{i(2\pi)^4}\frac{k^2+16t_0^2+2(k_1^2+k_2^2)}{(k^2+16t_0^2)^2+64t_0^2(k_1^2+k_2^2)}=0. \tag*{(69)} 1 − 32 G T ∫ i ( 2 π ) 4 d 4 k ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) = 0. (69)
From Eq. (67) , the inverse propagator with μ = ρ = 1 \mu=\rho=1 μ = ρ = 1 , ν = σ = 2 \nu=\sigma=2 ν = σ = 2 and i ≠ 3 i\ne3 i = 3 can be rewritten as
Γ t , i ≠ 3 12 , 12 ( p → 0 ) = − 4 G T − 16 ∫ d 4 k i ( 2 π ) 4 tr [ γ 1 γ 2 τ i 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 γ 1 γ 2 τ i 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 ] = − 4 G T − 16 ∫ d 4 k i ( 2 π ) 4 tr [ γ 1 γ 2 1 k̸ + 4 t 0 γ 1 γ 2 τ 3 γ 1 γ 2 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 ] , (70) \begin{aligned}
\Gamma_{t,i\ne3}^{12,12}(p\to0)=-\frac{4}{G_T}-16\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^1\gamma^2\tau_i\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\gamma^1\gamma^2\tau_i\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\right] \\
&=-\frac{4}{G_T}-16\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^1\gamma^2\frac{1}{\not{k}+4t_0\gamma^1\gamma^2\tau_3}\gamma^1\gamma^2\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\right], \tag*{(70)}
\end{aligned} Γ t , i = 3 12 , 12 ( p → 0 ) = − G T 4 − 16 ∫ i ( 2 π ) 4 d 4 k tr [ γ 1 γ 2 τ i k − 4 t 0 γ 1 γ 2 τ 3 1 γ 1 γ 2 τ i k − 4 t 0 γ 1 γ 2 τ 3 1 ] = − G T 4 − 16 ∫ i ( 2 π ) 4 d 4 k tr [ γ 1 γ 2 k + 4 t 0 γ 1 γ 2 τ 3 1 γ 1 γ 2 k − 4 t 0 γ 1 γ 2 τ 3 1 ] , (70)
where we used the anticommutation relation { τ i , τ 3 } = 0 \{\tau_i,\tau_3\}=0 { τ i , τ 3 } = 0 with i = 1 , 2 i=1,2 i = 1 , 2 and τ i 2 = 1 \tau_i^2=1 τ i 2 = 1 . Similarly, from Eq. (68) , the inverse propagator with μ = ρ = 1 \mu=\rho=1 μ = ρ = 1 and ν = σ = 2 \nu=\sigma=2 ν = σ = 2 can be rewritten as
Γ p 12 , 12 ( p → 0 ) = − 4 G T + 16 ∫ d 4 k i ( 2 π ) 4 tr [ γ 5 γ 1 γ 2 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 γ 5 γ 1 γ 2 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 ] = − 4 G T − 16 ∫ d 4 k i ( 2 π ) 4 tr [ γ 1 γ 2 1 k̸ + 4 t 0 γ 1 γ 2 τ 3 γ 1 γ 2 1 k̸ − 4 t 0 γ 1 γ 2 τ 3 ] , (71) \begin{aligned}
\Gamma_p^{12,12}(p\to0)={}&-\frac{4}{G_T}+16\int\frac{\mathrm{d}^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^5\gamma^1\gamma^2\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\gamma^5\gamma^1\gamma^2\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\right]\\
={}&-\frac{4}{G_T}-16\int\frac{\mathrm{d}^4k}{i(2\pi)^4}\operatorname{tr}\left[\gamma^1\gamma^2\frac{1}{\not{k}+4t_0\gamma^1\gamma^2\tau_3}\gamma^1\gamma^2\frac{1}{\not{k}-4t_0\gamma^1\gamma^2\tau_3}\right], \tag*{(71)}
\end{aligned} Γ p 12 , 12 ( p → 0 ) = = − G T 4 + 16 ∫ i ( 2 π ) 4 d 4 k tr [ γ 5 γ 1 γ 2 k − 4 t 0 γ 1 γ 2 τ 3 1 γ 5 γ 1 γ 2 k − 4 t 0 γ 1 γ 2 τ 3 1 ] − G T 4 − 16 ∫ i ( 2 π ) 4 d 4 k tr [ γ 1 γ 2 k + 4 t 0 γ 1 γ 2 τ 3 1 γ 1 γ 2 k − 4 t 0 γ 1 γ 2 τ 3 1 ] , (71)
where we used the anticommutation relation { γ μ , γ 5 } = 0 \{\gamma^\mu,\gamma_5\}=0 { γ μ , γ 5 } = 0 and γ 5 2 = 1 \gamma_5^2=1 γ 5 2 = 1 . Thus, the same expression for the inverse propagator with μ = ρ = 1 \mu=\rho=1 μ = ρ = 1 , ν = σ = 2 \nu=\sigma=2 ν = σ = 2 and i ≠ 3 i\ne3 i = 3 for t t t -mode is obtained.
After tedious calculations similar to the case of the pseudovector condensate a 0 ≠ 0 a_0\ne0 a 0 = 0 described in Appendix A, we finally obtain Γ t , i ≠ 3 / p 12 , 12 ( p → 0 ) \Gamma_{t,i\ne3/p}^{12,12}(p\to0) Γ t , i = 3/ p 12 , 12 ( p → 0 ) and Γ p 12 , 12 ( p → 0 ) \Gamma_p^{12,12}(p\to0) Γ p 12 , 12 ( p → 0 ) as
Γ t , i ≠ 3 / p 12 , 12 ( p → 0 ) = − 4 G T + 128 ∫ d 4 k i ( 2 π ) 4 k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) (72) \Gamma_{t,i\ne3/p}^{12,12}(p\to0)=-\frac{4}{G_T}+128\int\frac{\mathrm{d}^4k}{i(2\pi)^4}\frac{k^2+16t_0^2+2(k_1^2+k_2^2)}{(k^2+16t_0^2)^2+64t_0^2(k_1^2+k_2^2)} \tag*{(72)} Γ t , i = 3/ p 12 , 12 ( p → 0 ) = − G T 4 + 128 ∫ i ( 2 π ) 4 d 4 k ( k 2 + 16 t 0 2 ) 2 + 64 t 0 2 ( k 1 2 + k 2 2 ) k 2 + 16 t 0 2 + 2 ( k 1 2 + k 2 2 ) (72)
By using the gap equation in Eq. (69) , we obtain
Γ t , i ≠ 3 / p 12 , 12 ( p → 0 ) = 0. (73) \Gamma_{t,i\ne3/p}^{12,12}(p\to0)=0. \tag*{(73)} Γ t , i = 3/ p 12 , 12 ( p → 0 ) = 0. (73)
Because Γ t , i ≠ 3 / p 12 , 12 ( p ) \Gamma_{t,i\ne3/p}^{12,12}(p) Γ t , i = 3/ p 12 , 12 ( p ) corresponds to the inverse propagator p 2 − m t / p 2 p^2-m_{t/p}^2 p 2 − m t / p 2 , thus the mesons corresponding to the modes t i = 1 o r 2 12 t_{i=1\ \mathrm{or}\ 2}^{12} t i = 1 or 2 12 and p 12 p^{12} p 12 have no masses. Therefore, these mesons correspond to the massless Nambu-Goldstone bosons. Here, from Eq. (59) , p 12 ∝ ψ ˉ i γ 5 γ 1 γ 2 ψ = ψ ˉ γ 0 γ 3 ψ p^{12}\propto\bar{\psi}i\gamma_5\gamma^1\gamma^2\psi=\bar{\psi}\gamma^0\gamma^3\psi p 12 ∝ ψ ˉ i γ 5 γ 1 γ 2 ψ = ψ ˉ γ 0 γ 3 ψ . Thus, p p p -mode corresponds to the isoscalar-tensor mode with μ = 0 \mu=0 μ = 0 and ν = 3 \nu=3 ν = 3 . Thus, two isovector-tensor modes with i = 1 , 2 i=1,2 i = 1 , 2 and μ = 1 \mu=1 μ = 1 , ν = 2 \nu=2 ν = 2 and one isoscalar-tensor mode with μ = 0 \mu=0 μ = 0 and ν = 3 \nu=3 ν = 3 are the massless Nambu-Goldstone modes. We can check that the other meson modes have non-zero masses.
Summary and Concluding Remarks
In this paper, the possibility of the existence of the pseudovector condensed phase has been investigate in the quark matter with the finite quark chemical potential μ \mu μ . Here, the interaction strength for the pseudovector interaction in the NJL model G p v ( = − G ) G_{\mathrm{pv}}(=-G) G pv ( = − G ) has been taken as a free parameter. If G p v ≈ G S G_{pv} \approx G_{S} G p v ≈ G S , there is a possibility of the pseudovector condensed phase due to the vacuum effects in a certain region of the quark chemical potential, in which the chiral condensate disappears. It has been thus shown that the chiral condensate and the pseudovector condensate do not coexist. If the pseudovector interaction strength G p v G_{pv} G p v is taken rather large such as G p v > 1.3 G S G_{pv} > 1.3G_{S} G p v > 1.3 G S , then, the point with σ p v ≠ 0 \sigma_{pv} \ne0 σ p v = 0 and σ s = 0 \sigma_{s} = 0 σ s = 0 becomes the absolute minimum of the effective potential, while σ p v = 0 \sigma_{pv} = 0 σ p v = 0 and σ s ≠ 0 \sigma_{s} \ne0 σ s = 0 is rest as the local minimum. However, this situation is unphysical because the vacuum with μ = 0 \mu= 0 μ = 0 does not reveal the quark spin polarization due to the pseudovector condensate in the realistic case.
Also, the massless meson modes on the pseudovector condensate, which leads to the quark spin polarization ⟨ ψ † Σ 3 τ 3 ψ ⟩ ( = ⟨ ψ ˉ γ 5 γ 3 τ 3 ψ ⟩ ) \langle\psi^{\dagger}\Sigma_{3}\tau_{3}\psi\rangle(=\langle\bar{\psi}\gamma_{5}\gamma_{3}\tau_{3}\psi\rangle) ⟨ ψ † Σ 3 τ 3 ψ ⟩ ( = ⟨ ψ ˉ γ 5 γ 3 τ 3 ψ ⟩) , have been investigated in the Nambu-Jona-Lasinio model with the pseudovector interaction between quarks. If the spin polarization ⟨ ψ † Σ 3 τ 3 ψ ⟩ \langle\psi^{\dagger}\Sigma_{3}\tau_{3}\psi\rangle ⟨ ψ † Σ 3 τ 3 ψ ⟩ due to the pseudovector condensate ⟨ ψ ˉ γ 5 γ 3 τ 3 ψ ⟩ \langle\bar{\psi}\gamma_{5}\gamma_{3}\tau_{3}\psi\rangle ⟨ ψ ˉ γ 5 γ 3 τ 3 ψ ⟩ only exists and the chiral condensate ⟨ ψ ˉ ψ ⟩ \langle\bar{\psi}\psi\rangle ⟨ ψ ˉ ψ ⟩ does not exist, four massless modes have appeared, namely, two rho meson modes, ρ i = 1 μ = 3 \rho^{\mu=3}_{i=1} ρ i = 1 μ = 3 , ρ i = 2 μ = 3 \rho^{\mu=3}_{i=2} ρ i = 2 μ = 3 and two a a a meson modes, a i = 1 μ = 3 a^{\mu=3}_{i=1} a i = 1 μ = 3 , a i = 2 μ = 3 a^{\mu=3}_{i=2} a i = 2 μ = 3 .
The reason is as follows : Let us consider the chiral s u A ( 2 ) × s u V ( 2 ) \mathit{su}_{A}(2) \times\mathit{su}_{V}(2) su A ( 2 ) × su V ( 2 ) transformation, namely
ψ ⟶ e i θ a γ 5 τ a / 2 ψ ≈ ψ + i θ a γ 5 τ a 2 ψ ≈ ψ + θ a δ A a ψ , for s u A ( 2 ) -transformation ψ ⟶ e i φ a τ a / 2 ψ ≈ ψ + i φ a τ a 2 ψ ≈ ψ + φ a δ V a ψ , for s u V ( 2 ) -transformation (74) \begin{aligned}
\psi&\longrightarrow e^{i\theta_{a}\gamma_{5}\tau^{a}/2}\psi\approx\psi+i\theta_{a}\gamma_{5}\frac{\tau^{a}}{2}\psi\\
&\approx\psi+\theta_{a}\delta_{A}^{a}\psi,\qquad\text{for }\mathit{su}_{A}(2)\text{-transformation}\\
\psi&\longrightarrow e^{i\varphi_{a}\tau^{a}/2}\psi\approx\psi+i\varphi_{a}\frac{\tau^{a}}{2}\psi\\
&\approx\psi+\varphi_{a}\delta_{V}^{a}\psi,\qquad\text{for }\mathit{su}_{V}(2)\text{-transformation} \tag*{(74)}
\end{aligned} ψ ψ ⟶ e i θ a γ 5 τ a /2 ψ ≈ ψ + i θ a γ 5 2 τ a ψ ≈ ψ + θ a δ A a ψ , for su A ( 2 ) -transformation ⟶ e i φ a τ a /2 ψ ≈ ψ + i φ a 2 τ a ψ ≈ ψ + φ a δ V a ψ , for su V ( 2 ) -transformation (74)
where δ A a ψ = i γ 5 ( τ a / 2 ) ψ \delta_{A}^{a}\psi=i\gamma_{5}(\tau^{a}/2)\psi δ A a ψ = i γ 5 ( τ a /2 ) ψ and δ V a ψ = i ( τ a / 2 ) ψ \delta_{V}^{a}\psi=i(\tau^{a}/2)\psi δ V a ψ = i ( τ a /2 ) ψ , respectively, and a a a represents the isospin indices. The Noether currents j A , a μ j_{A,a}^{\mu} j A , a μ and j V , a μ j_{V,a}^{\mu} j V , a μ and the Noether charge Q a Q^{a} Q a are calculated as
j A , a μ = ∂ L ∂ ( ∂ μ ψ ) δ A a ψ = − ψ ˉ γ μ γ 5 τ a 2 ψ , j V , a μ = ∂ L ∂ ( ∂ μ ψ ) δ V a ψ = − ψ ˉ γ μ τ a 2 ψ , (75) \begin{aligned}
j_{A,a}^{\mu} &= \frac{\partial\mathcal{L}}{\partial(\partial_{\mu}\psi)}\delta_{A}^{a}\psi=-\bar{\psi}\gamma^{\mu}\gamma_{5}\frac{\tau^{a}}{2}\psi,\\
j_{V,a}^{\mu} &= \frac{\partial\mathcal{L}}{\partial(\partial_{\mu}\psi)}\delta_{V}^{a}\psi=-\bar{\psi}\gamma^{\mu}\frac{\tau^{a}}{2}\psi, \tag*{(75)}
\end{aligned} j A , a μ j V , a μ = ∂ ( ∂ μ ψ ) ∂ L δ A a ψ = − ψ ˉ γ μ γ 5 2 τ a ψ , = ∂ ( ∂ μ ψ ) ∂ L δ V a ψ = − ψ ˉ γ μ 2 τ a ψ , (75)
Q A a = ∫ d 3 x j A , a μ = 0 = − ∫ d 3 x ψ ˉ γ 0 γ 5 τ a 2 ψ , Q V a = ∫ d 3 x j V , a μ = 0 = − ∫ d 3 x ψ ˉ γ 0 τ a 2 ψ . (76) \begin{aligned}
Q_{A}^{a} &= \int d^{3}x\,j_{A,a}^{\mu=0}=-\int d^{3}x\,\bar{\psi}\gamma^{0}\gamma_{5}\frac{\tau^{a}}{2}\psi,\\
Q_{V}^{a} &= \int d^{3}x\,j_{V,a}^{\mu=0}=-\int d^{3}x\,\bar{\psi}\gamma^{0}\frac{\tau^{a}}{2}\psi. \tag*{(76)}
\end{aligned} Q A a Q V a = ∫ d 3 x j A , a μ = 0 = − ∫ d 3 x ψ ˉ γ 0 γ 5 2 τ a ψ , = ∫ d 3 x j V , a μ = 0 = − ∫ d 3 x ψ ˉ γ 0 2 τ a ψ . (76)
Then, under the chiral transformations, the following relations are obtained:
[ i Q A a , ψ ˉ i γ 5 τ b ψ ] = − δ a b ψ ˉ ψ , [ i Q A a , ψ ˉ ψ ] = ψ ˉ i γ 5 τ a ψ , [ i Q V a , ψ ˉ i γ 5 τ b ψ ] = 2 ϵ a b c ψ ˉ i γ 5 τ c ψ , [ i Q V a , ψ ˉ ψ ] = 0 . (77) \begin{aligned}
[ iQ_A^a, \bar{\psi}i\gamma_5\tau^b\psi] &= -\delta^{ab}\bar{\psi}\psi\,,\\
[ iQ_A^a, \bar{\psi}\psi] &= \bar{\psi}i\gamma_5\tau^a\psi\,,\\
[ iQ_V^a, \bar{\psi}i\gamma_5\tau^b\psi] &= 2\epsilon^{abc}\bar{\psi}i\gamma_5\tau^c\psi\,,\\
[ iQ_V^a, \bar{\psi}\psi] &= 0 \,. \tag*{(77)}
\end{aligned} [ i Q A a , ψ ˉ i γ 5 τ b ψ ] [ i Q A a , ψ ˉ ψ ] [ i Q V a , ψ ˉ i γ 5 τ b ψ ] [ i Q V a , ψ ˉ ψ ] = − δ ab ψ ˉ ψ , = ψ ˉ i γ 5 τ a ψ , = 2 ϵ ab c ψ ˉ i γ 5 τ c ψ , = 0 . (77)
Therefore, if the quark-antiquark chiral condensate ⟨ ψ ˉ ψ ⟩ \langle\bar{\psi}\psi\rangle ⟨ ψ ˉ ψ ⟩ exists, namely ⟨ ψ ˉ ψ ⟩ ≠ 0 \langle\bar{\psi}\psi\rangle\ne0 ⟨ ψ ˉ ψ ⟩ = 0 , then from the first equation in Eq. (77) ,
⟨ [ i Q A a , ψ ˉ i γ 5 τ b ψ ] ⟩ = − δ a b ⟨ ψ ˉ ψ ⟩ ≠ 0 (78) \langle[ iQ_A^a, \bar{\psi}i\gamma_5\tau^b\psi]\rangle= -\delta^{ab}\langle\bar{\psi}\psi\rangle\ne0 \tag*{(78)} ⟨[ i Q A a , ψ ˉ i γ 5 τ b ψ ]⟩ = − δ ab ⟨ ψ ˉ ψ ⟩ = 0 (78)
is obtained. Here, δ a b \delta^{ab} δ ab represents the Kronecker delta. Thus, the chiral s u A ( 2 ) su_A(2) s u A ( 2 ) -symmetry, whose generators are Q A a Q_A^a Q A a (a = 1 , 2 , 3 a=1,2,3 a = 1 , 2 , 3 ), is spontaneously broken and the NG modes are
ψ ˉ i γ 5 τ b ψ ≈ π b , (79) \bar{\psi}i\gamma_5\tau^b\psi\approx\pi^b \,, \tag*{(79)} ψ ˉ i γ 5 τ b ψ ≈ π b , (79)
namely, three pions for b = 1 b=1 b = 1 , 2 2 2 , 3 3 3 are the NG bosons. This statement is found in the many textbooks.
Similarly, keeping in mind about the pseudovector condensate, we obtain
[ i Q A a , ψ ˉ γ μ τ b ψ ] = − ϵ a b c ψ ˉ γ 5 γ μ τ c ψ , [ i Q A a , ψ ˉ γ 5 γ μ τ b ψ ] = − ϵ a b c ψ ˉ γ μ τ c ψ , [ i Q V a , ψ ˉ γ μ τ b ψ ] = ϵ a b c ψ ˉ γ μ τ c ψ , [ i Q V a , ψ ˉ γ 5 γ μ τ b ψ ] = ϵ a b c ψ ˉ γ 5 γ μ τ c ψ , (80) \begin{aligned}
[ iQ_A^a, \bar{\psi}\gamma^\mu\tau^b\psi] &= -\epsilon^{abc}\bar{\psi}\gamma_5\gamma^\mu\tau^c\psi\,,\\
[ iQ_A^a, \bar{\psi}\gamma_5\gamma^\mu\tau^b\psi] &= -\epsilon^{abc}\bar{\psi}\gamma^\mu\tau^c\psi\,,\\
[ iQ_V^a, \bar{\psi}\gamma^\mu\tau^b\psi] &= \epsilon^{abc}\bar{\psi}\gamma^\mu\tau^c\psi\,,\\
[ iQ_V^a, \bar{\psi}\gamma_5\gamma^\mu\tau^b\psi] &= \epsilon^{abc}\bar{\psi}\gamma_5\gamma^\mu\tau^c\psi\,, \tag*{(80)}
\end{aligned} [ i Q A a , ψ ˉ γ μ τ b ψ ] [ i Q A a , ψ ˉ γ 5 γ μ τ b ψ ] [ i Q V a , ψ ˉ γ μ τ b ψ ] [ i Q V a , ψ ˉ γ 5 γ μ τ b ψ ] = − ϵ ab c ψ ˉ γ 5 γ μ τ c ψ , = − ϵ ab c ψ ˉ γ μ τ c ψ , = ϵ ab c ψ ˉ γ μ τ c ψ , = ϵ ab c ψ ˉ γ 5 γ μ τ c ψ , (80)
where ϵ a b c \epsilon^{abc} ϵ ab c is the completely antisymmetric tensor. Therefore, if the pseudovector-type quark-antiquark condensate exists, namely ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ ≠ 0 \langle\bar{\psi}\gamma_5\gamma^\mu\tau^3\psi\rangle\ne0 ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ = 0 , then from the first and the fourth equations in Eq. (80) ,
⟨ [ i Q A a = 1 , ψ ˉ γ μ τ b = 2 ψ ] ⟩ = − ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ ≠ 0 , ⟨ [ i Q A a = 2 , ψ ˉ γ μ τ b = 1 ψ ] ⟩ = ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ ≠ 0 , ⟨ [ i Q V a = 1 , ψ ˉ γ 5 γ μ τ b = 2 ψ ] ⟩ = ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ ≠ 0 , ⟨ [ i Q V a = 2 , ψ ˉ γ 5 γ μ τ b = 1 ψ ] ⟩ = − ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ ≠ 0 (81) \begin{aligned}
\langle[ iQ_A^{a=1}, \bar{\psi}\gamma^\mu\tau^{b=2}\psi]\rangle&= -\langle\bar{\psi}\gamma_5\gamma^\mu\tau^3\psi\rangle\ne0 \,,\\
\langle[ iQ_A^{a=2}, \bar{\psi}\gamma^\mu\tau^{b=1}\psi]\rangle&= \langle\bar{\psi}\gamma_5\gamma^\mu\tau^3\psi\rangle\ne0 \,,\\
\langle[ iQ_V^{a=1}, \bar{\psi}\gamma_5\gamma^\mu\tau^{b=2}\psi]\rangle&= \langle\bar{\psi}\gamma_5\gamma^\mu\tau^3\psi\rangle\ne0 \,,\\
\langle[ iQ_V^{a=2}, \bar{\psi}\gamma_5\gamma^\mu\tau^{b=1}\psi]\rangle&= -\langle\bar{\psi}\gamma_5\gamma^\mu\tau^3\psi\rangle\ne0 \tag*{(81)}
\end{aligned} ⟨[ i Q A a = 1 , ψ ˉ γ μ τ b = 2 ψ ]⟩ ⟨[ i Q A a = 2 , ψ ˉ γ μ τ b = 1 ψ ]⟩ ⟨[ i Q V a = 1 , ψ ˉ γ 5 γ μ τ b = 2 ψ ]⟩ ⟨[ i Q V a = 2 , ψ ˉ γ 5 γ μ τ b = 1 ψ ]⟩ = − ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ = 0 , = ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ = 0 , = ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ = 0 , = − ⟨ ψ ˉ γ 5 γ μ τ 3 ψ ⟩ = 0 (81)
are obtained. Thus, from the first and the second equations, the chiral s u A ( 2 ) \mathit{su}_{A}(2) su A ( 2 ) -symmetry, namely Q A 1 Q^{1}_{A} Q A 1 and Q A 2 Q^{2}_{A} Q A 2 , is spontaneously broken, which leads to the NG modes as
ψ ˉ γ μ τ a = 2 ψ ≈ ρ 2 μ , ψ ˉ γ μ τ a = 1 ψ ≈ ρ 1 μ . (82) \bar{\psi}\gamma^{\mu}\tau^{a=2}\psi\approx\rho^{\mu}_{2},
\qquad
\bar{\psi}\gamma^{\mu}\tau^{a=1}\psi\approx\rho^{\mu}_{1}. \tag*{(82)} ψ ˉ γ μ τ a = 2 ψ ≈ ρ 2 μ , ψ ˉ γ μ τ a = 1 ψ ≈ ρ 1 μ . (82)
Namely, two rho meson modes with a = 1 a=1 a = 1 and 2 2 2 are the NG bosons. In addition to the chiral s u A ( 2 ) \mathit{su}_{A}(2) su A ( 2 ) -symmetry, the chiral s u V ( 2 ) \mathit{su}_{V}(2) su V ( 2 ) -symmetry, namely Q V 1 Q^{1}_{V} Q V 1 and Q V 2 Q^{2}_{V} Q V 2 , is also broken. Thus, from the third and the fourth equations in Eq. (81) , we obtain the other two NG modes as
ψ ˉ γ 5 γ μ τ a = 2 ψ ≈ a 2 μ , ψ ˉ γ 5 γ μ τ a = 1 ψ ≈ a 1 μ . (83) \bar{\psi}\gamma_{5}\gamma^{\mu}\tau^{a=2}\psi\approx a^{\mu}_{2},
\qquad
\bar{\psi}\gamma_{5}\gamma^{\mu}\tau^{a=1}\psi\approx a^{\mu}_{1}. \tag*{(83)} ψ ˉ γ 5 γ μ τ a = 2 ψ ≈ a 2 μ , ψ ˉ γ 5 γ μ τ a = 1 ψ ≈ a 1 μ . (83)
Namely, two a a a meson modes with a = 1 a=1 a = 1 and 2 2 2 are also the NG bosons. Therefore, under the pseudovector condensate, the chiral s u V ( 2 ) \mathit{su}_{V}(2) su V ( 2 ) -symmetry is broken as well as the chiral s u A ( 2 ) \mathit{su}_{A}(2) su A ( 2 ) -symmetry breaking. Thus, four NG modes are realized on the pseudovector condensate. This situations are exactly shown in section 4.2.
Next, if the chiral condensate and the pseudovector condensate coexist, namely σ 0 ≠ 0 \sigma_{0}\ne0 σ 0 = 0 and a 0 ≠ 0 a_{0}\ne0 a 0 = 0 , the symmetry with respect to Q A 1 Q^{1}_{A} Q A 1 , Q A 2 Q^{2}_{A} Q A 2 , Q A 3 Q^{3}_{A} Q A 3 is broken due to σ 0 ≠ 0 \sigma_{0}\ne0 σ 0 = 0 and the symmetry for Q A 1 Q^{1}_{A} Q A 1 , Q A 2 Q^{2}_{A} Q A 2 , Q V 1 Q^{1}_{V} Q V 1 and Q V 2 Q^{2}_{V} Q V 2 is also broken due to a 0 ≠ 0 a_{0}\ne0 a 0 = 0 . Since the symmetry with respect to five generators Q A 1 Q^{1}_{A} Q A 1 , Q A 2 Q^{2}_{A} Q A 2 , Q A 3 Q^{3}_{A} Q A 3 , Q V 1 Q^{1}_{V} Q V 1 and Q V 2 Q^{2}_{V} Q V 2 is broken, it is expected that the five NG modes should appear. We already found three massless modes by calculating the diagonal elements of the two-point vertex functions Γ α ( p ) \Gamma_{\alpha}(p) Γ α ( p ) , namely Γ π , i = 3 ( p → 0 ) = 0 \Gamma_{\pi,i=3}(p\to0)=0 Γ π , i = 3 ( p → 0 ) = 0 , Γ a , i ≠ 3 μ = 3 , ν = 3 ( p → 0 ) = 0 \Gamma^{\mu=3,\nu=3}_{a,i\ne3}(p\to0)=0 Γ a , i = 3 μ = 3 , ν = 3 ( p → 0 ) = 0 . Here, π 3 \pi^{3} π 3 mode related to Q A 3 Q^{3}_{A} Q A 3 and a i = 1 a n d 2 μ = 3 a^{\mu=3}_{i=1\ \mathrm{and}\ 2} a i = 1 and 2 μ = 3 modes related to Q V 1 Q^{1}_{V} Q V 1 and Q V 2 Q^{2}_{V} Q V 2 correspond to the NG modes. However, the rest two NG modes are naively missing, which related to Q A 1 Q^{1}_{A} Q A 1 and Q A 2 Q^{2}_{A} Q A 2 . The generators Q A 1 Q^{1}_{A} Q A 1 and Q A 2 Q^{2}_{A} Q A 2 appear two times in the symmetry breaking due to σ 0 ≠ 0 \sigma_{0}\ne0 σ 0 = 0 , which leads to π 1 \pi_{1} π 1 and π 2 \pi_{2} π 2 as NG modes, and a 0 ≠ 0 a_{0}\ne0 a 0 = 0 , which leads to ρ 2 μ \rho^{\mu}_{2} ρ 2 μ and ρ 1 μ \rho^{\mu}_{1} ρ 1 μ as NG modes. Thus, π 1 \pi_{1} π 1 -ρ 2 μ \rho^{\mu}_{2} ρ 2 μ mixing and π 2 \pi_{2} π 2 -ρ 1 μ \rho^{\mu}_{1} ρ 1 μ mixing may occur. Actually, as was shown in section 4, these occurs and ( σ 0 π 1 + a 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 (\sigma_{0}\pi_{1}+a_{0}\rho^{\mu=3}_{2})/\sqrt{\sigma_{0}^{2}+a_{0}^{2}} ( σ 0 π 1 + a 0 ρ 2 μ = 3 ) / σ 0 2 + a 0 2 and ( σ 0 π 2 − a 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 (\sigma_{0}\pi_{2}-a_{0}\rho^{\mu=3}_{1})/\sqrt{\sigma_{0}^{2}+a_{0}^{2}} ( σ 0 π 2 − a 0 ρ 1 μ = 3 ) / σ 0 2 + a 0 2 are realized as the massless NG modes, while it is well known that, in the case of σ 0 ≠ 0 \sigma_{0}\ne0 σ 0 = 0 , the pion and the a a a meson mix. [20 ]
Next, keeping in mind about the tensor condensate only, we obtain
[ i Q A a , ψ ˉ γ μ γ ν τ b ψ ] = δ a b i ψ ˉ γ ν γ μ ψ , [ i Q A a , ψ ˉ i γ 5 γ μ γ ν ψ ] = − ψ ˉ γ μ γ ν τ a ψ , [ i Q V a , ψ ˉ γ μ γ ν τ b ψ ] = ϵ a b c ψ ˉ γ μ γ ν τ c ψ , [ i Q V a , ψ ˉ i γ 5 γ μ γ ν ψ ] = 0. (84) \begin{aligned}
[ iQ_A^a,\,\bar{\psi}\gamma^\mu\gamma^\nu\tau^b\psi] &= \delta^{ab}i\bar{\psi}\gamma^\nu\gamma^\mu\psi,\\
[ iQ_A^a,\,\bar{\psi}i\gamma_5\gamma^\mu\gamma^\nu\psi] &= -\bar{\psi}\gamma^\mu\gamma^\nu\tau^a\psi,\\
[ iQ_V^a,\,\bar{\psi}\gamma^\mu\gamma^\nu\tau^b\psi] &= \epsilon^{abc}\bar{\psi}\gamma^\mu\gamma^\nu\tau^c\psi,\\
[ iQ_V^a,\,\bar{\psi}i\gamma_5\gamma^\mu\gamma^\nu\psi] &= 0 . \tag*{(84)}
\end{aligned} [ i Q A a , ψ ˉ γ μ γ ν τ b ψ ] [ i Q A a , ψ ˉ i γ 5 γ μ γ ν ψ ] [ i Q V a , ψ ˉ γ μ γ ν τ b ψ ] [ i Q V a , ψ ˉ i γ 5 γ μ γ ν ψ ] = δ ab i ψ ˉ γ ν γ μ ψ , = − ψ ˉ γ μ γ ν τ a ψ , = ϵ ab c ψ ˉ γ μ γ ν τ c ψ , = 0. (84)
Therefore, if the tensor-type quark-antiquark condensate exists, namely ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ ≠ 0 \langle\bar{\psi}\gamma^\mu\gamma^\nu\tau^3\psi\rangle\ne0 ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ = 0 , then from the second and the third equations in Eq. (84) ,
⟨ [ i Q A a = 3 , ψ ˉ i γ 5 γ μ γ ν ψ ] ⟩ = − ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ ≠ 0 , ⟨ [ i Q V a = 1 , ψ ˉ γ μ γ ν τ b = 2 ψ ] ⟩ = ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ ≠ 0 , ⟨ [ i Q V a = 2 , ψ ˉ γ μ γ ν τ b = 1 ψ ] ⟩ = − ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ ≠ 0 (85) \begin{aligned}
\langle[ iQ_A^{a=3},\,\bar{\psi}i\gamma_5\gamma^\mu\gamma^\nu\psi]\rangle&= -\langle\bar{\psi}\gamma^\mu\gamma^\nu\tau^3\psi\rangle\ne0 ,\\
\langle[ iQ_V^{a=1},\,\bar{\psi}\gamma^\mu\gamma^\nu\tau^{b=2}\psi]\rangle&= \langle\bar{\psi}\gamma^\mu\gamma^\nu\tau^3\psi\rangle\ne0 ,\\
\langle[ iQ_V^{a=2},\,\bar{\psi}\gamma^\mu\gamma^\nu\tau^{b=1}\psi]\rangle&= -\langle\bar{\psi}\gamma^\mu\gamma^\nu\tau^3\psi\rangle\ne0 \tag*{(85)}
\end{aligned} ⟨[ i Q A a = 3 , ψ ˉ i γ 5 γ μ γ ν ψ ]⟩ ⟨[ i Q V a = 1 , ψ ˉ γ μ γ ν τ b = 2 ψ ]⟩ ⟨[ i Q V a = 2 , ψ ˉ γ μ γ ν τ b = 1 ψ ]⟩ = − ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ = 0 , = ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ = 0 , = − ⟨ ψ ˉ γ μ γ ν τ 3 ψ ⟩ = 0 (85)
are obtained. Thus, from the first equation, the chiral s u A ( 2 ) su_A(2) s u A ( 2 ) -symmetry is spontaneously broken, namely Q A 3 Q_A^3 Q A 3 , which leads to the NG modes as
ψ ˉ i γ 5 γ μ γ ν ψ ≈ p μ ν . (86) \bar{\psi}i\gamma_5\gamma^\mu\gamma^\nu\psi\approx p^{\mu\nu} . \tag*{(86)} ψ ˉ i γ 5 γ μ γ ν ψ ≈ p μν . (86)
Namely, one tensor meson mode corresponds to the NG mode. In addition to the chiral s u A ( 2 ) su_A(2) s u A ( 2 ) -symmetry, the chiral s u V ( 2 ) su_V(2) s u V ( 2 ) -symmetry, namely Q V 1 Q_V^1 Q V 1 and Q V 2 Q_V^2 Q V 2 , is also broken. Thus, from the second and the third equations in Eq. (85) , we obtain the other two NG modes as
ψ ˉ γ μ γ ν τ a = 2 ψ ≈ t 2 μ ν , \bar{\psi}\gamma^\mu\gamma^\nu\tau^{a=2}\psi\approx t_2^{\mu\nu} , ψ ˉ γ μ γ ν τ a = 2 ψ ≈ t 2 μν ,
ψ ˉ γ μ γ ν τ a = 1 ψ ≈ t 1 μ ν . (87) \bar{\psi}\gamma^\mu\gamma^\nu\tau^{a=1}\psi\approx t_1^{\mu\nu} . \tag*{(87)} ψ ˉ γ μ γ ν τ a = 1 ψ ≈ t 1 μν . (87)
Namely, two tensor meson modes with a = 1 a=1 a = 1 and 2 2 2 are also the NG bosons. Therefore, under the tensor condensate, the chiral s u V ( 2 ) su_V(2) s u V ( 2 ) -symmetry is broken as well as the chiral s u A ( 2 ) su_A(2) s u A ( 2 ) -symmetry breaking similar to the case of the pseudovector condensate. Thus, three NG modes are realized on the tensor condensate. This situations are exactly shown in section 5.2.
If there exists the quark spin polarization under consideration in this paper, the chiral symmetry is broken by the condensation with respect to the third component of the isospin, τ 3 \tau_3 τ 3 . In medium, the Lorentz symmetry is also broken by the quark spin polarization because the quark spin is polarized along z z z -axis, namely ⟨ ψ † Σ μ = 3 τ 3 ψ ⟩ \langle\psi^\dagger\Sigma_{\mu=3}\tau_3\psi\rangle ⟨ ψ † Σ μ = 3 τ 3 ψ ⟩ . As for the Lorentz symmetry breaking, three NG modes should also appear. However, we impose a strong constraint, namely μ = 3 \mu= 3 μ = 3 for spatial component and i = 3 i = 3 i = 3 for the isospin component always pair. Thus, if the quark spin polarization, for example pseudovector condensate, rotates on planes including the third axis, ( i , 3 ) = ( 0 , 3 ) (i, 3) = (0, 3) ( i , 3 ) = ( 0 , 3 ) -, ( 1 , 3 ) (1, 3) ( 1 , 3 ) - and ( 2 , 3 ) (2, 3) ( 2 , 3 ) -plane in the isospin space, the pseudovector condensate simultaneously rotates on planes including the third (z z z -) axis, ( μ , 3 ) = ( 0 , 3 ) ( = ( t , z ) ) (\mu, 3) = (0, 3)(= (t, z)) ( μ , 3 ) = ( 0 , 3 ) ( = ( t , z )) -, ( 1 , 3 ) ( = ( x , z ) ) (1, 3)(= (x, z)) ( 1 , 3 ) ( = ( x , z )) - and ( 2 , 3 ) ( = y , z ) ) (2, 3)(= y, z)) ( 2 , 3 ) ( = y , z )) -plane on the spacetime. Therefore, three NG modes should disappear. When the generators for the symmetry transformation are linearly independent, the reducing the number of the NG modes occurs, which has been indicated in [18 ] . However, it is interesting to know how many NG modes appear precisely in medium by the Lorentz symmetry breaking. [19 ] It is a future work to show which NG modes are type-I or II. This is an interesting problem because the behavior of the dispersion relation of the NG mode is determined in the long-wavelength limit.
Acknowledgements
The authors would like to express their sincere thanks to Referee for his/her helpful and useful comments, especially about the number of the massless Goldstone modes and the appearance of the massless modes due to the meson mixing relating to the symmetry-breaking patterns. The authors also thank to Mr. K. Matsusaka and Mr. K. Yamashita, who were carried out the preliminary work with them about the massless modes on the pseudovector condensate and on the tensor condensate, respectively. The authors acknowledge to the member of the Many-Body Theory Group in Kochi University.
Detail calculations for the inverse propagators Γ ρ , i ≠ 3 33 ( p → 0 ) \Gamma_{\rho,i\ne3}^{33}(p \to0) Γ ρ , i = 3 33 ( p → 0 ) , Γ a , i ≠ 3 33 ( p → 0 ) \Gamma_{a,i\ne3}^{33}(p \to0) Γ a , i = 3 33 ( p → 0 )
By use of the anticommutation relation for the Dirac gamma matrices γ μ \gamma^\mu γ μ and γ 5 \gamma_5 γ 5 , and the isospin matrices τ i \tau_i τ i , Eq. (32) or (33) is transformed. Noting ( A B ) − 1 = B − 1 A − 1 (AB)^{-1} = B^{-1}A^{-1} ( A B ) − 1 = B − 1 A − 1 for two matrices A A A and B B B , we can calculate the inverse propagator:
Γ ρ / a , i ≠ 3 33 ( p → 0 ) = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( γ 3 τ i ) 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 ( γ 3 τ i ) 1 k̸ − 2 a 0 γ 5 γ 3 τ 3 ] = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( γ 3 τ i ) 1 ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) ( k̸ − 2 a 0 γ 5 γ 3 τ 3 ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) × ( γ 3 τ i ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) 1 ( k̸ − 2 a 0 γ 5 γ 3 τ 3 ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) ] = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( γ 3 τ i ) 1 k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) × ( γ 3 τ i ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) 1 k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ] = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 tr [ ( γ 3 τ i ) 1 ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) × ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) ( γ 3 τ i ) ( k̸ + 2 a 0 γ 5 γ 3 τ 3 ) × ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) 1 ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) ] = 2 G − 4 ∫ d 4 k i ( 2 π ) 4 1 [ ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 ] 2 × tr [ γ 3 τ i { ( k 2 − 4 a 0 2 ) k̸ + 8 a 0 2 k 3 γ 3 + [ 4 a 0 k 3 γ 5 k̸ + 2 a 0 ( k 2 − 4 a 0 2 ) γ 5 γ 3 ] τ 3 } γ 3 τ i × { ( k 2 − 4 a 0 2 ) k̸ + 8 a 0 2 k 3 γ 3 + [ − 4 a 0 k 3 k̸ γ 5 + 2 a 0 ( k 2 − 4 a 0 2 ) γ 5 γ 3 ] τ 3 } } . (88) \begin{aligned}
\Gamma^{33}_{\rho/a,i\ne3}(p\to0)
&= \frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(\gamma^3\tau_i)\frac{1}{\not{k}-2a_0\gamma_5\gamma^3\tau_3}(\gamma^3\tau_i)\frac{1}{\not{k}-2a_0\gamma_5\gamma^3\tau_3}\right]\\
&= \frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(\gamma^3\tau_i)\frac{1}{(\not{k}+2a_0\gamma_5\gamma^3\tau_3)(\not{k}-2a_0\gamma_5\gamma^3\tau_3)}(\not{k}+2a_0\gamma_5\gamma^3\tau_3)\right.\\
&\qquad\left.\times(\gamma^3\tau_i)(\not{k}+2a_0\gamma_5\gamma^3\tau_3)\frac{1}{(\not{k}-2a_0\gamma_5\gamma^3\tau_3)(\not{k}+2a_0\gamma_5\gamma^3\tau_3)}\right]\\
&= \frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(\gamma^3\tau_i)\frac{1}{k^2-4a_0^2-4a_0k_3\gamma_5\tau_3}(\not{k}+2a_0\gamma_5\gamma^3\tau_3)\right.\\
&\qquad\left.\times(\gamma^3\tau_i)(\not{k}+2a_0\gamma_5\gamma^3\tau_3)\frac{1}{k^2-4a_0^2+4a_0k_3\gamma_5\tau_3}\right]\\
&= \frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\operatorname{tr}\left[(\gamma^3\tau_i)\frac{1}{(k^2-4a_0^2+4a_0k_3\gamma_5\tau_3)(k^2-4a_0^2-4a_0k_3\gamma_5\tau_3)}\right.\\
&\qquad\times(k^2-4a_0^2+4a_0k_3\gamma_5\tau_3)(\not{k}+2a_0\gamma_5\gamma^3\tau_3)(\gamma^3\tau_i)(\not{k}+2a_0\gamma_5\gamma^3\tau_3)\\
&\qquad\left.\times(k^2-4a_0^2-4a_0k_3\gamma_5\tau_3)\frac{1}{(k^2-4a_0^2+4a_0k_3\gamma_5\tau_3)(k^2-4a_0^2-4a_0k_3\gamma_5\tau_3)}\right]\\
&= \frac{2}{G}-4\int\frac{d^4k}{i(2\pi)^4}\frac{1}{\left[(k^2-4a_0^2)^2-16a_0^2k_3^2\right]^2}\\
&\qquad\times\operatorname{tr}\left[\gamma^3\tau_i\left\{(k^2-4a_0^2)\not{k}+8a_0^2k_3\gamma^3+\left[4a_0k_3\gamma_5\not{k}+2a_0(k^2-4a_0^2)\gamma_5\gamma^3\right]\tau_3\right\}\gamma^3\tau_i\right.\\
&\qquad\left.\times\left\{(k^2-4a_0^2)\not{k}+8a_0^2k_3\gamma^3+\left[-4a_0k_3\not{k}\gamma_5+2a_0(k^2-4a_0^2)\gamma_5\gamma^3\right]\tau_3\right\}\right\}. \tag*{(88)}
\end{aligned} Γ ρ / a , i = 3 33 ( p → 0 ) = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( γ 3 τ i ) k − 2 a 0 γ 5 γ 3 τ 3 1 ( γ 3 τ i ) k − 2 a 0 γ 5 γ 3 τ 3 1 ] = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( γ 3 τ i ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) ( k − 2 a 0 γ 5 γ 3 τ 3 ) 1 ( k + 2 a 0 γ 5 γ 3 τ 3 ) × ( γ 3 τ i ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) ( k − 2 a 0 γ 5 γ 3 τ 3 ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) 1 ] = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( γ 3 τ i ) k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 1 ( k + 2 a 0 γ 5 γ 3 τ 3 ) × ( γ 3 τ i ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 1 ] = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k tr [ ( γ 3 τ i ) ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) 1 × ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) ( γ 3 τ i ) ( k + 2 a 0 γ 5 γ 3 τ 3 ) × ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) ( k 2 − 4 a 0 2 + 4 a 0 k 3 γ 5 τ 3 ) ( k 2 − 4 a 0 2 − 4 a 0 k 3 γ 5 τ 3 ) 1 ] = G 2 − 4 ∫ i ( 2 π ) 4 d 4 k [ ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 ] 2 1 × tr [ γ 3 τ i { ( k 2 − 4 a 0 2 ) k + 8 a 0 2 k 3 γ 3 + [ 4 a 0 k 3 γ 5 k + 2 a 0 ( k 2 − 4 a 0 2 ) γ 5 γ 3 ] τ 3 } γ 3 τ i × { ( k 2 − 4 a 0 2 ) k + 8 a 0 2 k 3 γ 3 + [ − 4 a 0 k 3 k γ 5 + 2 a 0 ( k 2 − 4 a 0 2 ) γ 5 γ 3 ] τ 3 } } . (88)
In the second line of the last equality, moving γ 3 τ i \gamma^3\tau_i γ 3 τ i by using the anticommutation relations, trace is calculated. Here, noting the following relations, namely
tr τ i = 0 , tr τ i 2 = 1 , tr ( γ 3 k̸ ) = k μ tr ( γ 3 γ μ ) = − 4 k 3 , tr ( γ 3 k̸ γ 3 k̸ ) = k μ k ν tr ( γ 3 γ μ γ 3 γ ν ) = k μ k ν tr [ ( 2 g μ 3 − γ μ γ 3 ) ( γ 3 γ ν ) ] = − 2 k 3 k ν tr ( γ 3 γ ν ) + k μ k ν tr ( γ μ γ ν ) = 8 k 3 2 + 4 k 2 , tr ( γ 3 k̸ γ 3 ) = tr ( γ 3 k̸ γ 5 ) = tr ( γ 5 γ μ ) = tr k̸ = 0 , (89) \begin{aligned}
\operatorname{tr}\tau_i&=0,\qquad\operatorname{tr}\tau_i^2=1,\\
\operatorname{tr}(\gamma^3\not{k})&=k_\mu\operatorname{tr}(\gamma^3\gamma^\mu)=-4k_3,\\
\operatorname{tr}(\gamma^3\not{k}\gamma^3\not{k})&=k_\mu k_\nu\operatorname{tr}(\gamma^3\gamma^\mu\gamma^3\gamma^\nu)
=k_\mu k_\nu\operatorname{tr}\left[(2g^{\mu3}-\gamma^\mu\gamma^3)(\gamma^3\gamma^\nu)\right]\\
&=-2k_3k_\nu\operatorname{tr}(\gamma^3\gamma^\nu)+k_\mu k_\nu\operatorname{tr}(\gamma^\mu\gamma^\nu)
=8k_3^2+4k^2,\\
\operatorname{tr}(\gamma^3\not{k}\gamma^3)&=\operatorname{tr}(\gamma^3\not{k}\gamma_5)
=\operatorname{tr}(\gamma_5\gamma^\mu)=\operatorname{tr}\not{k}=0, \tag*{(89)}
\end{aligned} tr τ i tr ( γ 3 k ) tr ( γ 3 k γ 3 k ) tr ( γ 3 k γ 3 ) = 0 , tr τ i 2 = 1 , = k μ tr ( γ 3 γ μ ) = − 4 k 3 , = k μ k ν tr ( γ 3 γ μ γ 3 γ ν ) = k μ k ν tr [ ( 2 g μ 3 − γ μ γ 3 ) ( γ 3 γ ν ) ] = − 2 k 3 k ν tr ( γ 3 γ ν ) + k μ k ν tr ( γ μ γ ν ) = 8 k 3 2 + 4 k 2 , = tr ( γ 3 k γ 5 ) = tr ( γ 5 γ μ ) = tr k = 0 , (89)
and after some tedious calculations for trace, Equation (A1) can be simply expressed as
Γ ρ / a , i ≠ 3 33 ( p → 0 ) = 2 G − 96 ∫ d 4 k i ( 2 π ) 4 2 k 3 2 + k 2 − 4 a 0 2 [ ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 ] 2 . (90) \Gamma^{33}_{\rho/a,i\ne3}(p\to0)=\frac{2}{G}-96\int\frac{d^4k}{i(2\pi)^4}\frac{2k_3^2+k^2-4a_0^2}{\left[(k^2-4a_0^2)^2-16a_0^2k_3^2\right]^2}. \tag*{(90)} Γ ρ / a , i = 3 33 ( p → 0 ) = G 2 − 96 ∫ i ( 2 π ) 4 d 4 k [ ( k 2 − 4 a 0 2 ) 2 − 16 a 0 2 k 3 2 ] 2 2 k 3 2 + k 2 − 4 a 0 2 . (90)