betaFeedback

WorksPhysics and Astronomy

Constraints on the in-medium nuclear interaction from chiral symmetry and Lattice-QCD

Abstract

In this paper we discuss the combined effects on nuclear matter properties of the quark confinement mechanism in nucleon and of the chiral effective potential resulting from the spontaneous breaking of the chiral symmetry in nuclear matter. Based on the Nambu-Jona-Lasinio predictions, it is shown that the chiral potential acquires a specific scalar field cubic dependence, which contributes to the three-body interaction. We also discuss the constraints induced by Lattice-QCD on the model parameters governing the saturation properties. We introduce the term "QCD-connected parameters" for these quantities. We demonstrate that chiral symmetry and Lattice-QCD provide coherent constraints on the in-medium nuclear interaction, suggesting a fundamental origin of the saturation mechanism.

Introduction

Relativistic theories of nuclear matter initiated by Walecka and collaborators [1, 2] attract a lot of interest for, at least, two reasons: i) this type of approach provides a very economical saturation mechanism and ii) a spectacular well-known success in predicting the correct magnitude of the spin-orbit potential since nucleons move in an attractive background scalar field and in a repulsive vector background field which contribute in an additive way (see a recent discussion for this specific point in Ref. [3]). If the origin of the repulsive vector field can be safely identified as associated with the omega vector-meson exchange, the real nature of the attractive Lorentz scalar field has been a controversial subject since there is no sharp scalar resonance with a mass of about 500-700 MeV, which would lead to a simple interaction based on a scalar particle exchange. More fundamentally the question of the very nature of these background fields has to be elucidated; in other words, it is highly desirable to clarify their relationship with the QCD condensates, in particular the chiral quark condensate ⟨qˉq⟩\langle\bar{q}q\rangle, and more generally with the low energy realization of chiral symmetry which is spontaneously broken in the QCD vacuum and is expected to be progressively restored when the density increases. Indeed the microscopic origin of low-energy nuclear interaction properties is related to fundamental properties of the theory of the strong interaction (QCD) and should be implemented in the modeling of nuclear matter.

To bridge the gap between relativistic theories of the Walecka type and approaches insisting on chiral symmetry, it has been proposed in Ref. [4] to identify the ”nuclear physics” scalar sigma meson of the Walecka model at the origin of the nuclear binding, let us call it σW\sigma_W, with the chiral invariant s=S−Fπs=S-F_{\pi} field associated with the radial fluctuation of the chiral condensate SS around the ”chiral radius” FπF_{\pi}, identified with the pion decay constant. In the present approach we take the point of view that the effective theory has to be formulated, as a starting point, in term of the field WW associated with the fluctuations of the chiral quark condensate and parameterized as

W=σ+iτ⃗⋅π⃗≡SU=(s+Fπ)U=(σW+Fπ)UW=\sigma+i\vec{\tau}\cdot\vec{\pi}\equiv SU=(s+F_{\pi})U=(\sigma_W+F_{\pi})U

with

U(x)=eiτ⃗⋅ϕ⃗(x)/Fπ.(1)U(x)=e^{i\vec{\tau}\cdot\vec{\phi}(x)/F_{\pi}}. \tag*{(1)}

The scalar field σ\sigma (SS) and pseudoscalar fields π⃗\vec{\pi} (ϕ⃗\vec{\phi}) written in cartesian (polar) coordinates appear as the dynamical degrees of freedom and may deviate from the vacuum value, ⟨σ⟩vac=⟨S⟩vac=Fπ∝⟨qˉq⟩vac\langle\sigma\rangle_{\rm vac}=\langle S\rangle_{\rm vac}=F_{\pi}\propto\langle\bar{q}q\rangle_{\rm vac}. The sigma and the pion, associated with the amplitude s≡σWs\equiv\sigma_W and phase fluctuations ϕ⃗\vec{\phi} of this condensate, are considered in our approach to be effective degrees of freedom. Their dynamics are governed by an effective chiral potential, V(σ,π⃗)V(\sigma,\vec{\pi}), having a typical Mexican hat shape associated with a broken (chiral) symmetry of the QCD vacuum.

There is however a well identified problem concerning the nuclear saturation with usual chiral effective theories [5, 6, 7, 8]: independently of the particular chiral model, in the nuclear medium the value of SS (≡Smedium\equiv S_{\rm medium}) will be different from the one in vacuum (≡Svacuum\equiv S_{\rm vacuum}, the minimum of the vacuum effective potential represented by a ”Mexican hat” potential). At SmediumS_{\rm medium} the chiral potential has a smaller curvature: V′′(Smedium)<V′′(Svacuum)V''(S_{\rm medium})<V''(S_{\rm vacuum}). This single effect results in the lowering of the sigma mass and destroys the stability, which is a problem for the applicability of such effective theories in the nuclear context. The effect can be associated with a s3s^3 tadpole diagram generating attractive three-body forces destroying saturation even if the repulsive three-body force from the Walecka mechanism is present.

The origin of this problem is most probably related to the fact that nucleons are not point particle, but in reality composite systems made of quarks. Hence the nucleon will react against the presence of the nuclear scalar field. This effect can be taken into account by introducing the nucleon response to the scalar field ss, κNS=d2MN∗(s)/ds2\kappa_{\mathrm{NS}} = d^2 M_N^*(s)/ds^2 with the nucleon mass MN∗(s)M_N^*(s) defined in Eq. (7), which is the central ingredient of the quark-meson coupling model (QMC), introduced in the original pioneering work of P. Guichon [9] and successfully applied to finite nuclei with an explicit connection to the Skyrme force [10]. This effect associated with the polarization of the quark substructure in presence of the nuclear scalar field, will unavoidably generate three-body forces which may bring the needed repulsion. In practice this response or more precisely the nucleon scalar susceptibility κNS\kappa_{\mathrm{NS}} generates a non-linear coupling of the scalar field to the nucleon or equivalently a decrease of the scalar coupling constant with increasing density. Hence to achieve saturation, in a set of successive works devoted to the study of ordinary nuclear matter and neutron stars [11, 12, 13, 14, 15], we have complemented the relativistic chiral approach in such a way that the effect of the nucleon response is able to counterbalance the attractive chiral tadpole diagram to get good saturation properties, especially the correct curvature coefficient - the incompressibility modulus which is an empirical parameter defined at saturation density. All these aforementioned approaches were based on a chiral effective potential of the simplest linear sigma model with a Mexican hat shape of the following form

Vχ,LσM(s)=12Mσ2s2+12Mσ2−Mπ2Fπs3+18Mσ2−Mπ2Fπ2s4,(2)V_{\chi,\mathrm{L}\sigma\mathrm{M}}(s) = \frac{1}{2}M_{\sigma}^{2}s^{2} + \frac{1}{2}\frac{M_{\sigma}^{2}-M_{\pi}^{2}}{F_{\pi}}s^{3} + \frac{1}{8}\frac{M_{\sigma}^{2}-M_{\pi}^{2}}{F_{\pi}^{2}}s^{4}, \tag*{(2)}

which displays a strong cubic tadpole term, also referred as the tadpole diagram [7, 16, 17]. Indeed in order to get a correct description of the saturation properties it requires systematically a value of the dimensionless nucleonic response parameter, defined as (see also Eq. (8)),

C≡κNSFπ22MN,C \equiv\frac{\kappa_{\mathrm{NS}}F_{\pi}^{2}}{2M_{N}},

to be larger than one [11, 12, 13, 14, 15]. Such values are also required by the analysis of Lattice-QCD (LQCD) data on the chiral properties of the nucleon, with mass MNM_N, scalar charge QS=∂MN/∂mQ_S = \partial M_N/\partial m, and chiral susceptibility χN=∂2MN/∂m2\chi_N = \partial^2 M_N/\partial m^2 [18, 19, 20, 21] (mm is the current quark mass governing the explicit chiral symmetry breaking). Moreover in a recent work based on a Bayesian analysis with lattice data as an input [22], we found that the response parameter is strongly constrained to a value C∼1.4C \sim1.4 very close to the value where the scalar susceptibilities changes its sign: C=1.5C = 1.5.

The problem associated with this large value of CC is that it seems impossible to find a realistic confining models for the nucleon able to generate CC larger than one. For instance in the MIT bag model used in the QMC scheme, one has CMIT≃0.5C_{\mathrm{MIT}} \simeq0.5. One possible reason for this discrepancy between models and phenomenological values of CC lies in the use of the Lσ\sigmaM which is probably too naive. Hence one should certainly use an enriched chiral effective potential from a model able to give a correct description of the low-energy realization of chiral symmetry in the hadronic world. A good easily tractable candidate is the Nambu-Jona-Lasinio (NJL) model. Indeed in Ref. [23], referred as [NJLCONF] (NJL plus confinement) in the following, an explicit construction of the background scalar field was performed in the NJL model using a bosonization technique based on an improved derivative expansion valid at low (space-like) momenta [24]. Various confining interactions have been incorporated (quark-diquark string interaction, linear and quadratic confining interaction) on top of the NJL model which seem to be sufficient to generate saturation although the response parameters CC remain relatively small on the order of C∼0.5C \sim0.5. The reason is that, for a given scalar mass, the NJL chiral effective potential generates a significantly smaller attractive tadpole diagram than the simplistic Lσ\sigmaM. We will discuss this point in more details and demonstrate that the repulsive three-body force generating saturation, is not only determined by the nucleon response CC but also by the cubic term of the NJL potential, hereafter described by the new parameter CχC_{\chi}. The parameters CC and CχC_{\chi} combine together in the three-body interaction. We will also demonstrate how a particular combination of CC and CχC_{\chi} is constrained by lattice data [18, 19, 20, 21], which constitutes one main result of this paper.

In this paper we mainly discuss the effect of the chiral effective potential, i.e., the contribution of the CχC_{\chi} parameter, on the nuclear matter equation of state and on the saturation mechanism, without explicitly specifying the underlying nucleon confinement model. As mentioned above, very simple confining models have been already presented in [NJLCONF] and in a longer forthcoming paper referred as [NJLFCM] [25], we will explicitly introduce an effective Hamiltonian inspired from the field correlator method (FCM) developed by Y. Simonov and collaborators [26, 27, 28, 29, 30]. Modulo some ansatz prescription this approach allows us to generate simultaneously, at a semi-quantitative level, a confining interaction with long distance (r≫Tgr \gg T_g) behaviour V(r)=σgrV(r) = \sigma_g r, where the string tension σg=0.18 GeV2\sigma_g = 0.18\ {\rm GeV}^2, together with an equivalent NJL model with scalar interaction strength G1=120πσgTg4/(4NcNF)∼10 GeV−2G_1 = 120\pi\sigma_g T_g^4/(4N_cN_F) \sim10\ {\rm GeV}^{-2} and cutoff Λ∼1/Tg∼600 MeV\Lambda\sim1/T_g \sim600\ {\rm MeV}, where the gluon correlation length, Tg=0.25T_g = 0.25 to 0.3 fm0.3\ {\rm fm} [31], is itself related to the gluon condensate, G2G_2, according to Tg2=9σg/(π3G2)T_g^2 = 9\sigma_g/(\pi^3G_2). Note that the string tension σg\sigma_g and the gluon correlation length TgT_g are two parameters measured in Lattice-QCD [31].

II The NJL chiral confining model

The general picture underlying our approach has been sketched in our previous papers (see, e.g., [NJLCONF]) and will be precised in our forthcoming work [NJLFCM]. It can be summarized as follows: nuclear matter is made of nucleons, themselves built from quarks and gluons which look like Y-shaped strings generated by a non perturbative confining force, with constituent quarks at the ends. These quarks acquire a large mass from the quark condensate, which is the order parameter associated with the spontaneous breaking of chiral symmetry in the QCD vacuum. When the density nn of nuclear matter increases, the QCD vacuum is modified by the presence of the nucleons: the value of the quark condensate decreases and the chiral symmetry is progressively restored. Hence what is usually called ”the nuclear medium” can be seen as the original ”vacuum shifted” by a lower value of the order parameter. The mass of the constituent quarks coincides with the in-medium expectation value, M=S‾(n)M=\overline{S}(n), of the chiral invariant scalar field SS, associated with the radial fluctuation mode of the chiral condensate. We define an ”effective” or ”nuclear physics” scalar field ss by rescaling the chiral invariant scalar field SS, according to:

S≡M0FπS≡M0Fπ(s+Fπ)⟶∂∂s=M0Fπ∂∂S(3)S \equiv\frac{M_{0}}{F_{\pi}}S \equiv\frac{M_{0}}{F_{\pi}}(s+F_{\pi}) \quad\longrightarrow\quad\frac{\partial}{\partial s}=\frac{M_{0}}{F_{\pi}}\frac{\partial}{\partial S} \tag*{(3)}

where M0∼350M_{0} \sim350 MeV is the constituent quark mass in vacuum: S‾(s=0)=M0\overline{S}(s=0)=M_{0}. The vacuum expectation value of the ”effective” scalar field, S‾=Fπ\overline{S}=F_{\pi}, coincides by construction with the value of the pion decay constant FπF_{\pi}. The details of this construction are given in Ref. [23]. The important point is that its fluctuating piece, i.e., the ss field, has to be identified with the usual ”nuclear physics sigma meson” of relativistic Walecka theories, σW\sigma_{W}.

The nucleon is assumed to be described by an underlying model where constituent quarks (or diquarks) move in a confining interaction. In the previous [NJLCONF] work, ad-hoc confining potentials have been used on top of the NJL model generating the chirally broken vacuum. In the forthcoming longer paper [NJLFCM] the shape of this effective confining potential and the parameters of the equivalent NJL model will be obtained simultaneously in a way inspired from the field correlator method (FCM)[26, 27, 28, 29, 30]. The nucleon mass will thus naturally depend on the scalar field whose expectation value, M=S‾(n)M=\overline{S}(n), is associated with the in-medium constituent quark mass, namely:

MN∗(S)=MN+GS(S−M0)+3CNM0(S−M0)2+…(4)M_{N}^{*}(S)=M_{N}+G_{S}(S-M_{0})+3\frac{C_{N}}{M_{0}}(S-M_{0})^{2}+\ldots \tag*{(4)}

In passing we can notice that this approach is in spirit identical with the approach of Bentz and Thomas [7] but with a different underlying picture of the nucleon; in this latter paper the nucleon was constructed from the same NJL model as a bound quark-diquark state and the effect of confinement was taken into account through the presence of an infrared cutoff in the NJL loop integrals. We also used in our previous [NJLCONF] paper [23] a simple quark-diquark NJL model but with confinement incorporated through a string interaction between the color antitriplet diquark state and the color triplet quark state as in a heavy QQ‾Q\overline{Q} meson.

The two dimensionless response parameters, GSG_{S} which can be seen as the scalar number of quarks in the nucleon, and the susceptibility parameter CNC_{N}, only depend on the constituent quark mass and on the confining force, i.e., the confinement mechanism:

GS=(∂MN∗(S)∂S)S=M0,CN=M06(∂2MN∗(S)∂S2)S=M0.(5)G_{S}=\left(\frac{\partial M_{N}^{*}(S)}{\partial S}\right)_{S=M_{0}}, \qquad C_{N}=\frac{M_{0}}{6}\left(\frac{\partial^{2}M_{N}^{*}(S)}{\partial S^{2}}\right)_{S=M_{0}} . \tag*{(5)}

One important purpose of the present paper is to obtain phenomenological constraints on these two fundamental parameters that we will call ”QCD-connected parameters”, whereas our forthcoming paper [25] will provide a model calculation of these parameters in terms of σg\sigma_{g} and TgT_{g} within the FCM approach.

A The NJL chiral effective potential

In the following, we connect the expansion (4) of the nucleon mass to previously published expansion in terms of the effective ”nuclear physics” scalar field ss [4, 11, 12, 13, 14, 15, 22], defined as:

s=FπM0(S−M0).(6)s=\frac{F_{\pi}}{M_{0}}(S-M_{0}) . \tag*{(6)}

We have the following expansion of the nucleon mass:

MN∗(s)=MN+gSs+12κNSs2+O(s3)=MN(1+gSFπMNsFπ+C(sFπ)2+⋯ ),(7)M_N^*(s) = M_N + g_Ss + \frac{1}{2}\kappa_{NS}s^2 + \mathcal{O}(s^3) = M_N\left(1+\frac{g_SF_\pi}{M_N}\frac{s}{F_\pi}+C\left(\frac{s}{F_\pi}\right)^2+\cdots\right), \tag*{(7)}

with:

gS=M0FπGS,C≡κNSFπ22MN=3M0MNCN.(8)g_S = \frac{M_0}{F_\pi}G_S,\qquad C \equiv\frac{\kappa_{NS}F_\pi^2}{2M_N}=\frac{3M_0}{M_N}C_N. \tag*{(8)}

Consequently the in-medium nucleon mass mainly depends on two effective dimensionless QCD-connected parameters, the scalar nucleon coupling constant, gSg_S, and the dimensionless scalar nucleon susceptibility, C≡κNSFπ2/2MNC \equiv\kappa_{NS} F_\pi^2/2M_N, which embeds the influence of the internal nucleon structure or said differently the response of the nucleon to the nuclear scalar field. Notice that the response parameter CC used in our previous work is numerically close to the QCD-connected susceptibility parameter CNC_N. Its presence generates a decreasing density dependence of the in-medium scalar coupling constant, gS∗(s)=∂MN∗/∂s=gS+κNSs+…g_S^*(s)=\partial M_N^*/\partial s=g_S+\kappa_{NS}s+\ldots, corresponding to a progressive decoupling of the nucleon from the chiral condensate, which is an essential ingredient of the saturation mechanism (recall that ss is a negative quantity varying between zero in the vacuum to −Fπ-F_\pi at full chiral restoration).

The nuclear matter energy density as a functional of the scalar field S\mathcal{S} or the ss field is given by

ε0=∫4 d3k(2π)3Θ(pF−k)(k2+MN∗2(s)−MN,vac)+Vχ(s)+εω+ρ+εFock+εpion−nucleon loops,(9)\varepsilon_0=\int\frac{4\,d^3k}{(2\pi)^3}\Theta(p_F-k)\left(\sqrt{k^2+M_N^{*2}(s)}-M_{N,\mathrm{vac}}\right)+V_\chi(s)+\varepsilon_{\omega+\rho}+\varepsilon_{\mathrm{Fock}}+\varepsilon_{\mathrm{pion-nucleon\ loops}}, \tag*{(9)}

where only the scalar field contribution at the Hartree level together with the kinetic energy are explicitly written, while omega and rho meson exchanges, Fock terms and pion-nucleon loops (or correlation energy in the terminology of Ref. [12]) can be incorporated as well according to Refs. [12, 13, 14]. Note that Vχ(s)V_\chi(s) is the chiral effective potential which is expressed in the LσM by Eq. (2).

Let us now consider the case of the NJL model defined by the Lagrangian:

L=ψˉ(iγμ∂μ−m)ψ+G12[(ψˉψ)2+(ψˉiγ5τ⃗ψ)2]−G22[(ψˉγμτ⃗ψ)2+(ψˉγμγ5τ⃗ψ)2+(ψˉγμψ)2].(10)\mathcal{L}=\bar{\psi}(i\gamma^\mu\partial_\mu-m)\psi+\frac{G_1}{2}\left[(\bar{\psi}\psi)^2+(\bar{\psi}i\gamma_5\vec{\tau}\psi)^2\right]-\frac{G_2}{2}\left[(\bar{\psi}\gamma^\mu\vec{\tau}\psi)^2+(\bar{\psi}\gamma^\mu\gamma_5\vec{\tau}\psi)^2+(\bar{\psi}\gamma^\mu\psi)^2\right]. \tag*{(10)}

It depends on four parameters: the coupling constants G1G_1 (scalar), G2G_2 (vector), the current quark mass mm and a (noncovariant) cutoff parameter Λ\Lambda. Three of these parameters (G1G_1, mm, and Λ\Lambda) are adjusted to reproduce the pion mass, the pion decay constant and the quark condensate. For G2G_2 we consider different scenarios: G1=G2G_1=G_2 and G2=0G_2=0. We refer the reader to [NJLCONF] and [NJLFCM] for more details. Using path integral techniques and after a chiral rotation of the quark field, it can be equivalently written in a semi-bozonized form involving a pion field ϕ⃗\vec{\phi} embedded in the unitary operator U=ξ2=exp⁡(iτ⃗⋅ϕ⃗(x)/Fπ)U=\xi^2=\exp(i\vec{\tau}\cdot\vec{\phi}(x)/F_\pi), a scalar field, S\mathcal{S}, a vector field, VμV^\mu, and an axial-vector field, AμA^\mu. It has the explicit form given in Eqs. (2), (7)-(11) of Ref. [23]. Subtracting the vacuum expectation values, the chiral effective potential can be expressed as:

Vχ,NJL(s)=−2NcNf(I0(S)−I0(M0))+(S−m)2−(M0−m)22G1.(11)V_{\chi,\mathrm{NJL}}(s)=-2N_cN_f\left(I_0(\mathcal{S})-I_0(M_0)\right)+\frac{(\mathcal{S}-m)^2-(M_0-m)^2}{2G_1}. \tag*{(11)}

The quantity, −2NcNfI0(S)-2N_cN_f I_0(\mathcal{S}), is nothing but the total (in-medium) energy of the Dirac sea of constituent quarks with the NJL loop integral I0(S)I_0(\mathcal{S}) given hereafter. The vacuum constituent quark mass M0M_0 corresponds to the minimum of the chiral effective potential, i.e., Vχ,NJL′(s=0)=0V'_{\chi,\mathrm{NJL}}(s=0)=0, where V′V' is the derivative with respect to the scalar field ss. It is consequently the solution of the gap equation

M0=m+4NcNfM0G1I1(M0),(12)M_0=m+4N_cN_fM_0G_1I_1(M_0), \tag*{(12)}

where I1(M0)I_1(M_0) is another NJL loop integral given in the set of equations below

I0(S)=∫0Λdp(2π)3Ep(S),I1(S)=∫0Λdp(2π)312Ep(S),I2(S)=∫0Λdp(2π)314Ep3(S),I3(S)=∫0Λdp(2π)338Ep5(S).(13)\begin{aligned} I_0(\mathcal{S})&=\int_0^\Lambda\frac{dp}{(2\pi)^3}E_p(\mathcal{S}),\qquad I_1(\mathcal{S})=\int_0^\Lambda\frac{dp}{(2\pi)^3}\frac{1}{2E_p(\mathcal{S})},\\ I_2(\mathcal{S})&=\int_0^\Lambda\frac{dp}{(2\pi)^3}\frac{1}{4E_p^3(\mathcal{S})},\qquad I_3(\mathcal{S})=\int_0^\Lambda\frac{dp}{(2\pi)^3}\frac{3}{8E_p^5(\mathcal{S})}. \tag*{(13)} \end{aligned}

where Ep(S)=S2+p2E_p(\mathcal{S})=\sqrt{\mathcal{S}^2+p^2}.

Effective chiral potential expanded in the ss field

For a comparison with usual RMF model using the Lσ\sigmaM chiral effective potentials of Eq. (2) or equivalently non-linear sigma couplings, we expand the effective potential to third order in ss as:

Vχ,NJL(s)=Vχ(0)+Vχ′(0)s+12Vχ′′(0)s2+16Vχ′′′(0)s3+…(14)V_{\chi,\mathrm{NJL}}(s) = V_{\chi}(0) + V'_{\chi}(0)s + \frac{1}{2}V''_{\chi}(0)s^2 + \frac{1}{6}V'''_{\chi}(0)s^3 + \ldots \tag*{(14)}

An explicit calculation of the derivatives of the potential yields

Vχ,NJL(s)=12Mσ2s2+12Mσ2−Mπ2Fπs3(1−Cχ,NJL)+…(15)V_{\chi,\mathrm{NJL}}(s) = \frac{1}{2}M_{\sigma}^{2}s^2 + \frac{1}{2}\frac{M_{\sigma}^{2}-M_{\pi}^{2}}{F_{\pi}}s^3(1-C_{\chi,\mathrm{NJL}}) + \ldots \tag*{(15)}

where FπF_{\pi} is the pion decay constant and Mπ=mM0G1Fπ2M_{\pi} = \sqrt{\frac{mM_{0}}{G_{1}F_{\pi}^{2}}}, the canonical pion mass calculated in the bosonized NJL model. The effective sigma mass MσM_{\sigma} (considering the axial-pion mixing) is defined as

Mσ2=4M02fπ2Fπ2+Mπ2,with:fπ2=Fπ21−4G2Fπ2(16)M_{\sigma}^{2} = 4M_{0}^{2}\frac{f_{\pi}^{2}}{F_{\pi}^{2}} + M_{\pi}^{2}, \qquad\text{with:}\qquad f_{\pi}^{2} = \frac{F_{\pi}^{2}}{1-4G_{2}F_{\pi}^{2}} \tag*{(16)}

(where the second relation is obtained in the NJL model [23]) and Cχ,NJLC_{\chi,\mathrm{NJL}} is a specific NJL parameter:

Cχ,NJL=23M02J3(M0)I2(M0).(17)C_{\chi,\mathrm{NJL}} = \frac{2}{3}\frac{M_{0}^{2}J_{3}(M_{0})}{I_{2}(M_{0})}. \tag*{(17)}

This form of the NJL chiral effective potential deviates from the original Lσ\sigmaM, see Eq. (2), through the presence of the model dependent parameter Cχ,NJLC_{\chi,\mathrm{NJL}} whose net effect is to decrease the attractive cubic tadpole term of the Lσ\sigmaM. The use of this Cχ,NJLC_{\chi,\mathrm{NJL}} parameter is particularly convenient, since taking Cχ,NJL=1C_{\chi,\mathrm{NJL}} = 1 is equivalent to the absence of the tadpole diagram as in the case of the QMC model [9, 10].

In the absence of vector interaction (G2=0G_{2} = 0), for typical value of FCM parameters, σg=0.18 GeV2\sigma_{g} = 0.18\ \mathrm{GeV}^{2}, Tg=0.286 fmT_{g} = 0.286\ \mathrm{fm}, one obtains G1=12.514 GeV−2G_{1} = 12.514\ \mathrm{GeV}^{-2}. The NJL cutoff behaves necessarily as Λ∼1/Tg\Lambda\sim1/T_{g} but there is a certain arbitrariness in setting its precise value: we take Λ=0.604 GeV\Lambda= 0.604\ \mathrm{GeV}. Taking m=5.8 MeVm = 5.8\ \mathrm{MeV} this enables us to obtain reasonable values for the pion decay constant, Fπ=91.9 MeVF_{\pi} = 91.9\ \mathrm{MeV}, the pion mass Mπ=140 MeVM_{\pi} = 140\ \mathrm{MeV}, and the quark condensate ⟨qˉq⟩=−(241.1 MeV)3\langle\bar{q}q\rangle= -(241.1\ \mathrm{MeV})^{3}. The resulting vacuum constituent quark mass, effective sigma mass and CχC_{\chi} parameter are M0=356.7 MeVM_{0} = 356.7\ \mathrm{MeV}, Mσ=716.4 MeVM_{\sigma} = 716.4\ \mathrm{MeV} and Cχ,NJL=0.488C_{\chi,\mathrm{NJL}} = 0.488. Figure 1 shows that the approximate expansion (15) reproduces very well the exact NJL potential. Comparing Lσ\sigmaM with NJL scalar potential in Figure 1, one sees that the attractive tadpole term is larger in the case of Lσ\sigmaM. The effect of the parameter Cχ,NJLC_{\chi,\mathrm{NJL}} is then to reduce the attractive tadpole diagram and make the scalar potential more repulsive. Using another parameter set, G1=7.705 GeV−2G_{1} = 7.705\ \mathrm{GeV}^{-2}, Λ=0.740 GeV\Lambda= 0.740\ \mathrm{GeV} and m=3.5 MeVm = 3.5\ \mathrm{MeV}, compatible with the π−a1\pi-a_{1} mixing with G2=G1G_{2} = G_{1} as suggested by the FCM [26, 27, 28, 29, 30], one obtains M0=365.3 MeVM_{0} = 365.3\ \mathrm{MeV} and a smaller value of Cχ,NJL=0.43C_{\chi,\mathrm{NJL}} = 0.43 but the reduction of the tadpole diagram is still significant.

Figure 1. Effective potential (in units of the string tension σ2\sigma^2, with σg=0.18 GeV2\sigma_g = 0.18\ \mathrm{GeV}^2) plotted against ∣s∣/Fπ|s|/F_\pi for the NJL model (full line), Lσ\sigmaM (dashed line) and original Walecka model that is limited to the quadratic term (dotted line), for a given effective sigma mass Mσ=716.4 MeVM_\sigma= 716.4\ \mathrm{MeV}. Also shown is the approximate form of the NJL potential when limited to the cubic term in the scalar field ss expansion (15) (dot-dashed line). The effect of the s3s^3 term is well seen when comparing to the Walecka model. Note that the approximate expansion (15) is almost identical to the exact NJL potential.

In the following, we set Cχ≡Cχ,NJLC_{\chi} \equiv C_{\chi,\mathrm{NJL}} and Vχ≡Vχ,NJLV_{\chi} \equiv V_{\chi,\mathrm{NJL}} for simplicity.

Impact on nuclear matter properties

At the Hartree approximation (RMF), the scalar field minimizing the total energy is the solution of the following self-consistent equation of motion:

Vχ′(s)=−gs∗(s)nswithns=4∫0kFdk(2π)3MN∗(s)MN∗2(s)+k2,(18)V'_{\chi}(s) = -g_{s}^{*}(s)n_{s} \qquad\text{with} \qquad n_{s} = 4\int_{0}^{k_{F}}\frac{\mathrm{d}k}{(2\pi)^{3}}\frac{M_{N}^{*}(s)}{\sqrt{M_{N}^{*2}(s)+k^{2}}}, \tag*{(18)}

where Vχ′(s)V'_{\chi}(s) is the derivative of the Mexican hat chiral effective potential, with respect to the scalar field ss. This equation constitutes an in-medium modified gap equation whose solution is controlled by the nucleonic scalar density nsn_{s}.

To second order in s/Fπs/F_{\pi} or equivalently to second order in the scalar density nsn_{s}, the in-medium gap equation can be formally solved with the result:

sˉ=−gSMσ2ns+gSMσ4(κNS−gSVχ′′′(0)2Mσ2)ns2=−gSMσ2ns+gSMσ4(2MNFπ2C−3gS2FπMσ2−Mπ2Mσ2(1−Cχ))ns2=−gSMσ2ns+gS2Mσ4Fπ(2C~s−32)ns2withC~s≃MNgSFπC+34Cχ.(19)\begin{aligned} \bar{s} &= -\frac{g_S}{M_\sigma^2}n_s+\frac{g_S}{M_\sigma^4}\left(\kappa_{NS}-\frac{g_S V_\chi'''(0)}{2M_\sigma^2}\right)n_s^2 \\ &= -\frac{g_S}{M_\sigma^2}n_s+\frac{g_S}{M_\sigma^4}\left(\frac{2M_N}{F_\pi^2}C-\frac{3g_S}{2F_\pi}\frac{M_\sigma^2-M_\pi^2}{M_\sigma^2}(1-C_\chi)\right)n_s^2 \\ &= -\frac{g_S}{M_\sigma^2}n_s+\frac{g_S^2}{M_\sigma^4F_\pi}\left(2\widetilde{C}_s-\frac{3}{2}\right)n_s^2 \quad\text{with} \quad\widetilde{C}_s\simeq\frac{M_N}{g_SF_\pi}C+\frac{3}{4}C_\chi. \tag*{(19)} \end{aligned}

For a qualitative discussion, we have supposed Mπ≪MσM_\pi\ll M_\sigma to get the approximate expression C~s\widetilde{C}_s. The scalar field contribution to the energy per nucleon is defined as Es/A=Vχ(s)/n+MN∗(s)−MNE_s/A=V_\chi(s)/n+M_N^*(s)-M_N. To leading order in density, its contribution is defined as E(2b)E^{(2b)}, which reads

E(2b)A=−gS2Mσ2ns+12gS2Mσ2ns2n=−12gS2Mσ2n+12gS2Mσ2(ns−n)2n.(20)\frac{E^{(2b)}}{A}=-\frac{g_S^2}{M_\sigma^2}n_s+\frac{1}{2}\frac{g_S^2}{M_\sigma^2}\frac{n_s^2}{n}=-\frac{1}{2}\frac{g_S^2}{M_\sigma^2}n+\frac{1}{2}\frac{g_S^2}{M_\sigma^2}\frac{(n_s-n)^2}{n}. \tag*{(20)}

In the first expression of Eq. (20), we have separated the effect of the scalar self-energy of the nucleon and the contribution of the effective potential at leading order in the densities nn and nsn_s. In the second form, we display explicitly the term proportional to (ns−n)2(n_s-n)^2, corresponding to an effective repulsive three-body force, which is exactly the Walecka saturation mechanism when the omega is added. This contribution, which survives for an point-like nucleon, is proportional to the square of the nucleon momentum. This is the so-called Z graph associated with the excitation of NNˉN\bar{N} pairs [32, 33].

To second order in density Es/AE_s/A provides an effective three-body contribution to the energy per nucleon:

E(3b)A≃gS22Mσ4(κNS−gSVχ′′′(0)3Mσ2)ns2=gS3Mσ4Fπ(2C~3−1)ns2withC~3≃MNgSFπC+12Cχ.(21)\frac{E^{(3b)}}{A}\simeq\frac{g_S^2}{2M_\sigma^4}\left(\kappa_{NS}-\frac{g_S V_\chi'''(0)}{3M_\sigma^2}\right)n_s^2=\frac{g_S^3}{M_\sigma^4F_\pi}\left(2\widetilde{C}_3-1\right)n_s^2 \quad\text{with} \quad\widetilde{C}_3\simeq\frac{M_N}{g_SF_\pi}C+\frac{1}{2}C_\chi. \tag*{(21)}

We can recover Eq. (44) of Ref. [16] with Cχ=0C_\chi=0.

We now give a qualitative discussion of the influence of the three parameters gSg_{S}, κNS\kappa_{NS} and Vχ′′′(0)V_{\chi}^{\prime\prime\prime}(0) or equivalently gSg_{S}, CC and CχC_{\chi}, taking various works as illustrative examples.

If we ignore both the response of the nucleon, i.e., κNS=0\kappa_{NS}=0 (or C=0C=0), and the contribution of the tadpole diagram to the chiral potential, i.e., Vχ′′′(0)=0V_{\chi}^{\prime\prime\prime}(0)=0 (or Cχ=1C_{\chi}=1), we recover the original Walecka model since the three-body contribution (21) is absent and the saturation mechanism is associated with the ZZ graph alone, see Eq. (20). It is known that in this case saturation requires a large gS/Mσg_{S}/M_{\sigma} value, which implies a large repulsion induced by gω/mωg_{\omega}/m_{\omega} in order to obtain the empirical value of the binding energy. As a consequence one gets a much too large incompressibility modulus KsatK_{\mathrm{sat}}. One possibility to cure this problem is to introduce density dependent coupling constants [34, 35].

In the QMC model originally proposed in Ref. [9] and providing a successful phenomenology [10], the response of the nucleon is incorporated, but without explicit connection with the chiral status of the scalar field. Hence no tadpole diagram is considered, i.e., Vχ′′′(0)=0V_{\chi}^{\prime\prime\prime}(0)=0 or Cχ=1C_{\chi}=1. The original QMC model is formulated in the MIT bag model, yielding C∼0.5C\sim0.5 and E(3b)∝2C~3−1=2C∼1E^{(3b)}\propto2\widetilde{C}_{3}-1=2C\sim1 which turns out to be sufficient to bring the needed repulsion to get nuclear saturation with a correct incompressibility modulus, although this approach does not satisfy chiral symmetry requirements.

Soon after the first version of the relativistic Walecka model, it has been realized [5, 6, 7, 8] that in relativistic theory with a mexican hat-like effective potential, the contribution of the Walecka ZZ graph is not large enough to stabilize nuclear matter against the effect of the attractive tadpole diagram. This is the typical situation of the original LσM where Vχ′′′(0)V_{\chi}^{\prime\prime\prime}(0) is large and positive (Cχ∼0C_{\chi}\sim0) and even of the NJL model (Cχ<1C_{\chi}<1) where the response of the nucleon is ignored, i.e., C=0C=0. Some phenomenological approaches, such as the so-called NL3 model [36], have introduced self-interactions of the scalar field in the form of an effective potential but without connection to chiral symmetry. In particular a repulsive cubic term, i.e., Vχ′′′(0)<0V_{\chi}^{\prime\prime\prime}(0)<0, is introduced in this model. From Table II of Ref. [36], one can obtain the equivalent Cχ∼1.47C_{\chi}\sim1.47 parameter, which corresponds to C~3∼0.74\widetilde{C}_{3}\sim0.74. One can thus re-interpret the original NL3 model with a negative value of the c2c_{2} parameter (see table II of Ref. [36]) as a way to simulate in an effective way the nucleon response with C∼0.74C\sim0.74. The way the non-linear potential has been introduced in the NL3 model was pragmatic, but it can now be understood in a more fundamental approach.

Constraining the chiral confining potential by Lattice-QCD

In this section, we connect the in-medium properties of the nucleon mass defined by Eq. (7) with the Lattice-QCD calculations performed in vacuum (s=0s=0). For this reason, the nucleon mass will be noted in the following MN(s)M_{N}(s) (without the *). The derivatives of the nucleon mass could however be obtained, on the one hand, from the derivatives of the nucleon mass (7) taken at s=0s=0 and providing gSg_{S} and κNS\kappa_{NS}, and, on the other hand, from the Lattice-QCD calculations.

Within an underlying microscopic confining model for the nucleon, i.e., [N J L C O N F] and [N J L F C M], generating the quark core wave functions, the axial charge, the πNN\pi NN coupling constant and the πNN\pi NN form factor can be obtained, allowing the calculation of the pion cloud contribution (pion self-energy) to the in-medium nucleon (and Delta resonance) mass, as in the Cloudy Bag model [37] or similar approaches using an alternative confinement potential [38]. The pion cloud contribution to the nucleon mass is expressed as

Σ(π)(M;m)=−32(gA(M)2Fπ(M))2∫dq(2π)3q2v2(q;M)(1ωq1ωq+ϵNq+32251ωq1ωq+ϵΔq),(22)\Sigma^{(\pi)}(M;m)=-\frac{3}{2}\left(\frac{g_{A}(M)}{2F_{\pi}(M)}\right)^{2}\int\frac{\mathrm{d}q}{(2\pi)^{3}}q^{2}v^{2}(\mathbf{q};M)\left(\frac{1}{\omega_{q}}\frac{1}{\omega_{q}+\epsilon_{Nq}}+\frac{32}{25}\frac{1}{\omega_{q}}\frac{1}{\omega_{q}+\epsilon_{\Delta q}}\right), \tag*{(22)}

with ωq=q2+Mπ2(M)\omega_{q}=\sqrt{q^{2}+M_{\pi}^{2}(M)} and Mπ2(M)=mM/G1Fπ2(M)M_{\pi}^{2}(M)=mM/G_{1}F_{\pi}^{2}(M), the other quantities being defined in Eq. (22) of Ref. [12]. Here the various quantities such as Mπ2(M)M_{\pi}^{2}(M) are in-medium quantities where the vacuum constituent quark mass M0M_{0} is replaced by M=S‾M=\overline{S} (see Eq. (34) of Ref. [23] and the text just before). Thus in this framework, the nucleon mass is split according to:

MN(s)=MN(M;m)=MNcore(M)+Σ(π)(M;m).(23)M_{N}(s)=M_{N}(M;m)=M_{N}^{\mathrm{core}}(M)+\Sigma^{(\pi)}(M;m). \tag*{(23)}

Nucleon response and its chiral properties

The derivatives of the nucleon mass with respect to the constituent quark mass gives the response parameters, which are defined in Eq. (5), i.e., GS=∂MN(M;m)/∂MG_{S}=\partial M_{N}(M;m)/\partial M and CN=(1/6)∂2MN(M;m)/∂M2C_{N}=(1/6)\partial^{2}M_{N}(M;m)/\partial M^{2}, where the derivatives are taken at M=M0(m)M=M_{0}(m), i.e. s=0s=0. To benefit from the lattice data, we can relate them to two chiral properties of the nucleon, the scalar charge, QS=∂MN(M=M0(m);m)/∂mQ_{S}=\partial M_{N}(M=M_{0}(m);m)/\partial m, and the chiral susceptibility, χN=∂2MN(M=\chi_{N}=\partial^{2}M_{N}(M= M0(m;m)/∂m2M_0(m;m)/\partial m^2. All what we need for this calculation are the derivatives of the constituent quark mass with respect to the current quark mass. These derivatives are obtained from the NJL model and read [NJLFCM]:

(∂M0∂m)=(Mπ2m)M0Mσ2,(∂2M0∂m2)≈−(Mπ2m)23M0Mσ4(1−Cχ).(24)\left(\frac{\partial M_0}{\partial m}\right)=\left(\frac{M_\pi^2}{m}\right)\frac{M_0}{M_\sigma^2},\qquad\left(\frac{\partial^2 M_0}{\partial m^2}\right)\approx-\left(\frac{M_\pi^2}{m}\right)^2\frac{3M_0}{M_\sigma^4}(1-C_\chi). \tag*{(24)}

where in the second expression a correction factor of order Mπ2/Mσ2M_\pi^2/M_\sigma^2 has been neglected. We now note that in Eq. (23) the current quark mass appears explicitly only in the pionic self-energy Σ(π)\Sigma^{(\pi)}. It appears also implicitly through the dependence of the constituent quark mass upon the current quark mass. Hence the scalar charge, QSQ_S [16], receives two different contributions:

QS=∂MN∂m=∂MN∂M(∂M0∂m)+dΣ(π)(M0;m)dm=FπM0gS(Mπ2m)M0Mσ2+(Mπ2m)dΣ(π)(M0;m)dMπ2=(Mπ2m)FπgSMσ2+(Mπ2m)dΣ(π)dMπ2≡QS(s)+QS(π).(25)\begin{aligned} Q_S=\frac{\partial M_N}{\partial m}=\frac{\partial M_N}{\partial M}\left(\frac{\partial M_0}{\partial m}\right)+\frac{d\Sigma^{(\pi)}(M_0;m)}{dm}=\frac{F_\pi}{M_0}g_S\left(\frac{M_\pi^2}{m}\right)\frac{M_0}{M_\sigma^2}+\left(\frac{M_\pi^2}{m}\right)\frac{d\Sigma^{(\pi)}(M_0;m)}{dM_\pi^2} \\ &=\left(\frac{M_\pi^2}{m}\right)\frac{F_\pi g_S}{M_\sigma^2}+\left(\frac{M_\pi^2}{m}\right)\frac{d\Sigma^{(\pi)}}{dM_\pi^2}\equiv Q_S^{(s)}+Q_S^{(\pi)}. \tag*{(25)} \end{aligned}

where we have employed the relation (3). The second term, QS(π)Q_S^{(\pi)}, is referred as the pion cloud contribution. It is obtained by taking only the linear quark mass dependence appearing in Mπ2=M0m/G1Fπ2M_\pi^2=M_0m/G_1F_\pi^2, thus ignoring all the implicit mm dependencies through the MM dependence of MπM_\pi, gAg_A, FπF_\pi and the form factor. We refer the first term, QS(s)Q_S^{(s)}, as the scalar field contribution despite it contains the implicit mm dependence of the pionic self-energy. In effect QS(s)Q_S^{(s)} receives itself two separate contributions:

QS(s)=∂MNcore∂M(∂M0∂m)+∂Σ(π)∂M(∂M0∂m)≡(Mπ2m)FπgSMσ2.(26)Q_S^{(s)}=\frac{\partial M_N^{\mathrm{core}}}{\partial M}\left(\frac{\partial M_0}{\partial m}\right)+\frac{\partial\Sigma^{(\pi)}}{\partial M}\left(\frac{\partial M_0}{\partial m}\right)\equiv\left(\frac{M_\pi^2}{m}\right)\frac{F_\pi g_S}{M_\sigma^2}. \tag*{(26)}

The second contribution contains the implicit mm dependence of the pion self-energy coming from the MM dependence of the various quantities (Fπ(M),Mπ(M),gA,(F_\pi(M),M_\pi(M),g_A, form factor) through the mm dependence of the constituent quark mass taken at its vacuum value M0M_0. Regarding this specific point it is generally assumed that the pion properties are protected by chiral symmetry and this is what we find in the model developped in [NJLFCM] where the pion mass displays a remarkable stability for a large domain of the constituent quark mass or equivalently of the nuclear scalar field ss. As a consequence the induced effect on gSg_S is extremely small. However the combined effect of the modification of the nucleon size and of the pion decay constant might induce a more important correction on the πNN\pi NN vertex (gAv(q)/2Fπ)2(g_Av(q)/2F_\pi)^2 but we do not consider this effect which certainly requires a more detailed study. It follows for Eq. (25) that:

m∂MN∂m=FπgSMπ2Mσ2+Mπ2dΣ(π)(M0;m)dMπ2≡σN(s)+σN(π)≡σN.(27)m\frac{\partial M_N}{\partial m}=F_\pi g_S\frac{M_\pi^2}{M_\sigma^2}+M_\pi^2\frac{d\Sigma^{(\pi)}(M_0;m)}{dM_\pi^2}\equiv\sigma_N^{(s)}+\sigma_N^{(\pi)}\equiv\sigma_N. \tag*{(27)}

Hence we recover the nucleon sigma term. This result is just the expression of the Feynman-Hellman theorem. This light quark sigma term has been abundantly discussed in our previous papers [12, 13, 23]. Using a dipole πNN\pi NN form factor with cutoff Λ=1\Lambda=1 GeV, the pionic contribution to the sigma term was found to be σN(π)=21.5\sigma_N^{(\pi)}=21.5 MeV [12] and a pionic self-energy Σ(π)(M;m)=420\Sigma^{(\pi)}(M;m)=420 MeV. The value of the non pionic contribution was found to be σN(s)∼29\sigma_N^{(s)}\sim29 MeV [12] to get a total sigma term σN=50.5\sigma_N=50.5 MeV. Evidently the relative weight of the two contributions may be altered by the precise values of the parameters, but according to our model FCM calculation [25] and from the lattice data constraints discussed below, this modification of the relative weight should be rather moderate and the value of the sigma term and its repartition is a rather strong constraint on the nucleon modelling.

For the scalar susceptibility one obtains from Eq. (25), ignoring again higher order correction Mπn/MσnM_\pi^n/M_\sigma^n

χN=∂2MN∂m2=∂MN∂M(∂2M0∂m2)+∂2MN∂M2(∂M0∂m)2+∂∂M(dΣ(π)dm)(∂M0∂m)+d2Σ(π)dm2=−(Mπ2m)23gSFπMσ4(1−Cχ)+(Mπ2m)2κNSFπ2Mσ4+(Mπ2m)21Mσ2ddMπ2(M0∂Σ(π)∂M)+d2Σ(π)dm2≡χN(s)+χN(sπ)+(Mπ2m)2d2Σ(π)(M0;m)d(Mπ2)2≡χN(s)+χN(sπ)+χN(π).(28)\begin{aligned} \chi_N=\frac{\partial^2M_N}{\partial m^2}=\frac{\partial M_N}{\partial M}\left(\frac{\partial^2M_0}{\partial m^2}\right)+\frac{\partial^2M_N}{\partial M^2}\left(\frac{\partial M_0}{\partial m}\right)^2+\frac{\partial}{\partial M}\left(\frac{d\Sigma^{(\pi)}}{dm}\right)\left(\frac{\partial M_0}{\partial m}\right)+\frac{d^2\Sigma^{(\pi)}}{dm^2} \\ &=-\left(\frac{M_\pi^2}{m}\right)^2\frac{3g_SF_\pi}{M_\sigma^4}(1-C_\chi)+\left(\frac{M_\pi^2}{m}\right)^2\kappa_{NS}\frac{F_\pi^2}{M_\sigma^4}+\left(\frac{M_\pi^2}{m}\right)^2\frac{1}{M_\sigma^2}\frac{d}{dM_\pi^2}\left(M_0\frac{\partial\Sigma^{(\pi)}}{\partial M}\right)+\frac{d^2\Sigma^{(\pi)}}{dm^2} \\ &\equiv\chi_N^{(s)}+\chi_N^{(s\pi)}+\left(\frac{M_\pi^2}{m}\right)^2\frac{d^2\Sigma^{(\pi)}(M_0;m)}{d(M_\pi^2)^2}\equiv\chi_N^{(s)}+\chi_N^{(s\pi)}+\chi_N^{(\pi)}. \tag*{(28)} \end{aligned}

where we have used Eq. (24). One can split the scalar susceptibility into a non pionic (χN(s))(\chi_N^{(s)}), a mixed scalar field-pionic (χN(sπ))(\chi_N^{(s\pi)}) and a purely pionic (χN(π))(\chi_N^{(\pi)}) piece. The first two contributions in the second line of Eq. (28) with Mσ2∼4M02M_\sigma^2\sim4M_0^2 (considering small MπM_{\pi} and G2G_{2} in Eq. (16)), gives χN(s)\chi_N^{(s)} as:

χN(s)=−(Mπ2m)2FπgSMσ4(3−2C~L)withC~L=MNgSFπC+32Cχ(29)\chi_N^{(s)}=-\left(\frac{M_{\pi}^{2}}{m}\right)^{2}\frac{F_{\pi}g_S}{M_{\sigma}^{4}}\left(3-2\widetilde{C}_L\right)\quad\text{with}\quad\widetilde{C}_L=\frac{M_N}{g_S F_{\pi}}C+\frac{3}{2}C_{\chi} \tag*{(29)}

As for the case of gSg_S, the nucleon susceptibility κNS\kappa_{NS} may receive a contribution from the pion-self-energy; again the contribution to the dimensionless CC parameter is very small if the vertex correction is omitted. The mixed scalar field-pionic susceptibility originating from the scalar field (i.e, the constituent quark mass) dependence of the pionic self-energy,

χN(sπ)=−(Mπ2m)2FπMσ2∂∂s(σN(π)Mπ2),(30)\chi_N^{(s\pi)}=-\left(\frac{M_{\pi}^{2}}{m}\right)^{2}\frac{F_{\pi}}{M_{\sigma}^{2}}\frac{\partial}{\partial s}\left(\frac{\sigma_N^{(\pi)}}{M_{\pi}^{2}}\right), \tag*{(30)}

was ignored in our previous works. Using a sharp cutoff in the expression of the nucleon pionic self-energy, it can be shown analytically that this term is negligible compared to the other contributions to the susceptibility.

In view of the comparison with lattice QCD result it is very important to notice that the chiral susceptibility is governed by the particular combination:

χN∼(3−2C~L).(31)\chi_N\sim\left(3-2\widetilde{C}_L\right). \tag*{(31)}

to be compared with the particular combination entering the expression of the three-body repulsive contribution (21) to the binding energy per nucleon:

E(3b)A∼(2C~3−1)ns2withC~3=MNgSFπC+12Cχ.(32)\frac{E^{(3b)}}{A}\sim\left(2\widetilde{C}_3-1\right)n_s^2\quad\text{with}\quad\widetilde{C}_3=\frac{M_N}{g_S F_{\pi}}C+\frac{1}{2}C_{\chi}. \tag*{(32)}

Limiting ourselves to the pure Lσ\sigmaM case Cχ=0C_{\chi}=0, inducing C~L=C~3\widetilde{C}_L=\widetilde{C}_3, the susceptibility χN\chi_N (31) and the three-body repulsive contribution (21) are directly related, as found in our previous works, e.g., [16]. This constitutes a very important result linking chiral properties of the nucleon to the saturation mechanism. In the general case where Cχ≠0C_{\chi}\ne0, there is still a strong link between the susceptibility and the three-body repulsive contribution.

Constraints from Lattice-QCD

Those chiral properties of the nucleon, associated with explicit chiral symmetry breaking, namely the first and second derivatives of the nucleon mass with respect to the current quark mass, are thus very sensitive to the modeling of the nucleon. We have also shown that the scalar coupling constant, gSg_S, and the nucleon response parameter, CNC_N (or CC or κNS\kappa_{NS}), depend on the quark substructure and the confinement mechanism as well as the effect of spontaneous chiral symmetry breaking. We will now show how they can be constrained by lattice data.

The nucleon mass, as well as other intrinsic properties of the nucleon (sigma term, chiral susceptibilities), are QCD quantities which are in principle obtainable from lattice simulations. The problem is that lattice calculations of this kind are still difficult for small quark masses, or equivalently small pion mass MπM_{\pi}. Here MπM_{\pi} represents the pion mass to leading order in the quark mass (i.e., ignoring the NLO chiral logarithm correction), Mπ2=2mB=−2m⟨q‾q⟩χL/F2M_{\pi}^{2}=2mB=-2m\langle\overline{q}q\rangle\chi_L/F^{2} (GOR relation). The quantities FF (the pion decay constant in the chiral limit) and BB are two low energy parameters appearing in chiral perturbation theory [39]. In practice MπM_{\pi} deviates numerically very little from the bosonised NJL pion mass MπM_{\pi}. Typically at the time of the publication of the pioneering work from the Adelaide group [19] (that we will call hereafter AD1), these LQCD limitations were m>50m>50 MeV and Mπ2>0.27M_{\pi}^{2}>0.27 GeV2^{2} (to be compared to the physical value, 0.020.02 GeV2^{2}). Hence a technique was needed to extrapolate the lattice data to the physical region. The difficulty of the extrapolation is linked to the non analytical behaviour of the nucleon mass as a function of mm (or equivalently Mπ2M_{\pi}^{2}) which comes from the pion cloud contribution. The idea of the Adelaide group, [18, 19, 20, 21] (papers referred hereafter as AD0, AD1, AD2 and AD3) was to separate the pion cloud self-energy, Σπ(Mπ,Λ)\Sigma_{\pi}(M_{\pi},\Lambda), from the rest of the nucleon mass and to calculate it with just one adjustable cutoff parameter Λ\Lambda entering the form factor. Actually different cutoff forms for the pion loops (Gaussian, dipole, monopole, sharp) were used with the adjustable parameter Λ\Lambda. This formulation of Chiral Perturbation Theory (ChiPT) is thus called the Finite Range Regulator (FRR) method. The remaining non pionic part is expanded in terms of powers of Mπ2M_{\pi}^{2} as follows:

MN(Mπ2)=a0+a2Mπ2+a4Mπ4+⋯+Σπ(Mπ,Λ)(33)M_N(M_{\pi}^{2})=a_0+a_2M_{\pi}^{2}+a_4M_{\pi}^{4}+\cdots+\Sigma_{\pi}(M_{\pi},\Lambda) \tag*{(33)}

where Σπ(Mπ,Λ)=Σ(π)(Mπ,Λ)+Σtad(π)(Mπ,Λ)\Sigma_{\pi}(M_{\pi},\Lambda)=\Sigma^{(\pi)}(M_{\pi},\Lambda)+\Sigma_{\mathrm{tad}}^{(\pi)}(M_{\pi},\Lambda).

In AD1, which incorporates in the analysis the effect of a tadpole contribution Σtad(π)(Mπ,Λ)\Sigma_{\mathrm{tad}}^{(\pi)}(M_{\pi},\Lambda), the best-fit value for a2a_{2} shows little sensitivity to the shape of the form factor, with a value a2≃1.5 GeV−1a_{2}\simeq1.5\ \mathrm{GeV}^{-1}, which corresponds to a non pionic piece of the light quark sigma commutator σN(s)=30 MeV\sigma_{N}^{(s)}=30\ \mathrm{MeV}. In AD0 (which is actually the preprint version of AD1) and in the more recent paper, AD3, the contribution of the tadpole was not considered. Depending on the precise method used in the lattice simulation, the preferred values for a2a_{2} was smaller, in the range a2≃1.0a_{2}\simeq1.0 to 1.2 GeV−11.2\ \mathrm{GeV}^{-1}. Notice that taking a2a_{2} in the range a2≃1.2a_{2}\simeq1.2 to 1.5 GeV−11.5\ \mathrm{GeV}^{-1} corresponds to a non pionic piece of the light quark sigma commutator σN(s)=24\sigma_{N}^{(s)}=24 to 30 MeV30\ \mathrm{MeV}.

In AD1, (which incorporates the effect of the pion tadpole) the best-fit value for a4a_{4} shows again little sensitivity to the shape of the form factor, with a values a4≃−0.5 GeV−3a_{4}\simeq-0.5\ \mathrm{GeV}^{-3}. In AD0 and AD3, depending on the precise method used in the lattice simulation, the preferred values for a4a_{4} was even smaller, in the range a4≃−0.2a_{4}\simeq-0.2 to −0.25 GeV−3-0.25\ \mathrm{GeV}^{-3}.

Ignoring the pion tadpole contribution to the nucleon mass, we assume that we can identify the pionic self-energy on the lattice with our model calculation described above. Consequently the first and second derivative of the non pionic piece of the lattice expansion,

QS,L(s)=∂MN(no pion)∂m=(Mπ2m)(a2+a4Mπ2)≃(Mπ2m)a2(34)Q_{S,L}^{(s)}=\frac{\partial M_{N}^{(\mathrm{no\ pion})}}{\partial m}=\left(\frac{M_{\pi}^{2}}{m}\right)(a_{2}+a_{4}M_{\pi}^{2})\simeq\left(\frac{M_{\pi}^{2}}{m}\right)a_{2} \tag*{(34)}
χN,L(s)=∂2MN(no pion)∂m2=(Mπ2m)2(2a4)(35)\chi_{N,L}^{(s)}=\frac{\partial^{2}M_{N}^{(\mathrm{no\ pion})}}{\partial m^{2}}=\left(\frac{M_{\pi}^{2}}{m}\right)^{2}(2a_{4}) \tag*{(35)}

can be identified with the non pionic piece of the scalar charge, see Eq. (26), and of the chiral susceptibility, see Eq. (30), derived above:

QS(s)≡QS,L(s) and χN(s)≡χN,L(s).(36)Q_{S}^{(s)}\equiv Q_{S,L}^{(s)}\ \text{and}\ \chi_{N}^{(s)}\equiv\chi_{N,L}^{(s)}. \tag*{(36)}

One arrives at the important result:

a2=FπgSMσ2(37)a_{2}=\frac{F_{\pi}g_{S}}{M_{\sigma}^{2}} \tag*{(37)}
a4=−FπgS2Mσ4(3−2C~L)withC~L=MNgSFπC+32Cχ.(38)a_{4}=-\frac{F_{\pi}g_{S}}{2M_{\sigma}^{4}}\left(3-2\widetilde{C}_{L}\right)\quad\text{with}\quad\widetilde{C}_{L}=\frac{M_{N}}{g_{S}F_{\pi}}C+\frac{3}{2}C_{\chi}. \tag*{(38)}

Our previous works [12, 13] coincide with these relations in the specific case of the LσM effective potential (Cχ=0C_{\chi}=0). They provide two constraints on the parameters of the confining model. Also notice that the model results on the rhs of the above equations should be rigorously understood with the various parameters calculated in the chiral limit which are in practice very close to their values at the physical current quark mass.

The very robust conclusion is that the lattice result is much smaller than the one obtained in a the simplistic linear sigma model (C=Cχ=0C=C_{\chi}=0), for which a4≃−3.5 GeV−3a_{4}\simeq-3.5\ \mathrm{GeV}^{-3}. Hence lattice data require a strong compensation from effects governing the three-body repulsive force needed for the saturation mechanism.

Discussion

The above results demonstrate that the lattice data a2a_{2} and a4a_{4}, themselves related to the chiral responses of the nucleon, bring severe constraints on the nuclear matter equation of state. This suggests to enter these quantities as an input of a Bayesian analysis to generate the probability distribution function for the nucleon response parameters gSg_{S} and CC. Such an analysis limited to the Hartree level has been performed in a recent work [22], but using the simplistic LσM, with an output for CC very close to C≃1.5C\simeq1.5, the obvious reason being the very small input value for a4a_{4} (see Eq. (38)). In a work in preparation [40], we will perform again the same kind of analysis but with the incorporation of the Fock terms (and in particular the pion and rho Fock terms in presence of short range correlation) first with the LσM and second with the enriched NJL chiral effective potential. As already mentioned, the problem of the analysis using the LσM chiral effective potential is a large value of the CC response parameter in strong disagreement with all the nucleon models calculation which predict a value of CC smaller and most of the time significantly smaller than one (recall the MIT bag value C≃0.5C\simeq0.5).

Just to have an insight on the effect of an enriched chiral effective potential we return to our original paper [11]. In this paper where the LσM was used we obtained correct saturation properties with C=1C=1 (see Figure 1 of [11]. We can retrospectively calculate the a2a_{2} and a4a_{4} parameters: we find a2=1.67 GeV−1a_{2} = 1.67\ \mathrm{GeV}^{-1} and a4=−1.48 GeV−3a_{4} = -1.48\ \mathrm{GeV}^{-3}. If the obtained a2a_{2} is not very far from the lattice values, a4a_{4} is in magnitude three times larger than the upper value compatible with lattice calculation. To see the effect of the NJL-like potential (via the parameter CχC_{\chi}), we simply incorporate the (1−Cχ)(1-C_{\chi}) correction in the cubic term of the Lσ\sigmaM chiral effective potential, fixing Cχ=0.44C_{\chi} = 0.44. Keeping all the other parameters at their original value, we take C=0.78C = 0.78, so as to keep the same value of the repulsive three-body force, i.e., C~3=C+Cχ/2=1\widetilde{C}_{3} = C + C_{\chi}/2 = 1 (21). The saturation points is only slightly modified (see Figure 2) but now C~L=C+3Cχ/2=1.44\widetilde{C}_{L} = C + 3C_{\chi}/2 = 1.44 (38) and the a4a_{4} parameter becomes very close to zero, a4=−0.1 GeV−3a_{4} = -0.1\ \mathrm{GeV}^{-3}, in much better agreement with lattice data.

Figure 2. Full line: original calculation of the EOS [11] with C=1C = 1, Cχ=0C_{\chi} = 0. Dotted line: New calculation with the same parameters but C=0.78C = 0.78, Cχ=0.44C_{\chi} = 0.44. The density is scaled by the normal nuclear matter density.

Conclusions

The nuclear matter properties originate from the fundamental theory of the strong interaction and the aim of this manuscript is to investigate how this microscopic origin can be implemented in the modeling of nuclear matter. Of particular importance on the QCD side are the quark confinement mechanism and the chiral potential associated to the chirally broken QCD vacuum.

In this article we use an enriched chiral effective potential, based on the NJL model, in place of the Lσ\sigmaM employed in our previous phenomenological works. This significantly increases the agreement with LQCD data together with expected model values of the nucleonic response parameter CC. Note that this conclusion should be confirmed by a more thorough analysis (work in preparation [40]).

Hence the fundamental QCD theory and nuclear matter modeling are linked by, on the one hand the LQCD data a2a_{2} and a4a_{4} and on the other hand what we have called the ”QCD connected parameters”, namely the response parameters GsG_{s} and CNC_{N}. Specifically we have shown that a particular combination of CC and CχC_{\chi} (C~L\widetilde{C}_{L}) is constrained by LQCD, which constitutes one of the main result of this paper. In addition a closely related combination (C~3=C+Cχ/2\widetilde{C}_{3} = C + C_{\chi}/2) governs the repulsive three-body force ensuring the mechanism mechanism.

Indeed these results provide a link between chiral properties of the nucleon and the saturation mechanism, already obtained in our previous works, but limited to the pure Lσ\sigmaM case. Further investigations of these results shall be perform to understand more globally how they modify the properties of nuclear matter. Works in this direction is being performed.

References

References

  1. [1]Brian D. Serot and John D. Walecka, The Relativistic Nuclear Many-Body Problem, Advances in Nucl. Phys. 16, 1 (1986).
  2. [2]B.D. Serot, J.D. Walecka, Int. J. Mod. Phys. E16, 15 (1997).
  3. [3]G. Chanfray and J. Margueron, Phys. Rev. C 102, 024331 (2020).
  4. [4]G. Chanfray, M. Ericson, P.A.M. Guichon, Phys. Rev. C 63, 055202 (2001).
  5. [5]J. Boguta, Phys. Lett. 120B, 34 (1983).
  6. [6]A.K. Kerman and L.D. Miller in “Second High Energy Heavy Ion Summer Study”, LBL-3675, 1974.
  7. [7]W. Bentz and A.W. Thomas, Nucl. Phys. A696, 138 (2001).
  8. [8]G. Chanfray, Nucl. Phys. A 721 (2003) 76c.
  9. [9]P.A.M. Guichon, Phys. Lett. B 200, 235 (1988).
  10. [10]P.A.M. Guichon and A.W. Thomas, Phys. Rev. Lett. 93, 132502 (2004); P.A.M. Guichon, H.H. Matevosyan, N. Sandulescu and A.W. Thomas, Nucl. Phys. A 772, 1 (2006); R. Stone, P.A. M. Guichon, P. G. Reinhard and A.W. Thomas, Phys. Rev. Lett. 116, 092511 (2016).
  11. [11]G. Chanfray, M. Ericson, Eur. Phys. J. A 25, 151 (2005).
  12. [12]G. Chanfray, M. Ericson, Phys. Rev. C 75, 015206 (2007).
  13. [13]E. Massot, G. Chanfray, Phys. Rev. C 78, 015204 (2008).
  14. [14]E. Massot, G. Chanfray, Phys. Rev. C 80, 015202 (2009).
  15. [15]E. Massot, J. Margueron and G. Chanfray, EPL 97, 39002 (2012).
  16. [16]M. Ericson and G. Chanfray, Eur. Phys. J. A 34, 215 (2007).
  17. [17]G. Chanfray and M. Ericson, Proceedings of the Eighteenth PANIC Conférence (Eliat, Israel 2008) ) eds I. Tserruya, A. Gal, D. Ashery. arXiv: 0901.2479 [nucl-th]; G. Chanfray and M. Ericson, Proceedings of the workshop “Achievements and New Directions in Subatomic Physics”, Adelaide, Australia, february 2010. Published in AIP Conf.Proc.1261:31-36, 2010. arXiv:1004.1265 [nucl-th].
  18. [18]D. B. Leinweber, A. W. Thomas, and R. D. Young, arXiv:hep-lat/0302020.DOI
  19. [19]D. B. Leinweber, A. W. Thomas, and R. D. Young, Phys. Rev. Lett. 92, 242002 (2004).
  20. [20]A. W. Thomas, P. A. M. Guichon, D. B. Leinweber and R. D. Young, Progr. Theor. Phys. Suppl. 156 (2004) 124; nucl-th/0411014.
  21. [21]W. Armour, C.R. Alton,D. B. Leinweber, A. W. Thomas, and R. D. Young, Nucl. Phys. A 840, 97 (2010).
  22. [22]R. Somasundaram, J. Margueron, G. Chanfray, H. Hansen, Eur.Phys.J. A 58, 5 (2022).
  23. [23]G. Chanfray, M. Ericson, Phys. Rev. C 83, 015204 (2011).
  24. [24]Lai-Him Chan, Phys. Rev. Lett.57, 1199 (1986).
  25. [25]G. Chanfray, H. Hansen and J. Margueron, work (NJLFCM) in preparation.
  26. [26]Yu. A. Simonov, Phys. At. Nucl. 60, 2069 (1997; Few-BodySyst. 25, 45 (1998).
  27. [27]Yu. A. Simonov, Phys. Rev. D 65, 094018 (2002).
  28. [28]Yu. A. Simonov, J. A. Tjon, Phys. Rev. D 62, 014501 (2000).
  29. [29]Yu. A. Simonov, J. A. Tjon, J. Weda, Phys. Rev. D 65, 094013 (2002).
  30. [30]Yu.A.Simonov, Phys. At.Nucl. 67, 846 (2004), hep-ph/0302090; Yu.A.Simonov, Phys. At. Nucl. 67, 1027 (2004), hep-ph/0305281; Yu.A.Simonov, Int. Mod. Phys. A 31, 165016 (2016), arXiv:1509.06930.
  31. [31]A. Di Giacomo, H.G. Dosch, V.I. Shevchenko, Yu. A. Simonov, Physics Reports 372, 319 (2002).
  32. [32]M.C. Birse, Phys. Rev. C 51 (1995) R1083; K. Saito, A.W. Thomas, Phys. Rev. C 52 (1995) 2789.
  33. [33]S.J.Wallace, F. Gross, J.A. Tjon, Phys. Rev. Lett. 74 (1995) 228; S.J.Wallace, Nucl. Phys. A 631 (1998) 137c.
  34. [34]S. Typel and H. H. Wolter, Nucl. Phys. A 656, 331(1999).
  35. [35]E. N. E. van Dalen and H. Muther, Phys. Rev. C 84, 024320 (2011), arXiv:1106.3157 [nucl-th].DOI
  36. [36]G. A. Lalazissis, J. Konig, and P. Ring, Phys. Rev. C 55, 540 (1997), arXiv:nucl-th/9607039.DOI
  37. [37]S. Theberge, A. W. Thomas, and G. A. Miller, Phys. Rev. D 22, 2838 (1980); S. Theberge, G. A. Miller, and A. W. Thomas, Can. J. Phys. 60, 59 (1982); A.W. Thomas, Adv. Nucl. Phys. 13, 1 (19XX).
  38. [38]S.N.Jena, M.R. Behera, S. Panda, Phys. Rev. D55, 291 (1997).
  39. [39]Heinrich Leutwyler (2012) Chiral perturbation theory.Scholarpedia, 7(10):8708.DOI
  40. [40]M. Chamseddine, J. Margueron, G. Chanfray and H. Hansen, work in preparation.

Paper details

vlnc:W_sw7e5frx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents