An extended MDI interaction for transport models and neutron stars Si-Pei Wang, Lie-Wen Chen Abstract The momentum-dependent interaction (MDI) model, which has been widely used in microscopic transport models for heavy-ion collisions (HICs), is extended to include three different momentum-dependent terms and three zero-range density-dependent terms, dubbed as MDI3Y model. Compared to the MDI model, the single-nucleon potential in the MDI3Y model exhibits more flexible momentum-dependent behaviors, e.g., (1) the nucleon optical model potential U 0 U_{0} in symmetric nuclear matter can have very different high energy behaviors while nicely describe the known data at lower energies; (2) the isoscalar nucleon effective mass m s ∗ m_{s}^{\ast} can have a peak structure at subsaturation densities; (3) the nuclear symmetry potential U sym U_{\mathrm{sym}} can display a non-monotonic momentum dependence. Furthermore, the inclusion of three zero-range density-dependent interactions follows the idea of Fermi momentum expansion, allowing more flexible variation for the largely uncertain high-density behaviors of nuclear matter equation of state (EOS), especially the symmetry energy. Moreover, we also obtain the corresponding Skyrme-like energy density functional through density matrix expansion of the finite-range exchange interactions. Based on the MDI3Y model, we construct four interactions with the same symmetry energy slope parameter L = 35 L=35 MeV but different momentum dependence of U sym U_{\mathrm{sym}} , by fitting the empirical nucleon optical potential, the empirical properties of symmetric nuclear matter, the microscopic calculations of pure neutron matter EOS and the astrophysical constraints on neutron stars. In addition, two interactions with L = 55 L=55 and 75 75 MeV are also constructed for comparison. Using these MDI3Y interactions, we study the properties of nuclear matter and neutron stars. These MDI3Y interactions, especially those with non-monotonic momentum dependence of U sym U_{\mathrm{sym}} , will be potentially useful in transport model analyses of HICs data to extract nuclear matter EOS and the isospin splitting of nucleon effective masses. I INTRODUCTION The EOS of nuclear matter, usually defined as energy (or pressure) vs density, is of fundamental importance in nuclear physics and astrophysics Danielewicz et al. 2002 ; Lattimer and Prakash 2004 ; Oertel et al. 2017 . Theoretically, it is still a big challenge to calculate the nuclear matter EOS from first principle, especially at suprasaturation densities, due to the complicated nonperturbative feature of QCD Brambilla et al. 2014 , although progress in microscopic many-body calculations, see, e.g., Refs. Tews et al. 2013 ; Wellenhofer et al. 2015 ; Drischler et al. 2020 ; Gandolfi et al. 2012 ; Wlazłowski et al. 2014 ; Roggero et al. 2014 ; Tews et al. 2016 ; Akmal et al. 1998 ; Baldo and Fukukawa 2014 ; Carbone et al. 2014 , has improved the predictions of the EOS of pure neutron matter (PNM) (main limited at lower densities). Therefore, our knowledge on nuclear matter EOS so far obtained is mainly from model analyses on data from terrestrial experiments or astrophysical observations. In the past decades, by analyzing the giant resonances of finite nuclei Blaizot 1980 ; Youngblood et al. 1999 ; Shlomo et al. 2006 ; Li et al. 2007 ; Garg and Colò 2018 ; Li et al. 2023 , as well as the collective flows Danielewicz et al. 2002 ; Le Fèvre et al. 2016 ; Cozma 2024 and the kaon production Aichelin and Ko 1985 ; Fuchs et al. 2001 ; Hartnack et al. 2006 ; Fuchs 2006 in HICs, the EOS of isospin symmetric nuclear matter (SNM) has been relatively well constrained even up to about five times the nuclear saturation density \rho_{0} . On the other hand, for the isospin-dependent part of nuclear matter EOS, referred to as the symmetry energy E_{\mathrm{sym}} , its density behavior below and around \rho_{0} has been reasonably well constrained by analyzing finite nuclei properties Chen et al. 2005a ; Centelles et al. 2009 ; Chen et al. 2010 ; Agrawal et al. 2012 ; Zhang and Chen 2013 ; Brown 2013 ; Roca-Maza et al. 2013 ; Zhang and Chen 2014 ; Danielewicz and Lee 2014 ; Zhang and Chen 2015 ; Danielewicz et al. 2017 ; Xu et al. 2020 ; Qiu et al. 2024 and observables in HICs at energies less than about 100 MeV/nucleon ( Chen et al. 2005b ; Li and Chen 2005 ; Kowalski et al. 2007 ; Shetty et al. 2007 ; Tsang et al. 2009 ; Wada et al. 2012 ; Morfouace et al. 2019 ; Zhang et al. 2020 ) , with its suprasaturation density behavior remaining among the most uncertain properties of nuclear matter (see, e.g., Ref. Sorensen et al. 2024 for a recent review). In terrestrial labs, the intermediate and high energy HICs induced by neutron-rich nuclei provides a unique tool to explore the suprasaturation density behaviors of the E_{\mathrm{sym}} , as the dense neutron-rich nuclear matter may be formed during the collisions. In nature, the neutron stars and their mergers offer ideal sites to study the high-density E_{\mathrm{sym}} since the interior of neutron stars are dominated by dense and extremely neutron-rich matter. To investigate the HICs and neutron star properties, one usually adopts the microscopic transport models and energy density functionals based on the same in-medium nuclear effective interactions, which generally depend on the medium baryon density and isospin asymmetry as well as the momentum and isospin of the interacting particles. Therefore, the in-medium nuclear effective interactions play a decisive role in extracting physics information from analyzing data of HICs and multimessenger signals of neutron stars and their mergers. In this work, we mainly focus on the momentum dependence of the in-medium nuclear effective interactions. Microscopic transport models such as the Boltzmann-Uehling-Uhlenbeck (BUU) equation Bertsch and Das Gupta 1988 and quantum molecular dynamics (QMD) model Aichelin 1991 have been developed to describe the dynamics of HICs and to extract the EOS of dense nuclear matter from the collision observables. In recent years, the Transport Model Evaluation Project has been comparing the predictions from different models under controlled conditions, and aims to extract reliable constraints on the EOS from HICs Xu et al. 2016 ; Zhang et al. 2018 ; Ono et al. 2019 ; Colonna et al. 2021 ; Wolter et al. 2022 ; Xu et al. 2024 . The basic input in the one-body transport models is the single-nucleon potential (nuclear mean-field potential) Bertsch and Das Gupta 1988 , which generally depends on the baryon density and the isospin asymmetry of the medium as well as the isospin and momentum of the particle and can be obtained from the in-medium nuclear effective interactions Behera and Satpathy 1979 ; Decharge and Gogny 1980 ; Wiringa 1988 ; Behera et al. 1998 . From the Brueckner theory, the momentum dependence mainly arises from the non-locality of the exchange interaction Brueckner 1955 ; Jeukenne et al. 1976 ; Mahaux et al. 1985 ; Bertsch and Das Gupta 1988 ; Li et al. 2018 . Experimentally, the momentum dependence of single-nucleon potentials is evident from the measured momentum/energy dependence of nucleon optical potential. It has been shown that the momentum dependence of nuclear mean-field potential affects not only the dynamics of HICs Aichelin et al. 1987 ; Gale et al. 1990 ; Zhang et al. 1994 ; Greco et al. 1999 ; Danielewicz 2000 ; Larionov et al. 2000 ; Persram and Gale 2002 ; Chen et al. 2005b ; Li and Chen 2005 ; Li et al. 2008 , but also the thermodynamic properties of nuclear matter Csernai et al. 1992 ; Mishra et al. 1993 ; Behera et al. 2002 ; Moustakidis 2007 ; Xu et al. 2008 ; Behera et al. 2011 ; Xu et al. 2015 ; Constantinou et al. 2015 ; Xu et al. 2019 . It should be mentioned that for the zero-range Skyrme interaction Vautherin and Brink 1972 , including higher-order derivative terms Carlsson et al. 2008 ; Raimondi et al. 2011 can introduce additional momentum dependence, which allows for its application in the transport model simulations for HICs, e.g., up to beam energy of about 2 GeV per nucleon if the N5LO order pseudopotential is used Wang et al. 2018 ; Yang et al. 2024 ; Wang et al. 2024 ; Wang et al. 2025 . However, the single-nucleon potential in these extended Skyrme interactions will eventually diverge at higher energies due to their polynomial structure, and consequently, the nucleon effective mass in the high-energy limit does not approach the bare mass as it should. At this point, the finite-range interactions are more realistic for HICs simulations with increasing energy. In 1980s, the momentum-dependent mean-field potential was first introduced into the BUU model by Gale, Bertsch, and Das Gupta (GBD) Gale et al. 1987 , and a more realistic parametrization was later proposed by Welke et al. Welke et al. 1988 based on a finite-range force with Yukawa form, which was referred to as the momentum-dependent Yukawa interaction (MDYI). To incorporate the isospin dependence of the mean-field potential, usually expressed as the Lane potential Lane 1962 , Bombaci Bombaci 2001 extended the GBD interaction into the BGBD interaction. On the other hand, the MDYI interaction was further improved by Das et al. Das et al. 2003 , leading to the so-called MDI interaction, which has been widely used in various transport models (see Refs. Li et al. 2008 ; Chen et al. 2014 for reviews). For the finite-range interaction with Yukawa form, a parametrization with two Yukawa couplings was introduced and applied in QMD approach by Maruyama et al. Maruyama et al. 1998 . Furthermore, the famous Michigan three-range Yukawa (M3Y) interaction was originally established from a bare nucleon-nucleon interaction such as Reid Bertsch et al. 1977 and Paris Anantaraman et al. 1983 , by fitting the Yukawa functions to the G-matrix. To improve the description of nuclear matter properties with M3Y interaction, density-dependent coupling constants were introduced Khoa and von Oertzen 1993 ; Khoa and von Oertzen 1995 and applied to the folding potential for nuclear reactions Khoa et al. 1996 ; Khoa et al. 1997 . On the other side, Nakada introduced zero-range DD terms in the M3Y interaction and developed the M3Y-P n interaction family Nakada 2003 ; Nakada 2008a ; Nakada 2010 ; Nakada 2013 with constraints from nuclear matter and finite nuclei. The properties of nuclear matter and neutron stars with these M3Y-type interactions have been investigated in Refs. Than et al. 2009 ; Loan et al. 2011 ; Davesne et al. 2025 . In the present work, we extend the MDI model to a new model called MDI3Y, by including three different MD terms and three zero-range DD terms. The three different MD terms are obtained from the exchange contribution of the central term in the M3Y interaction. This extension significantly enhances the flexibility in the momentum dependence of both the single-nucleon potential and the symmetry potential, which exhibits several new features compared to the traditional MDI interaction. Firstly, the nucleon optical model potential U_{0} in symmetric nuclear matter at \rho_{0} can have very different high energy behaviors while nicely describe the empirical nucleon optical potential below about 1 GeV Hama et al. 1990 ; Cooper et al. 1993 . Secondly, the isoscalar nucleon effective mass m_{s}^{\ast} can have a peak structure at subsaturation densities with a maximum value larger than the bare nucleon mass. Note that the behavior of m_{s}^{\ast} is only determined by fitting the single-nucleon potential to the nucleon optical potential Hama et al. 1990 ; Cooper et al. 1993 . The m_{s}^{\ast}/m>1 at low densities could have significant effects in the mean-field calculations of finite nuclei Brown et al. 1963 ; Barranco and Treiner 1981 ; Mughabghab and Dunford 1998 ; Shlomo 1992 ; Farine et al. 2001 ; Tondeur et al. 2000 ; Goriely et al. 2001 . Thirdly, the nuclear symmetry potential U_{\mathrm{sym}} can display a non-monotonic momentum dependence by independently adjusting the linear isospin splitting coefficient \Delta m_{1}^{\ast}(\rho_{0}) for nucleon effective mass, the isovector nucleon effective mass m_{v}^{\ast}(\rho_{0}) , and the infinite momentum limit value U_{\mathrm{sym}}^{\infty} of the U_{\mathrm{sym}} at \rho_{0} . The non-monotonic U_{\mathrm{sym}} implies that the neutron-proton effective mass splitting m_{\rm n\text{-}p}^{\ast} may have different signs at different energies, which could be tested in transport model simulations for HICs by investigating the light particle emission Li et al. 2006a ; Coupland et al. 2016 ; Zhang et al. 2014 ; Kong et al. 2015 ; Wang et al. 2023a . In addition, in the MDI3Y model, the many-body force is mimicked by three zero-range DD terms, following the idea of Fermi momentum expansion Patra et al. 2022 . Consequently, the density behaviors of the SNM EOS and E_{\mathrm{sym}} become highly flexible, allowing for reasonable descriptions of neutron star properties. We notice that when specific parameters are selected, the DD term in the MDI3Y model will reduce to the form in M3Y-P n model Nakada 2003 ; Nakada 2008a ; Nakada 2010 ; Nakada 2013 . Based on the MDI3Y model, we construct interactions with L=35 , 55 , and 75 MeV for the density slope parameter of the symmetry energy. Moreover, for the case of L=35 MeV, we also employ four different U_{\mathrm{sym}} , including both monotonic and non-monotonic types. In future, these new interactions can be used in transport model simulations for HICs at intermediate and high energies, through which one may obtain a more comprehensive understanding of SNM EOS and E_{\mathrm{sym}} as well as U_{0} , U_{\mathrm{sym}} and m_{\rm n\text{-}p}^{\ast} . This paper is organized as follow: In Sec. II , we introduce the theoretical framework of the MDI3Y model, and we present the expressions of energy density and single-nucleon potential under general nonequilibrium conditions. In Sec. III , we discuss the fitting procedure to construct MDI3Y interactions, along with the data/constraints used for fitting. In Sec. IV , we present several properties for the cold nuclear matter, including the bulk properties, the single-nucleon potential, the first-order and second-order symmetry potentials, and the nucleon effective masses. The neutron star properties for the MDI3Y interactions are also presented. In Sec. V , we summarize this work and make a brief outlook. In Appendix A , we present the relations between the parameters in the MDI3Y interaction and M3Y-P n interaction. In Appendix B , we present some details of the derivations for the energy density using Hartree-Fock approximation, and we also discuss the density matrix expansion for the MDI3Y interaction. For completeness, we provide the expressions related to the main text in Appendix C . II Theoretical framework II.1 The MDI3Y interactions We assume that the central part V^{C} of the interaction between two nucleons located at \vec{r}_{1} and \vec{r}_{2} is composed of three Yukawa interactions: \small V^{C}\left(\vec{r}_{1},\vec{r}_{2}\right)=\sum_{i=1}^{3}\left(W_{i}+B_{i}P_{\sigma}-H_{i}P_{\tau}-M_{i}P_{\sigma}P_{\tau}\right)\frac{e^{-\mu_{i}\left|\vec{r}_{1}-\vec{r}_{2}\right|}}{\mu_{i}\left|\vec{r}_{1}-\vec{r}_{2}\right|}, (1) where P_{\sigma} and P_{\tau} are the spin and isospin exchange operators, respectively. Following the idea of Fermi momentum expansion Patra et al. 2022 ; Wang et al. 2024 , three zero-range DD terms are introduced to mimic the many-body force as follows: V^{\mathrm{DD}}=\sum_{n=1,3,5}\frac{1}{6}t_{3}^{\left[n\right]}\left(1+x_{3}^{\left[n\right]}P_{\sigma}\right)\rho^{n/3}(\vec{r})\delta(\vec{r}_{1}-\vec{r}_{2}), (2) where \vec{r}=(\vec{r}_{1}+\vec{r}_{2})/2 . As a result, the EOS of nuclear matter can be expressed as a power series in \rho^{1/3} [see Eq. ( 18 ) and the relevant discussion in Sec. II.3 ]. It should be noted that including additional terms with larger n in Eq. ( 2 ) can provide more flexible high-density behavior of nuclear matter EOS, although the DD term up to n=5 is sufficient to satisfy the current constraints. In general, the complete nuclear effective interaction also includes spin-orbit and tensor components, which are important for nuclear structure calculations Nakada and Sato 2002 ; Nakada 2006 ; Nakada 2008b . In the present work, we focus on the spin-averaged quantities, where the spin-dependent terms make no contribution. Therefore, the MDI3Y interaction used in this work is written as v_{\mathrm{MDI3Y}}=V^{C}+V^{\mathrm{DD}}. (3) It should be noted that Eqs. ( 1 ) and ( 2 ) can be equivalently expressed as the combinations of different spin-isospin (S-T) channels, as the forms in the M3Y-P n interactions Nakada 2003 ; Nakada 2008a ; Nakada 2010 ; Nakada 2013 . In Appendix A , we provide the relations between the parameters for these two representations, which could be useful for future investigations about the S-T structures and properties of finite nuclei. II.2 Hamiltonian density and single-nucleon potential in one-body transport model The total energy of the nuclear system can be obtained from \displaystyle E= \displaystyle\sum_{i}\langle i|\frac{p^{2}}{2m}|i\rangle+\frac{1}{2}\sum_{i,j}\langle ij|v_{\mathrm{MDI3Y}}(1-P_{M}P_{\sigma}P_{\tau})|ij\rangle (4) \displaystyle= \displaystyle\int\mathcal{H}(\vec{r})\mathrm{d}^{3}r, where P_{M} is the space exchange operator. The m is the average nucleon mass in free-space throughout this paper. During the HICs, the nucleons are generally far from thermal equilibrium, and in transport models they are described by the phase-space distribution function (Wigner function) f_{\tau}\left(\vec{r},\vec{p}\right) , with \tau=1 (or n) for neutrons and -1 (or p) for protons. The f_{\tau}\left(\vec{r},\vec{p}\right) becomes the Fermi-Dirac distribution in momentum space when the system approaches to equilibrium. Based on Eq. ( 4 ), the total Hamiltonian density \mathcal{H}(\vec{r}) of the collision system can be expressed as (see Appendix B.1 for details) \mathcal{H}(\vec{r})=\mathcal{H}^{\mathrm{kin}}(\vec{r})+\mathcal{H}^{\mathrm{loc}}(\vec{r})+\mathcal{H}^{\mathrm{MD}}(\vec{r})+\mathcal{H}^{\mathrm{DD}}(\vec{r}), (5) where \mathcal{H}^{\mathrm{kin}}(\vec{r}) , \mathcal{H}^{\mathrm{loc}}(\vec{r}) , \mathcal{H}^{\mathrm{MD}}(\vec{r}) , and \mathcal{H}^{\mathrm{DD}}(\vec{r}) are, respectively, the kinetic, local, MD, and DD terms. The kinetic term can be expressed as \mathcal{H}^{\mathrm{kin}}\left(\vec{r}\right)=\sum_{\tau=n,p}\int{\rm d}^{3}p\frac{p^{2}}{2m}f_{\tau}\left(\vec{r},\vec{p}\right). (6) The local and MD terms come from the direct and exchange terms of the finite-range Yukawa interactions. Within Hartree-Fock approximation, the local term can be expressed as \displaystyle\mathcal{H}^{\mathrm{loc}}(\vec{r})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}\int\mathrm{d}^{3}s\frac{e^{-\mu_{i}s}}{\mu_{i}s}\bigg\{(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2}) (7) \displaystyle\times\Big[\rho_{n}(\vec{r}+\frac{\vec{s}}{2})\rho_{n}(\vec{r}-\frac{\vec{s}}{2})+\rho_{p}(\vec{r}+\frac{\vec{s}}{2})\rho_{p}(\vec{r}-\frac{\vec{s}}{2})\Big] \displaystyle+2(W_{i}+\frac{B_{i}}{2})\rho_{n}(\vec{r}+\frac{\vec{s}}{2})\rho_{p}(\vec{r}-\frac{\vec{s}}{2})\bigg\}, where \vec{s}=(\vec{r}_{1}-\vec{r}_{2}) . Under the approximation of unform nuclear matter with constant density by assuming the densities vary slowly in space, Eq. ( 7 ) can be obtained approximately as \mathcal{H}^{\mathrm{loc}}(\vec{r})=\frac{A_{l}}{2\rho_{0}}\left(\rho_{n}^{2}+\rho_{p}^{2}\right)+\frac{A_{u}\rho_{n}\rho_{p}}{\rho_{0}}, (8) which is exactly equivalent to the local term in the original MDI model Das et al. 2003 ; Chen et al. 2014 and will be also adopted in the following for the MDI3Y model calculations. Furthermore, the MD term can be expressed as \mathcal{H}^{\mathrm{MD}}(\vec{r})=\frac{1}{\rho_{0}}\sum_{i=1}^{3}\sum_{\tau,\tau^{\prime}}C_{\tau\tau^{\prime},i}\iint\mathrm{d}^{3}p\mathrm{~d}^{3}p^{\prime}\frac{f_{\tau}(\vec{r},\vec{p})f_{\tau^{\prime}}\left(\vec{r},\vec{p}^{\,\prime}\right)}{1+\left(\vec{p}-\vec{p}^{\,\prime}\right)^{2}/\Lambda_{i}^{2}}. (9) Based on Eq. ( 2 ), the DD term can be expressed as \displaystyle\mathcal{H}^{\mathrm{DD}}(\vec{r})= \displaystyle\sum_{n=1,3,5}\frac{1}{24}t_{3}^{[n]}\left[\left(2+x_{3}^{[n]}\right)\rho^{2}\right. (10) \displaystyle\left.-\left(2x_{3}^{[n]}+1\right)(\rho_{n}^{2}+\rho_{p}^{2})\right]\rho^{n/3}, where \rho=\rho_{n}+\rho_{p} . Comparing Eqs. ( 8 ) and ( 9 ) with Eqs. ( 68 ) and ( 73 ) in Appendix, we can obtain the relations among the parameters in the energy density of uniform nuclear matter and those in the two-body interaction: \displaystyle\Lambda_{i} \displaystyle=\hbar\mu_{i}~~(i=1,2,3), (11) \displaystyle\frac{1}{\rho_{0}}A_{l} \displaystyle=\sum_{i=1}^{3}\left[\frac{4\pi}{\mu_{i}^{3}}\left(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2}\right)\right], (12) \displaystyle\frac{1}{\rho_{0}}A_{u} \displaystyle=\sum_{i=1}^{3}\left[\frac{4\pi}{\mu_{i}^{3}}\left(W_{i}+\frac{B_{i}}{2}\right)\right], (13) \displaystyle\frac{1}{\rho_{0}}C_{l,i} \displaystyle=\frac{2\pi}{\mu_{i}^{3}}\left(M_{i}+\frac{H_{i}}{2}-B_{i}-\frac{W_{i}}{2}\right)~~(i=1,2,3), (14) \displaystyle\frac{1}{\rho_{0}}C_{u,i} \displaystyle=\frac{2\pi}{\mu_{i}^{3}}\left(M_{i}+\frac{H_{i}}{2}\right)~~(i=1,2,3), (15) where C_{l,i}\equiv C_{\tau,\tau,i} and C_{u,i}\equiv C_{\tau,-\tau,i} . The force-ranges \mu_{i}^{-1} correspond to the Compton wavelengths of the mesons with masses of \Lambda_{i}c . The A_{l} and A_{u} correspond to the direct term, while C_{l,i} and C_{u,i} describe the exchange term. It can be seen from Eq. ( 68 ) that the direct contributions from different Yukawa interactions are degenerate for uniform nuclear matter. This leads to the fact that in the MDI3Y model for spin-saturated nuclear matter considered in the present work, there are 17 model parameters, namely, A_{l} , A_{u} ; C_{l,i} , C_{u,i} , \Lambda_{i} ( i=1,2,3 ); t_{3}^{[n]} , x_{3}^{[n]} ( n=1,3,5 ), while the number of parameters in Eqs. ( 1 ) and ( 2 ) is 21 . To fully determine the parameters for the two-body interaction, one can impose some additional constraints, e.g., the S-T structure of nuclear matter energy density, which can be obtained from microscopic calculations for spin-unsaturated nuclear matter Baldo et al. 1997 ; Baldo et al. 2014 ; Davesne et al. 2016 ; Batail et al. 2023 . It is interesting to note that if the two-body interaction is taken as one Yukawa term, its parameters can be uniquely determined by the corresponding MDI interaction Xu and Ko 2010 . It should be pointed out that Eqs. ( 6 ), ( 8 ), ( 9 ) and ( 10 ) fully express the Hamiltonian density for nucleonic system under general nonequilibrium condition for the MDI3Y model. Furthermore, the Hamiltonian density of the MDI3Y model can be reduced to the original Hamiltonian density of the MDI model by using one Yukawa term contribution in Eq. ( 9 ) and changing the DD contribution (i.e., Eq. ( 10 )) to the original DD contribution in the MDI model. In addition, one can see that the Hamiltonian density in the MDI3Y model can be obtained from Hartree-Fock calculation by uniform nuclear matter approximation for the direct term contribution of the finite-range interaction. Based on the Landau Fermi liquid theory, the single-nucleon potential can be calculated as U_{\tau}\left(\vec{r},\vec{p}\right)=\frac{\delta\mathrm{H}^{\mathrm{pot}}}{\delta f_{\tau}\left(\vec{r},\vec{p}\right)}=\frac{\partial\left[\mathcal{H}^{\mathrm{loc}}\left(\vec{r}\right)+\mathcal{H}^{\mathrm{DD}}\left(\vec{r}\right)\right]}{\partial\rho_{\tau}(\vec{r})}+\frac{\delta\mathrm{H}^{\mathrm{MD}}}{\delta f_{\tau}\left(\vec{r},\vec{p}\right)}, (16) where \mathrm{H}^{\mathrm{pot}}=\int\mathrm{d}\vec{r}\left[\mathcal{H}^{\mathrm{loc}}(\vec{r})+\mathcal{H}^{\mathrm{DD}}(\vec{r})+\mathcal{H}^{\mathrm{MD}}(\vec{r})\right] is the potential part of the Hamiltonian with \mathrm{H}^{\mathrm{MD}}=\int\mathrm{d}\vec{r}\mathcal{H}^{\mathrm{MD}}(\vec{r}) being the MD part. Substituting Eqs. ( 8 )-( 10 ) into Eq. ( 16 ), we then obtain the single-nucleon potential in nuclear matter with the MDI3Y interactions as \displaystyle U_{\tau}(\vec{r},\vec{p})= \displaystyle A_{l}\frac{\rho_{\tau}}{\rho_{0}}+A_{u}\frac{\rho_{-\tau}}{\rho_{0}}+\sum_{n=1,3,5}\left\{\frac{1}{12}t_{3}^{\left[n\right]}\left[\left(2+x_{3}^{\left[n\right]}\right)\rho(\vec{r})-\left(2x_{3}^{\left[n\right]}+1\right)\rho_{\tau}(\vec{r})\right]\rho(\vec{r})^{\frac{n}{3}}\right\} (17) \displaystyle+\sum_{n=1,3,5}\left\{\frac{t_{3}^{\left[n\right]}}{24}\frac{n}{3}\left[\left(2+x_{3}^{\left[n\right]}\right)\rho(\vec{r})^{2}-\left(2x_{3}^{\left[n\right]}+1\right)\sum_{\tau=n,p}\rho_{\tau}(\vec{r})^{2}\right]\rho(\vec{r})^{\frac{n}{3}-1}\right\} \displaystyle+\frac{2}{\rho_{0}}\sum_{i=1}^{3}\int\mathrm{d}^{3}p^{\prime}\frac{C_{l,i}~f_{\tau}(\vec{r},\vec{p}^{\,\prime})+C_{u,i}~f_{-\tau}\left(\vec{r},\vec{p}^{\,\prime}\right)}{1+\left(\vec{p}-\vec{p}^{\,\prime}\right)^{2}/\Lambda_{i}^{2}}. In Appendix B.2 , we discuss the density matrix expansion for the MDI3Y interaction, with the exchange contribution from finite-range interaction being approximated by a Skyrme-like zero-range interaction with density-dependent parameters. The resulting energy density functional and single-nucleon potential can be expressed in terms of local densities and their spatial gradients, which could be used to study finite nuclei properties Hofmann and Lenske 1998 , transition density of neutron star Xu and Ko 2010 ; Xu et al. 2009a , as well as in transport models Wang et al. 2019 ; Wang et al. 2020 to self-consistently obtain the initial states of nuclei and perform Lattice BUU simulations for HICs. II.3 Equation of state of cold nuclear matter The EOS of isospin asymmetric nuclear matter with total nucleon density \rho=\rho_{n}+\rho_{p} and isospin asymmetry \delta=(\rho_{n}-\rho_{p})/\rho are defined as its binding energy per nucleon. For the uniform nuclear matter at zero temperature, f_{\tau}(\vec{r},\vec{p}) becomes a step function in momentum space, i.e., f_{\tau}(\vec{r},\vec{p})=g_{\tau}\theta(p_{F_{\tau}}-|\vec{p}|) , where g_{\tau} is the degeneracy and p_{F_{\tau}}=\hbar(3\pi^{2}\rho_{\tau})^{1/3} is the Fermi momentum of nucleons with isospin \tau . The EOS of cold nuclear matter can then be analytically expressed as \displaystyle E(\rho,\delta)= \displaystyle\frac{3}{5}\frac{a^{2}\hbar^{2}}{2m}F_{5/3}\rho^{2/3}+\frac{1}{4\rho_{0}}\left[A_{l}(1+\delta^{2})+A_{u}(1-\delta^{2})\right]\rho^{3/3} (18) \displaystyle+\sum_{i=1}^{3}\sum_{\tau,\tau^{\prime}}C_{\tau\tau^{\prime},i}\,E^{\mathrm{MD}}_{\tau\tau^{\prime},i}+\sum_{n=1,3,5}\frac{1}{48}t_{3}^{[n]}\left[3-(2x_{3}^{[n]}+1)\delta^{2}\right]\rho^{\frac{n}{3}+1}, where a=(3\pi^{2}/2)^{1/3} . In Eq. ( 18 ), we have defined the contribution of the MD term as \displaystyle E^{\mathrm{MD}}_{\tau\tau^{\prime},i}= \displaystyle\frac{\Lambda_{i}^{2}}{24\pi^{4}\hbar^{6}\rho_{0}\rho}\Bigg\{p_{F,\tau}p_{F,\tau^{\prime}}\Big[3(p_{F,\tau}^{2}+p_{F,\tau^{\prime}}^{2})-\Lambda_{i}^{2}\Big]+4\Lambda_{i}\bigg[(p_{F,\tau}^{3}-p_{F,\tau^{\prime}}^{3})\arctan\frac{p_{F,\tau}-p_{F,\tau^{\prime}}}{\Lambda_{i}} (19) \displaystyle-(p_{F,\tau}^{3}+p_{F,\tau^{\prime}}^{3})\arctan\frac{p_{F,\tau}+p_{F,\tau^{\prime}}}{\Lambda_{i}}\bigg]+\frac{1}{4}\bigg[\Lambda_{i}^{4}+6\Lambda_{i}^{2}(p_{F,\tau}^{2}+p_{F,\tau^{\prime}}^{2})-3(p_{F,\tau}^{2}-p_{F,\tau^{\prime}}^{2})^{2}\bigg]\ln\frac{(p_{F,\tau}+p_{F,\tau^{\prime}})^{2}+\Lambda_{i}^{2}}{(p_{F,\tau}-p_{F,\tau^{\prime}})^{2}+\Lambda_{i}^{2}}\Bigg\}. The E^{\mathrm{MD}}_{\tau\tau^{\prime}} can be expanded as a power series of p_{F_{\tau}} (here we consider the case of SNM with p_{F_{\tau}}=p_{F}=a\hbar\rho^{1/3} ): E^{\mathrm{MD}}_{\tau\tau^{\prime}}\propto\frac{8}{3}p_{F}^{3}-\frac{16p_{F}^{5}}{5\Lambda^{2}}+\frac{192p_{F}^{7}}{35\Lambda^{4}}-\frac{512p_{F}^{9}}{45\Lambda^{6}}+\mathcal{O}(p_{F}^{11}). (20) It can been seen from Eq. ( 18 ) that the inclusion of the DD term as in Eq. ( 10 ) leads to the EOS that can be expressed as a power series in \rho^{1/3} (equivalently the Fermi momentum p_{F} ), from \rho^{2/3} to \rho^{9/3} , while the contributions from DD terms and MD terms can be clearly distinguished. The EOS of isospin asymmetric nuclear matter can be expanded as a power series in \delta , i.e., E\left(\rho,\delta\right)=E_{0}\left(\rho\right)+E_{\mathrm{sym}}\left(\rho\right)\delta^{2}+E_{\mathrm{sym},4}\left(\rho\right)\delta^{4}+\mathcal{O}\left(\delta^{6}\right), (21) where E_{0}(\rho) is the EOS of the SNM. The symmetry energy E_{\mathrm{sym}}(\rho) and the fourth-order symmetry energy E_{\mathrm{sym},4}(\rho) can be expressed as E_{\mathrm{sym}}(\rho)=\left.\frac{1}{2!}\frac{\partial^{2}E(\rho,\delta)}{\partial\delta^{2}}\right|_{\delta=0}, (22) and E_{\mathrm{sym},4}(\rho)=\left.\frac{1}{4!}\frac{\partial^{4}E(\rho,\delta)}{\partial\delta^{4}}\right|_{\delta=0}, (23) respectively. The expressions of E_{0}(\rho) , E_{\mathrm{sym}}(\rho) , and E_{\mathrm{sym},4}(\rho) are presented in Appendix C . The pressure of the isospin asymmetric nuclear matter can be expressed as P\left(\rho,\delta\right)=\rho^{2}\frac{\partial E(\rho,\delta)}{\partial\rho}, (24) and the saturation density \rho_{0} is defined as the density where the pressure of SNM is zero (except for \rho=0 ), i.e., \left.P\left(\rho_{0},\delta=0\right)=\rho_{0}^{2}\frac{\mathrm{d}E(\rho,0)}{\mathrm{d}\rho}\right|_{\rho=\rho_{0}}=0. (25) Around \rho_{0} , both E_{0}(\rho) and E_{\mathrm{sym}}(\rho) can be expanded as a power series in a dimensionless variable \chi\equiv\frac{\rho-\rho_{0}}{3\rho_{0}} Chen et al. 2009 , i.e., \displaystyle E_{0}(\rho)= \displaystyle E_{0}\left(\rho_{0}\right)+L_{0}\chi+\frac{K_{0}}{2!}\chi^{2} (26) \displaystyle+\frac{J_{0}}{3!}\chi^{3}+\frac{I_{0}}{4!}\chi^{4}+\frac{H_{0}}{5!}\chi^{5}+\mathcal{O}\left(\chi^{6}\right), and \displaystyle E_{\mathrm{sym}}(\rho)= \displaystyle E_{\mathrm{sym}}\left(\rho_{0}\right)+L\chi+\frac{K_{\mathrm{sym}}}{2!}\chi^{2} (27) \displaystyle+\frac{J_{\mathrm{sym}}}{3!}\chi^{3}+\frac{I_{\mathrm{sym}}}{4!}\chi^{4}+\frac{H_{\mathrm{sym}}}{5!}\chi^{5}+\mathcal{O}\left(\chi^{6}\right). The coefficients of \chi^{n} in Eqs. ( 26 ) and ( 27 ) are commonly used to characterize the density dependence of E_{0}(\rho) and E_{\mathrm{sym}}(\rho) , respectively, which can be expressed as \displaystyle L_{0} \displaystyle=\left.3\rho_{0}\frac{\mathrm{d}E_{0}(\rho)}{\mathrm{d}\rho}\right|_{\rho=\rho_{0}},L=\left.3\rho_{0}\frac{\mathrm{d}E_{\mathrm{sym}}(\rho)}{\mathrm{d}\rho}\right|_{\rho=\rho_{0}}, (28) \displaystyle K_{0} \displaystyle=\left.(3\rho_{0})^{2}\frac{\mathrm{d}^{2}E_{0}(\rho)}{\mathrm{d}\rho^{2}}\right|_{\rho=\rho_{0}},K_{\mathrm{sym}}=\left.(3\rho_{0})^{2}\frac{\mathrm{d}^{2}E_{\mathrm{sym}}(\rho)}{\mathrm{d}\rho^{2}}\right|_{\rho=\rho_{0}}, (29) \displaystyle J_{0} \displaystyle=\left.(3\rho_{0})^{3}\frac{\mathrm{d}^{3}E_{0}(\rho)}{\mathrm{d}\rho^{3}}\right|_{\rho=\rho_{0}},J_{\mathrm{sym}}=\left.(3\rho_{0})^{3}\frac{\mathrm{d}^{3}E_{\mathrm{sym}}(\rho)}{\mathrm{d}\rho^{3}}\right|_{\rho=\rho_{0}}, (30) \displaystyle I_{0} \displaystyle=\left.(3\rho_{0})^{4}\frac{\mathrm{d}^{4}E_{0}(\rho)}{\mathrm{d}\rho^{4}}\right|_{\rho=\rho_{0}},I_{\mathrm{sym}}=\left.(3\rho_{0})^{4}\frac{\mathrm{d}^{4}E_{\mathrm{sym}}(\rho)}{\mathrm{d}\rho^{4}}\right|_{\rho=\rho_{0}}, (31) \displaystyle H_{0} \displaystyle=\left.(3\rho_{0})^{5}\frac{\mathrm{d}^{5}E_{0}(\rho)}{\mathrm{d}\rho^{5}}\right|_{\rho=\rho_{0}},H_{\mathrm{sym}}=\left.(3\rho_{0})^{5}\frac{\mathrm{d}^{5}E_{\mathrm{sym}}(\rho)}{\mathrm{d}\rho^{5}}\right|_{\rho=\rho_{0}}. (32) Obviously, we have L_{0}=0 by the definition of \rho_{0} in Eq. ( 25 ). K_{0} is known as the incompressibility coefficient of SNM which characterizes the curvature of E_{0}(\rho) at \rho_{0} . The higher-order coefficients J_{0} , I_{0} and H_{0} are commonly referred to as the skewness, kurtosis, and hyper-skewness coefficients of SNM. Correspondingly, L , K_{\mathrm{sym}} , J_{\mathrm{sym}} , I_{\mathrm{sym}} , and H_{\mathrm{sym}} are the slope, curvature, skewness, kurtosis, and hyper-skewness coefficient of E_{\mathrm{sym}}(\rho) at \rho_{0} , respectively. The expressions of these characteristic parameters are presented in Appendix C . II.4 Single-nucleon potential, symmetry potential, nucleon effective masses and its linear isospin splitting coefficient in cold nuclear matter In the cold nuclear matter, the single-nucleon potential [Eq. ( 17 )] can be analytically expressed as a function of \rho , \delta , and the magnitude of nucleon momentum p=\left|\vec{p}\right| , i.e., \displaystyle U_{\tau}(\rho,\delta,p)= \displaystyle A_{l}\frac{\rho_{\tau}}{\rho_{0}}+A_{u}\frac{\rho_{-\tau}}{\rho_{0}}+\sum_{i=1}^{3}\sum_{\tau^{\prime}}C_{\tau\tau^{\prime},i}\,U^{\mathrm{MD}}_{\tau^{\prime},i}+\sum_{n=1,3,5}\frac{1}{24}t_{3}^{[n]}\left[(3+\frac{n}{2})-(1+2x_{3}^{[n]})(\tau\delta+\frac{n}{6}\delta^{2})\right]\rho^{\frac{n}{3}+1}, (33) where \displaystyle U^{\mathrm{MD}}_{\tau^{\prime},i}=\frac{\Lambda_{i}^{3}}{2\pi^{2}\hbar^{3}\rho_{0}}\Bigg[\frac{p_{F,\tau^{\prime}}^{2}+\Lambda_{i}^{2}-p^{2}}{2p\Lambda_{i}}\ln\frac{(p+p_{F,\tau^{\prime}})^{2}+\Lambda_{i}^{2}}{(p-p_{F,\tau^{\prime}})^{2}+\Lambda_{i}^{2}}+\frac{2p_{F,\tau^{\prime}}}{\Lambda_{i}}-2\arctan\frac{p+p_{F,\tau^{\prime}}}{\Lambda_{i}}+2\arctan\frac{p-p_{F,\tau^{\prime}}}{\Lambda_{i}}\Bigg]. (34) The U_{\tau}(\rho,\delta,p) can be expanded as a power series in \tau\delta , i.e., U_{\tau}(\rho,\delta,p)=U_{0}(\rho,p)+\sum_{i=1,2,\cdots}U_{\mathrm{sym},i}(\rho,p)(\tau\delta)^{i}=U_{0}(\rho,p)+U_{\mathrm{sym},1}(\rho,p)(\tau\delta)+U_{\mathrm{sym},2}(\rho,p)(\tau\delta)^{2}+\cdots, (35) where U_{0}(\rho,p)=U_{\tau}(\rho,\delta=0,p) (36) is the single-nucleon potential in SNM, and U_{\mathrm{sym},i}(\rho,p)\left.\equiv\frac{1}{i!}\frac{\partial^{i}U_{n}(\rho,\delta,p)}{\partial\delta^{i}}\right|_{\delta=0}=\left.\frac{(-1)^{i}}{i!}\frac{\partial^{i}U_{p}(\rho,\delta,p)}{\partial\delta^{i}}\right|_{\delta=0}. (37) is the i th-order symmetry potential. Neglecting higher-order terms ( \delta^{2},\delta^{3},\cdots ) in Eq. ( 35 ) leads to the well-known Lane potential Lane 1962 : U_{\tau}(\rho,\delta,p)\approx U_{0}(\rho,p)+U_{\mathrm{sym}}(\rho,p)(\tau\delta). (38) In the MDI3Y interaction, U_{0}(\rho,p) can be expressed as U_{0}(\rho,p)=C_{0}(\rho)+C_{1}f_{\Lambda_{1}}(\rho,p)+C_{2}f_{\Lambda_{2}}(\rho,p)+C_{3}f_{\Lambda_{3}}(\rho,p), (39) where f_{\Lambda_{i}}(p)=\frac{3\Lambda^{3}_{i}}{4p_{F_{0}}^{3}}\bigg[\frac{p_{F}^{2}+\Lambda^{2}_{i}-p^{2}}{2p\Lambda_{i}}\ln\frac{(p+p_{F})^{2}+\Lambda^{2}_{i}}{(p-p_{F})^{2}+\Lambda^{2}_{i}}+\frac{2p_{F}}{\Lambda_{i}}-2\arctan\frac{p+p_{F}}{\Lambda_{i}}+2\arctan\frac{p-p_{F}}{\Lambda_{i}}\bigg], (40) C_{0}(\rho)=U_{0}(\rho,p=\infty)=\frac{A_{l}+A_{u}}{2}\frac{\rho}{\rho_{0}}+\sum_{n=1,3,5}\frac{t_{3}^{[n]}}{16}\left(\frac{n}{3}+2\right)\rho^{\frac{n}{3}+1}, (41) C_{i}=C_{l,i}+C_{u,i}~~(i=1,2,3), (42) and p_{F_{0}}=a\hbar\rho_{0}^{1/3} is the Fermi momentum of nucleons in the SNM at \rho_{0} . The (first-order) symmetry potential can be expressed as U_{\mathrm{sym}}(\rho,p)=D_{0}(\rho)+D_{1}g_{\Lambda_{1}}(\rho,p)+D_{2}g_{\Lambda_{2}}(\rho,p)+D_{3}g_{\Lambda_{3}}(\rho,p), (43) where g_{\Lambda_{i}}(\rho,p)=\frac{1}{4}\frac{p_{F}^{2}\Lambda^{2}_{i}}{p_{F_{0}}^{3}\,p}\ln\frac{(p+p_{F})^{2}+\Lambda^{2}_{i}}{(p-p_{F})^{2}+\Lambda^{2}_{i}}, (44) D_{0}(\rho)=U_{\mathrm{sym}}(\rho,p=\infty)=\frac{A_{l}-A_{u}}{2}\frac{\rho}{\rho_{0}}-\sum_{n=1,3,5}\frac{1}{24}t_{3}^{[n]}\left(2x_{3}^{[n]}+1\right)\rho^{\frac{n}{3}+1}, (45) and D_{i}=C_{l,i}-C_{u,i}~~(i=1,2,3). (46) The second-order symmetry potential can be expressed as U_{\mathrm{sym},2}(\rho,p)=-\sum_{n=1,3,5}\frac{t_{3}^{[n]}}{48}\frac{n}{3}\left(2x_{3}^{[n]}+1\right)\rho^{\frac{n}{3}+1}+C_{1}h_{\Lambda_{1}}(\rho,p)+C_{2}h_{\Lambda_{2}}(\rho,p)+C_{3}h_{\Lambda_{3}}(\rho,p), (47) with C_{i} defined in Eq. ( 42 ) and h_{\Lambda_{i}}(\rho,p)=\frac{1}{6}\frac{\rho}{\rho_{0}}\frac{\Lambda_{i}^{2}\left(p^{2}+\Lambda_{i}^{2}-p_{F}^{2}\right)}{\left[(p+p_{F})^{2}+\Lambda^{2}_{i}\right]\left[(p-p_{F})^{2}+\Lambda^{2}_{i}\right]}-\frac{1}{24}\frac{p_{F}^{2}\Lambda^{2}_{i}}{p_{F_{0}}^{3}\,p}\ln\frac{(p+p_{F})^{2}+\Lambda^{2}_{i}}{(p-p_{F})^{2}+\Lambda^{2}_{i}}. (48) The momentum dependence of the single-nucleon potential is characterized by the nucleon effective mass Jaminon and Mahaux 1989 ; Li et al. 2018 , and it can be expressed as \frac{m_{\tau}^{\ast}(\rho,\delta)}{m}=\left[1+\left.\frac{m}{p}\frac{\mathrm{d}U_{\tau}(\rho,\delta,p)}{\mathrm{d}p}\right|_{p=p_{F_{\tau}}}\right]^{-1} (49) in nonrelativistic models. The isoscalar nucleon effective mass m_{s}^{\ast} is the nucleon effective mass in SNM, while the isovector nucleon effective mass m_{v}^{\ast} represents the effective mass of proton (neutron) in pure neutron (proton) matter, and their expressions are provided in Appendix C . We introduce a subscript “0” to denote their values at \rho_{0} , e.g., m_{s,0}^{\ast} and m_{v,0}^{\ast} . Another commonly used quantity is the isospin splitting of nucleon effective mass at the corresponding Fermi momentum, defined as m_{\mathrm{n}\text{-}\mathrm{p}}^{\ast}(\rho,\delta)\equiv\left[m_{n}^{\ast}(\rho,\delta)-m_{p}^{\ast}(\rho,\delta)\right]/m . The m_{\mathrm{n}\text{-}\mathrm{p}}^{\ast}(\rho,\delta) can be expanded as a power series in \delta , i.e., m_{\mathrm{n}\text{-}\mathrm{p}}^{\ast}(\rho,\delta)=\sum_{n=1}^{\infty}\Delta m_{2n-1}^{\ast}(\rho)\delta^{2n-1}, (50) where the first coefficient \Delta m_{1}^{\ast}(\rho) is usually referred to as the linear isospin splitting coefficient, whose expression is presented in Appendix C . Since the effective mass isospin splitting is also momentum-dependent in general, we introduce the notation m_{\mathrm{n}\text{-}\mathrm{p}} for the splitting at arbitrary momentum. III Fitting strategy and new interactions The force-range parameters \Lambda_{i} in the MD term represent the ranges of the nuclear interaction. We set \Lambda_{1}=140 MeV/ c , \Lambda_{2}=600 MeV/ c , approximately corresponding to the masses of \pi meson and \sigma (or \rho , \omega ) meson, respectively, to mimic the effective long-range and intermediate-range components. The value of \Lambda_{3} , which corresponds to the short-range force, can be determined from the high-energy behavior of the single-nucleon potential, as will be demonstrated later. After fixing the values of \Lambda_{1} , \Lambda_{2} and \Lambda_{3} , there are still 14 parameters in the MDI3Y model, which require 14 independent quantities to completely construct an interaction. In Table 1 , we list the model parameters and the adjustable quantities for the MDI3Y model. The quantities \rho_{0} , E_{0}(\rho_{0}) , K_{0} , J_{0} and E_{{\mathrm{sym}}}(\rho_{0}) , L , K_{{\mathrm{sym}}} , J_{{\mathrm{sym}}} are related to the bulk properties of the isospin asymmetric nuclear matter. It is worth noting again that including additional DD terms in Eq. ( 2 ) may provide more flexible high-density behavior, which could be useful for extracting the nuclear EOS at suprasaturation densities with new available experimental probes in the future. The C_{1} , C_{2} , C_{3} [Eq. ( 42 )] determine the momentum dependence of U_{0} , which can be obtained by fitting the nucleon optical potential Hama et al. 1990 ; Cooper et al. 1993 , while D_{1} , D_{2} , D_{3} [Eq. ( 46 )] determine the momentum dependence of U_{\mathrm{sym}} . On the other hand, the behavior of U_{\mathrm{sym}} can also be characterized by the quantities \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho) , with \displaystyle\Delta U_{\mathrm{sym}}(\rho)\equiv \displaystyle U_{\mathrm{sym}}(\rho,p=0)-U_{\mathrm{sym}}(\rho,p=\infty) (51) \displaystyle= \displaystyle D_{1}\frac{p_{F}^{3}}{p_{F_{0}}^{3}}\frac{\Lambda_{1}^{2}}{p_{F}^{2}+\Lambda_{1}^{2}}+D_{2}\frac{p_{F}^{3}}{p_{F_{0}}^{3}}\frac{\Lambda_{2}^{2}}{p_{F}^{2}+\Lambda_{2}^{2}} \displaystyle+D_{3}\frac{p_{F}^{3}}{p_{F_{0}}^{3}}\frac{\Lambda_{3}^{2}}{p_{F}^{2}+\Lambda_{3}^{2}}. The values of D_{1} , D_{2} , D_{3} can be uniquely determined by \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho_{0}) from Eqs. ( 115 ), ( 114 ), and ( 51 ). Due to the lack of experimental data, the behavior of U_{\mathrm{sym}} still exhibits significant uncertainties. To construct different types of U_{\mathrm{sym}} , we can either fit the results from theoretical calculations or adjust the values of \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho_{0}) , and we will discuss these two approaches in detail later. Table 1: The 14 model parameters and the adjustable quantities for the MDI3Y model. Parameters Quantities A_{l} , A_{u} , t_{3}^{[1]} , x_{3}^{[1]} , t_{3}^{[3]} , x_{3}^{[3]} , \rho_{0} , E_{0}(\rho_{0}) , K_{0} , J_{0} , E_{{\mathrm{sym}}}(\rho_{0}) , L , t_{3}^{[5]} , x_{3}^{[5]} , C_{l,1} , C_{u,1} , K_{{\mathrm{sym}}} , J_{{\mathrm{sym}}} , C_{1} , C_{2} , C_{3} , [ D_{1} , D_{2} , C_{l,2} , C_{u,2} , C_{l,3} , C_{u,3} D_{3} or \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , \Delta U_{\mathrm{sym}}(\rho_{0}) ] In the construction of the MDI3Y interactions, we take the values of \rho_{0} , E_{0}(\rho_{0}) , and K_{0} to be 0.16 fm -3 , -16 MeV, and 230 MeV, respectively. Next, with different values of \Lambda_{3} , we use GEKKO optimization suite Beal et al. 2018 to minimize the weighted squared difference \chi^{2} between U_{0} in Eq. ( 39 ) and the nucleon optical potential data U_{\mathrm{opt}} up to 1 GeV obtained by Hama et al. Hama et al. 1990 ; Cooper et al. 1993 : \chi^{2}=\sum_{i=1}^{N_{d}}\left(\frac{U_{0,i}-U_{\mathrm{opt},i}}{\sigma_{i}}\right)^{2}, (52) with constraint of the HVH theorem Hugenholtz and van Hove 1958 ; Satpathy et al. 1999 , i.e., E_{0}\left(\rho_{0}\right)=\frac{p_{F_{0}}^{2}}{2m}+U_{0}\left(\rho_{0},p_{F_{0}}\right). (53) The N_{d} is the number of the experimental data points. Since there are actually no practical errors \sigma_{i} here, we assign equal weights to each data point. Figure 1: The energy dependence of single-nucleon potential in cold SNM with \Lambda_{3}=700 , 1318 , and 3000 MeV/ c , respectively, in the MDI3Y model. Also shown are the nucleon optical potential (Schrödinger equivalent potential, \mathrm{U}_{\mathrm{sep}} ) in SNM at \rho_{0} obtained by Hama et al. Hama et al. 1990 ; Cooper et al. 1993 . Figure 2: The predicted values of U_{0}^{\infty} for different \Lambda_{3} in the MDI3Y model. Note U_{0}^{\infty}=55 MeV with \Lambda_{3}=1318 MeV/ c . In Fig. 1 , we present the energy dependence of U_{0} calculated from Eq. ( 39 ) using the optimized C_{i} for different values of \Lambda_{3} . For illustration, we select \Lambda_{3}=700 , 1318 , and 3000 MeV/ c as representative cases. It can be seen from Fig. 1 that the Hama’s data can be successfully reproduced with different choices of \Lambda_{3} , while their predictions above 1 GeV could differ significantly. From Fig. 1 , it can be observed that when \Lambda_{3}=1318 MeV/ c , the predicted U_{0} provides a nice description of the extrapolation of Hama’s data. In Fig. 2 , we show the values of U_{0} at infinitely large nucleon momentum at \rho_{0} , denoted as U_{0}^{\infty} , for different \Lambda_{3} , with U_{0}^{\infty}=55 MeV when \Lambda_{3}=1318 MeV/ c . In Fig. 3 , we present the values of the isoscalar effective mass at \rho_{0} , denoted as m_{s,0}^{\ast} , for different \Lambda_{3} . The value of m_{s,0}^{\ast} is determined by the behavior of U_{0} around the Fermi momentum, where the predictions for different \Lambda_{3} are nearly identical, resulting in almost consistent values of m_{s,0}^{\ast} . For \Lambda_{3}=1318 MeV/ c , m_{s,0}^{\ast} is found to be 0.763m . Based on the above discussion, we set \Lambda_{3}=1318 MeV/ c throughout this paper to fit the extrapolation of the Hama’s data at high energies, and in this case, we obtain that C_{1}=106.476 MeV, C_{2}=-249.137 MeV, and C_{3}=67.8626 MeV. It should be pointed out that, by changing \Lambda_{3} , we could adjust the value of U_{0}^{\infty} , and thus we can investigate the high-energy behaviors of single-nucleon potential, which remains largely uncertain. So far, the last undetermined quantity related to SNM is the skewness parameter J_{0} , which characterizes the stiffness of E_{0}(\rho) at suprasaturation densities. In this work, we set the value of J_{0} to -460 MeV, which is the maximum value that satisfies the constraints on the SNM pressure from flow data in HICs Danielewicz et al. 2002 . Figure 3: The predicted values of m_{s,0}^{\ast} for different \Lambda_{3} with the MDI3Y interaction. The dashed line indicates \Lambda_{3}=1318 MeV/ c with m_{s,0}^{\ast}/m=0.763 . In the next step, we discuss the construction of U_{\mathrm{sym}} . Considering its significant uncertainties, we construct four different momentum dependence for U_{\mathrm{sym}} , including both the monotonic and non-monotonic types. The different U_{\mathrm{sym}} are constructed in two ways: (i) directly fitting the theoretical results up to 800 MeV/ c from relativistic Dirac-Brueckner-Hartree-Fock (DBHF) theory van Dalen et al. 2005 ; (ii) adjusting the values for \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho_{0}) . For the first approach, the fitting process is the same as that used for determining U_{0} : we minimize the difference between U_{\mathrm{sym}} in Eq. ( 43 ) and the results of DBHF calculation, and we obtain that D_{1}=258.339 MeV, D_{2}=-232.462 MeV, and D_{3}=344.078 MeV. The obtained U_{\mathrm{sym}} initially increases and then decreases with increasing momentum, as predicted by the DBHF calculation. We denote this U_{\mathrm{sym}} type as D017 since the \Delta m_{1}^{\ast}(\rho_{0}) is obtained as 0.171 , which nicely agrees with result m_{\mathrm{n}\text{-}\mathrm{p}}^{\ast}(\rho_{0},\delta)=0.187\delta from the relativistic BHF theory in the full Dirac space Wang et al. 2023b . For the second approach, we construct three types of U_{\mathrm{sym}} , including two cases for \Delta m_{1}^{\ast}(\rho_{0})=0.2 : a monotonically decreasing type denoted as D02, and a type that initially decreases and then increases denoted as D02r; as well as one case for \Delta m_{1}^{\ast}(\rho_{0})=-0.2 that initially increases and then decreases, denoted as Dm02. Typically, \Delta m_{1}^{\ast}(\rho_{0}) is related to the slope of U_{\mathrm{sym}} at p_{F_{0}} , while \Delta U_{\mathrm{sym}}(\rho_{0}) determines its asymptotic behavior at high momenta. And m_{v,0}^{\ast} can be adjusted accordingly to modify the structure of U_{\mathrm{sym}} . The values of \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho_{0}) for D017, D02, D02r, and Dm02 are listed in Table 3 , and the corresponding momentum dependence of U_{\mathrm{sym}} can be found in Fig. 9 . The value of D_{0} can be determined through: E_{\mathrm{sym}}(\rho_{0})=\frac{1}{3}\frac{p_{F_{0}}^{2}}{2m^{\ast}_{s,0}}+\frac{1}{2}U_{\mathrm{sym}}\left(\rho_{0},p_{F_{0}}\right), (54) once E_{{\mathrm{sym}}}(\rho_{0}) is given Brueckner and Dabrowski 1964 ; Dabrowski and Haensel 1972 ; Dabrowski and Haensel 1973 ; Xu et al. 2010 ; Xu et al. 2011 ; Chen et al. 2012 . The remaining four undetermined quantities E_{{\mathrm{sym}}}(\rho_{0}) , L , K_{{\mathrm{sym}}} , and J_{{\mathrm{sym}}} describe the magnitude and density dependence of E_{{\mathrm{sym}}}(\rho) . We first consider the case of L=35 MeV, and then we set E_{{\mathrm{sym}}}(\rho_{0})=32 MeV, K_{{\mathrm{sym}}}=-275 MeV, and J_{{\mathrm{sym}}}=720 MeV to satisfy the following constrains: (1) the EOS of PNM obtained from the various microscopic calculations (see Ref. Zhang and Chen 2023 and the references therein); (2) the largest mass of neutron stars observed to date from PSR J0740+6620 Cromartie et al. 2019 ; Fonseca et al. 2021 ; (3) the upper limit \Lambda_{1.4}\leqslant 580 of the dimensionless tidal deformability of a canonical 1.4 M_{\odot} neutron star from the gravitational wave event GW170817 Abbott et al. 2018 ; (4) the mass-radius (M-R) relation for PSR J0030+0451 Miller et al. 2019 ; Riley et al. 2019 with a mass of 1.4 M_{\odot} , PSR J0437-4715 Choudhury et al. 2024 with a mass of 1.4 M_{\odot} , PSR J0740+6620 Miller et al. 2021 ; Riley et al. 2021 with a mass of 2.0 M_{\odot} , and the central compact object (CCO) within HESS J1731-347 Doroshenko et al. 2022 with an unusually low mass of 0.77 M_{\odot} and small radius of 10.4 km. By combining the E_{{\mathrm{sym}}}(\rho_{0}) , L , K_{{\mathrm{sym}}} , and J_{{\mathrm{sym}}} with four types of U_{\mathrm{sym}} and the foregoing isoscalar quantities, we have constructed four different interactions, denoted as MDI3YL35D017, MDI3YL35D02, MDI3YL35D02r, and MDI3YL35Dm02, respectively. To further investigate the effects of E_{{\mathrm{sym}}} , we provide two additional quantities sets: L=55 MeV, E_{{\mathrm{sym}}}(\rho_{0})=33 MeV, K_{{\mathrm{sym}}}=-100 MeV, J_{{\mathrm{sym}}}=1200 MeV; and L=75 MeV, E_{{\mathrm{sym}}}(\rho_{0})=34 MeV, K_{{\mathrm{sym}}}=-80 MeV, J_{{\mathrm{sym}}}=1000 MeV. For these two quantity sets of the symmetry energy, we select the D02 type of U_{\mathrm{sym}} to construct two additional interactions, labeled as MDI3YL55D02 and MDI3YL75D02. Based on the MDI3Y model, we have thus constructed six interactions in this work. It is worth emphasizing again that these six interactions have the same description of the SNM properties. The four interactions MDI3YL35D017, MDI3YL35D02, MDI3YL35D02r, and MDI3YL35Dm02 exhibit similar density behavior of the symmetry energy, while the MDI3YL35D02, MDI3YL55D02, and MDI3YL75D02 share the same momentum dependence of U_{\mathrm{sym}} . In Table 2 , we list the 14 model parameters for these MDI3Y interactions. Table 2: The 14 model parameters of the MDI3YX interactions. x_{3}^{[n]} ( n=1,3,5 ) are dimensionless. For all interactions: \Lambda_{1}=140 MeV/ c , \Lambda_{2}=600 MeV/ c , \Lambda_{3}=1318 MeV/ c , C_{1}=C_{l,1}+C_{u,1}=106.476 MeV, C_{2}=C_{l,2}+C_{u,2}=-249.137 MeV, C_{3}=C_{l,3}+C_{u,3}=67.8626 MeV. X= L35D017 L35D02 L35D02r L35Dm02 L55D02 L75D02 A_{l}~(\rm{MeV}) -340.957 -153.841 -48.9242 -115.571 -139.574 -220.641 A_{u}~(\rm{MeV}) 107.377 -79.7389 -184.655 -118.008 -94.0055 -12.9389 C_{l,1}~(\rm{MeV}) 182.407 85.1073 60.0178 57.0134 85.1073 85.1073 C_{u,1}~(\rm{MeV}) -75.9315 21.3685 46.4581 49.4624 21.3685 21.3685 C_{l,2}~(\rm{MeV}) -240.799 -103.423 -53.0583 -190.268 -103.423 -103.423 C_{u,2}~(\rm{MeV}) -8.33773 -145.713 -196.078 -58.8684 -145.713 -145.713 C_{l,3}~(\rm{MeV}) 205.970 36.7689 -50.7926 131.961 36.7689 36.7689 C_{u,3}~(\rm{MeV}) -138.107 31.0936 118.655 -64.0988 31.0936 31.0936 t_{3}^{[1]}~(\mathrm{MeV}\,\mathrm{fm}^{4}) 16321.3 16321.3 16321.3 16321.3 16321.3 16321.3 t_{3}^{[3]}~(\mathrm{MeV}\,\mathrm{fm}^{6}) -9094.38 -9094.38 -9094.38 -9094.38 -9094.38 -9094.38 t_{3}^{[5]}~(\mathrm{MeV}\,\mathrm{fm}^{8}) 5024.65 5024.65 5024.65 5024.65 5024.65 5024.65 x_{3}^{[1]} -1.93719 -1.33714 -1.22204 -0.578062 -1.11707 -2.08202 x_{3}^{[3]} -7.99556 -5.87032 -5.43377 -5.02137 -5.64697 -7.57124 x_{3}^{[5]} -11.6092 -8.49551 -7.74781 -7.84116 -11.4709 -13.4967 IV The properties of cold nuclear matter IV.1 Bulk properties of cold nuclear matter We first show the results on bulk properties of nuclear matter at zero temperature within the MDI3Y model. Table 3 summarizes the characteristic quantities of nuclear matter obtained with the six MDI3Y interactions constructed above. It is seen that all the six interactions have the same isoscalar properties as imposed in prior. For the four interactions with L=35 MeV, they predict different values for the higher-order parameters I_{{\rm{sym}}} , H_{{\rm{sym}}} , E_{{\rm{sym}},4}(\rho_{0}) and U_{\mathrm{sym}}^{\infty} , due to the different momentum dependence of U_{\rm sym} by using different \Delta m_{1}^{\ast}(\rho_{0}) , m^{\ast}_{v,0}/m and \Delta U_{\mathrm{sym}}(\rho_{0}) . Table 3: Characteristic quantities of nuclear matter with the MDI3YX interactions. X= L35D017 L35D02 L35D02r L35Dm02 L55D02 L75D02 \rho_{0}\,(\rm{fm}^{-3}) 0.160 0.160 0.160 0.160 0.160 0.160 E_{0}(\rho_{0})\,(\rm{MeV}) -16.0 -16.0 -16.0 -16.0 -16.0 -16.0 K_{0}\,(\rm{MeV}) 230.0 230.0 230.0 230.0 230.0 230.0 J_{0}\,(\rm{MeV}) -460.0 -460.0 -460.0 -460.0 -460.0 -460.0 I_{0}\,(\rm{MeV}) 2228.7 2228.7 2228.7 2228.7 2228.7 2228.7 H_{0}\,(\rm{MeV}) -12841 -12841 -12841 -12841 -12841 -12841 m^{\ast}_{s,0}/m 0.763 0.763 0.763 0.763 0.763 0.763 U_{0}^{\infty}\,(\mathrm{MeV}) 55.0 55.0 55.0 55.0 55.0 55.0 E_{{\rm{sym}}}(\rho_{0})\,(\rm{MeV}) 32.0 32.0 32.0 32.0 33.0 34.0 L\,(\rm{MeV}) 35.0 35.0 35.0 35.0 55.0 75.0 K_{{\rm{sym}}}\,(\rm{MeV}) -275 -275 -275 -275 -100 -80.0 J_{{\rm{sym}}}\,(\rm{MeV}) 720.0 720.0 720.0 720.0 1200 1000 I_{{\rm{sym}}}\,(\rm{MeV}) 1454.8 237.43 -87.192 113.64 -658.56 1365.4 H_{{\rm{sym}}}\,(\rm{MeV}) -11342 -3885.8 -1851.6 -3178.5 1778.1 -15437 E_{{\rm{sym}},4}(\rho_{0})\,(\rm{MeV}) 0.1736 1.030 1.265 0.5356 1.030 1.030 \Delta m_{1}^{\ast}(\rho_{0}) 0.17 0.20 0.20 -0.20 0.20 0.20 m^{\ast}_{v,0}/m 0.606 0.575 0.565 0.790 0.575 0.575 \Delta U_{\mathrm{sym}}(\rho_{0})\,(\mathrm{MeV}) 193 55.0 -40.0 80.0 55.0 55.0 U_{\mathrm{sym}}^{\infty}\,(\mathrm{MeV}) -164.7 -17.08 80.34 -54.08 -15.08 -13.08 Figure 4: The pressure of SNM P_{\mathrm{SNM}}(\rho) as a function of nucleon density predicted by the MDI3Y interactions. The constraints from the flow data in HICs Danielewicz et al. 2002 are also included. Shown in Fig. 4 is the density dependence of pressure of SNM P_{\mathrm{SNM}}(\rho) predicted by MDI3Y interactions, and the constraints from flow data in HICs Danielewicz et al. 2002 are also plotted. It is seen that all the 6 MDI3Y interactions predict the same P_{\mathrm{SNM}}(\rho) , and they all satisfy the constraints from flow data. Figure 5: The EOS of PNM E_{\mathrm{PNM}}(\rho) as a function of nucleon density predicted by the MDI3Y interactions. Also included are the combined results from various microscopic calculations Huth et al. 2021 ; Zhang and Chen 2023 . Shown in Fig. 5 is the EOS of PNM, i.e., energy per nucleon E_{\mathrm{PNM}}(\rho) , as a function of nucleon density predicted by the MDI3Y interactions. Also included in Fig. 5 are the combined results from various microscopic calculations Huth et al. 2021 ; Zhang and Chen 2023 . It is seen that all interactions, except for MDI3YL75D02, are consistent with the microscopic calculations, whereas MDI3YL75D02 predicts too small E_{\mathrm{PNM}} below about 0.07 fm -3 . Figure 6: The symmetry energy E_{{\mathrm{sym}}}(\rho) as a function of nucleon density predicted by the MDI3Y interactions. Shown in Fig. 6 is the density dependence of the symmetry energy E_{{\mathrm{sym}}}(\rho) obtained with the MDI3Y interactions. The four interactions with L=35 MeV share the same values for E_{{\mathrm{sym}}}(\rho_{0}) , L , K_{{\mathrm{sym}}} , and J_{{\mathrm{sym}}} , which results in similar E_{{\mathrm{sym}}}(\rho) up to approximately 3\rho_{0} . However, due to the different U_{\mathrm{sym}} , the E_{{\mathrm{sym}}}(\rho) from the four interactions becomes significantly different at higher densities, which can also be observed from their different predictions for I_{{\mathrm{sym}}} and H_{{\mathrm{sym}}} in Table 3 . Furthermore, it is seen that the four interactions with L=35 MeV predict a soft E_{{\mathrm{sym}}} at intermediate densities but stiff one at high densities, which could result in a peak structure in the sound speed of neutron star matter Wang et al. 2024 . Moreover, a recent study Ye et al. 2024 has demonstrated that such a special density behavior of E_{{\mathrm{sym}}}(\rho) , which is soft around 2 - 3\rho_{0} but stiff above about 4\rho_{0} , can provide a solution to the hyperon puzzle. It should be noted that the symmetry energy soft around 2 - 3\rho_{0} but stiff above about 4\rho_{0} well agrees with the predictions from the relativistic mean-field model with isovector-scalar \delta meson and its coupling to isoscalar-scalar \sigma meson ( Zabari et al. 2019a ; Zabari et al. 2019b ; Kubis et al. 2020 ; Miyatsu et al. 2022 ; Li et al. 2022 ; Miyatsu et al. 2023 ) . In addition, the soft symmetry energy with negative density slope around 2\rho_{0} is also consistent with the constraints ( Xiao et al. 2009 ) obtained from analyzing charged pion ratio in heavy-ion collisions at FOPI ( Reisdorf et al. 2007 ) although the results are somewhat model dependent ( Feng and Jin 2010 ; Xie et al. 2013 ; Xu et al. 2013 ; Hong and Danielewicz 2014 ; Song and Ko 2015 ; Zhang and Ko 2017 ; Zhang and Ko 2018 ) . We also note that Liu et al. Liu et al. 2025 recently simulated the Au+Au collisions at 400 MeV per nucleon using the UrQMD model, and their analyses of the experimental data for collective flows Russotto et al. 2011 ; Russotto et al. 2016 and charged pion productions Reisdorf et al. 2010 support a soft E_{\mathrm{sym}}(\rho) at densities of about 1 - 2\rho_{0} . IV.2 Single-nucleon potential, symmetry potential and the nucleon effective mass The single-nucleon potential in SNM at \rho_{0} , U_{0}(\rho_{0},p) , predicted by the MDI3Y interactions, is plotted in Fig. 1 (in red) as a function of the nucleon kinetic energy E-m=\sqrt{p^{2}+m^{2}}+U_{0}(\rho,p)-m . Note the six MDI3Y interactions have the same \Lambda_{3}=1318 MeV/ c . Fig. 1 additionally includes the real part of the nucleon optical potential (Schrödinger equivalent potential) obtained by Hama et al. Hama et al. 1990 ; Cooper et al. 1993 based on Dirac phenomenology of nucleon-nucleus scattering data. It is seen that the predicted U_{0}(\rho_{0},p) agrees well with the empirical nucleon optical potential obtained by Hama et al. for nucleon kinetic energy up to 1 GeV and also matches its extrapolation at higher energies. The value of U_{0}(\rho_{0},p) at infinitely large nucleon momentum, U_{0}^{\infty} , is obtained as 55 MeV. Figure 7: Momentum dependence of the single-nucleon potential in SNM U_{0}(\rho,p) , at \rho=\rho_{0}/4 , \rho_{0}/2 , and \rho_{0} , respectively. The vertical lines indicate the corresponding Fermi momenta. Figure 7 displays the momentum dependence of U_{0}(\rho,p) at \rho=\rho_{0}/4 , \rho_{0}/2 , and \rho_{0} , respectively, with the corresponding Fermi momentum p_{F} indicated in the plot. The momentum dependence of U_{0}(\rho,p) at p_{F} can be characterized by the isoscalar nucleon effective mass m_{s}^{\ast}(\rho) , as shown in Eq. ( 113 ). It is seen that U_{0}(\rho_{0}/4,p) is decreasing with nucleon momentum at p_{F} , which consequently leads to m_{s}^{\ast}(\rho_{0}/4)/m>1 . It is interesting to mention that m_{s}^{\ast} is also related to the single-particle states in finite nuclei. From mean-field calculation, it has been found that m_{s}^{\ast}/m\approx 1 is necessary to reproduce the level density near the Fermi surface Brown et al. 1963 ; Barranco and Treiner 1981 ; Shlomo 1992 ; Mughabghab and Dunford 1998 and to accurately fit the masses of open-shell nuclei Tondeur et al. 2000 ; Goriely et al. 2001 , while m_{s}^{\ast}/m should be about 0.7 for deeper states Farine et al. 2001 . On the other hand, the value of m_{s}^{\ast}/m\approx 0.8\pm 0.1 is supported by both the analyses of isoscalar giant quadrupole resonances and various many-body calculations (see Ref. Li et al. 2018 for a review). This inconsistent large value of m_{s}^{\ast} near the Fermi surface could be explained by the coupling between particle modes and surface-vibration modes, which leads to modifications of the single-particle energies near the Fermi surface and reducing the energy gap between occupied and unoccupied shells Bertsch and Kuo 1968 ; Hamamoto and Siemens 1976 ; Bernard and Van Giai 1980 ; Ma and Wambach 1983 ; Litvinova and Ring 2006 . To simultaneously describe the single-particle dynamics near and far from the Fermi surface, a radius-dependent m_{s}^{\ast}/m has been introduced in Ref. Ma and Wambach 1983 , which is about 0.7 in the interior and has a peak value of about 1.2 - 1.5 before becomes 1 at the surface due to the coupling. This surface-peaked m_{s}^{\ast}/m was later confirmed in self-consistent Skyrme-Hartree-Fock (SHF) calculations by including new terms into the Skyrme force Farine et al. 2001 ; Zalewski et al. 2010 . Shown in Fig. 8 is m_{s}^{\ast}(\rho) as a function of nucleon density for the MDI3Y interactions. The values of m_{s}^{\ast}(\rho) at \rho=\rho_{0}/4 , \rho_{0}/2 , and \rho_{0} are also marked in Fig. 8 . It is worth noting that these MDI3Y interactions predict the value of m_{s}^{\ast}(\rho)/m>1 for \rho<\rho_{0}/3 with m_{s}^{\ast}(\rho_{0})/m=0.763 . Also shown in Fig. 8 are the m_{s}^{\ast}(\rho) obtained from interactions SLy4 Chabanat et al. 1998 , SkSP.1 Farine et al. 2001 , and M3Y-P7 Nakada 2013 . By introducing a zero-range term with both momentum and density dependence into the conventional Skyrme interaction, SkSP.1 interaction predicts a surface-peaked m_{s}^{\ast}(\rho) with the value of m_{s}^{\ast}(\rho_{0})/m=0.8 , which may allow for the simultaneous descriptions of both nuclei masses and isoscalar giant quadrupole resonance Farine et al. 2001 . Figure 8: The isoscalar nucleon effective mass m_{s}^{\ast}(\rho) as a function of nucleon density. The results for interactions SLy4 Chabanat et al. 1998 , SkSP.1 Farine et al. 2001 , and M3Y-P7 Nakada 2013 are also shown for comparison. Figure 9: Momentum dependence for the four types of U_{{\mathrm{sym}}}(\rho_{0},p) . Here E_{{\mathrm{sym}}}(\rho_{0})=30 MeV has been assumed. The corresponding results from several microscopic calculations are also shown as symbols (see the text for details). Shown in Fig. 9 are the four different momentum dependence of U_{{\mathrm{sym}}} constructed in Sec. III . It is worth emphasizing again that, in the construction of these interactions, the momentum dependence of U_{{\mathrm{sym}}} is solely determined by a set of \Delta m_{1}^{\ast}(\rho_{0}) , m_{v,0}^{\ast} , and \Delta U_{\mathrm{sym}}(\rho_{0}) , and is not affected by the properties of E_{{\mathrm{sym}}}(\rho) . Changing the value of E_{{\mathrm{sym}}}(\rho_{0}) results in a vertical shift of U_{{\mathrm{sym}}}(\rho_{0},p) , and in Fig. 9 , we have chosen E_{{\mathrm{sym}}}(\rho_{0})=30 MeV for illustration. Also shown in Fig. 9 are results from several microscopic calculations, including the non-relativistic BHF theory with and without rearrangement contribution from the three-body force Zuo et al. 2006 , the relativistic DBHF theory van Dalen et al. 2005 and the relativistic impulse approximation (RIA) Chen et al. 2005c ; Li et al. 2006b . The non-monotonic U_{{\mathrm{sym}}} for D017 are obtained by fitting the results of DBHF calculations. It should be noted that the behavior of U_{{\mathrm{sym}}} is related to the nucleon effective mass splitting m_{\mathrm{n}\text{-}\mathrm{p}} . Specifically, a decreasing U_{\mathrm{sym}} indicates that in the neutron-rich environments, neutrons have a larger effective mass than protons ( m_{\mathrm{n}\text{-}\mathrm{p}}>0 ), whereas an increasing U_{\mathrm{sym}} yields a negative m_{\mathrm{n}\text{-}\mathrm{p}} . The non-monotonic U_{\mathrm{sym}} implies that m_{\mathrm{n}\text{-}\mathrm{p}} can have different signs at different momenta. Very recently, Sun et al. Sun et al. 2025 discussed the use of Gogny-type interactions in the QMD models to reproduce the non-monotonic behavior of U_{\mathrm{sym}} Berger et al. 1991 ; Blaizot et al. 1995 ; Chen et al. 2012 . Actually, transport models predict that neutrons or protons with the same momentum are more likely to be accelerated from the compressed participant region if they experience a stronger repulsive potential and exhibit a smaller effective mass, suggesting that the spectra of emitted particles, such as the neutron over proton spectral ratio, could be a sensitive probe for m_{\mathrm{n}\text{-}\mathrm{p}} and the momentum dependence of U_{\mathrm{sym}} Rizzo et al. 2005 ; Giordano et al. 2010 ; Zhang et al. 2014 ; Coupland et al. 2016 ; Wang et al. 2023a . Figure 10: Effective masses for neutrons m_{n}^{\ast} [panel (a)] and protons m_{p}^{\ast} [panel (b)] in PNM as a function of nucleon density for the four different types of the U_{{\mathrm{sym}}} . The results for SLy4 Chabanat et al. 1998 , SkSP.1 Farine et al. 2001 , and M3Y-P7 Nakada 2013 are included for comparison. The m_{n}^{\ast} in PNM from many-body calculations Alp et al. 2025 using various chiral EFT interactions are also shown as solid symbols. Figure 10 shows the density dependence of the nucleon effective mass in PNM for the four different types of U_{{\mathrm{sym}}} , with the results of neutrons m_{n}^{\ast} and protons m_{p}^{\ast} displayed in Fig. 10 (a) and (b), respectively. Note that the m_{p}^{\ast} in PNM is equal to the isovector nucleon effective mass m_{v}^{\ast} by definition. For comparison, Fig. 10 also includes the results for SLy4 Chabanat et al. 1998 , SkSP.1 Farine et al. 2001 , and M3Y-P7 Nakada 2013 . The m_{n}^{\ast} from the recent many-body calculations Alp et al. 2025 using various chiral effective field theory (EFT) interactions (see the reference for details) are also plotted in Fig. 10 (a). For D017, D02, and D02r with \Delta m_{1}^{\ast}(\rho_{0})>0 , m_{n}^{\ast} is always larger than m_{p}^{\ast} . While for Dm02 with \Delta m_{1}^{\ast}(\rho_{0})=-0.2 , m_{n}^{\ast} is smaller than m_{p}^{\ast} at densities below about 2\rho_{0} , but it exceeds m_{p}^{\ast} at higher densities. From Fig. 10 (a), it can be seen that m_{n}^{\ast} exhibits a peak behavior at low densities for D017, D02, D02r, and Dm02, and the peak is most distinct for D017 and least evident for Dm02. In contrast, m_{n}^{\ast} is continuously decreasing for SLy4, SkSP.1 and M3Y-P7. For D017, D02, D02r, Dm02, and M3Y-P7, m_{n}^{\ast} and m_{p}^{\ast} will eventually approach to the bare mass when density becomes sufficiently large, since they are obtained from finite-range interactions. It is seen that m_{n}^{\ast} for D017 and Dm02 begin to increase when \rho\gtrsim 4\rho_{0} . From Fig. 10 (a), a common peak behavior of m_{n}^{\ast} is observed below approximately 0.5\rho_{0} for different chiral EFT interactions, although their predictions differ significantly with increasing density. Note that all the chiral EFT interactions predict an increasing m_{n}^{\ast} at larger densities, due to the contributions from three-nucleon interactions Alp et al. 2025 . From Fig. 10 (b), it can be observed that m_{p}^{\ast} exhibits a distinct peak behavior at low densities for Dm02 and SkSP.1. Compared to SLy4, the peak behavior for SkSP.1 arises from the momentum-dependent DD term, whose contribution also leads to a rapid decrease in m_{p}^{\ast} with increasing density. Figure 11: Momentum dependence of the second-order symmetry potential U_{\mathrm{sym},2}(\rho_{0},p) , obtained with the six MDI3Y interactions. Note that the curves of MDI3YL35D02 and MDI3YL35D02r overlap each other. In Fig. 11 , we show the second-order symmetry potential U_{\mathrm{sym},2} as a function of nucleon momentum, obtained with the six MDI3Y interactions. The momentum dependence of U_{\mathrm{sym},2} is determined by the C_{i} parameters as shown in Eq. ( 47 ), which have already been obtained in Sec. III by fitting U_{0} to the empirical nucleon optical potential, and all the U_{\mathrm{sym},2} exhibit the same momentum dependence. Furthermore, the value of U_{\mathrm{sym},2}(\rho_{0},p_{F_{0}}) can be obtained through the single-nucleon potential decomposition of L Chen et al. 2012 : \displaystyle L= \displaystyle\frac{2}{3}\frac{p_{F_{0}}^{2}}{2m_{s,0}^{\ast}}-\frac{1}{6}\frac{p_{F_{0}}^{3}}{(m_{s,0}^{\ast})^{2}}\left.\frac{\partial m_{s,0}}{\partial p}\right|_{p=p_{F_{0}}}+\frac{3}{2}U_{\mathrm{sym},1}(\rho_{0},p_{F_{0}}) (55) \displaystyle+p_{F_{0}}\left.\frac{\partial U_{\mathrm{sym},1}(\rho_{0},p)}{\partial p}\right|_{p=p_{F_{0}}}+3U_{\mathrm{sym},2}(\rho_{0},p_{F_{0}}). Note that Eq. ( 55 ) is shown here at \rho_{0} but it is actually valid for arbitrary density \rho Chen et al. 2012 . Considering that the magnitude of U_{\mathrm{sym},2} is comparable to that of U_{\mathrm{sym},1} , the Lane potential Lane 1962 could be a good approximation for U_{\tau} only when the isospin asymmetry \delta is small, and the contribution of U_{\mathrm{sym},2} to L should not be neglected, which has been demonstrated in Ref. Chen et al. 2012 . Moreover, considering that the \partial m_{s,0}/\partial p is usually small, the magnitude of U_{\mathrm{sym},2} could be used to evaluate \Delta m_{1}^{\ast}(\rho_{0}) once the values of E_{\mathrm{sym}}(\rho_{0}) and L can be relatively well constrained. IV.3 Neutron star properties To study the properties of neutron stars, we adopt the conventional neutron star model, in which the neutron star is assumed to consist of an outer crust, an inner crust, and a uniform core composed of neutrons, protons, electrons and possible muons. The core-crust transition density \rho_{t} , which separates the liquid core from the nonuniform inner crust, is derived from the so-called thermodynamical method (see, e.g. Ref. Xu et al. 2009a ). The critical density \rho_{\mathrm{out}} between the inner and the outer crust is taken to be \rho_{\mathrm{out}}=2.46\times 10^{-4}\,\mathrm{fm}^{-3} Carriere et al. 2003 ; Xu et al. 2009b ; Xu et al. 2009a . For the outer crust, where \rho<\rho_{\mathrm{out}} , we use the Baym-Pethick-Sutherland EOS Baym et al. 1971 ; Iida and Sato 1997 ; for the core, where \rho>\rho_{t} , the EOS is obtained using the MDI3Y interactions under the \beta -equilibrium and charge-neutralized conditions; for the inner crust, where \rho_{\mathrm{out}}<\rho<\rho_{t} , we construct the EOS by interpolation with the form Carriere et al. 2003 ; Xu et al. 2009b ; Xu et al. 2009a P=a+b\epsilon^{4/3}. (56) Given the EOS of neutron star matter P(\epsilon) , the M-R relation of static neutron stars can be obtained by solving the Tolman-Oppenheimer-Volkoff (TOV) equation Tolman 1939 ; Oppenheimer and Volkoff 1939 . Shown in Fig. 12 is the M-R relation of neutron stars predicted by the new MDI3Y interactions. Also shown in Fig. 12 are the simultaneous M-R determinations obtained from the observations of various compact stars, including PSR J0030+0451 Miller et al. 2019 ; Riley et al. 2019 and PSR J0437-4715 Choudhury et al. 2024 with a mass around 1.4 M_{\odot} , PSR J0740+6620 Miller et al. 2021 ; Riley et al. 2021 with a mass around 2.0 M_{\odot} , as well as the CCO within HESS J1731-347 Doroshenko et al. 2022 with an unusually low mass around 0.77 M_{\odot} and small radius around 10.4 km. All contours are plotted for 68.3\% credible intervals (CI). It is seen that the four interactions with L=35 MeV predict similar M-R relations, and they are all consistent with the measurements of PSR J0030+0451, PSR J0437-4715, PSR J0740+6620, and the CCO in HESS J1731-347 within 68.3\% CI, as required in the construction of the four interactions. These results imply that varying U_{\mathrm{sym}} may have negligible impacts on the maximum mass of neutron stars but noticeable changes on their radii. In particular, MDI3YL35D017 predicts smaller radii for low-mass neutron stars but larger radii for massive ones. As shown in Fig. 12 , the M-R relations obtained from MDI3YL55D02 and MDI3YL75D02 are compatible with the observations for both PSR J0030+0451 and PSR J0740+6620, but their radii are too large to meet the constraints for PSR J0437-4715 and the CCO in HESS J1731-347. Figure 12: M-R relation for static neutron stars from the MDI3Y interactions. The astrophysical constraints for PSR J0030+0451 Miller et al. 2019 ; Riley et al. 2019 , PSR J0437-4715 Choudhury et al. 2024 , PSR J0740+6620 Miller et al. 2021 ; Riley et al. 2021 , and the CCO in HESS J1731-347 Doroshenko et al. 2022 are also included. All contours are plotted for 68.3\% CI. Table 4: The core-crust transition density \rho_{t} , the radius R_{1.4} and dimensionless tidal deformability \Lambda_{1.4} of 1.4M_{\odot} neutron star, as well as the central density \rho_{\mathrm{cen}}^{\mathrm{TOV}} and mass M_{\mathrm{TOV}} of the maximum mass neutron star configuration, predicted by the MDI3YX interactions. X= L35D017 L35D02 L35D02r L35Dm02 L55D02 L75D02 \rho_{t} ( \mathrm{fm}^{-3} ) 0.0954 0.103 0.105 0.107 0.0880 0.0799 R_{1.4} (km) 11.47 11.42 11.39 11.34 13.07 13.69 \Lambda_{1.4} 252.2 226.0 217.2 219.3 516.3 579.5 \rho_{\mathrm{cen}}^{\mathrm{TOV}} ( \mathrm{fm}^{-3} ) 0.99 1.03 1.05 1.03 1.00 1.00 M_{\mathrm{TOV}}/M_{\odot} 2.31 2.30 2.29 2.30 2.22 2.20 In Table 4 , we list some neutron star properties, including the core-crust transition density \rho_{t} , the radius R_{1.4} and dimensionless tidal deformability \Lambda_{1.4} of 1.4M_{\odot} neutron star, as well as the central density \rho_{\mathrm{cen}}^{\mathrm{TOV}} and mass M_{\mathrm{TOV}} of the maximum mass neutron star configuration. It is seen from Table 4 that all the six MDI3Y interactions satisfy the constraint of \Lambda_{1.4}\leq 580 from the gravitational wave signal GW170817 Abbott et al. 2018 . From the results of the four interactions with L=35 MeV, one sees that U_{\rm sym} can have mild influence on the values of \rho_{t} and \Lambda_{1.4} . Shown in Fig. 13 is the squared sound speed C_{s}^{2}\equiv\mathrm{d}P/\mathrm{d}\epsilon of neutron star matter as a function of nucleon density, and the location of the \rho_{\mathrm{cen}}^{\mathrm{TOV}} is marked. It is seen that the causality condition C_{s}^{2}\leq c^{2} is satisfied for all the six interactions. An obvious peak structure can be observed in the four interactions with L=35 MeV, which results from the corresponding symmetry energy that is soft at intermediate densities but stiff at high densities by construction, as shown in Fig. 6 , and varying U_{\mathrm{sym}} can change the location and height of the peak. The peak disappears when L increases to 55 and 75 MeV, which is consistent with the conclusions in Ref. Wang et al. 2024 . Figure 13: The squared sound speed ( C_{s}^{2} ) of neutron star matter as a function of nucleon density predicted by the MDI3Y interactions. The central density ( \rho_{\mathrm{cen}}^{\mathrm{TOV}} ) of maximum mass neutron star configuration is marked as solid symbol. V SUMMARY AND OUTLOOK We have extended the traditional MDI model to the so-called MDI3Y model by including two additional finite-range Yukawa interactions and modifying the density-dependent terms following the idea of Fermi momentum expansion. This extension significantly enhances the flexibility of momentum behavior of the single-nucleon potential. For example, the single-nucleon potential can have very different high energy behaviors while nicely describe the known empirical data at lower energies; the isoscalar nucleon effective mass can exhibit a peak structure in the density dependence with m_{s}^{\ast}/m>1 at low densities; and the symmetry potential can display non-monotonic momentum dependence by independently adjusting the values of the linear isospin splitting coefficient \Delta m_{1}^{\ast}(\rho) of nucleon effective mass, the isovector nucleon effective mass m_{v}^{\ast} and the limit of the symmetry potential at infinite nucleon momentum U_{\mathrm{sym}}^{\infty} . At the same time, due to the extended density-dependent terms, the MDI3Y model exhibits highly flexible density dependence of nuclear matter EOS, especially for the symmetry energy. Based on the MDI3Y model, we have built a series of interactions with various symmetry potentials, including both monotonic and non-monotonic, and different symmetry energy behaviors. The parameters of these interactions are obtained under the constraints from microscopic calculations, terrestrial experiments and astrophysical observations. With these new interactions, we have studied the properties of nuclear matter and neutron stars. Moreover, the exchange contribution from the finite-range interaction can be approximated by that from a Skyrme-like zero-range interaction based on density matrix expansion, and the corresponding energy density and single-nucleon potential are expressed in terms of local densities. In future, the MDI3Y interactions constructed in the present work can be used to replace the traditional MDI interactions in transport models, allowing for a more comprehensive investigation of the symmetry energy, symmetry potential, and the isospin splitting of nucleon effective masses from heavy-ion collisions. In addition, the MDI3Y model can be further extended to include octet baryons to study the properties of hyperon stars and the hyperon production in heavy-ion collisions. Acknowledgements. We are grateful to Xin Li, Rui Wang, Jun-Ting Ye, Ying-Xun Zhang, and Zhen Zhang for useful discussions. This work was supported in part by the National Natural Science Foundation of China under Grant No 12235010, the National SKA Program of China (Grant No. 2020SKA0120300), and the Science and Technology Commission of Shanghai Municipality (Grant No. 23JC1402700). Appendix A RELATIONS BETWEEN PARAMETERS IN MDI3Y AND M3Y-P n INTERACTIONS The central part of the two-body interaction in M3Y-P n Nakada 2003 ; Nakada 2008a ; Nakada 2010 ; Nakada 2013 is written as \displaystyle V_{\mathrm{M3Y}}^{C}(\vec{r}_{1},\vec{r}_{2})= \displaystyle\sum_{i=1}^{3}\bigg[t_{i}^{(\mathrm{SE})}P_{\mathrm{SE}}+t_{i}^{(\mathrm{TE})}P_{\mathrm{TE}} (57) \displaystyle+t_{i}^{(\mathrm{SO})}P_{\mathrm{SO}}+t_{i}^{(\mathrm{TO})}P_{\mathrm{TO}}\bigg]\frac{e^{-\mu_{i}\left|\vec{r}_{1}-\vec{r}_{2}\right|}}{\mu_{i}\left|\vec{r}_{1}-\vec{r}_{2}\right|}, where the projection operators on the singlet-even (SE), triplet-even (TE), singlet-odd (SO), and triplet-odd (TO) states are \displaystyle P_{\mathrm{SE}} \displaystyle=\frac{1-P_{\sigma}}{2}\frac{1+P_{\tau}}{2},~~P_{\mathrm{TE}}=\frac{1+P_{\sigma}}{2}\frac{1-P_{\tau}}{2}, (58) \displaystyle P_{\mathrm{SO}} \displaystyle=\frac{1-P_{\sigma}}{2}\frac{1-P_{\tau}}{2},~~P_{\mathrm{TO}}=\frac{1+P_{\sigma}}{2}\frac{1+P_{\tau}}{2}. By substituting Eq. ( 58 ) into Eq. ( 57 ) and comparing with Eq. ( 1 ), we can obtain the relations between the parameters as Nakada 2003 \displaystyle t_{i}^{(\mathrm{SE})}= \displaystyle W_{i}-B_{i}-H_{i}+M_{i}, (59) \displaystyle t_{i}^{(\mathrm{TE})}= \displaystyle W_{i}+B_{i}+H_{i}+M_{i}, \displaystyle t_{i}^{(\mathrm{SO})}= \displaystyle W_{i}-B_{i}+H_{i}-M_{i}, \displaystyle t_{i}^{(\mathrm{TO})}= \displaystyle W_{i}+B_{i}-H_{i}-M_{i}. The first version of M3Y-P n interactions Nakada 2003 introduced the DD term as Eq. ( 2 ) with n=1 . In the subsequent versions of M3Y-P n Nakada 2008a ; Nakada 2010 ; Nakada 2013 , the DD term was replaced by \displaystyle V_{\mathrm{M3Y}}^{\mathrm{DD}}= \displaystyle\Big(t_{\rho}^{(\mathrm{SE})}P_{\mathrm{SE}}[\rho(\vec{R})]^{\alpha^{(\mathrm{SE})}} (60) \displaystyle+t_{\rho}^{(\mathrm{TE})}P_{\mathrm{TE}}[\rho(\vec{R})]^{\alpha^{(\mathrm{TE})}}\Big)\delta(\vec{r}_{1}-\vec{r}_{2}), with \alpha^{(\mathrm{TE})}=1/3 and \alpha^{(\mathrm{SE})}=1 . The Eq. ( 2 ) can be rewritten as \displaystyle V^{\mathrm{DD}}= \displaystyle\sum_{n=1,3,5}\frac{1}{6}t_{3}^{[n]}\Big[(1-x_{3}^{[n]})(P_{\mathrm{SE}}+P_{\mathrm{SO}}) (61) \displaystyle+(1+x_{3}^{[n]})(P_{\mathrm{TE}}+P_{\mathrm{TO}})\Big]\rho^{n/3}(\vec{R})\delta(\vec{r}_{1}-\vec{r}_{2}). Considering that V^{\mathrm{DD}} has no contributions to the SO and TO channels when calculating Eq. ( 4 ), Eq. ( 61 ) will reduce to Eq. ( 60 ) when taking x_{3}^{[1]}=1 , x_{3}^{[3]}=-1 , and t_{3}^{[5]}=0 . We then obtain t_{\rho}^{(\mathrm{TE})}=\frac{1}{3}t_{3}^{[1]} and t_{\rho}^{(\mathrm{SE})}=\frac{1}{3}t_{3}^{[3]} . Appendix B HARTREE-FOCK CALCULATIONS WITH FINITE-RANGE INTERACTION B.1 HAMILTONIAN DENSITY The contribution from the finite-range interaction V^{C} in Eq. ( 4 ) can be obtained as E_{r}=E_{r}^{D}+E_{r}^{E}, (62) where \displaystyle E_{r}^{D}= \displaystyle\frac{1}{2}\sum_{i,j}\langle ij|V^{C}|ij\rangle (63) \displaystyle= \displaystyle\frac{1}{2}\sum_{i=1}^{3}\int\!\!\!\!\int\mathrm{d}^{3}r_{1}\mathrm{d}^{3}r_{2}\frac{e^{-\mu_{i}|\vec{r}_{1}-\vec{r}_{2}|}}{\mu_{i}|\vec{r}_{1}-\vec{r}_{2}|}\bigg\{(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2}) \displaystyle\times\Big[\rho_{n}(\vec{r}_{1})\rho_{n}(\vec{r}_{2})+\rho_{p}(\vec{r}_{1})\rho_{p}(\vec{r}_{2})\Big] \displaystyle+2(W_{i}+\frac{B_{i}}{2})\rho_{n}(\vec{r}_{1})\rho_{p}(\vec{r}_{2})\bigg\}, and \displaystyle E_{r}^{E}= \displaystyle\frac{1}{2}\sum_{i,j}\langle ij|V^{C}(-P_{M}P_{\sigma}P_{\tau})|ij\rangle (64) \displaystyle= \displaystyle\frac{1}{2}\sum_{i=1}^{3}\int\!\!\!\!\int\mathrm{d}^{3}r_{1}\mathrm{d}^{3}r_{2}\frac{e^{-\mu_{i}|\vec{r}_{1}-\vec{r}_{2}|}}{\mu_{i}|\vec{r}_{1}-\vec{r}_{2}|}\bigg\{(M_{i}+\frac{H_{i}}{2}-B_{i}-\frac{W_{i}}{2}) \displaystyle\times\Big[\rho_{n}(\vec{r}_{1},\vec{r}_{2})\rho_{n}(\vec{r}_{2},\vec{r}_{1})+\rho_{p}(\vec{r}_{1},\vec{r}_{2})\rho_{p}(\vec{r}_{2},\vec{r}_{1})\Big] \displaystyle+2(M_{i}+\frac{H_{i}}{2})\rho_{n}(\vec{r}_{1},\vec{r}_{2})\rho_{p}(\vec{r}_{2},\vec{r}_{1})\bigg\}, represent the direct and exchange contributions, respectively. In the derivations of Eqs. ( 63 ) and ( 64 ), P_{\tau} is replaced by \delta_{\tau_{1}\tau_{2}} since there is no isospin mixing state, and the effect of P_{\sigma} is simply replaced by a factor of \frac{1}{2} for the spin-averaged quantities. The matrix element of the density operator \hat{\rho}_{\tau}=\sum_{i}|i\tau\rangle\langle i\tau| is expressed as \rho_{\tau}(\vec{r}_{1},\vec{r}_{2})=\langle\vec{r}_{1}|\hat{\rho}_{\tau}|\vec{r}_{2}\rangle , which reduces to the local density when \vec{r}_{1}=\vec{r}_{2} . To obtain the energy density functional from finite-range interaction, we introduce the coordinate transformation: \vec{r}=(\vec{r}_{1}+\vec{r}_{2})/2,~~\vec{s}=\vec{r}_{1}-\vec{r}_{2}. (65) The direct contribution can be expressed as E_{r}^{D}=\int\mathrm{d}^{3}r\mathcal{H}_{r}^{D}(\vec{r}), (66) where \displaystyle\mathcal{H}_{r}^{D}(\vec{r})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}\int\mathrm{d}^{3}s\frac{e^{-\mu_{i}s}}{\mu_{i}s}\bigg\{(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2}) (67) \displaystyle\times\Big[\rho_{n}(\vec{r}+\frac{\vec{s}}{2})\rho_{n}(\vec{r}-\frac{\vec{s}}{2})+\rho_{p}(\vec{r}+\frac{\vec{s}}{2})\rho_{p}(\vec{r}-\frac{\vec{s}}{2})\Big] \displaystyle+2(W_{i}+\frac{B_{i}}{2})\rho_{n}(\vec{r}+\frac{\vec{s}}{2})\rho_{p}(\vec{r}-\frac{\vec{s}}{2})\bigg\}. With the integral \int\mathrm{d}^{3}se^{-\mu s}/s=4\pi/\mu^{2} , direct contribution reduces to the Skyrme type for uniform nuclear matter: \displaystyle\mathcal{H}_{r}^{D}(\vec{r})= \displaystyle\bigg[\sum_{i=1}^{3}\frac{4\pi}{\mu_{i}^{3}}(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2})\bigg]\frac{\rho_{n}^{2}+\rho_{p}^{2}}{2} (68) \displaystyle+\bigg[\sum_{i=1}^{3}\frac{4\pi}{\mu_{i}^{3}}(W_{i}+\frac{B_{i}}{2})\bigg]\rho_{n}\rho_{p}, which is denoted as \mathcal{H}^{\mathrm{loc}}(\vec{r}) [Eq. ( 8 )] in the main text. Expressing the density matrix in terms of the Wigner function through the Fourier transform: \displaystyle f_{\tau}(\vec{r},\vec{p})=\frac{1}{(2\pi\hbar)^{3}}\int\rho_{\tau}(\vec{r}+\frac{\vec{s}}{2},\vec{r}-\frac{\vec{s}}{2})e^{-\mathrm{i}\vec{p}\cdot\vec{s}/\hbar}\mathrm{d}^{3}s, (69) \displaystyle\rho_{\tau}(\vec{r}+\frac{\vec{s}}{2},\vec{r}-\frac{\vec{s}}{2})=\int f_{\tau}(\vec{r},\vec{p})e^{\mathrm{i}\vec{p}\cdot\vec{s}/\hbar}\mathrm{d}^{3}p, the integrals in the exchange contribution can be rewritten as \displaystyle\int\mathrm{d}^{3}r\mathrm{d}^{3}s\frac{e^{-\mu s}}{\mu s}\rho_{\tau}(\vec{r}+\frac{\vec{s}}{2},\vec{r}-\frac{\vec{s}}{2})\rho_{\tau^{\prime}}(\vec{r}-\frac{\vec{s}}{2},\vec{r}+\frac{\vec{s}}{2}) (70) \displaystyle= \displaystyle\int\mathrm{d}^{3}r\mathrm{d}^{3}p\mathrm{d}^{3}p^{\prime}\mathrm{d}^{3}s\frac{e^{-\mu s}}{\mu s}f_{\tau}(\vec{r},\vec{p})e^{\mathrm{i}\vec{p}\cdot\vec{s}/\hbar}f_{\tau^{\prime}}(\vec{r},\vec{p}^{\,\prime})e^{-\mathrm{i}\vec{p}^{\,\prime}\cdot\vec{s}/\hbar} \displaystyle= \displaystyle\int\mathrm{d}^{3}r\mathrm{d}^{3}p\mathrm{d}^{3}p^{\prime}\frac{4\pi}{\mu^{3}}\frac{f_{\tau}(\vec{r},\vec{p})f_{\tau^{\prime}}(\vec{r},\vec{p}^{\,\prime})}{1+\left(\vec{p}-\vec{p}^{\,\prime}\right)^{2}/(\hbar\mu)^{2}}, where we have used the integral \int\mathrm{d}^{3}se^{\mathrm{i}(\vec{p}-\vec{p}^{\,\prime})\cdot\vec{s}/\hbar}\frac{e^{-\mu s}}{\mu s}=\frac{4\pi}{\mu^{3}}\frac{1}{1+\left(\vec{p}-\vec{p}^{\,\prime}\right)^{2}/(\hbar\mu)^{2}}. (71) The exchange contribution can then be expressed as E_{r}^{E}=\int\mathrm{d}^{3}r\mathcal{H}_{r}^{E}(\vec{r}), (72) where \displaystyle\mathcal{H}_{r}^{E}(\vec{r})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}\frac{4\pi}{\mu_{i}^{3}}\int\!\!\!\!\int\frac{\mathrm{d}^{3}p\mathrm{~d}^{3}p^{\prime}}{1+\left(\vec{p}-\vec{p}^{\,\prime}\right)^{2}/(\hbar\mu_{i})^{2}} (73) \displaystyle\times\bigg\{(M_{i}+\frac{H_{i}}{2}-B_{i}-\frac{W_{i}}{2}) \displaystyle\times\Big[f_{n}(\vec{r},\vec{p})f_{n}\left(\vec{r},\vec{p}^{\,\prime}\right)+f_{p}(\vec{r},\vec{p})f_{p}\left(\vec{r},\vec{p}^{\,\prime}\right)\Big] \displaystyle+2(M_{i}+\frac{H_{i}}{2})f_{n}(\vec{r},\vec{p})f_{p}\left(\vec{r},\vec{p}^{\,\prime}\right)\bigg\}, which is denoted as \mathcal{H}^{\mathrm{MD}}(\vec{r}) [Eq. ( 9 )] in the main text. B.2 DENSITY MATRIX EXPANSION Based on the density matrix expansion Negele and Vautherin 1972 and by analogy with the results for the MDI model Xu and Ko 2010 , the exchange energy density \mathcal{H}_{r}^{E}(\vec{r}) from finite-range interaction can be approximated by the Skyrme-like zero-range energy density \mathcal{H}_{\mathrm{SL}}^{E}(\vec{r}) , which is expressed in terms of densities \rho_{\tau}(\vec{r}) as well as kinetic energy densities \uptau_{\tau}(\vec{{r}})=\sum_{i}|\bm{\nabla}\langle\vec{r}|i\tau\rangle|^{2} : \displaystyle\mathcal{H}_{\mathrm{SL}}^{E}(\vec{r})= \displaystyle A[\rho_{n}(\vec{r}),\rho_{p}(\vec{r})] (74) \displaystyle+B[\rho_{n}(\vec{r}),\rho_{p}(\vec{r})]\uptau_{n}(\vec{{r}})+B[\rho_{p}(\vec{r}),\rho_{n}(\vec{r})]\uptau_{p}(\vec{{r}}) \displaystyle+C[\rho_{n}(\vec{r}),\rho_{p}(\vec{r})]|\bm{\nabla}\rho_{n}(\vec{r})|^{2} \displaystyle+C[\rho_{p}(\vec{r}),\rho_{n}(\vec{r})]|\bm{\nabla}\rho_{p}(\vec{r})|^{2} \displaystyle+D[\rho_{p}(\vec{r}),\rho_{n}(\vec{r})]\bm{\nabla}\rho_{n}(\vec{r})\cdot\bm{\nabla}\rho_{p}(\vec{r}), where \displaystyle A(\rho_{n},\rho_{p})= \displaystyle V_{\mathrm{NM}}(\rho_{n},\rho_{p})-\frac{3}{5}(3\pi^{2})^{2/3}[\rho_{n}^{5/3}B(\rho_{n},\rho_{p}) \displaystyle+\rho_{p}^{5/3}B(\rho_{p},\rho_{n})], (75) \displaystyle B(\rho_{n},\rho_{p})= \displaystyle-2[\rho_{n}V^{L}(\rho_{n})+\rho_{p}V^{U}(\rho_{p},\rho_{n})], (76) \displaystyle C(\rho_{n},\rho_{p})= \displaystyle-\frac{\partial F(\rho_{n},\rho_{p})}{\partial\rho_{n}}, (77) \displaystyle D(\rho_{n},\rho_{p})= \displaystyle-\frac{\partial F(\rho_{n},\rho_{p})}{\partial\rho_{p}}-\frac{\partial F(\rho_{p},\rho_{n})}{\partial\rho_{n}}. (78) We have defined that \displaystyle V_{\mathrm{NM}}(\rho_{n},\rho_{p})= \displaystyle\rho_{n}^{2}V_{\mathrm{NM}}^{L}(\rho_{n})+\rho_{p}^{2}V_{\mathrm{NM}}^{L}(\rho_{p}) (79) \displaystyle+2\rho_{n}\rho_{p}V_{\mathrm{NM}}^{U}(\rho_{n},\rho_{p}), with \displaystyle V_{\mathrm{NM}}^{L}(\rho_{\tau})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}(M_{i}+\frac{H_{i}}{2}-B_{i}-\frac{W_{i}}{2}) (80) \displaystyle\times\int\mathrm{d}^{3}s\rho_{\mathrm{SL}}^{2}(k_{\tau}s)\frac{e^{-\mu_{i}s}}{\mu_{i}s}, \displaystyle V_{\mathrm{NM}}^{U}(\rho_{\tau},\rho_{\tau^{\prime}})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}(M_{i}+\frac{H_{i}}{2}) (81) \displaystyle\times\int\mathrm{d}^{3}s\rho_{\mathrm{SL}}(k_{\tau}s)\rho_{\mathrm{SL}}(k_{\tau^{\prime}}s)\frac{e^{-\mu_{i}s}}{\mu_{i}s}. Other terms are defined as F(\rho_{\tau},\rho_{\tau^{\prime}})=\frac{1}{2}V^{L}(\rho_{\tau})\rho_{\tau}+\frac{1}{2}V^{U}(\rho_{\tau},\rho_{\tau^{\prime}})\rho_{\tau^{\prime}}, (82) \displaystyle V^{L}(\rho_{\tau})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}(M_{i}+\frac{H_{i}}{2}-B_{i}-\frac{W_{i}}{2}) (83) \displaystyle\times\int\mathrm{d}^{3}ss^{2}\rho_{\mathrm{SL}}(k_{\tau}s)g(k_{\tau}s)\frac{e^{-\mu_{i}s}}{\mu_{i}s}, \displaystyle V^{U}(\rho_{\tau},\rho_{\tau^{\prime}})= \displaystyle\frac{1}{2}\sum_{i=1}^{3}(M_{i}+\frac{H_{i}}{2}) (84) \displaystyle\times\int\mathrm{d}^{3}ss^{2}\rho_{\mathrm{SL}}(k_{\tau}s)g(k_{\tau^{\prime}}s)\frac{e^{-\mu_{i}s}}{\mu_{i}s}, where \displaystyle\rho_{\mathrm{SL}}(k_{\tau}s)= \displaystyle\frac{3}{k_{\tau}s}j_{1}(k_{\tau}s), (85) \displaystyle g(k_{\tau}s)= \displaystyle\frac{35}{2(k_{\tau}s)^{3}}j_{3}(k_{\tau}s), (86) with j_{1} and j_{3} being, respectively, the first- and third-order spherical Bessel functions and k_{\tau}=(3\pi^{2}\rho_{\tau})^{1/3} being the Fermi wave number. It is seen that A(\rho_{n},\rho_{p}) and D(\rho_{n},\rho_{p}) are symmetric in \rho_{n} and \rho_{p} , while B(\rho_{n},\rho_{p}) and C(\rho_{n},\rho_{p}) are not. The single nucleon potential from finite-range interaction can be expressed as U_{r,\tau}(\rho_{n},\rho_{p})=U_{r,\tau}^{D}(\rho_{n},\rho_{p})+U_{r,\tau}^{E}(\rho_{n},\rho_{p}), (87) with the direct contribution being \displaystyle U_{r,\tau}^{D}(\rho_{n},\rho_{p})=\frac{\partial\mathcal{H}_{r}^{D}(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}=\sum_{i=1}^{3}\int\mathrm{d}^{3}r^{\prime}\frac{e^{-\mu_{i}|\vec{r}-\vec{r}^{\,\prime}|}}{\mu_{i}|\vec{r}-\vec{r}^{\,\prime}|}\bigg[(W_{i}+\frac{B_{i}}{2}-H_{i}-\frac{M_{i}}{2})\rho_{\tau}(\vec{r}^{\,\prime})+(W_{i}+\frac{B_{i}}{2})\rho_{-\tau}(\vec{r}^{\,\prime})\bigg], (88) which reduces to (A_{l}\rho_{\tau}+A_{u}\rho_{-\tau})/\rho_{0} in the uniform nuclear matter, as given in Eq. ( 17 ). Since \mathcal{H}_{\mathrm{SL}}^{E}(\vec{r}) includes terms with density gradients, the exchange contribution can be calculated as Kolomietz and Shlomo 2017 \displaystyle U_{r,\tau}^{E}(\rho_{n},\rho_{p})= \displaystyle\frac{\partial\mathcal{H}_{\mathrm{SL}}^{E}(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}-\bm{\nabla}\cdot\frac{\partial\mathcal{H}_{\mathrm{SL}}^{E}(\rho_{n},\rho_{p})}{\partial(\bm{\nabla}\rho_{\tau})} (89) \displaystyle= \displaystyle\frac{\partial A(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}+\frac{\partial B(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}\uptau_{n}+\frac{\partial B(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}\uptau_{p} \displaystyle+\frac{\partial C(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}(\bm{\nabla}\rho_{n})^{2}+\frac{\partial C(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}(\bm{\nabla}\rho_{p})^{2}+\frac{\partial D(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}\bm{\nabla}\rho_{n}\cdot\bm{\nabla}\rho_{p} \displaystyle-2\delta_{\tau,n}\bigg[\frac{\partial C(\rho_{n},\rho_{p})}{\partial\rho_{n}}(\bm{\nabla}\rho_{n})^{2}+\frac{\partial C(\rho_{n},\rho_{p})}{\partial\rho_{p}}\bm{\nabla}\rho_{n}\cdot\bm{\nabla}\rho_{p}+C(\rho_{n},\rho_{p})\nabla^{2}\rho_{n}\bigg] \displaystyle-2\delta_{\tau,p}\bigg[\frac{\partial C(\rho_{p},\rho_{n})}{\partial\rho_{p}}(\bm{\nabla}\rho_{p})^{2}+\frac{\partial C(\rho_{p},\rho_{n})}{\partial\rho_{n}}\bm{\nabla}\rho_{n}\cdot\bm{\nabla}\rho_{p}+C(\rho_{p},\rho_{n})\nabla^{2}\rho_{p}\bigg] \displaystyle-\bigg[\frac{\partial D(\rho_{n},\rho_{p})}{\partial\rho_{n}}\bm{\nabla}\rho_{n}\cdot\bm{\nabla}\rho_{-\tau}+\frac{\partial D(\rho_{n},\rho_{p})}{\partial\rho_{p}}\bm{\nabla}\rho_{p}\cdot\bm{\nabla}\rho_{-\tau}+D(\rho_{n},\rho_{p})\nabla^{2}\rho_{-\tau}\bigg]. With the extended Thomas-Fermi approximation Brack et al. 1985 ; Jennings 1976 : \uptau_{\tau}=a\rho_{\tau}^{5/3}+b\frac{(\bm{\nabla}\rho_{\tau})^{2}}{\rho_{\tau}}+c\nabla^{2}\rho_{\tau}, (90) where a=\frac{3}{5}(3\pi^{2})^{2/3} , b=1/36 , and c=1/3 , the single-nucleon potential from finite-range interaction can then be written as U_{r,\tau}(\rho_{n},\rho_{p})=U_{r,\tau}^{D}(\rho_{n},\rho_{p})+U_{r,\tau}^{0}(\rho_{n},\rho_{p})+U_{r,\tau}^{\nabla}(\rho_{n},\rho_{p}), (91) where U_{r,\tau}^{0}(\rho_{n},\rho_{p})=\frac{\partial A(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}+a\frac{\partial B(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}\rho_{n}^{5/3}+a\frac{\partial B(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}\rho_{p}^{5/3}, (92) \displaystyle U_{r,\tau}^{\nabla}(\rho_{n},\rho_{p})=G_{\tau}^{nn1}(\bm{\nabla}\rho_{n})^{2}+G_{\tau}^{pp1}(\bm{\nabla}\rho_{p})^{2}+G_{\tau}^{np}\bm{\nabla}\rho_{n}\cdot\bm{\nabla}\rho_{p}+G_{\tau}^{nn2}\nabla^{2}\rho_{n}+G_{\tau}^{pp2}\nabla^{2}\rho_{p}, (93) with G_{\tau}^{nn1}=\frac{b}{\rho_{n}}\frac{\partial B(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}+(1-2\delta_{\tau,n})\frac{\partial C(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}-\delta_{\tau,p}\frac{\partial D(\rho_{n},\rho_{p})}{\partial\rho_{n}}, (94) G_{\tau}^{pp1}=\frac{b}{\rho_{p}}\frac{\partial B(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}+(1-2\delta_{\tau,p})\frac{\partial C(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}-\delta_{\tau,n}\frac{\partial D(\rho_{n},\rho_{p})}{\partial\rho_{p}}, (95) G_{\tau}^{np}=-2\frac{\partial C(\rho_{\tau},\rho_{-\tau})}{\partial\rho_{-\tau}}, (96) G_{\tau}^{nn2}=c\frac{\partial B(\rho_{n},\rho_{p})}{\partial\rho_{\tau}}-2\delta_{\tau,n}C(\rho_{n},\rho_{p})-\delta_{\tau,p}D(\rho_{n},\rho_{p}), (97) G_{\tau}^{pp2}=c\frac{\partial B(\rho_{p},\rho_{n})}{\partial\rho_{\tau}}-2\delta_{\tau,p}C(\rho_{p},\rho_{n})-\delta_{\tau,n}D(\rho_{n},\rho_{p}). (98) Appendix C THE CHARACTERISTIC QUANTITIES OF THE SNM EOS AND THE SYMMETRY ENERGY, THE FOURTH-ORDER SYMMETRY ENERGY, AND THE NUCLEON EFFECTIVE MASSES In this Appendix, we provide the expressions for some quantities appearing in the main text within the MDI3Y model. Based on Eq. ( 18 ), the EOS of SNM E_{0}(\rho) can be expressed as \displaystyle E_{0}(\rho)= \displaystyle\frac{3a^{2}\hbar^{2}}{10m}\rho^{2/3}+(A_{l}+A_{u})\frac{\rho}{4\rho_{0}}+\frac{t_{3}^{[1]}}{16}\rho^{4/3}+\frac{t_{3}^{[3]}}{16}\rho^{6/3}+\frac{t_{3}^{[5]}}{16}\rho^{8/3} (99) \displaystyle+\frac{1}{12\pi^{4}\hbar^{6}\rho_{0}\rho}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\Lambda_{i}^{2}\left[p_{F}^{2}(6p_{F}^{2}-\Lambda_{i}^{2})-8\Lambda_{i}p_{F}^{3}\arctan\frac{2p_{F}}{\Lambda_{i}}+\frac{1}{4}(\Lambda_{i}^{4}+12p_{F}^{2}\Lambda_{i}^{2})\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right]. The symmetry energy E_{\mathrm{sym}}(\rho) can be expressed as \displaystyle E_{\mathrm{sym}}(\rho)= \displaystyle\frac{a^{2}\hbar^{2}}{6m}\rho^{2/3}+(A_{l}-A_{u})\frac{\rho}{4\rho_{0}}-\frac{1}{48}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}-\frac{1}{48}t_{3}^{[3]}\left(2x_{3}^{[3]}+1\right)\rho^{6/3}-\frac{1}{48}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3} (100) \displaystyle-\frac{1}{6\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}C_{u,i}\,\Lambda_{i}^{2}\,p_{F}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}+\frac{1}{24\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\Lambda_{i}^{2}\,p_{F}\left(4-\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right). The Taylor expansion coefficients defined in Eqs. ( 28 )-( 32 ) are commonly used to characterize the density behaviors of E_{0}(\rho) and E_{\mathrm{sym}}(\rho) , which can be expressed as \displaystyle L_{0}(\rho)= \displaystyle\frac{3a^{2}\hbar^{2}}{5m}\rho^{2/3}+(A_{l}+A_{u})\frac{3\rho}{4\rho_{0}}+\frac{t_{3}^{[1]}}{4}\rho^{4/3}+\frac{3t_{3}^{[1]}}{8}\rho^{6/3}+\frac{t_{3}^{[5]}}{2}\rho^{8/3} (101) \displaystyle+\frac{a^{6}\rho}{12\pi^{4}\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\left[6\frac{\Lambda_{i}^{2}}{p_{F}^{2}}+3\frac{\Lambda_{i}^{4}}{p_{F}^{4}}-\frac{3\Lambda_{i}^{4}(4p_{F}^{2}+\Lambda_{i}^{2})}{4p_{F}^{6}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle K_{0}(\rho)= \displaystyle-\frac{3a^{2}\hbar^{2}}{5m}\rho^{2/3}+\frac{t_{3}^{[1]}}{4}\rho^{4/3}+\frac{9t_{3}^{[3]}}{8}\rho^{6/3}+\frac{5t_{3}^{[5]}}{2}\rho^{8/3} (102) \displaystyle+\frac{a^{9}\hbar^{3}\rho^{2}}{4\pi^{4}\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\left[-\frac{4p_{F}^{2}\Lambda_{i}^{2}+6\Lambda_{i}^{4}}{p_{F}^{7}}+\frac{\Lambda_{i}^{4}(8p_{F}^{2}+3\Lambda_{i}^{2})}{2p_{F}^{9}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle J_{0}(\rho)= \displaystyle\frac{12a^{2}\hbar^{2}}{5m}\rho^{2/3}-\frac{t_{3}^{[1]}}{2}\rho^{4/3}+5t_{3}^{[5]}\rho^{8/3} (103) \displaystyle+\frac{a^{12}\hbar^{6}\rho^{3}}{4\pi^{4}\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\left[\frac{2(40p_{F}^{4}\Lambda_{i}^{2}+110p_{F}^{2}\Lambda_{i}^{4}+27\Lambda_{i}^{6})}{p_{F}^{10}(4p_{F}^{2}+\Lambda_{i}^{2})}-\frac{56p_{F}^{2}\Lambda_{i}^{4}+27\Lambda_{i}^{6}}{2p_{F}^{12}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle I_{0}(\rho)= \displaystyle-\frac{84a^{2}\hbar^{2}}{5m}\rho^{2/3}+\frac{5t_{3}^{[1]}}{2}\rho^{4/3}-5t_{3}^{[5]}\rho^{8/3} (104) \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{3\Lambda_{i}^{2}}{4\pi^{2}\hbar^{3}\rho_{0}p_{F}^{3}}\Bigg[-4p_{F}^{2}\frac{32p_{F}^{6}+1272p_{F}^{4}\Lambda_{i}^{2}+626p_{F}^{2}\Lambda_{i}^{4}+81\Lambda_{i}^{6}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{2}}+\left(140p_{F}^{2}\Lambda_{i}^{2}+81\Lambda_{i}^{4}\right)\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg], \displaystyle H_{0}(\rho)= \displaystyle\frac{64168a^{2}\hbar^{2}}{m}\rho^{2/3}-20t_{3}^{[1]}\rho^{4/3}+20t_{3}^{[5]}\rho^{8/3} (105) \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{3\Lambda_{i}^{2}}{4\pi^{2}\hbar^{3}\rho_{0}p_{F}^{3}}\Bigg[4p_{F}^{2}\frac{14080p_{F}^{8}+72864p_{F}^{6}\Lambda_{i}^{2}+53840p_{F}^{4}\Lambda_{i}^{4}+13970p_{F}^{2}\Lambda_{i}^{6}+1215\Lambda_{i}^{8}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{3}} \displaystyle-5\left(364p_{F}^{2}\Lambda_{i}^{2}+243\Lambda_{i}^{4}\right)\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg], and \displaystyle L(\rho)= \displaystyle\frac{a^{2}\hbar^{2}}{3m}\rho^{2/3}+(A_{l}-A_{u})\frac{3\rho}{4\rho_{0}}-\frac{1}{12}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}-\frac{1}{8}t_{3}^{[3]}\left(2x_{3}^{[3]}+1\right)\rho^{6/3} (106) \displaystyle-\frac{1}{6}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3}-\frac{\rho}{4\rho_{0}}\sum_{i=1}^{3}C_{u,i}\left[\frac{8\Lambda_{i}^{2}}{4p_{F}^{2}+\Lambda_{i}^{2}}+\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right] \displaystyle+\frac{\rho}{16\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\left[\frac{4\Lambda_{i}^{2}(4p_{F}^{2}-\Lambda_{i}^{2})}{p_{F}^{2}(4p_{F}^{2}+\Lambda_{i}^{2})}+\frac{\Lambda_{i}^{4}}{p_{F}^{4}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle K_{\mathrm{sym}}(\rho)= \displaystyle-\frac{a^{2}\hbar^{2}}{3m}\rho^{2/3}-\frac{1}{12}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}-\frac{3}{8}t_{3}^{[3]}\left(2x_{3}^{[3]}+1\right)\rho^{6/3}-\frac{5}{6}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3} (107) \displaystyle-\sum_{i=1}^{3}C_{u,i}\frac{p_{F}\,\Lambda_{i}^{2}}{2p_{F_{0}}^{3}}\left[\frac{4p_{F}^{2}\left(-4p_{F}^{2}+\Lambda_{i}^{2}\right)}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{2}}-\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right] \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{p_{F}\,\Lambda_{i}^{2}}{4p_{F_{0}}^{3}}\left[\frac{4\left(-8p_{F}^{4}+6p_{F}^{2}\Lambda_{i}^{2}+\Lambda_{i}^{4}\right)}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{2}}-\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle J_{\mathrm{sym}}(\rho)= \displaystyle\frac{4a^{2}\hbar^{2}}{3m}\rho^{2/3}+\frac{1}{6}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}-\frac{5}{3}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3} (108) \displaystyle-\sum_{i=1}^{3}C_{u,i}\frac{p_{F}\,\Lambda_{i}^{2}}{2p_{F_{0}}^{3}}\left[\frac{4p_{F}^{2}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{3}}\left(48p_{F}^{4}-40p_{F}^{2}\Lambda_{i}^{2}-5\Lambda_{i}^{4}\right)+5\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right] \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{p_{F}\,\Lambda_{i}^{2}}{4p_{F_{0}}^{3}}\left[\frac{4}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{3}}\left(160p_{F}^{6}-192p_{F}^{4}\Lambda_{i}^{2}-70p_{F}^{2}\Lambda_{i}^{4}-7\Lambda_{i}^{6}\right)+7\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], \displaystyle I_{\mathrm{sym}}(\rho)= \displaystyle-\frac{28a^{2}\hbar^{2}}{3m}\rho^{2/3}-\frac{5}{6}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}+\frac{5}{3}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3} (109) \displaystyle-\sum_{i=1}^{3}C_{u,i}\frac{4p_{F}\,\Lambda_{i}^{2}}{p_{F_{0}}^{3}}\left[\frac{4p_{F}^{2}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{4}}\left(-112p_{F}^{6}+248p_{F}^{4}\Lambda_{i}^{2}+65p_{F}^{2}\Lambda_{i}^{4}+5\Lambda_{i}^{6}\right)-5\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right] \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{p_{F}\,\Lambda_{i}^{2}}{2p_{F_{0}}^{3}}\Bigg[\frac{4}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{3}}\left(-2560p_{F}^{8}+4128p_{F}^{6}\Lambda_{i}^{2}+2400p_{F}^{4}\Lambda_{i}^{4}+490p_{F}^{2}\Lambda_{i}^{6}+35\Lambda_{i}^{8}\right) \displaystyle-35\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg], \displaystyle H_{\mathrm{sym}}(\rho)= \displaystyle\frac{280a^{2}\hbar^{2}}{3m}\rho^{2/3}+\frac{20}{3}t_{3}^{[1]}\left(2x_{3}^{[1]}+1\right)\rho^{4/3}-\frac{20}{3}t_{3}^{[5]}\left(2x_{3}^{[5]}+1\right)\rho^{8/3} (110) \displaystyle-\sum_{i=1}^{3}C_{u,i}\frac{4p_{F}\,\Lambda_{i}^{2}}{p_{F_{0}}^{3}}\Bigg[\frac{4p_{F}^{2}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{5}}\left(2368p_{F}^{8}-15120p_{F}^{6}\Lambda_{i}^{2}-6100p_{F}^{4}\Lambda_{i}^{4}-955p_{F}^{2}\Lambda_{i}^{6}-55\Lambda_{i}^{8}\right)+55\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg] \displaystyle+\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{p_{F}\,\Lambda_{i}^{2}}{2p_{F_{0}}^{3}}\Bigg[\frac{4}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{5}}(112640p_{F}^{10}-224896p_{F}^{8}\Lambda_{i}^{2}-182560p_{F}^{6}\Lambda_{i}^{4}-56840p_{F}^{4}\Lambda_{i}^{6} \displaystyle-8190p_{F}^{2}\Lambda_{i}^{8}-455\Lambda_{i}^{10})+455\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg]. The fourth-order symmetry energy can be expressed as \displaystyle E_{\mathrm{sym},4}(\rho)\equiv\left.\frac{1}{4!}\frac{\partial^{4}E(\rho,\delta)}{\partial\delta^{4}}\right|_{\delta=0}= \displaystyle\frac{a^{2}\hbar^{2}}{162m}\rho^{2/3}+\sum_{i=1}^{3}\frac{p_{F}}{108p_{F_{0}}^{3}}\Bigg[C_{u,i}\left(2p_{F}^{2}+7\Lambda_{i}^{2}\right)+C_{l,i}\Lambda_{i}^{2}\frac{40p_{F}^{4}+42p_{F}^{2}\Lambda_{i}^{2}+7\Lambda_{i}^{4}}{\left(4p_{F}^{2}+\Lambda_{i}^{2}\right)^{2}} (111) \displaystyle-\frac{\Lambda_{i}^{2}}{4p_{F}^{2}}\left(16C_{u,i}p_{F}^{2}+7C_{l,i}\Lambda_{i}^{2}+7C_{u,i}\Lambda_{i}^{2}\right)\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\Bigg]. In the MDI3Y model, the nucleon effective mass can be generally expressed as \displaystyle\left[\frac{m_{\tau}(\rho,\delta,p)}{m}\right]^{-1}\equiv \displaystyle 1+\frac{m}{p}\frac{\mathrm{d}U_{\tau}(\rho,\delta,p)}{\mathrm{d}p} (112) \displaystyle= \displaystyle 1+\frac{m}{4\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}C_{l,i}\frac{\Lambda_{i}^{2}}{p^{3}}\left[4p\,p_{F,\tau}-(p^{2}+p_{F,\tau}^{2}+\Lambda_{i}^{2})\ln\frac{(p+p_{F,\tau})^{2}+\Lambda_{i}^{2}}{(p-p_{F,\tau})^{2}+\Lambda_{i}^{2}}\right] \displaystyle+\frac{m}{4\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}C_{u,i}\frac{\Lambda_{i}^{2}}{p^{3}}\left[4p\,p_{F,-\tau}-(p^{2}+p_{F,-\tau}^{2}+\Lambda_{i}^{2})\ln\frac{(p+p_{F,-\tau})^{2}+\Lambda_{i}^{2}}{(p-p_{F,-\tau})^{2}+\Lambda_{i}^{2}}\right]. The isoscalar nucleon effective mass m_{s}^{\ast} is the nucleon effective mass in SNM, i.e., \displaystyle\left(\frac{m_{s}^{\ast}}{m}\right)^{-1}\equiv \displaystyle 1+\frac{m}{p}\frac{\mathrm{d}U_{0}(\rho,p)}{\mathrm{d}p}\bigg|_{p=p_{F,\tau}=p_{F}} (113) \displaystyle= \displaystyle 1+\frac{m}{4\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}\left(C_{l,i}+C_{u,i}\right)\frac{\Lambda_{i}^{2}}{p_{F}^{3}}\left[4p_{F}^{2}-(2p_{F}^{2}+\Lambda_{i}^{2})\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], while the isovector nucleon effective mass m_{v}^{\ast} equals to the effective mass of proton (neutron) in pure neutron (proton) matter, i.e., \displaystyle\left(\frac{m^{\ast}_{v}}{m}\right)^{-1}\equiv 1+\frac{m}{p}\frac{\mathrm{d}U_{\tau}(\rho,\delta=-\tau,p)}{\mathrm{d}p}\bigg|_{p=p_{F,\tau}=0}=1-\frac{8m}{3\pi^{2}\hbar^{3}\rho_{0}}\sum_{i=1}^{3}C_{u,i}\frac{p_{F}^{3}\Lambda_{i}^{2}}{(2^{2/3}p_{F}^{2}+\Lambda_{i}^{2})^{2}}. (114) The linear isospin splitting coefficient \Delta m^{\ast}_{1}(\rho) can be expressed as \displaystyle\Delta m^{\ast}_{1}(\rho)\equiv\left.\frac{\partial m_{\mathrm{n}\text{-}\mathrm{p}}^{\ast}(\rho,\delta)}{\partial\delta}\right|_{\delta=0}=\frac{8}{3}p_{F}^{5}\frac{\mathcal{A}(\rho)}{\left[\mathcal{B}(\rho)\right]^{2}}, (115) where \displaystyle\mathcal{A}(\rho)=\sum_{i=1}^{3}\Lambda_{i}^{2}\left\{\left(2\tilde{C}_{l,i}+3\tilde{C}_{u,i}\right)+\tilde{C}_{l,i}\frac{\Lambda_{i}^{2}}{4p_{F}^{2}+\Lambda_{i}^{2}}-\left[\frac{1}{2}\left(\tilde{C}_{l,i}+3\tilde{C}_{u,i}\right)+\frac{3}{4}\left(\tilde{C}_{l,i}+\tilde{C}_{u,i}\right)\frac{\Lambda_{i}^{2}}{p_{F}^{2}}\right]\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right\}, (116) and \displaystyle\mathcal{B}(\rho)=p_{F}^{3}+\sum_{i=1}^{3}\Lambda_{i}^{2}\left[4(\tilde{C}_{l,i}+\tilde{C}_{u,i})p_{F}^{2}-(\tilde{C}_{l,i}+\tilde{C}_{u,i})\left(2p_{F}^{2}+\Lambda_{i}^{2}\right)\ln\frac{4p_{F}^{2}+\Lambda_{i}^{2}}{\Lambda_{i}^{2}}\right], (117) with \tilde{C}_{l,i}=\frac{3m}{8p_{F_{0}}^{3}}C_{l,i} and \tilde{C}_{u,i}=\frac{3m}{8p_{F_{0}}^{3}}C_{u,i} .