WorksPhysics and Astronomy
Hybrid Stars with Different Amount of Quark Matter
Abstract
In a previous work, we explored the possibility of neutron stars with large quark cores. Through Bayesian inference, we obtained large sets of equations of state imposing constraints regarding the properties of nuclear matter, mass-radius data from NICER, the GW170817 data from Virgo-LIGO collaboration, perturbative QCD and causality. We also imposed that the maximum mass neutron star (NS) must have a large quark core. In this work, we obtain a new set of equations of state (SmallQ set) with the same constraints as before. However, we relaxed the last constraint, imposing only that NS must have quark core, regardless of its size. We compare the results with the set from our previous paper constrained by NICER and GW170817 observations NJL-GW. For both sets, we use the Relativistic Mean Field model with nonlinear meson interactions for the hadron phase and the Nambu–Jona-Lasinio model with four- and eight-quark interaction terms for the quark phase. The SmallQ set has higher values of phase transition density, with 3 – 4.5 , compared to the NJL-GW, with 2 – 2.5 . The new set describes NSs with smaller radius and smaller tidal deformability, and is more compatible with NICER and GW170817 data. A difference in the quark matter composition is observed, with a larger quark strange fraction occurring in the SmallQ set. Within this set, NSs may contain little to no quark matter. It is shown that there is a non negligible fraction of EOS whose mass-radius curve presents a positive slope for 1.8 stars, while for nucleonic stars this fraction is negligible and for hyperonic stars the slope is always negative.
Introduction
Neutron stars (NSs) are the second most compact objects in the Universe. Their inner core can reach several densities above the nuclear saturation density (), which cannot be reproduced by terrestrial experiments or described by first-principle calculations. Such extreme conditions may allow for the presence of exotic matter, as hyperons or quark matter. The latter case forms what is known as hybrid stars.
The presence of exotic matter in NS’s core may present different signatures, such as in the mass-radius diagram and in the tidal deformability. These quantities can be compared with the observational data of X-ray observations from NICER [1, 2, 3, 4, 5, 6, 7] and the GW170817 data [8] generated by the Virgo-LIGO Collaboration. The observational data filter the EOS and may indicate whether a hybrid star is more likely than a hadron star [9, 10, 11]. Although the NS’s radius data still remains of large uncertainty, the J0740+6620 observation [3, 4, 5] already imposes a strong constraint on the mass, . This constraint may be difficult to reconcile with GW170817 since it requires a soft EOS while the constraint requires a stiff EOS.
Other quantities may also indicate the presence of exotic matter, such as the profile of nuclear matter properties (NMP) [11], polytropic index () [12], measure of conformability (, where is the trace anomaly and ) [13], and quantities related to binary NS mergers [14, 15]. These methods have been used in agnostic models, whose equations of state (EOS) are generated without assuming any microscopic behavior, and are constrained by NS data, ab-initio calculations at low (chEFT) and high density (pQCD). In [12, 13], the conditions and were used to indicate quark matter behavior.
Microscopic models generate the EOS from a Lagrangian, which contains information about the composition of matter. This is a clear advantage for probing quark matter in NSs compared with the agnostic models. In this work, we use the microscopic models approach to generate the EOS and compare it with the results from [12, 13]. We expect to understand which composition may generate the different regions of phase space obtained in those works.
It has been proposed that the composition of NS matter may also be inferred from the shape of the mass-radius or mass-tidal deformability relation. In fact, exotic matter may leave a strong imprint on the slope of the mass-radius diagram. In [16], the authors have shown that NSs with only nucleonic matter in their composition predominantly present a negative slope of the mass-radius curve at , while NSs with hyperonic matter present a positive slope at small masses, , and a negative slope for . In [11], we obtained a positive slope at for of the hybrid EOS. Even at , about of the stars still presented a positive slope. In addition to the radius, the tidal deformability derivatives may also indicate the presence of exotic matter inside a NS [17].
In a previous work [11], we used Bayesian inference to obtain five different sets of equations of state (EOS): one set contains only hadron matter, and the other four sets contain hybrid EOS. The hadron phase is described by the Relativistic Mean Field (RMF) model with non-linear meson interactions [18] and the quark phase is described by the Nambu–Jona-Lasinio (NJL) [19, 20, 21] with added four- and eight-quark interaction terms or by the Mean Field Theory of QCD (MFTQCD) [22]. Although the imposed constraints in the Bayesian inference had some differences between the sets, we have always imposed that the maximum mass NS generated by the EOS must have large quark cores. As a consequence, the phase transition occurs at low density: – for the three NJL sets, and – for the MFTQCD set. In addition, large quark cores require quite stiff nuclear EoS, much stiffer than predicted by nuclear EoS.
In [23], the STAR Collaboration analyzes the results of the Au+Au reactions at a collision energy of GeV in a dense medium. It is shown that hadron interactions dominate in these conditions, indicating that hadron matter likely constitutes the medium. These results seem to indicate that these reactions can reach – and that the quark-hadron phase transition has to occur at higher densities according to [24].
In this work, we obtained a new set of hybrid EOS (SmallQ set), with the hadron and quark phases described by the RMF and NJL models. We impose the same constraints from the NJL-GW set of our previous work [11], but we relax the constraints on the phase transition density and on the difference between the central pressure of the NS and the pressure at the phase transition density, which are both related to the quark core size. These resulted in NSs with small quark cores and a quark-hadron phase transition occurring at larger densities, which are more compatible with the results obtained in [24]. Similar to the NJL-GW set from our previous work, the Bayesian inference also constrains the SmallQ set to satisfy nuclear matter properties (NMP), pQCD constraints at from [25, 26], GW170817 data from the Virgo-LIGO Collaboration [8], and the pulsars J0030-0451 [2, 7], J0740-6630 [3, 5], and J0437-4715 [6, 27] from NICER. The pQCD constraint is imposed at because, for the large majority of our EOS, the central density of the maximum-mass NS lies below this value (the 90% CI upper limit of is , see Table IV), so that the whole density range realised in stable NSs is required to be consistent with pQCD.
| Set SmallQ | Set NJL-GW | |||||
| quant | median | min | max | median | min | max |
| 0.592 | 0.484 | 0.741 | 0.362 | 0.314 | 0.391 | |
| 182.666 | 121.281 | 276.928 | 55.989 | 33.814 | 77.024 | |
| 644.373 | 505.151 | 850.219 | 358.957 | 304.433 | 393.36 | |
| 806.523 | 630.096 | 977.917 | 395.183 | 334.473 | 458.06 | |
| 151.216 | 92.699 | 195.299 | 36.207 | 19.691 | 76.780 | |
| 0.350 | 0.295 | 0.460 | 0.455 | 0.380 | 0.500 | |
| 0.2487 | 0.1909 | 0.2795 | 0.2218 | 0.1839 | 0.2542 | |
| 2.168 | 2.047 | 2.282 | 2.108 | 1.993 | 2.212 | |
| 12.351 | 11.803 | 12.914 | 12.147 | 11.526 | 12.827 | |
| 0.806 | 0.719 | 0.898 | 0.871 | 0.767 | 0.968 | |
| 958.332 | 832.054 | 1094.927 | 1053.422 | 898.290 | 1211.631 | |
| 0.513 | 0.320 | 0.700 | 0.564 | 0.340 | 0.795 | |
| 0.126 | 0.032 | 0.447 | 1.176 | 0.777 | 1.555 | |
| 3.409 | 2.116 | 5.298 | 8.069 | 7.026 | 8.952 | |
| 12.964 | 12.725 | 13.298 | 13.448 | 13.193 | 13.698 | |
| 2.147 | 2.012 | 2.266 | 1.634 | 1.253 | 1.940 |
Table IV. The median and the 90% CI of the following quantities: the phase transition density (, in fm), pressure (, in MeV/fm), hadronic energy density (, in MeV/fm) and quark energy density (, in MeV/fm); measure of the strength of the phase transition (, in MeV/fm); onset of the quark strange (, in fm); fraction of the quark strange in the maximum mass NS (); maximum mass (, in ) and radius (, in km), density (, in fm), energy density (, in MeV/fm), speed of sound () quark core mass (, in ) and quark core radius (, in km) at the maximum NS mass; radius of the 1.4 (, in km); mass of stars with central pressure equal to (, in ).
By comparing the SmallQ (small quark core) set with the NJL-GW (large quark core) and RMF (no quark core, only hadron matter) sets from our previous paper, we aim to analyze the implications of the presence of quark matter inside NSs and identify possible properties capable of distinguishing NSs with different quark core sizes.
This article is organized as follows: Sec. II summarizes the theoretical models for the hadron and quark phases. Sec. III describes the Bayesian inference, including the prior and likelihood distributions used. In Sec. IV, we discuss the results of the EOS, the properties of NSs, the influence on the quark-hadron phase transition, and nuclear matter properties. A comparison with NS with large quark cores is discussed, as well as possible signatures of the presence of quark core. In Sec. V, we summarize the conclusions.
Equations of State
In this section, we describe the models used to obtain the equations of state (EOS) of the hybrid stars. For the hadron phase, we adopted the Relativistic Mean Field (RMF) model, and for the quark phase, we used the Nambu–Jona-Lasinio (NJL). The quark-hadron phase transition is build by using the Maxwell construction.
Hadron phase
In this work, the hadron phase is described by the Relativistic Mean Field (RMF) model with nonlinear meson interaction terms. In this model, nuclear interactions are mediated by the exchange of three mesons: the scalar-isoscalar, vector-isoscalar and vector-isovector mesons (, and , respectively). The Lagrangian is given by
where
where the Dirac spinor , which has bare mass , describes the proton and neutron fields. The Dirac matrices are represented by the , and is the isospin operator, where are the Pauli matrices. and are the coupling constants and masses of the mesons and (similar for ).
By calculating the equations of motion and applying the relativistic mean field approximation [28], the meson fields are expressed as
where for protons and neutrons, respectively; and and are the scalar and number densities. To apply this model in NSs condition, we impose chemical equilibrium and charge neutrality conditions, given by
where is the electric charge. The electron and muon are included as a free Fermi gas. The parameters , , , , , and are sampled in Bayesian inference.
Quark phase
To describe the quark phase in NSs, we use the 3-flavor Nambu–Jona–Lasinio (NJL) model with the ’t-Hooft term and four- and eight-quark interaction terms. The NJL model reproduces all the global symmetries of QCD and the spontaneous chiral symmetry breaking manifestations. It is a suitable model for low energy regimes, as in the NSs conditions. The Lagrangian is given by
where and are the quark current masses and chemical potential matrices, (with are the Gell-Mann matrices, , and the interaction terms are given by
These interaction terms were studied in [] and have an important role in describing the observational data of NSs.
By linearizing the Lagrangian and applying the mean field approximation, we obtain the gap equations
with , and are the effective mass and effective chemical potential, respectively, and the quark condensates are defined as and . We also obtain the grand canonical potential, in which, at zero temperature, is given by
where is the energy. The is a constant set to vanish the grand canonical potential in the vacuum, and is the Fermi momentum. Since the NJL is a non-renormalizable model, we use the 3-momentum cutoff scheme (). By imposing that , we obtain the expressions for the quark condensates at :
From the grand canonical potential, we obtain the EoS of the NJL model. To satisfy the NSs conditions, we impose the chemical potential equilibrium and charge neutrality condition, given by
where is the electric charge. The electron contribution is added as a free Fermi gas. Also, since the NJL is normalized in the vacuum, we set , where is a constant and has a strong influence of the phase transition location.
The parameters , , and are set to satisfy the mass and decay constant data from , , and in Table 1. The parameters , , , , , defined as
| (MeV) | (MeV) | (MeV) | ||
| 623.58 | 5.70 | 136.60 | 3.34 | -13.67 |
Table 1. Fixed parameters for NJL model, set to satisfy the experimental data from [29].
and are sampled through Bayesian analysis.
Bayesian Inference
Bayesian inference is a powerful statistical tool that estimates the parameters considering prior information and data constraints. By setting the values of the model parameters (), Bayesian inference applies the Bayes’ theorem, given by
where is the data used as a constraint. On the right side of the equation, we have the likelihood distribution (), which is the probability that the parameters satisfies the data, and the prior distribution (), that is the probability of the parameters are true before considering any information of the data. We also have the evidence () as a normalization factor. On the left side of the equation, we have the posterior distribution, which gives the combination of the probability of the prior and likelihood probabilities. This gives us the quality of the parameters in accordance with the imposed constraints. The values of are sampled and a large set of parameters with high posterior probability is returned. Different techniques of sampling can be applied. In this work, we used the Bayesian inference library BILBY [30] with the pyMultiNest package [31, 32], which is based on the nested sampling method.
To apply the Bayesian inference, in addition to the sampling method, we need to define the prior and likelihood functions. For the prior, we used a uniform distribution with the minimum and maximum values defined by Table II. For the likelihood, we consider the nuclear matter properties (NMP) and pure neutron matter (PNM) data (see Table III), the x-ray NICER data from J0030-0451 [2, 7], J0740-6630 [3, 5], J0437-4715 [6, 27], and the gravitational wave (GW) data from GW170817 [8], and the pQCD constraint from [25, 26] at . We also imposed some constraints to ensure a phase transition (see [11] for more details). The NMP are the coefficients of the expansion of the energy per nucleon around the saturation density ,
Table III. Constraints on the nuclear matter properties (NMP) and pure neutron matter (PNM) imposed in the likelihood. NMP (defined in Eq. (24)): saturation condition , binding energy per nucleon , incompressibility and symmetry energy at saturation. PNM: energy per neutron at , 0.10 and 0.15 fm from chiral effective field theory. All energies are in MeV.
where and is the isospin asymmetry. is the binding energy per nucleon, the incompressibility and the skewness of symmetric nuclear matter, while , , and are the symmetry energy and its slope, curvature and skewness at saturation. Saturation is imposed through , and is the energy per neutron of pure neutron matter. Only the quantities listed in Table III enter the likelihood; the remaining coefficients (, , and ) are predictions of the model.
In [11], we obtained five sets of EoS that reproduce NS with large quark core. In this work, we compare the NJL-GW from [11] with the new set obtained in this paper, the SmallQ set, that reproduces NS with small quark core. The differences between these two sets are the constraints related to the size of the quark cores. For the SmallQ set, we did not apply the range of phase transition constraint (). We also relaxed the quark inside constraint (). This likelihood establishes a minimum value for the difference between the pressure at NS () and . For the NJL-GW set, we defined this minimum value as MeV/fm, while for the SmallQ, we set MeV/fm. In this way, Bayesian inference still favors NS with a quark core, but with no restriction of its size.
| NJL | RMF | ||||
| Parameters | min | max | Parameters | min | max |
| 0 | 0.5 | 8 | 11 | ||
| 0 | 1 | 10 | 14 | ||
| 0 | 30 | 9.546 | 15 | ||
| 0 | 8 | BB | 1.500 | 3.500 | |
| 0 | 50 | CC | -4.627 | -1.500 | |
| (MeV/fm) | 0 | 30 | 0 | 0.016 | |
| 0 | 0.103 |
Table II. The lowest and highest values for the uniform distribution prior used for the NJL-GW and SmallQ sets. We defined and .
Results
In this section, we first discuss the EOS properties, followed by a discussion of several NS properties and hadron-quark phase transition properties. We finish by discussing the implications of the existence of a quark core on the properties of nuclear matter.
Results of the EoS
We start by comparing the EOS distributions obtained for SmallQ with the ones obtained for NJL-GW. Figure 1 shows the 90% CIs of pressure as a function of the energy density for the NJL-GW (purple), SmallQ (yellow), and RMF (black dash-dot) sets. The vertical lines mark the 90% CI maximum value of the central energy density of the maximum mass NS. The vertical bands represent the 90% CI of the hadron phase transition point (). Relaxing the constraint on the size of the quark core significantly increases the density localization of the phase transition point, shifting it to larger densities. This

Figure 1. Pressure as a function of energy density of the 90% CIs of the sets NJL-GW (purple), SmallQ (yellow) and RMF (black dash-dot). The vertical lines mark the 90% CI maximum of the central energy density in the maximum mass NS (). The vertical bands represent the 90% CI interval for the energy density at the phase transition ().
indicates that the data used to constrain the models prefer a smaller quark core. In addition, with respect to the NJL-GW Set, the central energy density decreases by about MeV/fm, and the pressure decreases in the region MeV/fm. However, due to the phase transition to quark matter at high energy densities, the pressure of the SmallQ Set remains greater than the pressure of the NJL-GW Set for MeV/fm, while the phase transition does not occur.
We also compare our results with the modelindependent results from [38, 39], which used a model independent speed of sound parametrization for the EOS. This is represented in Figure 1 by the gray solid (full) and dotted (90% CI) contours. Although both sets lie inside the full model-independent envelope, the SmallQ set is more compatible with the 90% CI contour than the NJL-GW set.
In Figure 2, we show the 90% CIs of the squared speed of sound versus the baryon density. Below and above , the two sets have the same behavior, with the NJL-GW set having slightly higher values in the former region. However, between the and , the sets produce different results. The NJL-GW curve has two bumps: the first one occurs due to the phase transition and the second occurs due to the onset of the quark strange. The SmallQ curve has only one bump because the phase transition occurs after the onset of the quark strange. Due to this difference in the phase transition point, the SmallQ Set results in larger values for the speed of sound in the intermediate region.

Figure 2. Speed of sound as a function of baryon density: the 90% CIs for the sets NJL-GW (purple) and SmallQ (yellow). The vertical lines mark the 90% CI maximum of the central energy density in the maximum mass NS (). The vertical bands represent the 90% CI interval for the energy density at the phase transition ().
Figure 3 shows three quantities related to the EOS as functions of density, which have been introduced in [12, 13, 40] to discuss quark deconfinement. From bottom to top: the normalized trace anomaly, defined as [40]; the polytropic index [12]; and the measure of conformability , where [13]. The regions defined by [13] or with [12] were proposed as identifying the presence of quark deconfined matter.

Figure 3. Measure of conformability (), polytropic index () and trace anomaly () as a function of density.
In the top plot, exhibits a decreasing behavior for both sets within their respective phase transition regions and enters the region as quark matter. However, for , the exhibits increasing behavior, returning to the region of due to the presence of a strong repulsive vector potential in the NJL model. In the middle plot, the functions decrease for
for both sets, and at , they are both in a quark deconfined phase. However, at , the values of increase above 1.75 and remain constant at . For the and results, the SmallQ set has more compatible results with the definitions in [12, 13], since the phase transition to quark deconfined matter occurs at and . The trace anomaly is shown in the bottom plot. For both sets, the majority of the values of are positive.
Discussion of the properties of NS
The EOS, pressure as a function of the energy density, is the function needed to integrate the TolmanOppenheimer-Volkov (TOV) equations and determine the NS properties. Figure 4 shows the 90% CIs of the mass-radius diagram for the NJL-GW and the SmallQ sets. In this figure, we also show the results of the RMF set, which contains only nucleonic matter. The NICER observational data of the PSR J0030+0451 [2, 7] (blue), PSR J0740+6620 [3, 5] (orange), PSR J0437+4715 [6] (green), and the HESS J1731-347 [41] (purple) with , , and are represented by solid, dashed, and dotted lines, respectively. The SmallQ set predicts smaller radii and can better describe the observational data than the NJL-GW set. This difference occurs because the SmallQ EoS is softer than the NJL-GW EoS in the hadron phase. A larger parameter stiffens the EoS. This coupling is correlated with the radius of the 1.4 NSs (), as shown in Figure 5. The corner plot also shows a strong correlation with the coupling which is imposed by the binding energy at saturation.

Figure 4. Mass-radius diagram of the 90% CI for the NJL-GW (purple), SmallQ (yellow) and RMF (black dash-dot) sets. The NICER observational data PSR J0030+0451 [2, 7] (blue), PSR J0740+6620 [3, 5] (orange), PSR J0437+4715 [6] (green), and and the HESS data J1731-347 [41] (purple) with , , are shown with solid, dashed and dotted lines, respectively.

Figure 5. Corner plot of , and of the sets NJL-GW (purple), SmallQ (yellow) and RMF (black).
In Figure 6, we show the corner plot of the quark parameters and , and the properties of maximum mass NS (), the density at which the quark strange appears (), and the quark strange fraction in the maximum mass NS (). The other parameters of the quark phase do not change significantly. We notice that the distribution of the parameter has a peak at 0 in the NJL-GW set. However, in the SmallQ set, this parameter is uniformly distributed. The has a strong correlation with the : increasing the value, the onset of the quark strange occurs at lower densities. As a consequence, the EoS becomes softer, and the phase transition density increases. This effect of can also be seen in [42]. The EOS of quark matter must be stiff to contain a large quark core. Therefore, it disfavors the onset of the s-quark, as in the NJL-GW set. Having relaxed this constraint, there is more freedom concerning the quark matter composition, and it is interesting to verify that more symmetric matter is favored with a larger s-quark fraction. In Figure 7, the 90% CI of the s-quark fraction is plotted as a function of the baryonic density. Except for the larger density, the s-quark fraction is generally larger for the SmallQ model.

Figure 6. Corner plot of , , , and for the sets NJL-GW (purple) and SmallQ (yellow).

Figure 7. Strange quark fraction, 90% CI, as a function of the baryonic density for SmallQ and NJL-GW sets. Vertical bands identify the 90% CI for the phase transition density.
The softening of the SmallQ EoS is compensated by an increase in the values, as this term stiffens the EoS and increases the values. Indeed, the corner plot shows a clear correlation between the and . This explains that, although decreases within the SmallQ set, increasing the values of the fraction of strangeness , the still increases in relation to the NJL-GW set.
Figure 8 shows the binary mass ratio as a function of the effective tidal deformability, defined as

Figure 8. Binary mass ratio as a function of tidal deformability for the 90% CIs of the NJL-GW (purple) and the SmallQ (yellow) sets. The LIGO-Virgo collaboration data GW170817 [43] is represented by the green contours, with 50% CI (solid), 90% CI (dashed) and 99% CI (dotted).
where and are the tidal deformability and the mass of the i-th NS of the binary system, respectively. The is calculated for for a fixed chirp mass of . This value corresponds to the LIGO-Virgo collaboration data from GW170817 [43], which is represented in Figure 8 for 50%, 90% and 99% CIs in green solid, dashed and dotted contours, respectively. In this figure, the NJL-GW set struggles to satisfy the GW170817 data, lying mostly within the 99% CI. However, the SmallQ model
provides a better description of this data and achieves the 50% CI contour.
In [44], analyze the effect of the slope on NS properties, considering a model-independent EoS. Properties such as maximum mass, radius, and the position of the peak, where considered. In [16], a similar analysis was carried out with EOS sets based on microscopic descriptions. The RMF model was used to compare the slope of nucleonic and hyperonic EoS. EOS with hyperons tend to have a positive slope at low masses. However, all the hyperonic EoS have a negative slope, at .
In Table V, we show the sign of the slope at , , for the NJL-GW and SmallQ hybrid sets and the RMF hadronic set from [11]. The hadronic set has a predominantly negative slope at all masses. For the hybrid sets, the positive slope is predominant at . At , 47% (52%) of the EoS in the SmallQ set have a positive (negative) slope, while 62% of the EoS in the NJL-GW set has a positive slope. At , the sign is inverted, and 78% (89%) of the EoS have a negative slope. Notice, however, that there is still a non negligible fraction of hybrid EOS with a positive slope, while for the hyperonic set no EOS is showing a positive slope at . Thus, a positive slope at low masses could indicate the presence of quark matter inside NS. Moreover, a positive slope at could certainly be a signature of the presence of quark matter. This behavior was also observed in our previous work [11].
Discussion on the Phase Transition
One of the main differences between the NJL-GW and SmallQ sets is the restriction of the value. The quantity is the central pressure of the maximum mass NS, while the is the pressure of the phase transition. The difference is constrained in the Bayesian inference as a means of controlling the size of the quark core. For the NJL-GW set, we impose that , resulting in NSs with large quark cores. For the SmallQ set, however, we relaxed this restriction and used . This means that this set still only includes hybrid stars, but allows for any size of quark core. In Figure 9, the distribution of are shown for the NJL-GW (purple) and SmallQ (yellow) sets. We notice that the SmallQ distribution is narrower than the NJL-GW one and the curve is concentrated at lower values.

Figure 9. Distribution of value for the NJL-GW (purple) and SmallQ (yellow) sets.
Another difference between the NJL-GW and SmallQ sets is the constraint for the phase transition density (). For the NJL-GW set, we imposed that the must have values between and . On the other hand, for the SmallQ set, we did not constraint the . The top panel of Figure 10 shows the distribution of the phase transition density in purple and yellow curves for the NJL-GW and the SmallQ sets, respectively. The constraint imposed for the NJL-GW set is shown in black dotted line. Notably, while the values of for the NJL-GW set remains small, satisfying the constraint, the SmallQ results in larger values of phase transition density, with between and (90% CI, see Table IV). This also affects the jump in the energy density at the phase transition (). The distribution of this quantity is shown in the bottom panel of Figure 10, in which we see that the SmallQ set has larger values for than the NJL-GW set, indicating that the former set has a stronger phase transition.
Figure 10. Distribution of phase transition density (top) and jump in the energy transition (bottom) for the NJL-GW (purple) and SmallQ (yellow) sets. The constraint in imposed in the NJL-GW set is represented by the black dotted line.
Given the constraints, in particular, the astrophysical observations, the hadronic EOS are favored with respect to the EOS with a large quark core. Giving the freedom to the model to determine the optimal size of the quark core, the model chooses a small quark core which defines a behavior similar to the one of the hadronic EOS. The transition to a quark phase within an NJL description is dictated by the density at which chiral restoration occurs, which occurs above twice times the saturation density. A model that predicts a restoration of chiral symmetry at smaller densities would probably favor larger quark cores. This will be investigated in the future.
Figure 11 shows the distributions of the maximum mass NS (), the mass in which the central density is equal to the phase transition density (), and the fraction of the quark core mass by the total mass of the maximum mass NS (). The SmallQ set has slightly larger values for the maximum mass NS than the NJL-GW set. However, both sets result in different distributions. The NJL-GW set results in smaller values of with a wider distribution. This set indicates that NS with (median value – see Table IV) can have quark matter in their core. On the other hand, the SmallQ set has a narrower distribution with larger values for the . In Table IV, the minimum of the 90% CI is (Table IV), indicating that NS with a quark core must have at least this mass. The sets also show a different behavior for the distribution of the quark core fraction in the maximum mass NS. The fraction of the quark mass for the NJL-GW set can vary from to 80%, while for the SmallQ set, most part of the distribution lies in . Notably, by relaxing the constraints, the Bayesian inference indicates a preference for NSs with smaller quark core.
Figure 11. Distribution of the maximum mass (), phase transition mass () and the fraction of the quark core mass () by the total mass in the maximum mass NS.
Table V. Number of EOSs with positive/negative slope in the mass-radius diagram at for the SmallQ set, NJL-GW and RMF sets from [11] and nucleonic and hyperonic sets from [16].
Nuclear Matter Properties
In this section, we discuss the nuclear matter properties (NMP). Table VI presents the 90 % CIs of all the NMP quantities for the SmallQ, NJL-GW and RMF sets. These quantities are defined in Eq. (24). Figure 12 shows the distributions of , , and for the NJL-GW (purple), SmallQ (yellow) and RMF (black) sets. Since the distributions of , EA and are constrained in the Bayesian inference, they do not show a significant difference in the plots and are therefore omitted here. Analyzing the figure, the SmallQ set distributions tend to lie between the NJL-GW and RMF results. This effect is because the phase transition of the SmallQ set occurs at larger densities than the NJL-GW set, and therefore, the effect of the presence of quarks at high densities is smaller. Since the SmallQ EoS contains more nucleonic matter, the distributions of this set shift towards the RMF ones. Although is constrained, there are some differences in these distributions. While the NJL-GW and SmallQ sets satisfy the constraint, the RMF shifts to larger values. This can be understood in the following way: the nucleonic RMF EOS has to be stiff to reach , and the skewness should not be too large in order to satisfy pQCD constraints. In the hybrid EOS, the hadron phase is more compatible with constraints, however the skewness is much larger to compensate. In addition, the stiffness of the quark phase can help reaching . In this case, the transition to quark matter helps satisfying pQCD constraints even with a large . We observe that SmallQ has slightly larger values of than NJL-GW for the 90 % CI minimum. However, the 90 % CI maximum remains the same. For the , although the RMF distribution is wider than that of the hybrid sets, we notice that the median value is similar for all of them. These values are also consistent with MeV, obtained by averaging different results from the literature in [45].
In the plot, the RMF distribution implies that this quantity has a negative sign, in agreement with [46], in which the authors obtained MeV by averaging the result of 16 studies. However, positive values of , as obtained from NJL-GW, are compatible with some results from [47], in which the CREX and PREX-II data were considered. The SmallQ set accepts both positive and negative values of .
Conclusion
In this work, we compare two sets of hybrid EoS, one that reproduces NSs with large quark cores, and the other, with small quark cores. The EoS are obtained using the Relativistic Mean Field (RMF) model with non-linear meson interaction terms for the hadron phase, and the Nambu–Jona-Lasinio (NJL) model with four- and eight-quark interaction terms for the quark phase. We apply the Maxwell construction to reproduce the phase transition. We use Bayesian inference to impose nuclear matter properties, pure neutron matter results from chEFT, the pQCD constraint at , as developed in [25, 26], the X-ray NICER data of J0030-0451 [2, 7], J0740-6630 [3, 5], J0437-4715 [6, 27], and the gravitational wave data from GW170817 [8]. We also apply some constraints to ensure a phase transition. For the NJL-GW set, we impose that the quark core of the must be large, while for the SmallQ set, we only impose that the must have a quark core, regardless of its size.
Due to the early quark-hadron phase transition in the NJL-GW set, the hadron phase is stiffer compared with the SmallQ set. As a consequence, the SmallQ set reproduces NSs with smaller radius and tidal deformability, which are more compatible with the NICER data and the GW170817 data. This also affects the composition of the quark phase, resulting in a larger quark strange fraction for the SmallQ set. Also, the phase transition density is larger than the onset of the quark strange density, in contrast with the NJL-GW set. This implies larger values of the speed of sound squared at intermediate densities for the SmallQ set. Furthermore, the SmallQ set is more compatible with the 90% CI agnostic model of [39], and also with the conditions of the quark phase behavior ( and ) set for the agnostic models in [12, 13].
NJL-GW set indicates the presence of quark matter inside the NS, while the SmallQ set indicates that these stars may contain little to no quark matter. Both sets have a predominantly positive mass-radius derivative for low masses. This result was already analyzed in the previous work [11] and in [16], which also indicate for exotic (quark and hyperon) matter and for hadron matter. In addition, there is still a non negligible fraction of EOSs that present a positive slope for stars, while for nucleonic stars this fraction is negligible and for hyperonic stars it was shown that for masses of this order or larger no EOS would have a positive slope. Observational data with more accurate radius values may favor the presence of exotic matter if it confirms that the NSs, such as the J0740+6620, has larger radius than the NSs (J0030+0451 and J0437-4715).
Data Availability
The final posterior of the model parameters, the equation of states and the TOV results of the SmallQ set can be obtained from the link [48]. The results for NJL-GW and RMF sets are available at [49].
Corner Plots of the Posterior
Figure 13 shows the corner plot of the hadron (left plot) and quark parameters (right plot), for the NJL-GW (purple) and SmallQ (yellow) sets. Relaxing the constraints related to the size of the quark core at affects all the hadron parameters, especially the and . These parameters are related to the , as discussed in Section IV B. The quark parameters are less affected, compared to the hadron parameters. We see notable differences in the and parameters. The former is related to the onset of the quark strange, and the latter, to the phase transition density and , also discussed in Section IV B.

Figure 13. Corner plot of hadron (left) and quark (right) parameters.
Figure 12. Nuclear matter properties distributions for the NJL-GW (purple), SmallQ (yellow) and RMF (black) sets.
| quant | Set SmallQ median | Set SmallQ min | Set SmallQ max | Set NJL-GW median | Set NJL-GW min | Set NJL-GW max | Set RMF median | Set RMF min | Set RMF max |
| 0.159 | 0.158 | 0.161 | 0.159 | 0.158 | 0.161 | 0.159 | 0.158 | 0.161 | |
| -16.001 | -16.033 | -15.967 | -16.000 | -16.032 | -15.968 | -16.000 | -16.032 | -15.968 | |
| 233.810 | 197.247 | 270.421 | 228.829 | 185.733 | 271.410 | 250.940 | 228.158 | 283.916 | |
| -296.730 | -434.109 | -23.609 | 71.412 | -273.811 | 573.546 | -445.834 | -510.475 | -346.521 | |
| 31.640 | 29.217 | 34.296 | 31.149 | 28.706 | 33.708 | 32.138 | 29.678 | 34.717 | |
| 48.425 | 40.484 | 60.597 | 54.256 | 48.117 | 62.705 | 46.898 | 34.300 | 66.925 | |
| -14.559 | -88.959 | 79.321 | 91.128 | 6.448 | 169.655 | -140.665 | -184.537 | -69.975 | |
| 1264.450 | 784.071 | 1551.287 | 1002.040 | 202.071 | 1506.580 | 1136.710 | 407.590 | 1563.186 |
Table VI. Median and 90% CI of the NMP (defined in Eq. (24)) for the SmallQ, NJL-GW and RMF sets: saturation density (fm), binding energy per nucleon , incompressibility , skewness , symmetry energy , and its slope , curvature and skewness (all in MeV).
References
References
- [1]T. E. Riley et al., A NICER View of PSR J0030+0451: Millisecond Pulsar Parameter Estimation, Astrophys. J. Lett. 887, L21 (2019), arXiv:1912.05702 [astro-ph.HE].DOI
- [2]M. C. Miller, F. K. Lamb, A. J. Dittmann, S. Bogdanov, Z. Arzoumanian, K. C. Gendreau, S. Guillot, A. K. Harding, W. C. G. Ho, J. M. Lattimer, R. M. Ludlam, S. Mahmoodifar, S. M. Morsink, P. S. Ray, T. E. Strohmayer, K. S. Wood, T. Enoto, R. Foster, T. Okajima, G. Prigozhin, and Y. Soong, Psr j0030+0451 mass and radius from nicer data and implications for the properties of neutron star matter, The Astrophysical Journal Letters 887, L24 (2019).
- [3]M. C. Miller, F. K. Lamb, A. J. Dittmann, S. Bogdanov, Z. Arzoumanian, K. C. Gendreau, S. Guillot, W. C. G. Ho, J. M. Lattimer, M. Loewenstein, S. M. Morsink, P. S. Ray, M. T. Wolff, C. L. Baker, T. Cazeau, S. Manthripragada, C. B. Markwardt, T. Okajima, S. Pollard, I. Cognard, H. T. Cromartie, E. Fonseca, L. Guillemot, M. Kerr, A. Parthasarathy, T. T. Pennucci, S. Ransom, and I. Stairs, The radius of psr j0740+6620 from nicer and xmm-newton data, The Astrophysical Journal Letters 918, L28 (2021).DOI
- [4]T. E. Riley et al., A NICER View of the Massive Pulsar PSR J0740+6620 Informed by Radio Timing and XMM-Newton Spectroscopy, Astrophys. J. Lett. 918, L27 (2021), arXiv:2105.06980 [astro-ph.HE].DOI
- [5]T. Salmi, D. Choudhury, Y. Kini, T. E. Riley, S. Vinciguerra, A. L. Watts, M. T. Wolff, Z. Arzoumanian, S. Bogdanov, D. Chakrabarty, K. Gendreau, S. Guillot, W. C. G. Ho, D. Huppenkothen, R. M. Ludlam, S. M. Morsink, and P. S. Ray, The radius of the high-mass pulsar psr j0740+6620 with 3.6 yr of nicer data, The Astrophysical Journal 974, 294 (2024).DOI
- [6]D. Choudhury, T. Salmi, S. Vinciguerra, T. E. Riley, Y. Kini, A. L. Watts, B. Dorsman, S. Bogdanov, S. Guillot, P. S. Ray, D. J. Reardon, R. A. Remillard, A. V. Bilous, D. Huppenkothen, J. M. Lattimer, N. Rutherford, Z. Arzoumanian, K. C. Gendreau, S. M. Morsink, and W. C. G. Ho, A nicer view of the nearest and brightest millisecond pulsar: Psr j0437−4715, The Astrophysical Journal Letters 971, L20 (2024).DOI
- [7]S. Vinciguerra, T. Salmi, A. L. Watts, D. Choudhury, T. E. Riley, P. S. Ray, S. Bogdanov, Y. Kini, S. Guillot, D. Chakrabarty, W. C. G. Ho, D. Huppenkothen, S. M. Morsink, Z. Wadasingh, and M. T. Wolff, An updated mass–radius analysis of the 2017–2018 nicer data set of psr j0030+0451, The Astrophysical Journal 961, 62 (2024).
- [8]B. P. Abbott et al. (LIGO Scientific, Virgo), GW170817: Measurements of neutron star radii and equation of state, Phys. Rev. Lett. 121, 161101 (2018), arXiv:1805.11581 [gr-qc].
- [9]L. Passarella, M. Guerrini, G. Pagliara, A. Lavagno, and A. Drago, Is the coexistence of strange quark stars and hadronic stars favored by astrophysical data? a bayesian analysis (2026), arXiv:2606.21435 [nucl-th].arxiv.org/abs/2606.21435
- [10]M. Albino, T. Malik, M. Ferreira, and C. Providência, Hybrid star properties with the njl and mean field approximation of qcd models: A bayesian approach, Phys. Rev. D 110, 083037 (2024).DOI
- [11]M. Albino, T. Malik, M. Ferreira, and C. Providência, Bayesian inference of hybrid stars with large quark cores, Phys. Rev. D 113, 083019 (2026).DOI
- [12]E. Annala, T. Gorda, A. Kurkela, J. Nättilä, and A. Vuorinen, Evidence for quark-matter cores in massive neutron stars, Nature Physics 16, 907–910 (2020).DOI
- [13]E. Annala, T. Gorda, J. Hirvonen, O. Komoltsev, A. Kurkela, J. Nättilä, and A. Vuorinen, Strongly interacting matter exhibits deconfined behavior in massive neutron stars, Nature Communications 14, 10.1038/s41467-023-44051-y (2023).
- [14]P. Hammond, A. Clevinger, M. Albino, V. Dexheimer, S. Bernuzzi, C. Brown, W. Cook, B. Daszuta, J. Fields, E. Grundy, C. Providência, D. Radice, and A. Steiner, Investigating the impact of higher-order phase transitions in binary neutron-star mergers, Phys. Rev. D 113, 044057 (2026).arxiv.org/abs/2508.10698
- [15]L.-J. Guo, W.-C. Yang, Y.-L. Ma, and Y.-L. Wu, Probing hadron–quark transition through binary neutron star merger, Research in Astronomy and Astrophysics 25, 035017 (2025).arxiv.org/abs/2308.01770
- [16]M. Ferreira and C. Providência, Learning about neutron star composition from the slope of the mass-radius diagram, Phys. Rev. D 112, 083058 (2025).DOI
- [17]A. Bauswein, A. Nikolaidis, G. Lioutas, H. Kochankovski, P. Char, C. Mondal, M. Oertel, L. Tolos, N. Chamel, and S. Goriely, Stellar properties indicating the presence of hyperons in neutron stars, Phys. Rev. Res. 8, 013253 (2026).
- [18]T. Malik, M. Ferreira, M. B. Albino, and C. Providência, Spanning the full range of neutron star properties within a microscopic description, Phys. Rev. D 107, 103018 (2023).
- [19]Y. Nambu and G. Jona-Lasinio, Dynamical model of elementary particles based on an analogy with superconductivity. ii, Phys. Rev. 124, 246 (1961).
- [20]S. P. Klevansky, The nambu—jona-lasinio model of quantum chromodynamics, Rev. Mod. Phys. 64, 649 (1992).DOI
- [21]T. Hatsuda and T. Kunihiro, Qcd phenomenology based on a chiral effective lagrangian, Physics Reports 247, 221 (1994).DOI
- [22]D. Fogaça and F. Navarra, Gluon condensates in a cold quark–gluon plasma, Physics Letters B 700, 236–242 (2011).DOI
- [23]B. E. Aboona et al. (STAR Collaboration), Onset of constituent quark number scaling in heavy-ion collisions at rhic, Phys. Rev. Lett. 135, 072301 (2025).DOI
- [24]G.-C. Yong, B.-A. Li, Z.-G. Xiao, and Z.-W. Lin, Probing the high-density nuclear symmetry energy with the Ξ−/Ξ0 ratio in heavy-ion collisions at √sNN ≈ 3 gev, Phys. Rev. C 106, 024902 (2022).DOI
- [25]T. Gorda, O. Komoltsev, and A. Kurkela, Ab-initio qcd calculations impact the inference of the neutron-star-matter equation of state, The Astrophysical Journal 950, 107 (2023).
- [26]O. Komoltsev, T. Gorda, and A. Kurkela, QCD likelihood function.
- [27]D. J. Reardon, M. Bailes, R. M. Shannon, C. J. Flynn, J. Askew, N. D. R. Bhat, Z.-C. Chen, M. Curlyo, Y. Feng, G. B. Hobbs, A. Kapur, M. Kerr, X. Liu, R. N. Manchester, R. M. Mankow, S. Mishra, C. J. Russell, M. Shamohammadi, L. Zhang, and A. Zic, The neutron star mass, distance, and inclination from precision timing of the brilliant millisecond pulsar j0437-4715, The Astrophys- ical Journal Letters 971, L18 (2024).
- [28]N. K. Glendenning, Compact Stars, edited by I. Appenzeller, G. Börner, M. Harwit, R. Kippenhahn, P. A. Strittmatter, and V. Trimble, Astronomy and Astrophysics Library (Springer US).
- [29]K. Olive, Review of particle physics, Chinese Physics C 38, 090001 (2014).
- [30]G. Ashton, M. Hübner, P. D. Lasky, C. Talbot, K. Ackley, S. Biscoveanu, Q. Chu, A. Divakarla, P. J. Easter, B. Goncharov, F. H. Vivanco, J. Harms, M. E. Lower, G. D. Meadors, D. Melchor, E. Payne, M. D. Pitkin, J. Powell, N. Sarin, R. J. E. Smith, and E. Thrane, Bilby: A user-friendly Bayesian inference library for gravitational-wave astronomy, The Astrophysical Journal Supplement Series 241, 27 (2019).
- [31]Buchner, J., Georgakakis, A., Nandra, K., Hsu, L., Rangel, C., Brightman, M., Merloni, A., Salvato, M., Donley, J., and Kocevski, D., X-ray spectral modelling of the agn obscuring region in the cdfs: Bayesian model selection and catalogue, A&A 564, A125 (2014).DOI
- [32]J. Buchner, Nested sampling methods, Statistics Surveys 17, 169 (2023).
- [33]S. Huth et al., Constraining Neutron-Star Matter with Microscopic and Macroscopic Collisions, Nature 606, 276 (2022), arXiv:2107.06229 [nucl-th].DOI
- [34]M. Dutra, O. Lourenço, S. S. Avancini, B. V. Carlson, A. Delfino, D. P. Menezes, C. Providência, S. Typel, and J. R. Stone, Relativistic mean-field hadronic models under nuclear matter constraints, Phys. Rev. C 90, 055203 (2014).DOI
- [35]B. G. Todd-Rutel and J. Piekarewicz, Neutron-rich nuclei and neutron stars: A new accurately calibrated interaction for the study of neutron-rich matter, Phys. Rev. Lett. 95, 122501 (2005).
- [36]S. Shlomo, V. M. Kolomietz, and G. Colò, Deducing the nuclear-matter incompressibility coefficient from data on isoscalar compression modes, Eur. Phys. J. A 30, 23 (2006).
- [37]R. Essick, P. Landry, A. Schwenk, and I. Tews, Detailed examination of astrophysical constraints on the symmetry energy and the neutron skin of 208Pb with minimal modeling assumptions, Phys. Rev. C 104, 065804 (2021).DOI
- [38]E. Annala, T. Gorda, E. Katerini, A. Kurkela, J. Nättilä, V. Paschalidis, and A. Vuorinen, Multimessenger Constraints for Ultradense Matter, Phys. Rev. X 12, 011058 (2022), arXiv:2105.05132 [astro-ph.HE].
- [39]S. Altiparmak, C. Ecker, and L. Rezzolla, On the sound speed in neutron stars, The Astrophysical Journal Letters 939, L34 (2022).
- [40]Y. Fujimoto, K. Fukushima, L. D. McLerran, and M. Praszalowicz, Trace Anomaly as Signature of Conformality in Neutron Stars, Phys. Rev. Lett. 129, 252702 (2022), arXiv:2207.06753 [nucl-th].DOI
- [41]V. Doroshenko, V. Suleimanov, G. Pühlhofer, and A. Santangelo, A strangely light neutron star within a supernova remnant, Nature Astronomy 6, 1 (2022).DOI
- [42]M. Ferreira, R. C. Pereira, and C. Providência, Hybrid stars with large strange quark cores constrained by gw170817, Phys. Rev. D 103, 123020 (2021).DOI
- [43]B. P. Abbott et al. (The LIGO Scientific Collaboration and the Virgo Collaboration), Gw170817: Measurements of neutron star radii and equation of state, Phys. Rev. Lett. 121, 161101 (2018).DOI
- [44]M. Ferreira and C. Providência, Constraining neutron star matter from the slope of the mass-radius curves, Phys. Rev. D 110, 063018 (2024).
- [45]M. Oertel, M. Hempel, T. Klähn, and S. Typel, Equations of state for supernovae and compact stars, Rev. Mod. Phys. 89, 015007 (2017).DOI
- [46]B.-A. Li, B.-J. Cai, W.-J. Xie, and N.-B. Zhang, Progress in constraining nuclear symmetry energy using neutron star observables since gw170817, Universe 7, 10.3390/universe7060182 (2021).DOI
- [47]B. T. Reed, F. J. Fattoyev, C. J. Horowitz, and J. Piekarewicz, Density dependence of the symmetry energy in the post–prex–crex era, Phys. Rev. C 109, 035803 (2024).
- [48]M. Albino, T. MALIK, M. Ferreira, and C. Providência, Hybrid stars with different amount of quark matter, 10.5281/zenodo.23471790 (2026).arxiv.org/abs/2610.28408
- [49]M. Albino, T. Malik, M. Ferreira, and C. Providência, A bayesian inference of hybrid stars with large quark cores, 10.5281/zenodo.17534438 (2025).arxiv.org/abs/2511.02653


