MS Thesis

Download as pdf or txt
Download as pdf or txt
You are on page 1of 46

Masters Thesis

An In-Depth Study of Neutron Stars and Equations of State


Marc Salinas
Department of Physics & Astronomy
California State University Long Beach
February 25, 2020

1
Abstract
Neutron stars are among the densest objects in the universe. The uncertainty of the inter-
nal structure of these stars have led to various methods for modeling the behavior of matter
at high density. In order to study the structure of compact stars, the Tolman-Oppenheimer-
Volkoff (TOV) equations are solved to yield Mass-Radius curves of different neutron star
structures. Such different structures investigated in this paper include stars of pure nuclear
matter, stars of pure quark matter, and hybrid stars. These classes of stars are probably
the only place in the universe where deconfined quark matter could exist. Because of the
possibility of quark deconfinement, Quantum Chromodynamics (QCD) plays an important
role in modeling the core of these stars. Although QCD is unsolved, we can still use some
of the main principles to obtain some possible Equations of State (EoS) to be used in con-
junction with the TOV equations. Although the EoS of the neutron star core is the bulk
of this paper, we investigate the strange matter hypothesis, the masquerade of hybrid stars
as nuclear stars, and the flavor camouflage in phase transitions, all through the use of the
Vector Interaction Enhanced Bag Model (vBag). In the end, the results of this paper can be
used in conjunction with observational astronomical data to constrain the equation of state
for neutron stars. Since these compact stars are likely the only objects in the universe where
the extreme density allows for quark deconfinement, it also provides us with one way to test
out the QCD and QFT framework for high density nuclear matter.

2
Contents
1 Introduction 4
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 The Tolman-Oppenheimer-Volkoff Equations 4


2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Derivation of the TOV equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 Crust Calculation 9
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4 Nuclear Equation of State 12


4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.2 Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.3 Parabolic Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.4 Parametrizations and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5 Quark Matter Equation of State 23


5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.2 The Thermodynamic Bag Model (tdBag) . . . . . . . . . . . . . . . . . . . . . . . . 24
5.3 Extending the tdBag Model (vBag) . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.4 Quark matter in Neutron Stars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

6 Hybrid Stars 34
6.1 Phase Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2 Camouflage of Phase Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

7 Future Work 43
7.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

A TOV Derivation 44

3
1 Introduction
1.1 Introduction
In this paper, we will begin by building the foundation needed to start studying compact objects,
including but not limited to, neutron stars. The well known Tolman-Oppenheimer-Volkoff equa-
tions are derived from Einsteins Field Equations and used throughout the entirety of this work.
We also explore equations of state (EoS) for nuclear and quark matter in that order. The parabolic
approximation, a method in solving for the equation of state for nuclear matter is used and ap-
plied to the work of [7]. Mass radius curves for various different EoS and the effect of different
parametrizations are investigated as well.

The framework for studying quark matter is developed shortly after by deriving the thermodynamic
bag model equations and witnessing the effects of varying bag pressure. This provides us with the
foundation for studying the vBag EoS. As the cornerstone of the research in this paper, it will
be used in a variety of ways thereafter. Its inclusion of vector interactions and chiral symmetry
breaking provides an additional parameter that can be tuned to study the quark EoS in more detail.
vBag is used to explore the strange matter hypothesis as was done in [9], in more detail. vBag
also provides us with methods in studying hybrid stars through various different phase transitions
tackled at the end of this paper. We will demonstrate the ability of these hybrid stars to be almost
indistinguishable from a purely hadronic star as the phase transitions are hidden. We also explore
flavor camouflage, the inability to distinguish the quark matter degrees of freedom even if a phase
transition is known to have occurred. Lastly, we discuss the implications of this research and the
future work that could be done.

2 The Tolman-Oppenheimer-Volkoff Equations


2.1 Introduction
In order to start analyzing mass-radius relationships in dense stars, a relativistic approach must
be taken. The Tolman-Oppenheimer-Volkoff (TOV) Equations take in an EoS, given some central
density, and integrate from the core to the surface to return the radius R(M, ε, P ). The TOV
equations can be shown to come out of Einsteins field equations when studying a static-spherically
symmetric object in space. A brief derivation will be given in the proceeding section [11]. A more
in depth derivation can be found in Appendix A.

2.2 Derivation of the TOV equations


The most general metric tensor to describe static spherically symmetric matter is of the form:
 2 2v 
ce 0 0 0
 0
 −e2λ 0 0 

2
 0 0 −r 0 
0 0 0 −r2 sin2 θ

4
Einstein’s field equations (EFE) given below must be solved for an appropriate stress energy tensor.
8πG
Guv = Tuv (1)
c4
The Einstein tensor Guv and the Ricci tensor Ruv involved is given by:
1
Guv = Ruv − guv R
2
The Ricci tensor is then given by the following in terms of the Christoffel symbols.

Rαβ = ∂ρ Γρβα − ∂β Γρρα + Γρρλ Γλβα − Γρβλ Γλρα (2)

After evaluating all the appropriate Christoffel symbols, the Ricci tensor components can be found
in terms of the metric tensor components:
1 1 1
R00 = − ∂i (g ii g00,i ) + g ii g 00 (g00,i )2 − g ii g jj gii,j g00,j (3)
2 4 4

1 1 1 1
Rii = − ∂i (g 00 g00,i ) + g 00 g kk g00,k (2gki,i − gii,k ) − (g 00 g00,i )2 + ∂j (g jj (2gji,i − gii,j ))
2 4 4 2
1 1
− ∂i (g jj gjj,i ) + g jj g kk gjj,k (2gki,i − gii,k ) (4)
2 4
1 jj kk
− g g (gji,k + gjk,i − gik,j )(gkj,i + gki,j − gji,k )
4
Plugging in the appropriate components of the metric tensor and taking derivatives we have a
valid expression of the Ricci tensor in terms of polar coordinates and unsolved constants.

2v 0
R00 = e2(v−λ) (v 02 + − v 0 λ0 + v 00 )
r
2λ0
R11 = − v 02 + v 0 λ0 − v 00
r (5)
R22 = e−2λ (rλ0 − rv 0 − 1) + 1
R33 = sin2 θe−2λ (rλ0 − rv 0 − 1) + sin2 θ
= R22 sin2 θ

The Ricci scalar can also easily be evaluated by contracting the Ricci tensor with the metric tensor
to give:

4λ0 4v 0
 
−2λ 2 02 0 0 00 2
R=e 2
+ 2v − + − 2v λ + 2v − 2 (6)
r r r r

5
We are now free to evaluate the Einstein tensor:

c2 2(v−λ) 2λ
G00 = e (e + 2rλ0 − 1)
r2
1 − e2λ + 2rv 0
G11 =
r2 (7)
G22 = re (rv 02 − λ0 + v 0 − rv 0 λ0 + rv 00 )
−2λ

G33 = r sin2 θe−2λ (rv 02 − λ0 + v 0 − rv 0 λ0 + rv 00 )


= G22 sin2 θ

In order to solve for the parameters e2v and e2λ , we must take the Newtonian limit and solve for
the parameters. We begin with the geodesic equation:

d2 xµ ρ
µ dx dx
σ
+ Γ ρσ =0
dτ 2 dτ dτ
For very slow movement the spatial velocity terms go to zero and we are left with:

d 2 xµ 0
µ dx dx
0
+ Γ 00 =0
dτ 2 dτ dτ
d2 xµ 1 µv dx0 dx0
+ η (g 0v,0 + g0v,0 − g 00,v ) =0
dτ 2 2 dτ dτ
Since in the Newtonian limit the gravitational field is static, the time derivatives go to zero.

d2 xµ 1 µv dx0 dx0
− η g00,v =0
dτ 2 2 dτ dτ
d 2 x0
Setting µ = 0, we see that = 0 for the same reason above. Looking at the spatial part we
dτ 2
have:
 2
d2 xi 1 ii dt
2
− η g00,i =0
dτ 2 dτ
d2 τ
Using the (+ − −−) metric and multiplying both sides by we get an important result:
dt2
d2 xi 1
2
= − g00,i
dt 2
We can take the g00 component to be some small deviation δ from the Minkowski metric at large
distances, so that g00 = (1+δ)η00 , where η00 = c2 . Plugging this back in to the previous expression
we arrive at:

d 2 xi c2
= − ∂i δ
dt2 2

6
We know from Newtonian mechanics that
d 2 xi
= −∂i V
dt2
2V
Where V is the Newtonian gravitational potential. Setting the two equal we see that δ = 2 and
c
finally:
2V
g00 = c2 e2v = c2 (1 + )
c2
GM
For V = − :
r
2GM
e2v = 1 − (8)
rc2
Outside the star the Stress-Energy tensor vanishes and the EFE read Guv = 0
1
Guv = Ruv − guv R = 0
2
1
g uα Ruv − g uα guv R = 0
2
1
Rvα − δvα R = 0
2
1
R= R
2

R = g uv Ruv = 0

Ruv = 0

2v 0
R00 = e2(v−λ) (v 02 + − v 0 λ0 + v 00 ) = 0
r
2λ0
R11 = − v 02 + v 0 λ0 − v 00 = 0
r
Solving for v 00 in the R11 equation and plugging it into the R00 equation we get v 0 + λ0 = 0. Since
the g00 and g11 components must approach the Minkowski metric in the Newtonian limit, we know
c2 e2v → c2 as r → ∞ and −e2λ → −1 as r → ∞. This means that both λ and v must both
approach zero as r → ∞. This condition allows us to show the functions do not differ by some
constant and thus v + λ = 0 must also hold true. This gives us the useful relation:

v = −λ

7
and the solution:
 −1
2λ 2GM
e = 1− (9)
rc2
The right hand side of EFE involves the stress-energy tensor. Here we use the tensor of a perfect
fluid given by:
1
Tuv = ( + p)uu uv − pguv
c2
The components come out to be:

T00 = g00
T11 = −pg11
(10)
T22 = −pg22
T33 = −pg33

With all the proper components, all that’s left is to work through all the algebra and simplify to
arrive at the TOV equations. Together the two equations can be used with an EoS to solve for
the mass and radius for different stars:

   −1
dp 4πGrp(r) GM 2GM
= − (r) + p(r) + 2 2 1−
dr c4 r c rc2 (11)
dM 4πr2 (r)
=
dr c2
Lastly, for the purpose of running code more efficiently we can make these TOV equations dimen-
sionless by means of some simple substitutions:

Let m̃ = mm , r̃ = rrs , p̃ = pps , and ˜ = s . Where rs will refer to the schwarzschild radius of the sun,
while ps and s will refer to the schwarzschild pressure and energy density of the sun given by:

m c2
ps = s = (12)
4πrs3
Making the substitutions listed above we get simpler and more practical unitless TOV equations
that will be used in running neutron star profiles:

dp̃ −(˜ + p̃)(m̃ + r̃3 p̃)


=
dr̃ 2r̃(r̃ − m̃) (13)
dm̃
= r̃2 ˜
dr̃

8
3 Crust Calculation
3.1 Introduction
The crust of the neutron star is not of extreme importance in this paper and so it will be dis-
cussed very briefly. It is also the least dense part of the star and can be constrained much more
easily through heavy ion collision experiments. In calculating the crust, the well known BBP and
BPS methods are used for the inner and outer crust respectively. The Feynman-Metrpolois-Teller
(HTWW) EoS is also used for extremely low densities. The equations of state for both layers of
crust are taken from the two papers by Baym [4, 3].

The crust of the neutron star can be broken down into several different layers starting with an
outermost layer consisting of a lattice of 56 Fe. As the density increases, the lowest favorable energy
states become increasingly neutron rich and lattices of heavier elements such as 62 Fe and 84 Se can
start to exist. In addition to the lattice of several different nuclei you also have a gas of free
neutrons and electrons penetrating the lattice. This region of free neutrons is referred to simply
as the free neutron regime [3].

The BBP model for the crust describes matter up to a density of about 5 × 1014 g/cm3 . However,
between about 104 and 4.3 × 1011 g/cm3 , in the free neutron regime, electrons are relativistic and
the lattice coulomb energy becomes very important in determining the nuclear sizes [3]. The BPS
model takes this lattice coulomb energy into account as well as uses more recent extrapolated
nuclear mass data to come up with a more accurate equation of state of the crust at higher
densities.

3.2 Results
The following equation of state tables are taken from the BPS paper [3]:

Table 1: Feynman Metropolis Teller (HTWW)

ρ P nb
(g cm−3 ) (dynes cm−2 ) (cm−3 )
7.86 ≤ 1.01E9 4.73E24
7.90 1.01E10 4.76E24
8.15 1.01E11 4.91E24
11.6 1.21E12 6.99E24
16.4 1.40E13 9.90E24
45.1 1.70E14 2.72E25
212 5.82E15 1.27E26
1150 1.90E17 6.93E26

9
Table 2: BPS Equation of State

ρ P nb ρ P nb
(g cm−3 ) (dynes cm−2 ) (cm−3 ) (g cm−3 ) (dynes cm−2 ) (cm−3 )
.. .. ..
1.044E4 9.744E18 6.295E27 . . .
2.622E4 4.968E19 1.581E28 6.601E9 4.362E27 3.972E + 33
6.587E4 2.431E20 3.972E28 8.312E9 5.662E27 5.000E + 33
1.654E5 1.151E21 9.976E28 1.046E10 7.702E27 6.294E + 33
4.156E5 5.266E21 2.506E29 1.318E10 1.048E28 7.924E + 33
1.044E6 2.318E22 6.294E29 1.659E10 1.425E28 9.976E + 33
2.622E6 9.755E22 1.581E30 2.090E10 1.938E28 1.256E + 34
6.588E6 3.911E23 3.972E30 2.631E10 2.503E28 1.581E + 34
8.293E6 5.259E23 5.000E30 3.313E10 3.404E28 1.990E + 34
1.655E7 1.435E24 9.976E30 4.172E10 4.628E28 2.506E + 34
3.302E7 3.833E24 1.990E31 5.254E10 5.949E28 3.155E + 34
6.589E7 1.006E25 3.972E31 6.617E10 8.089E28 3.972E + 34
1.315E8 2.604E25 7.924E31 8.332E10 1.100E29 5.000E + 34
2.624E8 6.676E25 1.581E32 1.049E11 1.495E29 6.294E + 34
3.304E8 8.738E25 1.990E32 1.322E11 2.033E29 7.924E + 34
5.237E8 1.629E26 3.155E32 1.664E11 2.597E29 9.976E + 34
8.301E8 3.029E26 5.000E32 1.844E11 2.892E29 1.105E + 35
1.045E9 4.129E26 6.294E32 2.096E11 3.290E29 1.256E + 35
1.316E9 5.036E26 7.924E32 2.640E11 4.473E29 1.581E + 35
1.657E9 6.860E26 9.976E32 3.325E11 5.816E29 1.990E + 35
2.626E9 1.272E27 1.581E33 4.188E11 7.538E29 2.506E + 35
4.164E9 2.356E27 2.506E33 4.299E11 7.805E29 2.572E + 35
.. .. ..
. . .

10
Table 3: BBP Equation of State

ρ P nb ρ P nb
(g cm−3 ) (dynes cm−2 ) (cm−3 ) (g cm−3 ) (dynes cm−2 ) (cm−3 )
.. .. ..
4.460E11 7.890E29 2.670E35 . . .
5.228E11 8.352E29 3.126E35 2.210E13 3.931E31 1.315E37
6.610E11 9.098E29 3.951E35 2.988E13 6.178E31 1.777E37
7.964E11 9.831E29 4.759E35 3.767E13 8.774E31 2.239E37
9.728E11 1.083E30 5.812E35 5.081E13 1.386E32 3.017E37
1.196E12 1.218E30 7.143E35 6.193E13 1.882E32 3.675E37
1.471E12 1.399E30 8.786E35 7.732E13 2.662E32 4.585E37
1.805E12 1.638E30 1.077E36 9.826E13 3.897E32 5.821E37
2.202E12 1.950E30 1.314E36 1.262E14 5.861E32 7.468E37
2.930E12 2.592E30 1.748E36 1.586E14 8.595E32 9.371E37
3.833E12 3.506E30 2.287E36 2.004E14 1.286E33 1.182E38
4.933E12 4.771E30 2.942E36 2.520E14 1.900E33 1.484E38
6.248E12 6.481E30 3.726E36 2.761E14 2.242E33 1.625E38
7.801E12 8.748E30 4.650E36 3.085E14 2.751E33 1.814E38
9.611E12 1.170E31 5.728E36 3.433E14 3.369E33 2.017E38
1.246E13 1.695E31 7.424E36 3.885E14 4.286E33 2.280E38
1.496E13 2.209E31 8.907E36 4.636E14 6.103E33 2.715E38
1.778E13 2.848E31 1.059E37 5.094E14 7.391E33 2.979E38
.. .. ..
. . .

The 3 three tables can be thought of together as one crust equation of state in which we can later
use to implement in different stellar models. Some plots of the combined crust equation of state
(HTWW+BPS+BBP) are given below for reference. To get the chemical potential at a given
pressure and energy density, we use the thermodynamic relation  + P = µb nb .

Figure 1: HTWW+BPS+BBP EoS

11
4 Nuclear Equation of State
4.1 Introduction
Nuclear matter is the bulk material of the neutron star and is thus one of the most important
pieces in constructing a possible equation of state. With the crust EoS given in the previous
section, the nuclear equation of state becomes important at around 1.5 to 2 times saturation den-
sity n0 = 0.16fm−3 . In order to‘glue’ two different equations of state we must assume some sort
of phase transition between matter in the crust to the nuclear matter in the outer core. In the
simplest case we will consider a Maxwell phase transition from crust matter to nuclear core matter.

In terms of choosing a good candidate equation of state, there are several different models and
approximations to choose from. Most nuclear equations of state are constrained by heavy ion
collision experiments up to and around saturation density. The saturation density of n0 = 0.16fm−3
is one of the experimental values used to constrain the EoS. This number comes from constraining
the radius parameter r0 of a nucleus, which is directly related to the number density by n0 =
3/(4πr03 ). The radius parameter is obtained by an analysis of electron-nucleus scattering in terms of
density distributions [11]. Other parameters such as the binding energy per nucleon B/A = −16.3
MeV and the symmetry energy coefficient (change in energy from the asymmetry of the neutron
and proton) asym = 32.5 MeV are also used to constrain the nuclear EoS. For very complex models,
more and more parameters are needed to fix certain coupling constants and/or coefficients. The
compression modulus k and the effective mass m∗ is very commonly used to constrain models
even further and carries a fair amount of freedom to them. The effective mass can range from
m∗ /m = 0.6 to m∗ /m = 0.8. The compression modulus, which defines the curvature of the
equation of state at saturation density, ranges from around 120 to 300 MeV. As we will see later,
the compression modulus plays an important role in determining the maximum masses for neutron
stars.

4.2 Approaches
There are several different approaches that can be taken in constructing a valid nuclear EoS. There
are non-relativistic potential models, field theoretical models, and hybrid models [8]. For this paper
we will focus less on the models themselves and more on the methodology of extracting an equation
of state from them. We will be employing the use of the parabolic approximation for the energy
per particle as it has the advantage of being computationally effortless and very flexible in the fact
that it only takes a change of parametrizations to switch between different models. This has the
added benefit of being able to compare several different nuclear equations of state without having
to solve a whole new set of field equations which can be computationally intensive as the models
get more and more complex.

4.3 Parabolic Approximation


In the parabolic approximation framework the energy per particle E(n, β) is given by:

E(n, β) = E0 (n) + β 2 Es (n) (14)

12
Here β is an asymmetry parameter depending on the proton fraction x = np /n [16]. The total
baryon number density n or nb is given by n = np + nn , where np and nn are the proton number
density and neutron number density respectively. E0 is the binding energy per nucleon of symmetric
nuclear matter (nn = np ), and Es is the asymmetry energy given by [7, 16]:

Es (n) = E(n, β = 1) − E(n, β = 0) (15)


Equation (14) has been shown by Lagaris and Pandharipande [5] to be an excellent approximation
of the binding energy per nucleon as the higher order beta terms β n≥4 are very small compared
to the mass of the nucleon. In this case, all we need is the energy per particle and the following
thermodynamic relations to obtain the energy density , the pressure P , and the chemical potential
µ [2, 16]:

B = nE(n, β)
 
2 ∂E(n, β)
PB = n
∂n S,N (16)
 
∂
µB =
∂n S,V

Here the pressure is taken at constant entropy S and particle number N . The chemical potential
is taken at constant entropy and volume V . The chemical potential for the neutron and proton are
obtained by evaluating the derivative of the energy density in respect to the neutron and proton
number densities as follows:

∂
µn =
∂nn
∂(nE(n, β))
=
∂nn

= E0 + β 2 Es + n (E0 + β 2 Es )
∂nn
∂E0 ∂β ∂Es
= E0 + β 2 Es + n + 2nEs β + nβ 2
∂nn ∂nn ∂nn
(17)
∂E0 ∂x ∂Es
= E0 + β 2 Es + n − 4nEs β + nβ 2
∂nn ∂nn ∂nn
∂E 0 ∂E s
= E0 + β 2 Es + n + 4xEs β + nβ 2
∂nn ∂nn
∂E0 ∂Es
= E0 + n + 2βEs − β 2 Es + nβ 2
∂nn ∂nn
   
∂ 2 2 ∂
µn = 1 + n E0 (n) − β − 2β − β n Es (n)
∂n ∂n

13
∂
µp =
∂np
∂(nE(n, β))
=
∂np

= E0 + β 2 Es + n (E0 + β 2 Es)
∂np
∂E0 ∂β ∂Es
= E0 + β 2 Es + n + 2nEs β + nβ 2
∂np ∂np ∂np
(18)
∂E0 ∂x ∂Es
= E0 + β 2 Es + n − 4nEs β + nβ 2
∂np ∂np ∂np
∂E0 ∂Es
= E0 + β 2 Es + n − 4(1 − x)Es β + nβ 2
∂np ∂np
∂E0 ∂Es
= E0 + n − 2βEs − β 2 Es + nβ 2
∂np ∂np
   
∂ 2 2 ∂
µp = 1 + n E0 (n) − β + 2β − β n Es (n)
∂n ∂n

We can rewrite the above expressions for the neutron and proton chemical potentials by denoting
a sign change: (upper sign) for neutron and (lower sign) for proton.
   
∂ 2 2 ∂
µn,p = 1 + n E0 (n) − β ∓ 2β − β n Es (n) (19)
∂n ∂n
In order to solve for various different equations of state, all that remain are models for the binding
energy E0 (n) in symmetric nuclear matter and the asymmetry energy Es (n).

4.4 Parametrizations and Results


As previously mentioned, there are many different classes of models to consider when calculating
the nuclear equation of state. In this paper, we will explore two types of potential models as well
as a relativistic Hartree-Fock approximation (HFA) model. Let us start with the parametrization
of two different potential models. In the potential model formulation the total energy density can
be written as [7]:

 = kin kin
n + p + V (np , nn , T ) (20)

14
The potential V (np , nn , T ) is parametrized as [7]:

   
1 3 1
V (np , nn , T ) = An0 − + x0 (1 − 2x) u22
3 2 2
   
2 3 1
Bn0 − + x3 (1 − 2x) uσ+12
3 2 2
+    
2 0 3 1
1+ B − + x3 (1 − 2x)2 uσ−1 (21)
3 2 2
d3 k
 Z
2 X
+ u (2Ci + 4Zi )2 g(k, Λi )(fn + fp )
5 i=1,2 (2π)3
d3 k
Z 
+ (Ci − 8Zi )2 g(k, Λi )[fn (1 − x) + fp x]
(2π)3
Here x is the proton fraction, u = n/n0 , and n0 is the nuclear saturation density. The constants
A, B, σ, C1 , C2 , B 0 , x0 , x3 , Z1 , and Z2 are fixed through experimentation and empirical knowledge.
We can solve for the binding energy per nucleon in symmetric matter by setting the proton fraction
x = 1/2 and fn = fp , where fi is the Fermi-Dirac distribution function. This simplifies our
expression and we have the potential for symmetric matter as:

1 Bn0 uσ+1 X Z ∞ d3 k
2
Vnm = An0 u + + 4u Ci g(k, Λi )fn (22)
2 1 + B 0 uσ−1 i=1,2 0 (2π)3
We can find the potential for pure neutron matter in the same fashion, by setting x = 0 and fp = 0:

2
1 Bn0 uσ+1 (1 − x3 ) 2 X
Z ∞ 3
dk
Vnem 2
= An0 u (1 − x0 ) + 3 + u (6Ci − 8Zi ) g(k, Λi )fn (23)
3 2 5 i=1,2 (2π)3
1 + B 0 uσ−1 (1 − x3 ) 0
3
The kinetic part of the energy density for nuclear matter is fairly easily to evaluate. For a Fermi
gas of protons and neutrons at zero temperature, the total kinetic energy density is:

3 p2F,N 3 p2F,P
kin = nN + nP (24)
5 2m 5 2m
Here nN and nP are the neutron and proton number densities respectively. pF,i are the Fermi
momenta, which we can write in terms of the number densities. Note the zero temperature
approximation is valid here as neutron star temperatures are small (KeV), compared to the energies
on the nuclear scale (MeV). To solve for the Fermi momenta of a degenerate Fermi gas, we start
with the number of particles N given by [14]:
Z EF
N= a(E)dE (25)
0

gV dp
a(E) = 3
4πp2 (26)
3~ dE
15
Here a(E) is the density of states and EF is the Fermi energy. With the density of states given by
Eq.26, we can solve for the number of particles N of the system in terms of the Fermi momentum.
The degeneracy of the system is given by g for a free non-relativistic particle confined in a volume
V . We can also replace the number density n with the number of particles since N/V = (2π)3 n.
Putting everything together we have an expression for the Fermi momentum:
1/3
6π 2 n

pF = ~ (27)
g
Plugging in for the Fermi momenta in Eq.24 we have:
  2 2/3  2 2/3 
kin 3 ~2 6π nN 6π nP
 = nN + nP (28)
5 2m g g
Furthermore, we can express the neutron and proton number densities in terms of the proton
fraction or the asymmetry parameter β. In this case, we use the latter for reasons we will see soon.
2/3  5/3  5/3 
3 ~2 6π 2 n

kin 1+β 1−β
 = n + (29)
5 2m g 2 2
If we factor out the 1/2, we can express the kinetic energy density in terms of the Fermi energy
of the nucleon, which has a degeneracy factor of 2g arising from the Isospin. Hence the reason for
using the asymmetry parameter. The kinetic energy density is now given as a function of β and
the number density n. (Note: We could also have said gP = gN = 2 and gnucleon = 4)
2/3
~2 6π 2 n

EF,nuc = (30)
2m 2g
 
kin 3n 5/3 5/3
 (n, β) = (1 + β) + (1 − β) EF,nuc (31)
10
Lastly, for the sake of convenience, we can adopt Prakash’s notation [7], and rewrite the kinetic
energy density in terms of a new parameter u = n/n0 , the Fermi energy of the nucleon at saturation
(0)
density EF , and the saturation density n0 :
 
kin 3 (0) 5/3 n0 5/3 5/3
 (n, β) = EF u (1 + β) + (1 − β) (32)
5 2
We can find the total energy density for symmetric nuclear matter and pure neutron matter
by combining Eq.22 and Eq.32 evaluated at β = 0 and β = 1 respectively:

3 (0) 1 Bn0 uσ+1 X Z ∞ d3 k


5/3 2
nm = EF n0 u + An0 u + + 4u Ci g(k, Λi )fn (33)
5 2 1 + B 0 uσ−1 i=1,2 0 (2π)3

16
2

3

1 Bn0 uσ+1 (1 − x3 )
nem = 22/3 (0)
E n0 u 5/3 2
+ An0 u (1 − x0 ) + 3
5 F 3 2
1 + B 0 uσ−1 (1 − x3 ) (34)
Z ∞ 3 3
2 X dk
+ u (6Ci − 8Zi ) g(k, Λi )fn
5 i=1,2 0 (2π)3

With the energy densities for symmetric nuclear matter and pure neutron matter we can solve for
the equation of state for different models that vary based on the function g(k, Λi ), and the several
different constant parameters mentioned above. The function g(k, Λi ) serves to simulate finite
range effects. For models called BPAL Equations of State (EoS), the function g(k, Λi ) takes the
form of [1 + (k/Λi )2 ]−1 [7]. At zero temperature, the integrals in Eq.33 and Eq.34 are integrated
up to the Fermi energy and the Fermi-Dirac distribution functions turn into step functions. Eq.33
and Eq.34 then reduce to the following expressions for the BPAL EoS:

Bn0 uσ+1
 1/3
u1/3

3 (0) 5/3 1 2
X
3 u
nm = EF n0 u + An0 u + + 3n0 u Ci Ri − arctan (35)
5 2 1 + B 0 uσ−1 i=1,2
Ri Ri

2

3

1 Bn0 uσ+1 (1 − x3 )
nem = 2 2/3 (0)
E n0 u 5/3 2
+ An0 u (1 − x0 ) + 3
5 F 3 2
1 + B 0 uσ−1 (1 − x3 ) (36)
3 
1/3 1/3

3 X (2u) (2u)
+ n0 u (3Ci − 4Zi )Ri3 − arctan
5 i=1,2
Ri Ri

We can take the same approach with the BPAL EoS and use a different function for g(k, Λi ). If we
set g(k, Λi ) = 1 − (k/Λi )2 , the finite range effects are approximated by effective local interactions
and the energy density takes the form of Skyrme’s effective interactions with R1 = 1.5 and R2 = 3
[7]. These models are thus called SL EoS (Skyrme Like):

Bn0 uσ+1 X  3 u2/3



3 (0) 5/3 1 2 2
nm = EF n0 u + An0 u + + n0 u Ci 1 − (37)
5 2 1 + B 0 uσ−1 i=1,2
5 Ri2

2

3

1 Bn0 uσ+1 (1 − x3 )
nem = 2 2/3 (0)
E n0 u 5/3 2
+ An0 u (1 − x0 ) + 3
5 F 3 2
1 + B 0 uσ−1 (1 − x3 ) (38)
3
2/3
 
2 X 3 (2u)
+ n0 u2 (3Ci − 4Zi ) 1 −
5 i=1,2
5 Ri2

With the energy densities for 2 different models, we can utilize our parabolic approximation (Eq.
14) to get the energy per particle of the system and other useful thermodynamic quantities such as

17
the baryonic energy density, pressure, and chemical potential. We will first solve for the asymmetry
energy Es for both the BPAL and SL models respectively using Eq.15:

2 σ
  Bu (1 − x3 )
BP AL 2/3 3 (0) 2/3 1 3
Es =2 E u + Au(1 − x0 ) +
5 F 3 2
1 + B 0 uσ−1 (1 − x3 )
3
1/3
(2u)1/3 Buσ
 
3 X
3 (2u) 3 (0) 1
+ (3Ci − 4Zi )Ri − arctan − EF u2/3 − Au −
5 i=1,2 Ri Ri 5 2 1 + B 0 uσ−1
 1/3
u1/3

3 u
X
−3 Ci Ri − arctan
i=1,2
Ri Ri
(39)

2 σ

3 (0) 2/3

1 Bu (1 − x3 ) 2 X

3 (2u)2/3

EsSL = 22/3 E u + Au(1 − x0 ) + 3 + u (3Ci − 4Zi ) 1 −
5 F 3 2 0 σ−1 5 i=1,2 5 Ri2
1 + B u (1 − x3 )
3
Buσ X  3 u2/3

3 (0) 1
− EF u2/3 − Au − −u Ci 1 −
5 2 1 + B 0 uσ−1 i=1,2
5 Ri2
(40)
For computational convenience, the asymmetry energy for both models can approximately be
expressed as [8]:
3 (0)
Es = (22/3 − 1) EF [u2/3 − F (u)] + S0 F (u) (41)
5
Where F (u) takes different functional forms that simulate the constant parameters in Eq.39 and
Eq.40. In order to keep track of the different models, we will use the notation BPALn1 n2 and
SLn1 n2 [7]. n1 will refer to the different√values for the compression modulus K0 and n2 = 1, 2, and 3
will refer to the functions for F (u) = u, u, or, 2u2 /(1 + u) respectively. In this paper we will be
using this approximation method and will calculate the asymmetry energy from these functional
relations as in [8]. Table 4 and 5 below highlight the different parameter values for each given
model [7]. (All constants are in untis of MeV with the exception of σ and B 0 ).

Table 4: Potential Model Parameters for nuclear matter

EOS K0 A B B0 σ C1 C2
BPAL1 120 75.94 -30.88 0 0.498 -83.84 23
BPAL2 180 440.94 -213.41 0 0.927 -83.84 23
BPAL3 240 -46.65 39.45 0.3 1.663 -83.84 23
SL1 120 3.706 -31.155 0 0.453 -41.28 23
SL2 180 159.47 -109.04 0 0.844 -41.28 23
SL3 240 -204.01 72.704 0.3 1.235 -41.28 23

18
Table 5: Potential Model Parameters for neutron-rich matter

EOS K0 x0 x3 Z1 Z2
BPAL11 -0.689 0.577 -14.0 16.69
BPAL12 120 -1.361 -0.244 -13.91 16.69
BPAL13 -1.903 -1.056 -1.83 5.09
BPAL21 0.086 0.561 -18.40 46.27
BPAL22 180 -0.410 -0.105 -9.38 24.05
BPAL23 -1.256 -1.358 -11.67 -10.90
BPAL31 0.376 0.246 -12.23 -2.98
BPAL32 240 0.927 -0.227 -11.51 8.38
BPAL33 1.654 -1.112 3.81 13.16
SL12 120 -3.548 -0.5 -13.355 2.789
SL22 180 -0.410 -0.105 9.38 -4.421
SL32 240 -0.442 -0.5 -13.387 2.917

With a given set of parameters we can solve for both the asymmetry energy and the energy of
symmetric nuclear matter to solve for the energy per particle in Eq.14. We begin solving for the
EoS of neutron star matter by considering the two conditions that must be fulfilled. The condition
of charge neutrality and beta equilibrium respectively are given below:
X
q i ni = 0 (42)
i

n → p + e− + v̄e
(43)
µn = µB = µp + µe = µp + µµ
The chemical potential of the neutrino is omitted due to the fact that most of all the neutrinos
have escaped during the early lifetime of the star. Once the star has reached a cold state and
equilibrated, the photons and neutrinos have been all emitted and thus µv = 0. Another important
thing to note is Eq.42 is a statement of local charge neutrality of the star. This may not necessarily
be the case in general as the only requirement is global charge neutrality, which we will explore
later in this paper. In the meantime, we will assume Eq.42 holds at every point in the star.
In the case for a purely nuclear star and no quark deconfinement, we have neutrons, protons,
electrons, and muons to consider. The number density takes the form n = gk 3 /(6π 2 ), where g is
the spin degeneracy and k is the Fermi momentum (pf = ~kf ). Since the Fermi momentum can
be expressed in terms of chemical potentials, we can rewrite the condition for charge neutrality as
follows:

np − ne − nµ = 0
kp3 − ke3 − kµ3 = 0
(44)
(µ2p − m2p )3/2 − (µ2e − m2e )3/2 − (µ2µ − m2µ )3/2 = 0
(µ2p − m2p )3/2 − (µ2e − m2e )3/2 − (µ2e − m2µ )3/2 = 0

19
Looking at our conditions for beta equilibrium and utilizing our expressions for the proton and
neutron chemical potentials we can also find a convenient expression for the electron chemical
potential [16].

µe = µn − µp
(45)
µe = 4βEs

Since we know the proton fraction is just np = nx, we have np = n2 (1 − β). Applying these
expressions to our condition for charge neutrality we have an equation for β given some number
density n.
n
(1 − β) − (16β 2 Es (n)2 − m2e )3/2 − (16β 2 Es (n)2 − m2µ ) = 0 (46)
2
We can solve the equation above using the bisection method giving us β for a given number density.
Now in order to create a viable equation of state for use in solving the TOV equations, we must
start with a chemical potential around the mass of the nucleon and solve for the corresponding
energy density and pressure. We repeat this process to a high enough chemical potential where
the star is no longer stable.

Given some baryonic chemical potential µB , we can solve Eq.19 using the bisection method as well
for a number density that satisfies the relation. Since the number density also depends on β as we
showed above through charge neutrality, a double bisection method is used to solve for both β and
n for a given µB . Once we have both those quantities, the baryonic energy density and pressure are
easily evaluated. Note this is only one approach taken to solve these set of equations. One could
have easily started with the electron chemical potential and worked backwards. For computational
speed, a table with n(β, E0 , Es ) can be generated through the use of Eq.46. Then when solving
for n given some µB (Eq.19), interpolation routines can be used. This avoids a double bisection
that may be computationally intensive. This method of tabulation and interpolation was used in
producing some of the EoS in this paper.

Additionally, we must also take into account the energy densities of the leptons, as the total energy
density in neutron star matter is a mixture of baryons and leptons [16]. The total energy density
and pressure is thus given by:

(n, β) = B (n, β) + e (n, β) + µ (n, β)


(47)
P (n, β) = PB (n, β) + Pe (n, β) + Pµ (n, β)

After calculating the EoS for several different models, two checks were put in place to validate the
accuracy of the calculations done in this paper. The first check was to make sure the EoS was
thermodynamically consistent. This was shown by the verification of the relation below at every
step in the chemical potential µB .

nB µB =  + P (48)

20
The second check was to compare the values for maximum mass and radius with those of Prakash’s
results in [7]. The resulting equations of state below have shown to be both thermodynamically
consistent and in good agreement with [7].

Figure 2: BPAL EoS with varying parameters for compression modulus.

Figure 3: BPAL EoS with varying parameters for compression modulus.

21
Figure 4: SL EoS with varying parameters for the nuclear symmetry potential energy dependence
on density.

Figure 5: SL EoS with varying parameters for the nuclear symmetry potential energy dependence
on density.

Lastly, we use our EoS to solve the TOV equations and obtain relationships for the mass and
radius of several different potential neutron stars. An important thing to note is the equations of

22
state we have calculated do not include any crust in the calculations. The EoS discussed are purely
for nuclear matter at higher densities. Crust can be taken into account easily via a simple Maxwell
construction from the crust phase to the hadronic phase. This will be demonstrated further in
this paper and the difference with and without crust will be apparent. Although the mass-radius
curves for a purely nuclear star without crust would be unrealistic, as we know heavier elements
are formed on the surface, we plot them below for the sake of completeness.

Figure 6: Several mass-radius relationships for the nuclear EoS

5 Quark Matter Equation of State


5.1 Introduction
As we have seen, neutron stars harbor one of the densest environments in the universe. It is no
surprise that at these extreme densities, matter can start to behave in strange ways. One of the
main focuses of neutron star research is to understand what sort of properties and behaviors are
attributed in this high density regime. One possibility is the collapse of hadrons into their respec-
tive constituent quarks. This phenomenon known as quark deconfinement, can only be possible
at extremely high densities, pressure, and/or temperature. We will begin to examine how this
behavior can impact the equation of state for neutron stars as well as examine the possibility of
stars composed of pure quark matter as opposed to the pure nuclear matter. A pure quark star,
as we will see, can in fact be more stable than their nuclear counterpart due to the strange matter
hypothesis.

The strange matter hypothesis states that the absolute ground state of matter is deconfined quark
matter consisting of equal parts of up, down, and strange quarks [11]. The hypothesis itself is

23
difficult to prove or disprove as examining matter in the deconfined state is an extremely difficult
task and can possibly only be done through heavy ion collision experiments, QCD, or through the
study of neutron stars. The energy of deconfined strange matter must be estimated to determine
whether the energy threshold is lower that that of Iron (∼ 930 Mev). This is quite difficult due
to the fact that Quantum Chromodynamics (QCD) still remains mostly unsolved and the neutron
star regime still presents challenges to modern QCD Lattice simulations [9].

In order to start studying quark matter in neutron stars, we must decide on how to model the EoS.
In this paper we begin with the thermodynamic Bag model (tdBag). This bag model stems from
the well known MIT Bag Model, where confinement is simulated by placing fermions or bosons
in contained fields of constant energy per volume or so called bag pressure B [1]. The tdBag
model is applied to bulk matter and describes a Fermi gas of free quarks confined with the bag
pressure B [9]. Another model we will examine, is the vBag model [17]. The vBag model is an
interaction enhanced Bag model that is an extension of the tdBag model by incorporating effects of
dynamical chiral symmetry breaking and vector repulsion [9, 17]. We will begin with the standard
tdBag model and examine some simple examples of equations of state and then study the more
advanced vBag model and its applications to strange stars.

5.2 The Thermodynamic Bag Model (tdBag)


We start studying the quark matter equation of state by considering the bulk of quark matter as
a grand canonical ensemble with thermodynamic potential Ω(µ, V, T ). The partition function for
the system is given by [14]:
X
Z= e−βE (49)
E

Here β = 1/kB T and is not to be confused with the asymmetry parameter. The total P energy of
the system can be expressed by the sum of the individual energies of each particle E = ε = nε ε
X P
Z= e−β ε nε ε (50)

To have an easier time evaluating the thermodynamic relations necessary for


P calculating equations
of state, we resort to the grand partition function and the fact that N = ε nε [14]:


X
Q(µ, V, T ) = eβµN Z
N =0
X∞  X 
P
βµN −β nε ε (51)
Q(µ, V, T ) = e e ε

N =0 nε
XY
Q(µ, V, T ) = (eβ(µ−ε) )nε
nε ε

24
Since for the case of fermions, the state nε can either be 0 or 1 (occupied or unoccupied), the grand
partition function reduces to a much simpler form.
Y
Q(µ, V, T ) = (1 + eβ(µ−ε) ) (52)
ε

The thermodynamic potential is then given in terms of the grand partition function as simply
Ω(µ, V, T ) = −1/β ln Q. We can use our result for the grand partition function of fermions and
use the density of states a(ε) to solve for a continuous bulk of quark matter.

Y 
1 β(µ−ε)
Ω(µ, V, T ) = − ln (1 + e )
β ε
1X
=− ln (1 + eβ(µ−ε) ) (53)
β ε
1 ∞
Z
=− a(ε) ln (1 + eβ(µ−ε) )dε
β 0
To solve this integral, we use integration by parts and our zero temperature approximation. Any
integral over the fermi distribution function fi becomes a step function Θ(µ − εF ). Thus we only
need to integrate up to the Fermi energy εF [13]. (Note the density of states a(ε) = dN/dε)

1 ∞
Z
Ω(µ, V, T ) = − a(ε) ln (1 + eβ(µ−ε) )dε
β 0
 ∞ Z ∞ 
1
β(µ−ε)
= − N (ε) ln (1 + e ) + β N (ε)f (ε)dε
β 0 0
  ∞ Z ∞ (54)
1 β(µ−ε)

= − N (ε) ln (1 + e ) − N (ε)f (ε)dε
β 0 0
Z εF
Ω(µ, V, T = 0) = − N (ε)Θ(µ − εF )dε
0

With a simple expression for the thermodynamic potential and grand partition function, we can
easily obtain the pressure, energy density, and number density through the following relations.
Here f denotes the flavor of the quark while the particle number N is given by N = gV k 3 /(6π 2 )
[2, 14].

25
Z εF Z εF
∂Ω ∂ gV k 3 g
pf = − = dε = 2 k 3 dε
∂V ∂V 0 6π 2 6π 0
Z εF
g
= 2 (ε2 − m2f )3/2 dε
6π mf
1 ∞
  Z Z kF q
1 µ ∂ ∂ 1 X ε g
f = − ln Q = β(ε−µ)
= εf (ε)a(ε)dε = 2 k 2 k 2 + m2f dk
V β ∂µ ∂β V ε 1+e V 0 2π 0
Z εF q
g
= 2 ε2 ε2 − m2f dε
2π mf
1 ∞ 1 εF
Z Z Z kF
1 ∂ 1 X g
nf = ln Q = f (ε) = a(ε)f (ε)dε = a(ε)dε = 2 k 2 dk
V β ∂µ V ε V 0 V 0 2π 0
Z εF q
g
= 2 ε ε2 − m2f dε
2π mf
(55)

The integral for the number density is straightforward through u-substitution, while the pressure
and energy density integrals are more difficult. They are solvable analytically through multiple
integration by parts and integral tables, but instead we use Mathematica in this paper for the
remaining integrals. Using the fact that εF = µ at T = 0 and a change of variables zf = mf /µf ,
we arrive at the following final expressions.

q
g 1
 q ε F + ε2F − m2f 
2 2 2 2 4
pf = εF (2εF − 5mf ) εF − mf + 3mf ln
6π 2 8 mf
q
2
gµ4f q 3 4 1 + 1 − zf
  
2 5 2
= 1 − zf 1 − zf + zf ln
24π 2 2 2 zf
 
g 1 2 2
q
4 m f
f = εF (2εF − mf ) ε2F − m2f + mf ln q (56)
2π 2 8 εF + ε2F − m2f
q
gµf4 q 
1

1 1 + 1 − zf2 
2 4
= 1 − zf2 1 − zf − zf ln
8π 2 2 2 zf
3
gµf
nf = (1 − zf2 )3/2
6π 2
Lastly, we take into account the bag pressure and sum over all flavors f , of quark matter to arrive
at equations of state for the simple thermodynamic bag model (tdBag). The total pressure, energy
density, and baryon number density are shown below. For quarks the degeneracy factor g = 6 due

26
to the color and spin (g = 2spin × 3color ) [11].
q
X gµ4f q
 
5 2

3 4 1 + 1 − zf2 
2
p = −B + − zf 1 − zf + zf ln
1
f
24π 2 2 2 zf
q
2
X gµ4f q 1 4 1 + 1 − zf
 
2 1 2 (57)
=B+ 2
1 − zf 1 − zf − zf ln
f
8π 2 2 zf
1 X gµ3f
nB = (1 − zf2 )3/2
3 f 6π 2
With the equations of state for the tdBag model, we can start to cater it specifically to neutron
star matter by taking into account the two properties of charge neutrality and beta (chemical)
equilibrium, as we did earlier for nuclear matter. In general, we can write the chemical potential
for any particle i with baryon number b and charge q in terms of the baryon chemical potential and
electron chemical potential as in Eq.58. We have omitted the chemical potential for strangeness
because it is not conserved on the timescale of the star [11].

µi = bµB − qµe (58)


Charge neutrality is the same as in Eq.42, yielding the condition for two and three flavor quark
matter respectively:

2 1
nu − nd − ne − nµ = 0
3 3 (59)
2 1 1
nu − nd − ns − ne − nµ = 0
3 3 3
Using Eq.57 for the number densities, the charge neutrality equation can be written in terms of
the corresponding chemical potentials. Since each chemical potential is in itself an unknown, we
can rewrite them using the equation for chemical equilibrium (Eq.58).
µB − 2µe µB + µe µB + µe
µu = µd = µs = µµ = µe (60)
3 3 3
In conjunction with our charge neutrality equations we have the following equations (2f and
3f respectively) dependant on only two variables: the baryochemical potential µB and electron
chemical potential µe . This leaves us with the ability to solve for µe for any given µB and vice
versa, as long as µB > 3mu + 2me so the solution is real only.

 2 3/2  2 3/2
µB − 2µe µB + µe
2gµ − m2u − gd − m2d − 3ge (µ2e − m2e )3/2 − 3gµ (µ2e − m2µ )3/2 = 0
3 3
 2 3/2  2 3/2  2 3/2
µB − 2µe µB + µe µB + µe
2gµ − m2u − gd − m2d − gs − m2s
3 3 3
− 3ge (µ2e − m2e )3/2 − 3gµ (µ2e − m2µ )3/2 = 0
(61)

27
Solving Eq.61 using the bisection method is relatively straightforward and is the only numerical
technique required for solving the tdBag model. With both µB and µe , we can work backwards
to obtain the individual chemical potentials of the u, d, and, s quarks as well as the muon. With
the respective chemical potentials the equation of state is obtained by summing over all flavors
(Eq.57) and adding the leptonic contributions to the pressure and energy density. The results of
some equations of state for 2f and 3f quark matter are shown below.

Figure 7: EoS of state for 2 flavor and 3 flavor quark matter, where Bef f is the varying effective
bag pressure. We can see the effect of the bag pressure is to lower the overall pressure at low
chemical potential.

28
Figure 8: EoS of state for 2 flavor and 3 flavor quark matter, where Bef f is the varying effective
bag pressure.

5.3 Extending the tdBag Model (vBag)


With the tdBag model out of the way, we are now in a position to tackle a much more complicated
equation of state for quark matter, by taking into account dynamical chiral symmetry breaking
and vector repulsion. Chiral symmetry breaking is a result of strong interactions breaking the
symmetry of the QCD Lagrangian, where the up and down quarks are essentially massless. The
difference in masses of the up and down quark are what end up breaking chirality [10]. The vBag
model which we will explore in this section takes into account these phenomena and modifies the
tdBag model to provide a more realistic approach to deconfined quark matter.

The vBag model takes the same form as the tdBag model equation of state, except there is an
additional parameter Kv that acts as a coupling constant for the vector interactions. The chemical
potentials in the new model are also replaced by an effective chemical potential µ∗f changing the

29
equations for pressure, energy density, and number density slightly [17, 9]:

µf = µ∗f + Kv nf (µ∗f )
∗ Kv 2 ∗
Pf (µf ) = PFkin
G,f (µf ) + nf (µf ) − Bχf
2 (62)
kin ∗ Kv 2 ∗
f (µf ) = F G,f (µf ) + nf (µf ) + Bχf
2
nf (µf ) = nf (µ∗f )
An important difference in the new vBag equations of state, is the addition of a self consistent
equation for the effective chemical potential and the chiral bag pressure Bχ . In a similar fashion to
the tdBag model, our approach is the same, we must find a µe that satisfies the condition for charge
neutrality given some baryonic chemical potential µB . The only difference in the vBag formulation
is the number density in our charge neutrality condition now depends on the effective chemical
potential instead. From this point forward unless otherwise specified, we will only examine the case
for quark matter with muons and strangeness, as it is computationally more difficult and follows
the same recipe for matter without muons or strange quarks. The charge neutrality condition
reads:

2nu (µ∗u ) − nd (µ∗d ) − ns (µ∗s ) − 3ne (µe ) − 3nµ (µe ) = 0 (63)


The effective chemical potentials are then given by Eq.62. Using the conditions for chemical
equilibrium (Eq.60) as we did for the tdBag model we have the following self consistent equations
to solve, dependent on µB , µe , and Kv . The coupling constant is fixed for a given model, so all
that is needed is to solve for the electrical chemical potential for any given µB . The result is a
double bisection method. The first (main) bisection finds a µe to satisfy charge neutrality, while
also using a bisection method to solve the self consistent effective chemical potential equations,
better shown by Eq.64.

2nu [µ∗u (µB , µe )] − nd [µ∗d (µB , µe )] − ns [µ∗s (µB , µe )] − 3ne (µe ) − 3nµ (µe ) = 0 (64)
Once the above equation has been solved for the electron chemical potential, we can solve for the
total pressure, energy density, and baryon number density, where PFkin kin
G,f , εF G,f , and nf are given
by the tdBag equations (Eq.57).

 
X
∗ Kv X 2 ∗
P = PFkin
G,f (µf ) + PFkin
G,e (µe ) +PFkin
G,µ (µe ) + 2 2
n (µ ) + ne + nµ − Bef f
f =u,d,s
2 f =u,d,s f f
 
X
kin ∗ kin kin Kv X 2 ∗ 2 2
= F G,f (µf ) + F G,e (µe ) + F G,µ (µe ) + nf (µf ) + ne + nµ + Bef f (65)
f =u,d,s
2 f =u,d,s
1 X
nB = nf (µ∗f )
3 f =u,d,s

With the novel vBag EoS, we are at the liberty to explore the effects of the coupling constant Kv
on the quark EoS. An important subtlety to note is the fact that if the coupling constant is set to

30
zero, we arrive again at the standard tdBag equation of state, which is to be expected. Plots of
three flavor quark equations of state are shown below (Fig.9 and Fig.10).

Figure 9: EoS for vBag. The effective bag constant is arbitrarily set to Bef f = 100. We see the
effect of the coupling constant decreasing the overall pressure as the chemical potential increases.
The tdBag model is represented here by setting the coupling constant to zero.

Figure 10: EoS for vBag. The effective bag constant is arbitrarily set to Bef f = 100. We see the
effect of the coupling constant lowering the energy density as the chemical potential increases. The
tdBag model is represented here by setting the coupling constant to zero.

31
5.4 Quark matter in Neutron Stars
With several different equations of state for quark matter to work with, we are in a position to
study the possibility of pure quark stars as was done recently in [9]. The strange matter hypothesis
argues the possibility of strange matter being the absolute ground state of matter. However, in
order for strange matter to be more stable than confined hadronic matter, deconfined strange mat-
ter would have to have an energy per particle lower than that of Iron (∼ 930 MeV) as mentioned
earlier. In this section, we comb through various EoS by varying both the coupling constant and
the effective bag pressure to find the minimum E/N for 3 flavor strange matter that is indeed
lower than that of Iron. Another requirement is that 3 flavor strange matter below the 930 MeV
criteria must have a 2 flavor counterpart with an energy density higher than 930 MeV so that
strange matter is the absolutely favored state. In general, strange matter has a lower energy per
particle due to the sharing of baryon number over three Fermi seas instead of two [11].

Another criteria to look at when considering strange stars, is at current neutron star data. We
can take a phenomenological approach and constrain our model to meet the neutron star mass
and radii constraints. We can look at several different stars in the Kv and Bef f space to see which
models satisfy the three requirements of mass, radii, and E/N constraints. We provide contour
plots below showing exactly this, as well as the window of models satisfying all of the above criteria.

Figure 11: Contours of constant mass in the vBag model parameter space. Two-flavor and three-
flavor lines of constant energy/baryon = 930 MeV are shown by the dotted lines [9].

32
Figure 12: Contours of constant radius at maximum mass in the vBag model parameter space.
Two-flavor and three-flavor lines of constant energy/baryon = 930 MeV are shown by the dotted
lines. Contours of the maximum mass for two heavy neutron stars (J1614-2230 at 1.97 M and
GW170817 at 2.17 M ) are also shown [9].

Models that fall within the E/N window and meet certain mass-radius constraints are ideal can-
didates for strange stars. Another criteria for the stability of, not only strange stars, but any
neutron star, is the stability against radial oscillations. This provides another layer of realism to
the applicability of the equation of states being examined. A thorough examination of stability
analysis on vBag parameters was done in [9]. For now we plot some of the mass-radius curves, of
vBag EoS that yield stars (of a given central density) stable against radial oscillations.

33
Figure 13: Mass radius plot of strange stars in the vBag model (Kv 6= 0). Two other models
(Kv = 0) representing the tdBag model are shown as well, taking both the zero temperature
approximation and the massless quark approximation (MQA).

The results of this study and of [9] are relevant and provide meaningful conclusions. First, it sheds
light on the possibility of strange stars by addressing the strange matter hypothesis. Second, it
questions the ability of using mass-radius data for distinguishing pure neutron stars from strange
stars. The latter case will be discussed and shown in more detail in the next section. In the
meantime, further study of quark matter in compact stars could shift our current understanding
of high density physics and provide insight into the true ground state of matter.

6 Hybrid Stars
6.1 Phase Transitions
In this section, we will take all that we have examined in the previous chapters on nuclear and
quark equations of state. We will explore the possibility of hybrid stars, where nuclear matter
(deconfines) into quark matter through different mechanisms. Realistically, one would like one
equation of state described by one theory of matter, with deconfinement mechanisms built in in-
trinsically. However, this poses great difficulty in the fact that we have yet to really understand
the problem of confinement and the mechanisms involved. QCD theory plays a vital role in un-
derstanding this problem and can help us to better understand the confinement transition.

34
When simulating deconfinement, there are a couple of methods to consider. The first straight-
forward method is a first order phase transition via Maxwell construction. Maxwell construction
states that two phases are in equilibrium when at some pressure the two chemical potentials are
equal. In the case for neutron stars, we are dealing with a hadronic phase and a quark phase so
the condition for a phase transition is:

PH (µH ) = PQ (µQ )
(66)
µH = µQ

This means that at the phase transition, the hadronic phase and quark phase will rest on top
of each other similar to oil on water. This implies a discontinuous jump in baryon density and
energy density. This discontinuity is a direct consequence from looking at the condition for charge
neutrality as a local phenomena instead of a global one. Local charge neutrality is not a necessary
condition for matter, as we know. An example of this is a neutral atom, where positive charges are
concentrated in the nucleus and negative ones in the orbital. This breaks local charge neutrality
but minimizes the total energy through global charge neutrality [11]. These characteristics will
be shown by examining the Maxwell phase transition first, followed by the Gibbs phase transition
where charge conservation becomes global.

There are many ways to perform a Maxwell construction. We can take the equal area approach and
integrate P (V ) for both phases to find the pressure where the areas are equal. Another method
better suited to this case is to find the baryonic chemical potential where both pressures are equal.
Since both the vBag model and parabolic approximation specifically calculate an EoS given some
baryonic chemical potential µB , we can take advantage of this and perform a bisection on the
equation PH (µB ) − PQ (µB ) = 0. The result µH,Q
B , if there is one, will give us the chemical potential
at which the phase transition occurs. All that is left to do is calculate the EoS for hadronic matter
up to µH,Q
B , and then start calculating the quark equation of state from µB
H,Q
up to the desired
chemical potential. This ‘gluing’ of EoS is characteristic of the Maxwell phase transition. We
perform maxwell construction via this exact recipe on the vBag EoS and a nuclear EoS below.

35
Figure 14: EoS of state for Maxwell transition showing the discontinuity in density as pressure
increases. Nuclear EoS used is BPAL33 and Quark EoS used is vBag with KV = 1.0 and Bef f = 120

Figure 15: EoS of state for Maxwell transition showing the discontinuity in energy density and
the continuity of pressure as a function of chemical potential. Nuclear EoS used is BPAL33 and
Quark EoS used is vBag with KV = 1.0 and Bef f = 120

36
Figure 16: Mass radius curve for the Maxwell transition EoS used above as well as it’s appropriate
pure star counterparts.

There are numerous different combinations of EoS that can be created through Maxwell construc-
tion, each with a ‘unique’ Mass-Radius curve. However, even this ‘uniqueness’ will be called into
question later on as was done in [18]. In the meanwhile, we introduce a solution to the discon-
tinuous and unrealistic behavior of the Maxwell transition from nuclear to quark matter. As we
mentioned earlier, the main issue of the Maxwell phase transition, is the condition of local charge
neutrality. We begin to remedy this by imposing global charge neutrality instead. If it is more
energetically favorable to isolate charges in separate regions then nature will take the lower energy
state even if charge is not locally conserved. Lastly, another issue with the Maxwell case is the
fact that it only takes into account one chemical potential, the baryonic chemical potential µB .
We have already seen that the EoS in neutron star matter depends on a minimum of two chem-
ical potentials: the baryonic chemical potential and the electric chemical potential µe . Now the
condition for chemical equilibrium becomes:

PH (µB , µe ) = PQ (µB , µe ) (67)


This condition greatly increases the computational difficulty of the problem as we are now searching
for the intersection of two planes in the P, µB , µe space. Imposing the condition of global charge
neutrality, where ρ is the charge density, we have [11]:

(1 − χ)ρH (µB , µe ) + χρQ (µB , µe ) = 0 (68)


The global charge neutrality condition allows for a mixed phase and creates a smoother transition
from one phase to another. There is now the existence of a pure nuclear phase (χ ≤ 0), a mixed

37
phase, and a pure quark phase (χ ≥ 1). The jump discontinuity in energy density and number
density also disappear as they take the following form in the mixed phase:

(1 − χ)nH (µB , µe ) + χnQ (µB , µe ) = nB


(69)
(1 − χ)εH (µB , µe ) + χεQ (µB , µe ) = ε

The procedure for calculating this type of phase transition (Gibbs construction) begins with varying
the baryonic chemical potential. In order to determine what phase we are in, we must calculate
χ to see if it lies between 0 and 1. Given some µB , we can find a µe that satisfies PH (µB , µe ) −
PQ (µB , µe ) = 0 via the bisection method. Once we have the two chemical potentials that satisfy
the condition for chemical equilibrium we can calculate χ.

ρH
χ=
ρH − ρQ
3np − 3ne − 3nµ (70)
χ=
3np − 2nu + nd + ns

If χ is negative, then we are in the nuclear phase and we calculate the nuclear EoS as we have
done before. This procedure is repeated everytime as µB increases until χ falls within the range of
the mixed phase. Once the mixed phase begins we continue the same procedure until χ becomes
greater than unity. The only difference in the mixed phase is the calculation of the EoS. The
energy densities and number densities are altered by Eq.69. Once the mixed phase transitions into
the pure quark phase we no longer need to calculate χ and the EoS is given by pure quark matter
up to the desired chemical potential. In practice the whole process, can be computationally
expensive, especially if the EoS becomes increasingly more complex. In order to make sure we
have the correct EoS in the end, we must check as we did earlier to make sure everything remains
thermodynamically consistent. Since we now have two chemical potentials the general relation
must hold true:

nB µB − ρµe = ε + P (71)
Plugging in our expressions for ρ, nB , and ε, we can show that each phase must also obey the
thermodynamic relation independently, that is:

nH µB − ρH µe = εH + P
(72)
nQ µB − ρQ µe = εQ + P

Alternatively, we an multiply the hadronic equation by (1 − χ), the quark equation by χ, and sum
them to arrive at Eq.71 again. Making sure these relations hold at every chemical potential is
crucial to making sure the resulting EoS in the Gibbs phase transition is correct. We show the
result of successfully performing this type of first order phase transition as well as its Maxwell
counterpart.

38
Figure 17: EoS of state for Maxwell and Gibbs transition. Nuclear EoS used is BPAL33 and Quark
EoS used is vBag with KV = 1.0 and Bef f = 120
.

Figure 18: EoS of state for Maxwell and Gibbs transition. Nuclear EoS used is BPAL33 and Quark
EoS used is vBag with KV = 1.0 and Bef f = 120
.

Now that we have tackled the nuclear equation of state, the quark equation of state, and two
different types of phase transitions, we can examine some the consequences the theory would have

39
on astrophysical observations. We explore the question posed earlier on the subject of uniqueness
of the Mass-Radius curves and break down the implications of this in the following section.

6.2 Camouflage of Phase Transitions


As was done in [18], we will analyze several different models of hybrid stars and their pure nuclear
counterpart. The nuclear equation of state that will be used in this section differs from the models
discussed earlier. We will use the data for the symmetry energy and energy of symmetric nuclear
matter extracted from a relativistic Dirac-Brueckner-Hartree-Fock (DBHF) model. The quark
matter EoS used will be the vBag EoS and various parametrizations. Although the crust was
mentioned in the very beginning of this paper, we will now be in a position to make use of it as we
are now making an effort to model astrophysical observations and compare results. We perform a
standard Maxwell transition from the crust to the much improved DBHF nuclear equation of state
and lastly a Gibbs or Maxwell transition to quark matter. This will provide us with a complete
model for the hybrid star equation of state. We first plot the difference in Mass-Radius curves
with and without crust.

Figure 19: Mass radius curves for DBHF nuclear EoS with and without crust. We can see the
crust barely has an effect on the mass of the star, but has a significant impact on the measured
radius.

In order to find a suitable EoS for a hybrid star that can ‘masquerade’ as a pure nuclear star, we
must come through multiple sets of hybrid stars including those with both Maxwell and Gibbs
phase transitions. Once a candidate parameter set is found, we can start to tweak the Kv and

40
Bef f parameters in the vBag model space to maximize the similarities in the M-R plane. This
would include the stiffening and softening on the EoS, as well as changing when and where the
phase transition occurs. Out of numerous parameter sets, the following EoS demonstrated in the
plots below are ideal candidates for masquerading as pure nuclear stars.

Figure 20: Mass radius curves for demonstrating the masquerade of 2 flavor quark matter in a
hybrid star. The parameters for vBag are (Kv = 6.0 and Bef f = 56.4 for Maxwell and Kv = 8.5
and Bef f = 70.0 for Gibbs)

Figure 21: Mass radius curves for demonstrating the masquerade of 3 flavor quark matter in a
hybrid star. The parameters for vBag are (Kv = 23.08 and Bef f = 72.3 for Maxwell and Kv = 31.0
and Bef f = 100.0 for Gibbs)

41
Fig. 20 and 21 show how phase transitions can be hidden and masquerade as a pure nuclear star.
We can also examine the case where, even if a phase transition can be identified, we still would
lack information as to what type of phase transition is taking place and the flavors involved. This
‘flavor camouflage’ was tackled in [18], and similar results will be shown here. Fig 22 and 23
demonstrate flavor camouflage for Maxwell and Gibbs phase transitions respectively.

Figure 22: Mass radius curves demonstrating the flavor camouflage in Maxwell phase transitions.
Parameter set used is (Kv = 4.0 and Bef f = 62.5 for 2f and Kv = 17.2 and Bef f = 76.0 for 3f)

Figure 23: Mass radius curves demonstrating the flavor camouflage in Gibbs phase transitions.
Parameter set used is (Kv = 4.0 and Bef f = 75.0 for 2f and Kv = 18.5 and Bef f = 101.0 for 3f)

42
As we have seen, there is still much work to do in the realm of neutron star physics. The ambiguity
of our models for high density nuclear matter and quark matter, clearly limit any inferences or
conclusions that we can make on observational data as we have just shown. Looking for kinks in
neutron star mass radius curves can provide us with insight to the nature of the phase transition
taking place, but will fail to provide us with the flavor content in the transitioned phase. In the
case where there is no kink, we are left even more dumbfounded as we may not even be able to
resolve if a phase transition has taken place at all. Although we have shed light mostly on the
issues that come with the uncertainty of our models, there is still plenty of work to be done in
refining what we already know.

7 Future Work
7.1 Conclusions
In this paper, we have only begun to scratch the surface of neutron star physics, and set the foun-
dation for future work. In the first section, through the frame of general relativity, we assumed a
static spherically symmetric star to derive the TOV equations. For most cases, this is a very good
approximation. However, this calculation would vary significantly for rotating compact objects
such as pulsars. The spinning of the star would alter the shape of the star, as it accrues more
mass on its equatorial plane, creating more of an obloid sphere. Studying pulsars, could prove
most useful in studying the composition of neutron stars. Measuring the period, the curvature of
the star, along with its mass and radius, could provide further constraints on the EoS compared
to a static star. Pulsars have also been known to ‘glitch’ as mentioned in Glendenning [11]. This
sudden change in the rotational period could be caused by some rigid structures embedded in the
star. Investigation into the manifestation of these physical structures in neutron stars could help us
model the glitches observed and shed light on the behavior of matter at extreme densities. These
rigid crystalline structures are another area of interest as they can take many different forms such
as rods, slabs, and droplets. It would be an interesting phenomena to explore in later research,
but it is not the only subject of interest.

If we do choose to study static neutron stars, there are still other improvements that could be
made to provide more accurate models. As we know, the equation of state increases in uncertainty
as the the density increases past the nuclear saturation density n0 = 0.163fm−3 . In the extremely
dense regions where QCD begins to become an increasingly important consideration, there is a
remarkable amount of freedom in choosing models. Taking a theoretical approach to ‘solve’ QCD
and be able to apply it to the neutron star EoS is a valid direction one might choose to take.
This would involve an attempt to model the confinement and deconfinement mechanism which is
both extremely fascinating and rather difficult. This would eliminate the need to model a phase
transition as it would already be embedded intrinsically into the model. One could also take a
field theoretical approach to purely study the EoS of highly dense matter. It could involve phe-
nomenology using heavy ion collision experiments as a template of which to begin basing models
off of, or it could involve a holistic approach based on first principles.

Ultimately, the goal of all these different and more advanced approaches is to constrain and/or

43
resolve the behavior of matter at extreme densities. Some of these methods might prove more
insightful than others given the current state of resources or data at disposal. One could also
implement Bayesian statistics to all the methods recently described above. It could prove a useful
tool in helping to ‘grade’ the models we use and provide us with information as to which parameters
are more likely to be valid. There are of course many other approaches that have not been
mentioned, as this list is not meant to be exhaustive. Gravitational wave phenomena and neutron
star mergers, are another side to the story that could possibly provide stronger constraints than
any other method at the moment. Whatever the case may be and whichever route is taken, there
is still much work to be done in studying neutron stars.

A TOV Derivation

44
References
[1] A. Chodos, R.L. Jaffe, K. Johnson, C.B. Thorn, and V.F. Weisskopf, Physical Review D 9,
3471 (1974).

[2] C. Kittel and H. Kroemer, Thermal Physics, 2nd ed. (W.H Freeman and Company, New York,
NY, 1980).

[3] G. Baym, C. Pethick, and P. Sutherland, The Astrophysical Journal 170, 299 (1971).

[4] G. Baym, H.A. Bethe, and C.J. Pethick, Nuclear Physics A 175, 225 (1971).

[5] I.E. Lagaris and V.R. Pandharipande, Nuclear Physics A 369, 470 (1981).

[6] J.I. Kapusta and C. Gale, Finite-Temperature Field Theory (Cambridge University Press,
Cambridge, 2010).

[7] M. Prakash, I. Bombaci, M. Prakash, P.J. Ellis, J.M. Lattimer, and R. Knorren, Physics
Reports 280, 1 (1997).

[8] M. Prakash, T.L. Ainsworth, and J.M. Lattimer, Physical Review Letters 61, 2518 (1988).

[9] M. Salinas, T. Klähn, P. Jaikumar, Particles, 2(4), 447-456, 2019

[10] M.D. Schwartz, Quantum Field Theory and The Standard Model (Cambridge University Press,
Cambridge, 2017).

[11] N.K. Glendenning, Compact Stars Nuclear Physics, Particle Physics and General Relativity
(Springer, New York, 2000).

[12] P. Haensel, A.Y. Potekhin, and D.G. Yakovlev, Neutron Stars 1: Equation of State and
Structure (Springer, New York, 2011).

[13] R.L. Jaffe, Degenerate Fermion Systems Lecture Notes, (MIT 2007)

[14] R.K. Pathria and P.D. Beale Statistical Mechanics, 3rd ed. (Elsevier Ltd, Burlington, MA,
2011).

[15] R.R. Silbar and S. Reddy, American Journal of Physics 72, 892 (2004).

[16] T. Klähn, D. Blaschke, S. Typel, E. N. E. van Dalen, A. Faessler. C. Fuchs, T. Gaitanos, H.


Grigorian, A. Ho, E. E. Kolomeitsev, M.C. Miller, G. Röpke, J. Trümper, D. N. Voskresensky,
F. Weber, and H. H. Wolter, Physical Review C 74, 035802 (2006)

[17] T. Klähn and T. Fischer, The Astrophysical Journal 810, 134 (2015).

[18] W. Wei, M. Salinas, T. Klähn, P. Jaikumar, The Astrophysical Journal 887, 2 (2019).

[19] W.H. Press, S.A. Teukolsky, W.T. Vetterling, and B.P. Flannery, Numerical Recipes in C:
The Art of Scientific Computing, 2nd ed. (Cambridge University Press, Cambridge, 1989).

45
[20] Y. Jaluria, Computer Methods for Engineering with MATLAB Applications, 2nd ed. (CRC
Press, Boca Raton, FL, 2011).

46

You might also like