Você está na página 1de 66

Astro 160: The Physics of Stars

Served by Roger Griffith


Nutritional Facts:
Serving size: 1 Semester (16 weeks)
Servings per container: many problems and solutions

Problem set 2
Problem # 1
In class, I derived the relationship between the luminosity and mass of stars under the assumption that
energy is transported by radiative diffusion and that the opacity is due to Thomson scattering. We will
carry out many related estimates so it is important to become familiar with this process. Consider a star
in hydrostatic equilibrium in which energy transport is by radiative diffusion. The star is composed of
ionozed hydrogen and is supported primarily by gas pressure.
(a). Derive an order of magnitude estimate of the luminosity L of a star of mass M and radius R if the
opacity is due to free-free absorption, fo which 1023 T 7/2 cm2 g1 ( is in cgs).
We know that the radiation flux is given by
Frad

caT 3
T

where we know that a is the radiation constant, c is the speed of light, T is the temperature, is the
opacity, which in our case is given by free-free absorption, is the mass density and T is the temperature
gradient. We have the following relationships

M
R3

dT
TR Tc
TC

dR RR RC
R

1023 T 7/2

given these relationships we can find


Frad

caT 15/2 R5
M2

we also know that the luminosity can be written as


L = 4r2 Frad Frad =

L
4r2

which gives us
caT 15/2 R7
M2
we can find the temperature by using the virial theorem which can be written as
L

GMm p
3Rk
1

where k is now the boltzman constant. Substituting this expression into the above equation yields
L caR

1/2

11/2

Gm p
k

15/2

this gives us an order of magnitude estimate of the luminosity of a star with mass M and radius R.
(b). If all stars have roughly the same central temperature, and are supported by gas pressure, what is
the mass-luminosity scaling (proportianality) relationship for stars?
we now know that the luminosity scales as
L M 11/2 R1/2
we can find the relationship between the mass M and the radius R of a star by using hydrostatic
equilibrium.
dP
GM
= 2
dr
r
M

Pc
R
M
T

R
M R
since T is constant, substituting this into the luminosity relationship yields
L M5

(c). Give a quantitative argument as to whether free-free opacity dominates electron scattering opacity
in stars more massive that the sun or in stars less massive that the sun.
We can solve this problem by looking at the defenition for the opacity in free-free absorption, which
can be written as with T constant
M
3 MR
R
thus we find
1
2
M
this expression tells us that the lower the mass of the star the higher the opacity, thus in lower mass
stars the free-free opacity dominates.

Problem # 2
The central density and temperature of the sun are c 150 g cm3 and Tc 1.5 107 K. For the
conditions at the center of the sun, answer the following questions. Assume that the sun is composed
solely of ionized hydrogen.
(a). What is the mean free path of an electron due to electron-electron Coulomb collisions? What is
the typical time between collisions?
2

We know that the mean free path is given by


1
ne

l=

we know that for a completely ionized hydrogen gas that


ne n p

mp

and the interaction cross section is given by


= r2
where r is the Coulomb radius found comparing the thermal energy to the Coulomb energy
e2
kT
r

e2
kT

e4
(kT )2

using these relationships we find the mean free path to be


mp
l
c

kT
e2

2

and the collision time is given by


tcol =

l
ve

and the velocity can be found by using


3
1
kT = me v2 v =
2
2

3kT
me

thus the time is given as


te =

me m p
3kT c

kT
e2

2

(b). What is the mean free path of an proton due to proton-proton Coulomb collisions? What is the
typical time between collisions? Which occurs more rapidly, electron-electron or proton-proton Coulomb
collisions?
The mean free path of proton-proton collisions would be the same as for the electron-electron collosion
because the gas is completely ionized. The mean free path is given by
mp
l
c

kT
e2

2

The collision time would be the same except now that the mass is the mass of the proton not the
electron. i.e
tp =

mp mp
3kT c
3

kT
e2

2

we can now see the collision times for the electron-electron collision occurs more rapidly due to the
mass being so much smaller.
te t p

(c). Which opacity is more important for photons, Thomson scattering or free-free absorption?
We know that
T =
and

2T
ne T
=
0.80
c
mp

F = 1023 T 7/2 1.15

free-free absorption dominates the opacity for photons in this case? not sure why this is. We know that
Thomson scattering is the primary way that photons move the energy out.
(d). What is the mean free path of a photon? How does this compare to the mean free path of an
electron (this should give you a feel for why photons are far more effective at moving energy around in
stars)? What is the typical time between photon absorptions/scattering?
we know that the mean free path of a photon is given by

2
1
e2
8
l=
where T =
= 6.65 1025 cm2
ne T
3 40me c2
which yields
mp
8.3 103 cm
2T
 
m p kT 2
8.9 107 cm
=
c e2

l photon =
lelectron

The typical time for a photon collision is given by


t=

lp
2.8 1013 s
c

(e). For a photon undergoing a random walk because absorption/scattering, how long would it take to
move a distance Rsun given the results in (d)? For comparison, it would take 2.3 seconds moving at the
speed of light to travel a distance Rsun in the absence of scattering/absorption.
We know that the diffusion time can be acquired with
tdi f f =

R2 nk
thermal energy R2 nkT
R2 2k

L
lc aT 4 lcaT 3
m p lcaT 3

we know that the average time for a photon to leave the star is given by
tdi f f

R2sun

104 yr
l ph c
4

Problem # 3
How old is the sun? In this problem we illustrate how the naturally occuring radioactive isotopes of
uranium, U 235 and U 238 can be used to determine the age of the rocks. Both isotopes decay via a sequence
of -decays and -decays to form stabel isotopes of lead: the decay chain of U 235 ends up with Pb207 , and
the decay chain of U 238 ends up with Pb206 . As a result, the number of uranium nuclei in a rock decays
exponetially with time in accord with:
N5 (t) = N5 (0)e5t and N8 (t) = N8 (0)e8t
To avoid clutter, the last digit of the mass number of the isotope has been used as a subscript label. The
decay constants 5 and 8 for the two isotopes corresponds to half-lives of
T5 =

ln 2
ln 2
= 4.5 109 yrs
= 0.7 109 yrs T8 =
5
8

The magnitudes of these half-lives are ideally suitable to the determination of the ages of the rocks
which are over a billion years old. Now consider a set of rock samples which were formed at the same
time, but with different chemical compositions. They differ in chemical composition because different
chemical elements are affected differently by the processes of rock formation. However rock formation
processes do not favour one isotope over another. For example, on formation, the relative abundances of
U 235 and U 238 should be the same in every sample. But these abundances will change with time as the
deacy of U 235 and U 238 produce nuclei of Pb207 and Pb206 .
Consider the ratio of the increase in the number of Pb207 nuclei relative to the increase of Pb206
nuclei. Show that this ratio is the same for all rock samples which were formed at the same time,
and that it is given by
N7 (t) N7(0) N5 (t) e5t 1
=
N6 (t) N6(0) N8 (t) e8t 1
We know that the ratio of the two isotopes can be written as
N7 (t) N7(0) N5 (t) N5(0)
=
N6 (t) N6(0) N8 (t) N8(0)
and given the first expression given in this problem, which can also be written as
N5 (0) = N5 (t)e5t and N8 (0) = N8 (t)e8t
substituting this into our previous expression yields
N7 (t) N7(0) N5 (t) e5t 1
=
N6 (t) N6(0) N8 (t) e8t 1
which is what we were asked to show.
Consider a graph in which the measured abundances in the rock samples of Pb207 and Pb206 are
plotted, N7 (t) along the y-axis and N6 (t) on the x-axis. Show that a straight line will be obtained if
all the samples were formed at the same time.

We know that
N7 (t) =
where

N5 (t) e5t 1
N6 (t)
N8 (t) e8t 1

N5 (t) e5t 1
= constant
N8 (t) e8t 1

Given that the current ratio of naturraly occurring U 235 to U 238 is 0.0071, evaluate the gradient of
the straight line for rock samples of age (a) 1 billion years, (b) 3 billion years and (c) 5 billion years.
We know that the gradient of the straight line is just the constant in front of N6 (t) so we just have to plug
in numbers
(a). t = 1 billion years.
We know that
5 9.90 1010 yr1 8 1.5 1010 yr1
given these and the fact that we know the ratio between U 235 and U 238 we can find the gradient, for 1
bilion years we get
e5t 1
0.0071 t
= 0.0715
e 8 1
For 3 billion years we get
e5t 1
= .231
0.0071 t
e 8 1
and finally for 5 billion years we get
0.0071

e5t 1
= .891
e8t 1

Problem # 4 Radiative Atmospheres


In this problem we will solve for the structure of the outer part of a star assuming that energy is
transported solely by radiative diffusion (which is not the case in the sun, but is the case in stars more
massive than the sun). The star has a mass M and a luminosity L. Assume that the luminosity and mass
are approximately constant at the large radii of interest, that gas pressure dominates, and that the opacity is
due to electron scattering. Do not assume that the atmosphere is thin (i.e even though Mr constant = M,
because rchanges, the gravitional acceleartion is not constant).
Write down the equations for hydrostatic equlibrium and energy transport by radiative diffusion. Use
these to calculate dPrad /dP, the change in radiatio pressure with pressure in the atmosphere. What does
this result imply for how the ratio of gas pressure to radiation pressure changes as a function of the
distance in the atmosphere? Show that your result for dPrad /dP implies that T 3 and P 4/3 for
radiative atmospheres (in the language that we will use in the next week, this means that the radiative part
of the star is an n=3 polytrope).
since we know what the radiation pressure is we can find what the change is with respect to r
1
Prad = aT 4
3

1 d
4
dPrad
= a (T 4 ) = T 4 T
dr
3 dr
3

and we know that the radiation flux is given by


F=
thus we can write
F=

4 caT 3
T
3

dPrad
=F
dr
c

dPrad c
dr

the hydrostatic equilibrium equation is


dP
GM
= 2
dr
r
deviding these two expressions yield
Fr2
dPrad
=
dP
cGM
we know that the Flux and luminosity are related by
L = 4r2 F

F=

L
4r2

thus we find

dPrad
L
=
dP
4cGM
this result implies that the ratio of the gas pressure to radiation pressure is independent of the distance
in the atmosphere. To show T 3 we can just use scaling arguments
Prad
L
T4
L

Pg
M
T M
since we assumed that L and M are constant than this gives
T3
To show that P 4/3 we can also use scaling argument, we also know that the radiation pressure
scales as some constant times the gas pressure
Prad
L

Pg
M
thus we find

but we know that

Pt = Pg + Prad Pg = Pt Prad

Pr
1 Pr Pg
Pt Pr
Pg T

T 1/3 Pg 4/3

thus we know that


Pg Pt Pg

Pt 2Pg 4/3
7

Problem set 3

Problem # 1
(a). Show that heat transfer by radiative diffusion implies a non-zero gradient for the radiation pressure
which is proportional to the radient heat flux. Bearing in mind that the magnitude of the force per unit
volume in a fluid due to the pressure is equal to the pressure gradient, find the radient heat flux density
which can, by itself, support the atmosphere of a star with surface gravity g. Hence show that a star of
mass M has a maximum luminosity given by
Lmax =

4cGM

where is the opacity near the surface. Obtain a numerical estimate for this luminosity by assuming
that the surface is hot enough for the opacity to be dominated by electron scattering. (This maximum
luminosity is called the Eddington luminosity.
To show that the heat transfer by radiative defusion implies a non-zero gradient we must begin with
Fr =

4 aT 3
T
3

Fr Prad

knowing these relationships we can do


1
Prad = aT 4
3

dPr 4 3 dT
dT
3 1 dPr
= aT

=
dr
3
dr
dr
4 aT 3 dr

thus this implies that there is a non-zero gradient.


To show that

4cGM

we must begin with the equation derived from problem 4 in the last problem set, i.e
Lmax =

dPr
L
=
dP
4cGM
but since we know that
P = Pg + Pr

Pg Pr P Pr

dPr
=1
dPr

and we find that


4cGM
3.3 104Lsun
L=

M
Msun

We cannot obtain a numerical estimate because we do not know tha mass. We could use Msun but this
would not be correct.

(b). Assume that radiative diffusion dominates energy transport in stars and that the opacity is due to
Thomson scattering. Use a scaling argument to estimate the mass M (in Msun ) at which the luminosity of
a star is Ledd .
We can do an order of magnitude estimate with respect to the sun by
LM

L
=
Lsun

M
Msun

3

and substituting the Eddington luminosity for L we find that M is given by


M=

4cGM
Lsun T

1/2

3/2

Msun 180 Msun

Problem # 2
The physical quantities near the center of a star are given in the following table. Neglecting radiation
pressure and assuming the average gas particle mass m is 0.7 amu, determine whether energy transport is
convective or radiative.
r
0.1Rsun

m(r)
0.028Msun

Lr
24.2Lsun

Tr
2.2107 K

(r)
3.1 104kg m3

0.040 m2 kg1

Using equation ?? from Phillips




L(r)
m(r)

crit

1 16Gc Pr

and the following relationships


1
(r)
Pr = aT 3 P Pg =
kb T
3
m

5
T = 0.04 m2 /kg
3

we find


2 16Gc aT 3 m
5 T 3(r)kb
crit
W
W
> .07
0.175
kg
kg
L(r)
m(r)

which implies that the energy transport of this star is primarily due to convection.
Problem # 3
The surface of a star (the photosphere) is the place where the mean free path of the photons is
comparable to the scale-height h of the atmosphere . At smaller radii (deeper in the star), the density is
higher and h , which implies that the photons bounce around many times; at larger radii is smaller,
h, and the photons are rarely absorbed and so travel on straight lines to us. Thus h is a good
approximation to the place in the atmosphere of a star where most of the light we see originates.
a) The temperature at the photosphere of the sun is 5800 K. Estimate the mass density in the photosphere. Assume that Thompson scattering dominates the opacity.

Knowing that h we can derive the following relationship


n = =

nT
T

1
kb T
=h=
n
mg

thus we find that the density is given by and assuming m m p


=

mg

8.26 104 g/cm3


kb T T

b) In reality, the surface of the sun is so low that hydrogen is primarily neutral. There are thus not that
many free electrons to Thompson scatter off of. The opacity at the surface of the sun is instead due to the
H ion and is given by 2.5 10311/2 T 9 cm2 g1 . Using this (correct) opacity, repeat the estimate
from a) of the density at the photosphere of the sun.
Substituting the opacity given into the above expression yields

3/2

mg

=
2.5 1013kT 10
b

mg

2.5 1031kb T 10

2/3

9.8 108 g/cm3

c) Just beneath the photosphere, energy is transported by convection, not radiation, for the reasons
discussed in class (in fact, the photosphere is the place where photons travel so freely out of the star that
energy transport by radiation finally dominates over convection). Estimate the convective velocity near
the photosphere given your density from b).
The convective heat flux is given by
1
Fc = v3c
2

vc =

2F c

1/3

and knowing that


Fc =

L
4r2

we find that the convective velocity is given by


vc =

2L
4r2

1/3

1.09 106 cm/s

d) What is the characteristic timescale for convective blobs to move around near the pho-tosphere?
How does this compare to the observed timescale for granulation on the surface of the sun, which was a
few min in the movie we watched in class?
since we know that the characteristic time scale is given by
tblob =

vc vc

we know that h which is the scale height of the sun is given by


h=

kT
= 1.7 107 cm
gm
10

thus we find that the blob timescale is


tblob 15.6 s
which is a lot shorter than the timescale given by the movie which was approximately 2 minutes.
e) Is the assumption ds/dr 0 valid near the surface of the sun? Why or why not?
Since we know that the temperature gradient near the surface of the sun is very high and energy is
mostly transported by photons impies that we cannot make the assumption ds/dr 0 .
Problem # 4 Convective atmospheres
In HW 2, you calculated the structure of a stellar atmosphere in which energy is transported by radiative
diffusion; you showed that such an atmosphere satisfies P 4/3 . Here we will consider the problem of
a convective atmosphere, which is much more relevant to sun-like stars. For simplicity, assume that the
atmosphere is composed of fully ionized hydrogen. The solar convection zone contains very little mass
(only 2of the mass of the sun). Thus, lets consider a model in which we neglect the mass of the
convection zone in comparison to the rest of the sun. For the reasons discussed in class, we can model
the convection zone as having P = K with = 5/3 and K a constant. Rc is the radius of the base of the
convection zone.
a) Solve for the density, temperature, and pressure as a function of radius in the convection zone. Do
not assume that the convection zone is thin (i.e., even though Mr = constant = M , because r changes
significantly in the convection zone, do not assume that the gravitational acceleration is constant).
To solve for the Pressure we can begin with
GM
dP
= 2
dr
r
thus we find

P = K

 1/
P
=
K

 3/5
GM
P
dP
= 2 dr
K
r

and integrating over the following limits we find


Z P 
P
Prc

dP

3/5

Z R
GM
Rc

r2

dr

this integral yields




i
5 h 2/5
1
1
2/5

= GM
P Pc
R Rc
2K 3/5

thus we find that the pressure is given by


P=

5/2


2 3/5
1
1
2/5
+ Pc
K
GM

5
R Rc

To solve for the density we can just plug this solution into
3/2



 3/5
1
2 3/5
1
1
P
2/5
= 3/5
+ Pc
K
GM

=
K
5
R Rc
K
11

and finally the temperature can be found by using


P = nkb T =

2
kb T
mp

T (r) =

m pP
2kb

and substituting the P and from the previous expressions we find






m p K 3/5 2 3/5
1
1
2/5
T (r) =
+ Pc
K
GM

2kb
5
R Rc
b) In detailed solar models, the pressure at the base of the convection zone is 5.2 1013dyne/cm2
and the density is 0.175 g cm3 . Using your solution from a), estimate the radius of the base of the
convection zone Rc . Compare this to the correct answer of Rc 0.71Rsun
If we solve the density equation for Rc we find
i 5K 3/5
1 h
1
2/5

= (K 3/2 )2/3 Pc
Rc R
2GM

and plugging in values we find that

1
= 1.998 109m1
Rc

Rc 5.11 108m = 0.72Rsun

c) In your model, what is the temperature of the sun at 0.99Rsun , 0.9Rsun , and at the base of the solar
convection zone. This gives you a good sense of how quickly the temperature rises from its surface value
of 5800 K as one enters the interior of the sun.
To find the temperature as a function of radius we would use the temperature equation derived from
part (a). i.e


m p K 3/5
2 3/5
25
2/5
T (r = 0.99Rsun) =
4.1 104 K
K
GMr
+ Pc
2kb
3
66 Rsun


m p K 3/5
15
2 3/5
2/5
5.1 105 K
+ Pc
K
GMr
T (r = 0.90Rsun) =
2kb
3
18 Rsun
T (r = 0.72Rsun) =

m p K 3/5 2/5
Pc 1.8 106 K
2kb

Problem set 4

Problem # 1
I mentioned in class that there are two ways to estimate the energy carried by convection. The first is
that the energy flux is Fc 1/2v3c Fc,1 where vc is the characteristic velocity of the convective motions.
12

This is the KE flux carried by moving blobs. The other estimate is that Fc Evc where E is the
difference in the thermal energy of a rising hot blob (or sinking cool blob) relative to the background star
(where E is per unit mass). I claimed in lecture that these two expressions are equivalent, to order of
magnitude (which is the accuracy of mixing length theory). In this problem, you will prove my claim.
(a). Calculate the acceleration a due to buoyancy of a rising hot blob (or sinking cool blob) in terms of
the fractional density difference / relative to the background star. Dont worry about the sign of the
acceleration or ?, just their magnitudes.
We know that the accelaration of the blob due to bouyancy is given by



b
ab = g
1 = g

since b .

(b). Use (a) to calculate the convective velocity vc in terms of /. Recall that in lecture we estimated
vc using the work done by the buoyancy force.
We know that the work done by the bouyancy force can be found by
1
W = mv2c =
2

Z l
0

thus we find that the convective velocity is given as


s
vc =

F dl = aml

2g

which can also be expressed as

v2c
=
2gl

we can also write this as given that l H, thus


v2c = 2g

kb T
kb T
=2
mg
m

(c). Use (b) to calculate E, the difference in the thermal energy (per unit mass) of a rising hot blob
(or sinking cool blob) relative to the background star, in terms of vc .
We can write the last expression as
T = v2c

m
2kb

and from the equation of state, which is given as


E =

1 kb T
( V )
m

where ( V ) is the mass. Using these two expression and what we found from part (b) we can see
that
1 2 E
E v2c
E
= vc =

V
2
m
m
2
13

d) Combine your previous results to show that Fc,1 Fc,2 .


From (a) we know that

1
E
Fc,1 v3c Fc,2 vc
2
m

using this and our solution we find


E
1
E
1 3
vc vc v2c
2
m
2
m
Problem # 2
Estimate the convective velocity vc and the dimensionless entropy gradient (ds/dr)(H/c p) in the convection zones of 0.1 and 10 Msun stars. Assume that the material undergoing convection is at about the
mean density of the star and that gas pressure dominates. You can either use a scaling argument to estimate
the density, temperature, luminosity, etc. of such stars or look up in a book (e.g., Carrol & Ostlie) any
properties of 0.1 and 10M? stars that you need to make your estimate (e.g., radius and luminosity). But
you cant just look up vc and (ds/dr)(H/c p).
From class we know that
Fc =
but we know that



H ds 3/2

C p dr

3Cs3

thus

Fc



H ds 1/2

vc = Cs
C p dr

L
4R2



2/3
H ds
L
1
=

C p dr
2
4R
Cs2

which reduces to



2/3

H ds
mp
LR


C p dr = 3M
kb T

where

Cs2 =

kT
mp

3M
4R3

and for the convective velocity we find






H ds 1/2
LR 1/3


=
vc = Cs
C p dr
3M

from Carrol and Ostley we find that for 10Msun and .1M sun we find that the radius, and luminosity are
approximately
M 10Msun R 6Rsun L 5700Lsun
M 0.1Msun R 0.2Rsun L .0034Lsun
given these values we find
M = 10Msun
M = 0.1Msun



H ds
6


vc 5.3 104 cm/s
C p dr 3.61 10


H ds
10


vc 698 cm/s
C p dr 5.8 10
14

Problem # 3 Polytropes
(a). The mass M of a star is given by
M=

Z R
0

4r2 (r)dr

Use the Lane-Emden equation for polytropes, and the dimensionless density and radius defined in lecture,
to rewrite this in terms of the central density of the star as


3M
n=
an
c = a
4R3
where an is a dimensionless number, the ratio of the central density to the mean density of the star. an is
a function that you should determine that depends only on the solution to the Lane-Emden equation (you
cannot actual evaluate an in general without numerically solving for [], so your answer will just be in
terms of the solution to the Lane-Emden equation).
Since we know that
=
given these two relations we can find
(r) = n c

(r)

1/n

r
a

r2 = a2 2 a = R dr = Rd

and from the Lane-Amden equation we know




d
2 d

= 2 n
d
d
given these following relationships we find that
M=

Z 1
0



Z 1
d
2 d
4R c d = 4R c

d
3 2

thus we find that an is given by


1
an =
3



Z 1
d
2 d

d
0 d

and we can finally show that


c =

3M
an
4R3

(b). Show that the central pressure of a polytrope can be written as


4G2c a2
n+1
where a 6= an is the constant (with units of length) defined in lecture (note that the polytropic relation

P = K can be used to write K = Pc c . Use this result and (a) to derive an expression for the central
pressure of a polytropic model of the form


GM 2
Pc =
cn
R4
Pc =

15

where cn is again a dimensionless function that you should write down. Also show that the central
pressure of a polytrope can be written as
4/3

Pc = dn GM 2/3 c

where dn depends on an and cn . The values of an , dn , and dn can be determined by numerically solving the
Lane-Emden equation. The most useful cases for our purposes are = 4/3 (n = 3) and = 5/3 (n = 3/2)
polytropes. For n = 1.5, an = 5.99 and cn = 0.77 while for n = 3, an = 54.183 and cn = 11.05. We will use
these quite a bit during this course. Note how, as mentioned in class, the results for the central pressure and
density of polytropes above are very similar to what you would get from an order of magnitude estimate,
except that for polytropes we get an exact correct numerical factor given by an , cn and dn .
We know that

K = Pc
c

and also

#
1/n1 1/2

"

(n + 1)Kc
a=
4G
given these relationships we can now find

a2 4G
1/n1
1/n1
= Pc
= Kc
c c
n+1
but since we know that = 1/n + 1 we find
1/n1

Pc
c c

(1+1/n) 1/n1
c

= Pc c

= Pc 2
c

and finally we find


Pc =

a2 4G2c
n+1

using this results we can now derive


Pc =
which becomes

a2 4G2c
GM 2
= 4 cn
n+1
R

a2 4Ga2n
Pc =
n+1

3M
4R3

looking at these two expressions we can see that

2

GM 2
cn
R4

9a2 a2n 1
cn =
4R2 n + 1
now looking at
4/3

Pc = dn GM 2/3 c
and from (b) we find
2/3 4/3

a2 4Gc c
Pc =
n+1

a2 4G
=
n+1
16

an 3M
4R3

2/3

4/3

which becomes
4/3
Pc = GM 2/3 c cn

4
an

4/3

92/3

and after some fun algebra, which will be omitted here we find
dn = cn

4
3an

4/3

(c). What are the values of dn for n = 3 and 1.5 polytropes, respectively?
using the above result we find
n = 3 an = 5.99 cn = 0.77 dn = 0.477
n = 3/2 an = 54.183 cn = 11.05 dn = 0.363
(d). Use your expressions for the central pressure and density to give an expression for the central
temperature of a polytrope. Assume gas pressure dominates.
to find these expression we will assume that gas pressure dominates, i.e
Pc =

c
kb Tc
m

thus

1/3

dn GM 2/3 c m
Pc m
=
c kb
kb
and skipping some algebra we find that the central temperature is given by
Tc =

dn GM
Tc =
Rkb

3an
4

1/3

mp
2

so all of the expressions can be written as


dn GM
Tc =
Rkb

3an
4

1/3

mp
2

Pc =

GM 2
cn
R4

c =

3M
an
4R3

(e). Calculate the central temperature, pressure, and density for = 4/3 (n = 3) and = 5/3 (n = 3/2)
polytropes for M = Msun and R = Rsun (i.e., for the sun). Assume fully ionized hydrogen for simplicity.
Which polytrope better approximates the true interior temperature, pressure, and density of the sun? Can
you explain physically why this is the case?
using the above expressions for temperature, density and pressure we find for
=
we find

5
an = 5.99 cn = 0.77 dn = 0.477
3

c 8.44 g/cm3 Pc 8.67 1015 Tc 6.2 106 K

and for
=

4
an = 54.183 cn = 11.05 dn = 0.363
3
17

we find

c 77.26 g/cm3 Pc 1.24 1015 Tc 1.02 107 K

thus we can see that the = 4/3 polytrope best represents the values observed in the sun, this is mainly
due to the fact that the center of the sun is radiative and not convective. Since we now know that
P 4/3 radiative
P 5/3 convective
Problem # 4
Consider a pre-main sequence star (gas cloud) of mass M undergoing Kelvin-Helmholz contraction.
In class, we showed that fully convective stars move down the Hayashi line with Te f f constant. But
stars with M > 0.3Msun do not end up fully convective on the main sequence and so must go through
a phase of KH contraction in which energy transport is dominated by photons. Assume throughout this
problem that gas pressure dominates and that free-free ab- sorption dominates the opacity (because the
temperature is lower during KH contraction than on the main sequence, free-free absorption tends to
be even more important). Motivated by HW #2 Problem 1, assume that the luminosity of a star in
which photons carry the energy out and the opacity is dominated by free-free absorption is given by
L Lsun (M/Msun )11/2 (R/Rsun )1/2 .
(a). Determine how the radius, luminosity, and effective temperature vary as a function of time and
mass M for a radiative star undergoing KH contraction. Dont worry about the constants in these relations;
all you need to calculate are proportionalities (i.e., how do the various quantities depend on time and mass
M). Do the luminosity and effective temperature increase or decrease as the star contracts?
since we know that

Pg > Pr = f f

thus we know that


L f f Lsun

M
Msun

11/2 

R
Rsun

1/2

Lrad

1 GM 2 dR
2 R2 dt

since we are doing proportionalities we find


M2 R
M 11/2 R1/2
2
R tT
so we find that the radius scales as
R
plugging this into

1
M 7t 2

L M 11/2 R1/2 M 9t

thus

L M 9t

and to find the temperature


L

R2 Te4f f

Te f f

18

L
R2

1/4

M 23/4 t 5/4

thus
Te f f M 23/4 t 5/4
We can see that as time and mass increase the luminosity increases as well as the effective temperature.
(b). Estimate the radius of a star (in Rsun ) of a given mass M (in Msun ) at the time when energy transport
by photons takes over from convection during the KH phase. At what luminosity does this occur (again as
a function of mass M)? Assume based on lecture that the luminosity of a fully convective star is
L 0.2Lsun (M/Msun )4/7 (R/Rsun )2
if we set the free-free luminosity equal to the convective luminosity we find
Lsun

M
Msun

11/2 

R
Rsun

1/2

0.2Lsun (M/Msun )4/7 (R/Rsun )2

some algebra yields


R 5

5/2

M
Msun

69/35

Rsun

(c). Sketch the paths of 1Msun pre-stellar gas clouds during their KH contraction phase in the HR
diagram. Include both the convective and radiative parts of their evolution and the correct transition point
between the two based on b). Be sure to properly label your axes (L in Lsun and Te f f in K). Note that on
the main sequence a 3 Msun star has L 40Lsun and Te f f 10000 K (you know the values for the sun).
The KH contraction phase ends when the star contracts to the point where its luminosity and temperature
have these values.
since we now have a relationship for the radius we can find what the luminosity is by
L 0.2Lsun

M
Msun

4/7 

R
Rsun

2

2/5

0.2Lsun

which yields
L(Msun) 0.724Lsun
and a plot is given by

19

M
Msun

4/7 

M
Msun

128/35

Figure 1: We plot the path that a 1 solar mass star would trace when moving from the Hayashi track
to the main sequence.

Problem set 5

Problem # 1
In lecture we discussed the slow, nearly hydrostatic, contraction of pre-stellar gas clouds as they approach the main sequence - Kelvin Helmholz (KH) contraction.
(a). Argue that, for KH contraction to occur, the timescale
for KH contraction tKH must be longer than
p
the gravitational free-fall time of the cloud, t f f 1/ Ghi, where hi is the mean density of the cloud.
What happens if tKH < t f f ?
Since we know that
1
tf f p
Ghi

tKH

"

M
Msun

20

1/2

Rsun
R

#3

(2 107yrs)

froma a purely physical argument we know that things cannot fall faster than gravity can pull it. Thus
tKH t f f
and saying t f f tKH would be an unphysical statement due to the knowledge we have about gravity.
We also know that when things collapse that the radius gets smaller, hence collapse and from the relationships of time in both free-fall and Kelvin-Helmholtz contraction we can see that as R that t f f thus the
only possible solution is that tKH t f f .

(b). Estimate the critical radius Rc (inRsun) at which tKH t f f for a given cloud of mass M (in Msun ).
Assume, as we did in class, that the cloud is fully convective at early times. Show that for R < Rc , the
cloud undergoes KH contraction according to your criterion from a). Recall that the luminosity of a fully
convective star is L 0.2Lsun (M/Msun )4/7 (R/Rsun )2 .
We can find the critical radius by setting the above expression equal to each other, i.e
"
#3

1
M 1/2 Rsun
p
=
(2 107 yrs)
Msun
Rc
Ghi
re-arranging this expression for Rc we find


M 1/2
Rsun (Ghi)1/6 (2 107 yrs)1/3
Rc =
Msun

and letting
hi =

3M
M
3
3
4Rc
Rc

we find
Rc

2/3
7
1/3 G1/9
4/9 Rsun (2 10 yrs)
M
1/3
Msun

M 4/9 7.6 103

if we put the mass off the sun we get


Rc 4.79 1010cm
Since we know that
t f f R3/2

tKH

1
R3

we can see that


R
R

tKH t f f
tKH t f f

from these two relationships we can see that for KH contraction to occur that tKH t f f and also that
R < Rc .
(c). What is the central temperature of the (fully convective) cloud (in K) as a function of its mass M
(in Msun ) when R = Rc ?
We know that the temperature of a fully convective object is given by


dn GM 3an 1/3 m p
Tc =
Rkb
4
2
21

and since we know that fully convective stars have a polytropic index of = 5/3. Knowing this we
find
n = 3 an = 5.99 dn = 0.477
thus
M
Rc
2.79 1015 M 5/9

Tc 2.16 1015

and if we want the temperature of a collapsing gas cloud with respect to Msun we get
Tc 1.9 104K
Problem # 2
The globular cluster M13 in Hercules contains about 0.5 million stars with an average mass of about
half the solar mass. Use Jeans criteria to check whether this cluster could have formed in the early universe
just after the time when the universe was cool enough for the electrons and nuclie to form neutral atoms;
at this time the density of the universe was 1027 kg m3 and the temperature was T 104 K .
Using the Jeans mass equation
Mj =

kb
Gm p

3/2

T 3/2

using the values given we find


M j 1.37 1042 g
and the mass of M13 is
M13 0.5 106 Msun 9.95 1038 g
and we can see that
M j M13
which means that this cluster could not have formed in the early universe. Things only collapse if the
mass is greater than the Jeans mass
Problem # 3
The binding energy per nucleon for 56 Fe is 8.8 MeV per nucleon. Estimate the energy released per
kilogram of matter by the sequnce of reactions which fuse hydrogen to iron.
We know that the enery released will be given as
Etot =

Eb
Nnucleon
nucleon

and the number of nucleons are given by


1 kg
M

5.98 1026 nucleon


Nnucleon =
mp
mp
thus
Etot 5.27 1027 MeV
(b). Consider two hypothetical stars of the same mass M and the same luminosity L (that is constant in
time). The stars are initially pure hydrogen. In star A, fusion proceeds until the entire star is converted into
22

He. In star B, fusion proceeds until the entire star is converted into Fe. Which star has a longer lifetime,
and by how much?
We know that

EHe
EFe
LFe =
tHe
tFe
and since we know that these two luminosities are theoretically equal
LHe =

EHetFe = EFetHe
which gives
tHe =

6.4
EHe
tFe =
tFe 0.72tFe
EFe
8.5

thus we see that


tHe < tFe
we can see that the time for all of the hydrogen to fuse into helium is less then the time for all of the
hydrogen to fuse into iron so the star that is converted to iron has a longer lifetime.
Problem # 4
(a) What is the classical distance of closest approach for two protons with an energy of 2 keV (the
mean thermal energy at the center of the sun)? Estimate the probability that the protons tunnel through the
Coulomb barrier trying to keep them apart. Answer the same two questions for two 4 He nuclei and for a
proton and a 4 He nucleus with the same energy of 2 keV.
The classical distance of closest approach is given by
rc =

e2 Z1 Z2
7.2 1011 cm
E0

We know that the propbability for a particle-particle interaction is given by


 1/2
E
EG

P=e
where

EG = 22 2 Z12 Z22 (mr c2 )

is the Gamow energy. For a proton-proton interaction we find


EG .493 MeV
thus the probability is given as
P 1.51 107
for a He-He interaction we find
e2 Z1 Z2
rc =
2.8 1010 cm
E0
EG = 322 2 2m p c2 31.6 MeV
and the probabilty is given by
P 2.50 1055
23

for the proton-He interaction we find


rc =

e2 Z1 Z2
1.44 1010 cm
E0
EG 3.15 MeV

thus the probability is given by


P 5.8 1018

(b) What energy E would be required for i) the two 4 He nuclei, ii) the proton and the 4 He nucleus, and
iii) two 12 C nuclei to have the same probability of penetrating the Coulomb barrier as the two protons?
For particles with energies equal to the mean thermal energy of the plasma, what temperatures do these
correspond to?
Since we know that
E=

EG
(ln P)2

for the He-He interaction we find


EHeHe 0.125 MeV
for the proton-He interaction we find
E pHe .013 MeV
for the carbon-carbon interaction the Gamow energy is given by
EG 259222 6m p c2 7.7 GeV
and thus
Ecc 31 MeV
we know that

E
kb

T
so
THeHe

1.45 109K

TpHe

1.5 108K

Tcc

3.6 1011K

Problem # 5
Calculations of nuclear reaction rates are done in the center of mass (COM) frame, so it is useful to
remember a few results about the COM. Consider two particles of mass m1 and m2 with positions x1 and
x2 and velocities v1 and v2 .
(a) .What is the velocity of the COM?
We kbnow that the center of mass is given by
com =

m1 r 1 + m2 r 2
m1 + m2
24

so the velocity would be


vcom =

m1 v1 + m2 v2
m1 + m2

(b). What are the velocities of each of the two particles in the COM reference frame (i.e., in the frame
for which the COM is at the origin)?
We know that the relative velocities are given by
vrel1 = v1 vcom
vrel2 = v2 vcom
a bit of algebra yields
vrel1 =
vrel2 =

m2
(v1 v2 )
m1 + m2
m1
(v2 v1 )
m1 + m2

(c). What is the total KE of the two particles in the COM frame? Show that this is equal to the KE of
the reduced mass moving at the relative velocity, as claimed in class.
We know that the total kinetic energy is given by
1
1
m1 v2rel1 + m2 v2rel2
2
2
m1 m2
=
(m2 (v1 v2 )2 + m1 (v2 v1 )2 )
2(m1 + m2 )2
m1 m2
(v1 v2 )2
=
2(m1 + m2 )

Ktot =

Ktot =

1
mr (v1 v2 )2
2

where mr is the reduced mass.

Problem set 7

Problem # 1 The Main Sequence for Fully Convective Stars


In this problem we will determine the main sequence for fully convective low mass stars. We showed
in lecture that fully convective stars have Te f f 4000(L/Lsun )1/102 (M/Msun )7/51 K (I actually derived a
coefficient of 2600 K in lecture but commented that more detailed calculations get something similar but
with the coefficient closer to the value of 4000 K used here). We can also write this result as
25

L 0.2(M/Msun )4/7 (R/R)2sun

Lsun Lconv

I called this luminosity Lconv since it is derived from the properties of energy transport alone (convective interior + radiative atmosphere with H opacity). The luminosity of a star is also given by
L f usion = 4r2 (T, )dr
where is due to the proton-proton chain for low mass stars (this was given in lecture). As discussed in
class, the main sequence is determined by the requirement that the energy escaping the star (in this case by
convection) is equal to the energy generated in the star (in this case by pp fusion), i.e., that Lconv = L f usion
.
a) Use scaling arguments to derive the power-law relations R(M), L(M), Tc (M), and L(Te f f ) (the HR
diagram) for fully convective stars, like we did for other examples in lecture. Approximate T with
an appropriate choice of (recall that low mass stars will have somewhat lower central temperatures than
the sun, closer to 106 K, as you will see in part b).
We know that

L f us R3 (, T ) R3 2 T

Lconv M 4/7 R2

we can find what is by




EG 1/3
2
= +
3
4kT
given that we know what the temperature is and also what EG for p-p reaction
T 5 106 K

EG 500 keV
we find that

= 5.92 6.0

we also know

M
R3

we know that in steady state


L f usion = Lconv
thus we can find


M2
M R R
Tc6 T 6 M 5/21 R5/6
R6
we know from the Virial temperature, assuming gas pressure dominates
4/7 2

M
R

thus we find
Tc M 25/77
knowing this we can now find
R M 52/77
with this and the relationship for the convective luminosity we find
L M 4/7 R2 M 148/77
26

with this we can now find what the effective temperature as a function of mass is, i.e
L R2 Te4f f
thus
Te4f f

L
M 4/7
R2

which yields
Te f f M 1/7
to find what the luminosity as a function of the effective temperature is (HR diagram)
M Te7f f
which yields
148/11

L Te f f

In a) you just determined a scaling relation between stars of different mass, but not the absolute values
of L, Te f f , etc. In class, we did the latter by scaling to the sun. Note, however, that it is not reasonable
to estimate the properties of low mass stars by scaling from the properties of the sun, since the sun is not
a fully convective star! Instead we need to actually determine the structure of some fully convective star.
This is what we will do in the rest of the problem. We can significantly improve on the above scaling
arguments by using the fact that fully convective stars are n = 3/2 polytropes. It turns out that for a
polytrope, in equation (1) can be Taylor expanded near the center to yield
L f usion

2.4c M
(3 + )3/2

where I have again approximated T and where c is evaluated at the center of the star. I am not
asking you to prove equation (2). You will have to trust me. Note that for a typical value of for the pp
chain, equation (2) says that L f usion 0.1c M . This makes sense because fusion only takes place at the
center of the star (not all of the mass participates).
b) Use the results for n = 3/2 polytropes from HW 4, Problem # 3, to write the central temperature of
the star Tc , central density c , and pp energy generation at the center of the star c in terms of the mass
M and radius R. Assume X = 0.7 and = 0.6 (typical for stars just reaching the main sequence). Note
that you should give expressions for Tc , c , and c here, with constants and real units, not just scaling
relationships. So that the constants in front of your expressions are reasonable, please normalize M to
Msun and R to Rsun .
The general expressions given by HW 4 problem #3 are
dn GM
Tc =
Rkb

3an
4

1/3

m p

Pc =

GM 2
cn
R4

c =

we found that for a n = 3/2 polytrope


an = 5.99 cn = 0.77 dn = 0.477
27

3M
an
4R3

thus we find that




m pGMsun
Tc = 0.322
kb Rsun

M/Msun
R/Rsun

Msun
c = 1.43 3
Rsun

M/Msun
(R/Rsun )3

plugging in all the constants yields




M/Msun
Tc 7.43 10 K
R/Rsun
6

we also know

M/Msun
c 8.41
(R/Rsun )3

2/3 15.7T71/3

c = 5 105 c X 2 T7

we also know that we can approximate this as

c = AT7 X 2 = AT76 X 2
setting this two expressions equal to each other we can find what A is, i.e lettingT 107 K we find
A = 5 105 e15.7 0.076
thus we find

c 0.076cT76 X 2 0.037cT76

substituting Tc and c gives

M
c 0.053
Msun

7 

Rsun
R

9

c) Use equation (2), the results of b), and Lconv = L f usion on the main sequence to determine the R(M),
L(M), Tc (M), and L(Te f f ) relations for fully convective stars. If you use the same , your expressions here
should be the same as in a) except that you should now be able to determine the absolute normalization
for R(M), L(M), etc., i.e., you have determined the true luminosity and radius of a ful ly convective star
from first principles. In doing this problem, remember that is temperature dependent so make sure you
check that your value of is reasonable given the resulting central temperature that you calculate.
Using the results from b and also
L f usion

2.4cM
.09cM
(3 + )3/2

( 6)

thus we know that


Lcon = L f usion


M
0.2
Msun

4/7 

R
Rsun

2

Lsun

M
= 0.0047
Msun

7 

Rsun
R

rearranging this we find

which yield

R
Rsun

11

Msun
= 0.023
Lsun

M
Msun



R
M 52/77
0.67
Rsun
Msun
28

52/7

9 

M
Msun

Msun

to find for the temperature we can use




M/Msun
Tc 7.43 10 K
R/Rsun
6

using our previous resulst gives




M
Tc 1.1 10 K
Msun
7

25/77

to find for the luminosity we use




M
Lconv = 0.2
Msun

4/7 

R
Rsun

2

Lsun

plugging in for the radius we find




148/77

Lsun

148/77

Lsun = 4R2 Te4f f

M
Lconv = 0.09
Msun
for the effective temperature we find
L = 4R

Te4f f

M
0.09
Msun

which can be simplified to






M 148/77
M 104/77 4
18
0.09
Lsun = 1.56 10
Te f f
Msun
Msun
thus
Te f f

M
= 3868 K
Msun

which can also be written as


M
=
Msun

Te f f
3868 K

1/7

7
148/11

L(Te f f ) = 4.85 1050 Lsun Te f f

d) What are your predicted luminosities, radii, and effective temperatures for main sequence stars with
M = 0.1 and 0.3Msun ? Compare your values to the values of L = 0.01Lsun , R = 0.3Rsun , and Te f f = 3450
K for M = 0.3Msun and L = 103 Lsun , R = 0.11Rsun, and Te f f = 3000 K for M = 0.1Msun that I found in a
graduate textbook (based on detailed models).
Given our relationships we find for
M = 0.1Msun L = 1.1 103Lsun R = 0.14Rsun Te f f = 2783 K
and for
M = 0.3Msun L = .009Lsun R = 0.298Rsun Te f f = 3256 K
29

Problem # 2 Very Massive Stars


Consider very massive stars with M 50 100Msun . Recall that I showed in lecture and you showed
on HW 3, Problem # 1, that in such stars, radiation pressure due to photons (a relativistic particle) is more
important than gas pressure. Fusion is by the CNO cycle. Assume for now that energy is transported
primarily by photons and that the opacity is due to Thomson scattering (reasonable for hot massive stars).
a) Use scaling arguments to derive the power-law relations R(M), L(M), Tc(M), and L(Te f f ) (the HR
diagram) for very massive stars, like we did for other examples in lecture.
Using radiative diffusion along with
1
Prad = aT 4
3
we also know

thus

and radiative diffusion says

dPr 4 3
= aT
dT
3

4
dT
GM
dP dP dT
=
= aT 3
= 2
dR dT dR 3
dR
R
M
dT
3 2
dR T R
dT
M
L

3 2
2
3
R T
dR T R

which gives us

LM

using the Virial theorem, where Prad dominates rather than Pgas we find
T4 M

Tc

M 1/2
R

where the left hand term is from the radiation pressure, but since we know that is an energy density
we must devide by the density to find what the energy is per particle. Now using the steady state for
luminosity we find
L MT 18
where we chose = 18 as a more appropriate value rather than the value given for the sun = 20, this
is motivated by the fact that more massive stars have somewhat higher temperatures, thus reducing . We
find
1 R3
T 18
M
and using the result from Virial temperature we find
M 9 R3

R18 M
thus we find
R M 10/21
and we also find for the central temperature
Tc M 1/42
30

and to find the effective temperature we know


Te4f f

L
M
2 M 1/21
2
R
R

simplifying gives
Te f f M 1/84
and finally the luminosity as a function of Te f f is given by
L(Te f f ) Te84
ff
b) Estimate the fraction of the mass in the star that is undergoing convection (recall that fusion by
the CNO cycle is very concentrated at small radii because of the strong temperature dependence). For
comparison, detailed calculations show that the fraction of the mass that undergoes core convection
increases from 10 % at 2Msun to 75% at 60Msun .
The condition for convection is given by
d ln T
1 Ptot L Lr /L
1

>
d ln P 4 Prad LEdd Mr /M

since we know that


=

4
Ptot Prad
3

gives us
1 Lr M
1
>
4 LEdd Mr 4
which simplyfies to
Lr
Mr
>
Ledd
M
we know that in the limit that M 150Msun Lr LEdd ,
Mr
<1
M
which means that the fraction of the mass of the star that is undergoing convection approaches 1, which is
100% of the mass is undergoing convection. Its a little strange that stars that are much less massive than
the sun and the stars that are much more massive than the sun are both almost fully convective.
n=3/2 polytrope
The previous case yielded a result for a n=3 polytrope, we find that for
n = 3/2 polytrope

and in the limit where M 150Msun

Mr 5 Lr
<
M
8 LEdd
Lr Ledd
Mr 5
<
M
8
31

5
3

This seems rather strange in the sense that stars that are approximately 60Msun have a convective core
that encompasses 75% of the mass, which means that the convective core decreases after M > 60Msun ?

c) Calculate the main sequence lifetime of a very massive star as a function of its mass M . Be sure to
take into account the results of b).
We know that the main sequence lifetime of a star is given by
tMS

Etot
LEdd

where
Etot = NQ

Q 7 MeV

where that is the total energy per reaction, we also know


Mr
M
=
(n = 3 polytrope)
mp mp
Mr 5 M
=
(n = 3/2 polytrope)
N =
mp 8 mp
N =

we also know that

4cGM
T
so we find the main-sequence lifetime to be given as
LEdd =

T Q
(n = 3 polytrope)
m p 4cG
5T Q

(n = 3/2 polytrope)
m p 32cG

tMS
tMS
we know that

T 0.4 cm2 /g

Q 7 MeV 1.12 105 ergs

thus we find that the main-sequence lifetime for both types of polytropes are given by
2.21 106yr < tMS < 3.39 106yr
seems reasonable.

Problem set 8

Problem # 1 Fermi gas


32

a) Above what density is a gas of room temperature fermions degenerate? Below what temperature
would gas with the density of air be degenerate?
as

We know that if the density of the gas is ng nQ where nQ is the quantum concentration, nQ is defined
nQ

2mkT

h2

3/2

(1)

and for a gas at room temperature to be degenerate


ng nQ =

2mkT

h2

3/2

1.46 1026cm3

where we used
T = 300K

m = 28m p

due to the fact that air is mostly composed of N2 . If we assume that the questio is only speaking about
free electrons we get


2me kT 3/2
ng nQ =
1.25 1019 cm3
2
h
using the same temperature as before.

To find the temperature at which gas with a density of air would be degenerate can by using the above
expression, except now we must find what the density of air is at STP and use this, i.e
nair =

P
= 2.52 1019 cm3 = nQ
kT

and now using Equation 1 we find


2/3

T=

nQ h2
2mkT

9.2 103 K

using m = 28m p
b) Compare the relative importance of the thermal energy, the electrostatic (Coulomb) energy between
electrons and ions, and electron degeneracy (electron Fermi energy) in room temperature silver (Z = 47;
10g cm3 ). Which dominates?
We can write the thermal energy as
3
Eth kT .039 eV
2
We can write the Coulomb energy as
Ecoul Z 2 e2
we know
r n1/3
33

1
r

 1/3
m

thus
2 2

Ecoul Z e

 1/3

12.43 keV
m
using 10 g/cm3 and m 100m p . The Fermi energy can be written as
Ef =

3
8

2/3

h2
2me

47
m

2/3

75 eV

where we used m 100m p and the 47 comes from the fact that there are 47 electrons in a silver atom.
We can see that
Ecoul E f > Eth
for room temperature silver.
Problem # 2 Deuterium Fusion in Contracting Protostars
Small amounts of Deuterium are made in the Big Bang. D is destroyed in the interiors of stars via
the reaction p + D 3 He + . The S value for D-burning is2.5 104 keV-barn = 4 1037 erg cm2 ,
each reaction releases 5.5 MeV, and the cosmic abundance of D from the Big Bang is nD 2 105 nH .
Lets focus on a low mass fully convective star undergoing KH contraction; such a star can be reasonably
well modeled as an n = 3/2 polytrope. Assume that the star has cosmic composition ( 0.6). Note that
in this problem, you should not use the approximation T . Instead, you will need to keep the full
expression for .
a) What is the Gamow energy for D fusion? Write down the resulting thermally averaged cross-section
hvi for D fusion.
The Gamow energy can be written as
EG = Z12 Z22

mr
MeV
mp

using Z1 = Z2 = 1 and mr = 32 m p we find the Gamow energy to be


EG .67 MeV
The thermally averaged cross-section is given as
1/6

2.6S(E)EG 3(EG /4kT )1/3


hvi =
e
k2/3 T 2/3
using all the constants given and the Gamow energy we find
hvi =

3.7 1015 cm3 3742(K/T )1/3


e
K2/3 T 2/3 s

hvi =

3.7 1015 cm3 3742(K/T )1/3


e
K2/3 T 2/3 s

in terms of M and R we find

b) In class we derived a quantitative model for the Kelvin-Helmholtz contraction of a low mass star as
it approaches the main sequence. Use these results to calculate the local contraction time tc R/|dR/dt|
34

as a function of the mass and radius of the star. This is the amount of time that a star of a given mass M
spends at a given radius R. Does the contraction time get shorter or longer as the star contracts?
From lecture we derived the following relationship

3 GM 2 dR
= 0.2Lsun (M/Msun )4/7 (R/Rsun )2
L=
7 R2 dt

thus we find

 
 
 

Msun 2
R 2
M 4/7
R 2
M
Rsun
Msun
Rsun

10/7 
4
Msun
R
= 3.34 105cm/s
M
Rsun

dR
7 0.2 R2sun Lsun
=
dt
3 GM 2sun

and we can find the local contraction time to be


tc = Rsun

R
Rsun





1
M 10/7 Rsun 3
7
= 6.7 10 yr
(dR/dt)
Msun
R

As the star contracts the contraction time gets longer.


c) What is the lifetime tD of a D nucleus at the center of the star in terms of the local density and
temperature (the lifetime is the average time before a D nucleus is destroyed by fusion into 3 He)? Use the
properties of n = 3/2 polytropes to write tD as a function of M and R. Does the D lifetime get shorter or
longer as the star contracts?
We know that average lifetime of a deuteron is given by
tD =

l
1
=
v n p v

which gives us
htD i =
but we also know
c =
thus we find
htDi =
we also know

m p
1
=
n p hvi c hvi

3M
M
an = 1.43 3
3
4R
R

m p R3
m p
1 R3
=
= 7.019 1025 g
c hvi 1.43hviM
hvi M

dn GM
Tc =
Rkb

3an
4

1/3

m p = 2.60 1016

cm M
K
g R

thus we find
htD i = 7.019 1025 g

3
1/3 R
1 R3
gs
= 1.89 1010 3 T 2/3 e3742(K/T )
hvi M
cm
M

35

written in terms of M and R we find




R
htD i = 1.23 10 s
Rsun
6

7/3 

Msun
M

1/3

1/3

e19.19((Msun /M)(R/Rsun ))

the deuteron lifetime gets shorter as the star contracts.


d)For any mass M show that there is a critical radius RD at which tD = tc . This represents the radius
(time) at which D starts to undergo significant fusion. Give the numerical value of RD for M = 0.03 and 0.1
Msun . For each of these two cases, also determine the central temperature of the star Tc and the D lifetime
tD when R = RD . Does D fusion occur before or after the star reaches the main sequence?
We know that
tD = tc
which yields


R
1.23 10 s
Rsun
6

7/3 

Msun
M

1/3

19.19((Msun /M)(R/Rsun ))1/3

M
= 6.7 10 yr
Msun
7

10/7 

Rsun
R

3

and so we find


RD
Rsun

16/3

= 1.71 10

21

M
Msun

37/21

1/3

e19.19((Msun /M)(R/Rsun ))

we can solve this numerically to find


RD = 0.44Rsun
RD = 1.11R

M = 0.03Msun
M = 0.1Msun

to find the central temperature we can use




M
Tc = 7.4 10 K
Msun
6



Rsun
RD

we find
Tc 5.0 105 K M = 0.03Msun
Tc 6.67 105K M = 0.1Msun
and to solve for the deuteron lifetime we find

 

RD 7/3 Msun 1/3 19.19((Msun /M)(R/Rsun ))1/3
6
htD i = 1.23 10 s
e
Rsun
M
so we find
tD 4.67 106yr M = 0.03Msun
tD 4.1 105yr M = 0.1Msun
e) Can D fusion halt (at least temporarily) the KH contraction of the star? Explain your answer quantitatively.
36

Since we know that


=

Qrd
d

d = md n d

we also know that


L = M =

rd =

nd
tD

MQ
MQrd
=
d
2m ptD

since we know that Q = 8.8 106ergs, we find


M(8.8 106ergs)
L=
2m ptD
so for M = .03Msun we find

L 5.37 1035 ergs/s

and for M = 0.1Msun we find

L 4.08 1037 ergs/s

thus we can see that for both of these stars deuteron fusion can stop the KH contraction temporarily.
Problem # 3 The R(M) Relation for Degenerate Objects
Consider an object supported entirely by the pressure of non-relativistic degenerate electrons. Because
P = K5/3 such an object can be modeled (rigorously) as an n = 3/2 polytrope.K is a constant that depends
on the electron mean molecular weight e .
a) Use your results for how the central pressure Pc and density c of an n = 3/2 polytrope depends
on the radius R and mass M of the object to derive the R(M) relation for degenerate objects (the radius
also depends on e ). Note that you should give an expression with proper constants and not just a scaling
relationship. Normalize the mass M to Msun and the radius R to Rsun (this should sound pretty familiar by
now).
We know that
h2
Pdeg = Pc =
5me
rearranging this equation for c yields
1/3
c

h2
=
5me

3
8

2/3 

3
8

c
e m p

2/3 

5/3

1
m p

4/3

= dn GM 2/3 c

5/3

1
dn GM 2/3

from the last problem set we showed




M
c = 8.41
Msun



Rsun
R

3

using this we find




R
M 1/3
5/3
= 0.04e
Rsun
Msun
b) Use a) to estimate the radius of Jupiter. How does your result compare to the correct value?
Using part a) with e 1.17 which is the value given for the sun on Google and M = MJ , we find
R 0.30Rsun
37

c) The results you have derived in a) should show that as M 0, R . This is not correct, however,
because Coulomb interactions become important in the equation of state of low-mass objects (brown
dwarfs and planets). Estimate the density at which the Coulomb energy per particle becomes comparable
to the Fermi energy. What mass and radius does this correspond to? Explain why this is a very rough
estimate of the maximum radius of a degenerate object.
If we know
E f = Ecoul
then
1 e2
=
40 r
but we know that

3
8

2/3

h2 2/3
n
2me

1
n1/3
r

thus
1 2 1/3
e n =
40

2/3

3
8

3
8

2

h2 2/3
n
2me

so we find the density to be given by


n=

2me 2
e
40 h2

3 

6.15 1028 m3

to find the mass we can use


c = nm p = 1.43

M
R3

and using R from part b) we find


1.43

Msun (M/Msun )
= nm p
(0.03Rsun)3 (M/Msun )1

which can be simplified to


M
=
Msun

nm p (.03Rsun)3
1.43Msun

1/2

which can also be expressed as


M 190Mearth
using this we can now find the radius to be
R 0.361Rsun

Problem set 9

38

5.7 104

Problem # 1
Use the chemical potential for a non-degenerate, non-relativistic gas (derived in class; also 2.21
in Phillips) to show that in the limit n nQ (the non-degenerate limit), the full quantum mechanical
distribution function reduces to the classical Maxwell-Boltzmann distribution function. A good check that
you have things correct is that the QM dist. fcn you start with has some h s in it (Plancks constant), but
the classical dist. fcn you end up with should, of course, be independent of h.
We know that the chemical potential is defined as


n
= mc + kT ln
gnQ
2

(2)

and that the quantum distribution function is defined as


n(p) =

g/h3
e(E p )/kT 1

to show that in the classical regime


e(E p )/kT 1
We can write Equation 1 as
nQ 1 (mc2 )/kT
= e
n
g

2 )/kT

e(mc

n nQ

thus the quantum distribution function can be written as


n(p) =

g/h3
e(E p )/kT

We also know that


p2
E p = mc +
2m
2

nQ =

2mkT
h2

3/2

and using equation 1 we find


n(p) =

g/h3
(mc2 +p2 /2mmc2 kT

ln(nQ /n))/kT

gn
h3 eE/kT n

which after some simplification reduces to the Classical Boltzman distribution function


1
n(p) = n
2mkT

3/2

eE/kT

Problem # 2
Consider a cloud of gas that has a total mass M . Assume that all of the gas in the cloud is converted
into stars with the initial mass function given in class dN/dm m where = 2.35 and where this
formula is valid between m = 0.5Msun and m = 150Msun . Note that dN/dm has units of number of stars
per unit mass.

39

a) What is the ratio of the number of stars formed with masses within dm m1 of m1 and masses
within dm m2 of m2 ? What is the ratio of the number of 150Msun stars formed to the number of 0.5Msun
stars formed?
We know that the Initial Mass Function IMF is given as
dN
m m2.35
dm

= 2.35

we also know that for dm m1 of m1 we find


1
dN(m1 ) = m
1 dm1 m1

we also know that for dm m2 of m2 we find


1
dN(m2 ) = m
2 dm2 m2

and the fraction is given by


dN(m1 )
=
dN(m2 )

m1
m2

1

2208 m1 = 0.5Msun m2 = 150Msun

b) Estimate the mass of a cloud M so that approximately one 150Msun star forms in the cloud. If the
temperature of the cloud at the time of formation was 10 K, what was the density of the gas out of which
the cloud formed?
From part a) we found that
Ns 2208Nb
where Nb is for stars that are the number of 150Msun and Ns is for stars that have 0.5Msun and to
get the total mass we must multiply the total number of small stars to the average mass of the stars, and
from lecture we are told that
hM i 0.5Msun
Mcluster = Ns hM i 1104Msun
to find the density of the gas in which this cloud formed we can use the Jeans density


3kT 3
3
3.14 1022 kg m3
J =
4M 2 2Gm
Problem # 3
A stellar atmosphere consists almost entirely of hydrogen. Assume that 50 % of the hydrogen molecules
are dissociated into atoms and that the pressure is 100 Pa. Given that the binding energy of the hydrogen
molecule is 4.48 eV, estimate the temperature. Set all degeneracies to 1. As the hint at the back of the book
suggests, you should derive the Saha equation for the dissociation of H2 into hydrogen, i.e., the reaction
+ H2 H + H .
We know that

+ H2 H + H

and the Saha equation gives


(H2 ) 2(H)
40

but we know that the chemical potential is given by






nQ,H gH
nQ,K2 gH2
2
2
(H) = mH c kT ln
(H2) = mH2 c kT ln
nH
nH2
where

mH c2 = m p c2 + me c2 H

thus we find

mH2 c2 = 2m p c2 + 2me c2 H2 2H

nQ,K2 gH2
mH2 c kT ln
nH2
2

which becomes

H
2 = ln
kT

"




nQ,H gH
2
= 2 mH c kT ln
nH

nQ,H2 gH2
nH2



nH
nQ,H gH

2 #

we are given that


gH2 = gH = 1

nQ,H2 nQ,H

thus we find
nH H /kT
nH2
=
e 2
= nH
nH
nQ

h2
2mkT

2mkT
h2

3/2

3/2

eH2 /kT

but we are given that


nH =

P
3kT

1
nH2
=
nH
2

h2
2mkT

3/2

thus
P
1
=
2 3kT

eH2 /kT

this can only be solved analyticaly, we find that the temperature is given by
T 2260 K
Problem # 4 Lines from Hydrogen
Consider a pure hydrogen gas. In this problem we will calculate the fraction of H atoms that have
an electron in the n = 2 state (a result I plotted in class), and use that to understand some aspects of the
observed lines of H from stars. Recall that the energy levels of the H atom are given by E = 13.6/n2 eV
and the degeneracies are gn = 2n2 .
a) Use the Saha equation to solve for the fraction of hydrogen atoms that are ionized as a function of
temperature T . If n is the total number density of hydrogen atoms (both neutral and ionized) then what
we are after is n p /n since an ionized hydrogen atom is just a proton. Your result for n p /n will depend
on n (because, as discussed in class, the ionization of a gas depends weakly on density in addition to the
primary dependence on temperature). For densities appropriate to the photosphere of the sun, make a plot
of n p /n as a function of temperature T . If you are familiar with graphing using IDL, Mathematica, etc.
feel free to use that. Otherwise, you can just plug values into your calculator and make the plot by hand. In
your calculation, assume that all of the neutral hydrogen atoms are in the n = 1 (ground) state. The reason
this is an ok approximation is as follows. According to the reasoning in class, which you will confirm
here, Hydrogen is 1/2 ionized at T 1.5 104 K. At that temperature, nearly all of the neutral H atoms
41

are in the ground state (check it if you dont believe me!), so for temperatures at which H is largely neutral
(T 1.5 104 K), it is reasonable to say that almost everything is in the ground state.
We know from the Saha equation
ne n p ge g p
=
nH
gH
we know

2mkT
h2

3/2

e/kT

ne n p n = nH + n p 13.6 eV

thus

and so

n2p
=
nH

2mkT
h2

3/2

ge g p
=1
gH

e/kT

n2p = nH = (n n p)

which becomes a quadratic equation of the form


n2p + n p n = 0
with the solution of n p being

2 + 4n
np =
2
and since we know that this must be a positive thus we will take the positive solution

+ 2 + 4n
np =
2
and finally we are looking for

n p + 2 + 4n
=
n
2n

the plot is given by

42

np vs ntotal

0.8

np/n

0.6

0.4

0.2

0.0
5.0103

1.0104

1.5104
Temperature

2.0104

2.5104

3.0104

we can see that at a temperature of T 1.4 1.5 107K roughly 50% of the hydrogen atoms are
ionized.
b) Use your result from a) to calculate the fraction of all H atoms that have an electron in the n = 2
state of hydrogen. If n2 is the number density of atoms with electrons in the n = 2 state, then what we
are after here is n2 /n. You will need to use the Boltzmann factor in addition to your result from the Saha
equation in a). For densities appropriate to the photosphere of the sun, make a plot of n2 /n as a function
of temperature T . If you are familiar with graphing using IDL, Mathematica, etc. feel free to use that.
Otherwise, you can just plug values into your calculator and make the plot by hand.
We know that

n np
np
nH
=
= 1
n
n
n
and from the Boltzman equation we know that
n2 g2 (E2 E1 )/kT
= e
n1 g1
we also know
nH = n1 + n2

and

what we are looking for is

n2 = nH n1

nH
n1
= 1+
n2
n2



n1 1
n2
= 1+
nH
n2


np 
n2
n2 nH
n1 1 
1
=
= 1+
n
nH n
n2
n
43

thus we find the fraction of all H atoms that have an electron in the n = 2 state of hydrogen given by
!





np 
n2
+ 2 + 4n
n1 1 
g1 (E2 E1 )/kT 1
1
1
= 1+ e
= 1+
n
n2
n
g2
2n
where has been explicitily defined already. The plot is given by
n2 vs ntotal

4105

np/n

3105

2105

1105

0
1.0104

1.5104

2.0104
2.5104
Temperature

3.0104

3.5104

4.0104

We can see that the fraction of hydrogen atomes in the energy state n = 2 peaks at 1.5 104 K.

c) The Balmer lines of hydrogen are produced by transitions between the n = 2 states of Hydrogen and
the n = 3, 4, .... states. What are the wavelengths of the H(n = 2 3)and H(n = 2 4) lines of H?
Use your result from b) to explain why A stars show the most prominent H lines of hydrogen (relative to
more massive stars such as O stars and less massive stars such as M stars).
We know that
E = h =
thus
=
and

hc

hc
hc
=
E (E2 E1 )

E=

13.6 ev
n2

thus for the n = 2 n = 3 transition we get


=

hc
656.3 nm
(3.4 1.51)eV
44

and for the n = 2 n = 4 transition we get


=

hc
486.7 nm
(3.4 0.85)eV

From the plot given in part b) we can see that the fractional number of atoms in the n = 2 energy state
peaks at around 1.5 104K, which is approximately the surface temperature of A stars, we can also see
that for O type stars that have surface temperatures much greater than 15,000 K that there are 0% of
hydrogen atoms in the n = 2 energy state, most of the atoms are already ionized. The situation is similar
for M stars that have surface temperatures that are much lower than 15,000 K. We can see that at these
temperatures there are approximately 0 atoms with electrons in the n = 2 energy state.
d) The Lyman lines of hydrogen are produced by transitions between the n = 1 states of Hydrogen and
the n = 2, 3, 4, .... states. What is the wavelength of the Ly (n = 1 2) line of H? Roughly what fraction
of H atoms have electrons in the ground (n = 1) state of H in the atmosphere of an M-star? Would you
expect to see prominent Ly lines from an M-star? Why or why not?
Using
=
we find
=

hc
hc
=
E (E2 E1 )

hc
hc
=
121.6 nm
E (13.6 3.4)eV

We would not expect to see any Ly lines from M stars, even though all of the hydrogen atoms are in
the ground state, there is not enough thermal energy to excite the electrons from n = 1 to n = 2.

Problem set 10

Problem # 1
Consider a gas with total mass density and temperature T . Recall that the mean molecular weight
is defined by P kT /m p where P is the total ideal gas pressure (ions and electrons), while the electron
mean molecular weight e is defined by ne /e m p .
Since we know that the total pressure is given by
kT
PT = PI + Pe =
mp
and thus

but we know that

1
1
+
I e

1
1
1
= +
I e
X
1
=
I
A

XZ
1
=
e
A
45

kT
m p

where X is the mass fraction of the species, Z is the number of electrons, and A is the atomic number
of the species, thus we find
A
1 A
e =
(3)
=
X 1+Z
XZ
a) What are the values of and e if the gas consists of
i) ionized H,
Since we know that
X =1 A=1 Z=1
then
=

1
e = 1
2

ii) 75 % (by mass) ionized H and 25 % (by mass) ionized He,


We have to treat this case seperately since we have two species contributing to the mean molecular
weight
X = 0.75 Y = 0.25 A = 4 Z = 2
thus
1
X Y
1
13
= + = (X +Y ) =
I
A A A
16

Z
7
1
= (X +Y ) =
e A
8

and so we find
=

8
16
e =
27
7

iii) ionized He,


Since this is a pure fully ionized gas we can use Equation 1 with
X =1 A=4 Z=2
we find
=

4
e = 2
3

iv) ionized O,
Using Equation 1 with
X = 1 A = 16 Z = 8
we find
=

16
9

e = 2

v) ionized Fe
Using Equation 1 with
X = 1 A = 56 Z = 26
we find
=

56
27

e =
46

28
13

b) Which gas has the largest ideal gas pressure? Which gas has the largest electron degeneracy pressure? Assume that and T are the same in all cases.
Since we know that gas pressure goes as
1

than the smallest will give us the highest pressure, thus the element that has the highest gas pressure
Pg

is
=

1
2

Hydrogen gas

for degeneracy pressure we know


5/3
Pd ne

1
e

5/3

thus the gas that gives the highest degeneracy pressure is the one with the lowest value for e and this
is
e = 1

Hydrogen gas

Problem # 2 The Helium Main Sequence


In certain stages of stellar evolution, stars are largely composed of He and He fusion dominates the
stellar luminosity. One can approximate such stars as lying on a He main sequence. In this problem
we will calculate the properties of the He main sequence assuming that a star is composed of pure He,
that energy transport is via radiation, that electron scattering dominates the opacity, and that gas pressure
dominates. The energy generation rate for He fusing to Carbon is
= 5 1011 2 T83 exp(44/T8 ) ergs s1 g1
and 7.65 MeV is released converting 3 He nuclei into 1 C nucleus. Note that throughout this problem
you should not just give scaling (proportionality) laws for the desired relations; you should also determine
reasonable normalizations.
a) Calculate the relationship between mass M and luminosity L for the He main sequence.
For a star that has the given properties: energy transport is via radiation, electron scattering dominates
the opacity ( = T ), and that gas pressure dominates (P Pg ) we find the luminosity given as
L M 3 4 e
This equation gives the evolution of the lumunosity on the MS as chemical composition changes. We
can scale this to the sun to find
 
 


4
e
M 3
L = Lsun
Msun
sun
esun
we also know that

4
e = 2 sun = 0.6 esun 1.14
3
and so we find that the lumonosity for the Helium main sequence can be expressed as
He =

L 42.7Lsun
47

M
Msun

3

b) Estimate
the core temperature of a 1 solar mass He star. You do not need to do the full integral
R
L f usion = dM , but can approximate this as L f usion 0.1M(r = 0).
In steady state we can express the lumunosity of energy transport be equal to the luminosity due to
fusion
L f usion Ltransport
where we can use part a) and the approximation given to find
42.7Lsun

M
Msun

3

= 0.1M = 0.1M5 1011 2 T83 exp(44/T8 ) ergs s1 g1

if we let

M = Msun 150 g/cm3

we find
T83 e44/T8 = 1.36 1013
solving this numerically yields
Tc 1.5 108 K

T8 1.52

c) Given your result for Tc for a 1 Msun star from b), calculate the power-law relation Tc (M) by imposing
the steady state requirement that L f usion = L photons and using T (where L photons is the energy carried
out of the star by photons from a).
Since we know that

L photons M 3

we find

L f usion M

M 2 T

where in part b) we are given


=2

= 3 +

44
26
T8

knowing that the density and temperature scale as

M
R3

M
R

gives
M
2

M
R3

2 

M
R

26

R M 13/16

using this along with our expression for the temperature gives
T

M
M
13/16 M 3/16
R M

to scale to a one solar mass star, from part b) we find




M
T 1.5 10 K
Msun
8

48

3/16

d) Use your results above to determine the R(M)and Te f f (L)relations for the He main sequence. Then
sketch the relative positions of the H & He main sequences in the HR diagram.
From the Virial temperature we know
Tc =

Gm p M
k R

where k is Boltzmans constant, this expression can be scaled to the sun





Gm p Msun
M
Rsun
Tc =
k Rsun Msun
R
using the result from c) and plugging in all the constants along with = 4/3 gives


M
1.5 10 K
Msun
8

3/16

M
= 3.08 10 K
Msun
7



Rsun
R

and we find


R
Rsun

M
= 0.21
Msun

13/16

To find the relationship between the luminosity and the effective temperature we can use
L = 4R2 Te4f f
but we know from our previous expression
2

R = (0.21Rsun)

M
Msun

13/8

plugging this in our expression of the luminosity gives


L = 4(0.21Rsun)

M
Msun

13/8

Te4f f

now we can use the result from a)




M
Msun

L
42.6Lsun

1/3

using this we get


Lsun

L
Lsun

= 4(0.21Rsun)

L
42.6Lsun

13/24

thus we find that


Te f f

L
2.1 10 K
Lsun
4

The HR diagram for this star is given by

49

11/96

Te4f f

which seems rather odd, We would expect this main sequence He buring star to be above the main
sequence line. This can be explained by our initial assumption that went into deriving this relationship.
We assumed that this was a pure ball of He gas.
e) At what mass does the luminosity of the star exceed the Eddington luminosity?
We know that the Eddington luminosity is given by
LEdd

4cGM
=

L f usion = 42.7Lsun

M
Msun

3

setting these two expressions equal to each other gives


4cGMsun

given the opacity defined as


=
thus
=

1
T
e m p

M
Msun

= 42.7Lsun

ne T

e = 2

M
Msun

3

ne =

e m p

T
0.2 g/cm
2m p

M 38.95Msun
is the mass that will exceed the Eddington luminosity
f ) What is the He main sequence lifetime as a function of stellar mass? Compare this to the corresponding H burning lifetime.
50

We know that the main sequence lifetime is given by


t=

E
L

E = NQ =

0.1M
0.1Msun M
(7.65MeV) =
(7.65MeV )
12m p
12m p Msun

thus
0.1Msun
t=
12m p (42.7Lsun )

Msun
M

2

Msun
(7.65MeV) 2.34 10 yrs
M
7

2

thus the main sequence lifetime is




Msun
t 2.34 10 yrs
M
7

2

this is much much shorter than the H burning lifetime which is t 1010 yrs for M 1Msun . Also can
be written as
tHe 0.2% the time of the main sequence Hydrogen burning
Problem # 3 The Thin Shell Instability
As we discussed in lecture, during several phases of stellar evolution, fusion takes place in a thin shell.
Consider such a shell located a distance Rs from the center of a star. The mass interior to Rs is M , the
mass of the shell itself is Mshell and the thickness of the shell is H dR R, where H is the scale-height
at radius Rs (recall that H is the distance over which the density, pressure, temperature, etc. change).
a) Use hydrostatic equilibrium to show that the pressure at the base of the shell is given by
P(Rs)

GMMshell
4R4s

HE gives
dP
GM
= 2
dr
R
which can be written in differential form
P(Rs + dR) P(Rs)
GM
= 2
dR
Rs
but we know that
s =

Ms
Ms

Vs
4R2s dR

and since we know that H dR R (this comes from the definition of H) then P(Rs + dR) P(Rs)
than
GMMs
P(Rs)
=
dR
4R4s dR
which simplifies to
P(Rs) =

GMMs
4R4s

b) Use your result in a), together with the strong temperature dependence of fusion reactions, to explain
why fusion in a thin shell is unstable and will runaway, as in a bomb. Hint: How will P, , T, and dR of the
shell change if there is a small perturbation that increases the amount of fusion in the shell?
51

Since we know that

M
dR3
If we apply a small pertubation that increases fusion then we know T , , and dR and since the
energy generation has such a high power temperature dependence we know that in order for this to be stable
than the density must decrease to compensate (assume constant pressure). But we can see that the density
dependence is a function of the radius and cannot decrease by 20 orders of magnitude to compensate, and
thus this becomes a runaway reaction, i.e like a bomb.
L M2 T 44/T8

This unstable fusion occurs primarily when stars are on the asymptotic giant branch (fusion of He in
a thin shell outside a C/O core) and may be part of the reason that such stars lose so much mass on their
way to becoming white dwarfs.

Problem set 11

Problem # 1
Consider a 0.5 Msun WD. Approximate it as an n = 3/2 polytrope, reasonably appropriate since we
are below the Chandrasekhar mass. Estimate the ratio of the energy transported by photons (radiative
diffusion) to the energy transported by degenerate electrons (thermal conduction) at the center of the WD.
Scale the central temperature of the WD to 108 K, an appropriate number for a newly formed WD. Assume
that the opacity is due to electron scattering. Show that the energy transported by electron conduction
dominates that transported by photons.
We know that the radiative flux for photons is given by


4 acT 3
T = r T
Fr =
3 0
where o is the opacity and r is the conductivity (any process that transports energy). We also know that
flux due to thermal conduction of degenerate electrons is given by
Fdeg = deg T
where deg is the conductivity due to degenerate electrons, thus the ratio of the energy transported by
photons (radiative diffusion) to the energy transported by degenerate electrons (thermal conduction) is
given by
Fr
r
=
Fdeg deg
we have defined the degenerate electron conductivity to be
deg cls

EF
kT

3/2

k 2 h3 T ni
k 2 h3 T c
=
32e4 m2e
32e2 m2e i m p
52

ni =

c
i m p

and the radiative diffusion conductivity to be


4 acT 3 4 acT 3 4 acT 3 e m p
=
=
3 o c
3 neT
3 c T

r =

We also know that c (n = 3/2) polytrope equation is given by


c =

M
4M
an 1.43
3
3R
R

an = 5.99

where the radius is given by


R 0.013Rsun

M
Msun

1/3 

e 5/3
2

m
me

1

0.016Rsun

if we use M = 0.5Msun and e = 2. The ratio can be written as


128acT 2 e4 m2e m2p e i
r
=
deg
3k2 h3 T 2c
we also know
1
1
1
= +
i e

n
1
Xi
=
i i=0 A

(4)

n
1
Xi Z
=
e i=0 A

where we find

96
e = 2
7
assuming 50% C and 50 % O. If we use a temperature of T 108 K on Equation 1 we find
i =

r
0.084 deg 11.9r
deg
Thus we can see that the energy transported by electron conduction dominates this process.
Problem # 2
Assume that stars are formed with the Salpeter initial mass function (dN/dM M 2.35 ) between 0.5
and 150 Msun , that stars with Mi < 8Msun become 0.5Msun WDs, that stars with 30Msun > Mi > 8Msun
become 1.4 Msun NSs, and that stars with Mi > 30Msun become 7Msun BHs (the typical WD, NS, and BH
masses chosen here are well-motivated observationally). Assume further that all NSs and BHs are formed
via SN explosions.
a) What fraction of stars undergo SN explosions at the end of their lives?
We know that the Salpeter initial mass function is given as
dN =

M 2.35 dM

Thus the fraction of the stars that undergo SN explosions would be given by the sum of the fractions
of the stars that become NS and BH, this is given by
R 150

SN explosions = R8150
0.5

M 2.35 dM
M 2.35 dM

53

= 2.3%

b) What fraction of stars will become WDs? NSs? BHs?


The fraction of stars that become WD is given by
R8

M 2.35 dM
WD stars = R 0.5
= 97.7%
150 2.35
dM
0.5 M
The fraction of NS is given by
R 30

2.35 dM
8 M
NS stars = R 150
= 1.97%
2.35 dM
M
0.5

and the fraction of BH is given by


R 150

BH = R30
150
0.5

M 2.35 dM
M 2.35 dM

= 0.35%

c) Estimate the fraction of the mass of a stellar population that is returned to the interstellar medium
(via stellar winds or explosions) after 10 Gyrs. You do not need to do a rigorous, accurate to many
significant digits, calculation.
We know that the total mass of a specific type of star can be calculated using
M = dNM dN = AM 2.35 dM
where A is a normalization constant
MT = A

Z b
a

M 1.35 dM

we also know that the total mass that is redistributed to the ISM can only come from the fraction of stars
that have M Msun and for WD this means that the total mass that is redistributed to the ISM (mass loss)
is given by
Mloss = A
this becomes

1
Mloss = A
0.35

Z b

1
a0.35

M 2.35 (M Mend )dM

1
b0.35

Mend

1.35

1
a1.35

1
b1.35



(5)

we chose integration limits motivated by the knowledge that the sun M = Msun has an approximate lifetime
of 10 Gyr, thus stars wiuth this mass range are the only ones contributing to this enrichment of the ISM.
The mass fraction is given by
Mloss
M f rac =
Mtot
where


Z b
A
1
1
1.35
Mtot = A M
dM =

(6)
.35 a.35 b.35
a
Using
a = 1Msun b = 8Msun Mend = 0.5Msun
on Equation 2 and
a = 0.5Msun b = 8Msun
54

on Equation 3 gives
M f rac (W D) 48%
This tells us that approximately 48 % of the mass from these stars is given back to the ISM, since we know
that this population comprises 97 % then the total contribution to the ISM will be
f rac(W D) = 0.48 97.7 47%
For NS we can say
a = 8 b = 30 Mend = 1.4Msun
Using this in equations 2 and 3 we find
M f rac (NS) = 65%
this tells us that approximately 65 % of this mass is given to the ISM and since we know that this population
comprises 1.97 % then the total mass fraction contribution to the ISM will be
f rac(NS) = .65 .0197 1.2%
and finally For BH we can sa
a = 30 b = 150 Mend = 7Msun
Using this in equations 2 and 3 we find
M f rac (BH) = 87.5%
this tells us that approximately 87.5 % of this mass is given to the ISM and since we know that this
population comprises 0.3 % then the total mass fraction contribution to the ISM will be
f rac(NS) = .875 .003 .2%
Thus we can conclude that the total fraction of the mass that is redistributed to the ISM is given by
f racT 48.4%
Problem # 3
Consider a white dwarf with a mass of M = 0.5Msun and an effective temperature of 104 K.
a) Estimate the radius, luminosity, central temperature, and age of the WD. You are free to use any of
the results on WD cooling quoted in lecture.
We know that the radius of a WD can be estimated using

 

M 1/3  e 5/3 m 1
0.016Rsun 1.11 109 cm
R 0.013Rsun
Msun
2
me
The luminosity can be calculated using
L = 4R2 Te4f f = 5.83 1030erg s1 = 2.3 103Lsun
The central temperature is given by


L Msun
Tc = 10 K
5Lsun M
8

2/7

55

= 1.35 107 K

and finally the age can be calculated using




L Msun
t = 10 yrs
5Lsun M
6

5/7

= 1.49 108yrs

b) Estimate the thickness of the photosphere of the WD. What is the number density in the photosphere? Assume for simplicity that the opacity in the photosphere is approximately equal to the electron
scattering opacity.
We know that the thickness of the photosphere is given by the scale height
R2 kT
R2 kT
=
1.5 104cm 1 105 RW D
h=
mGM

H m p GM
we used H 1 since we assume that the atmosphere of the WD is comprised primarily of hydrogen. We
know that the mean free path is equal to the scale height in the photosphere, thus
l=
thus
n=

1
=h
n

1
9.75 1019 cm3
hT

c) Use the Saha equation for the ionization of hydrogen to estimate the temperature at which hydrogen
is 1/2 ionized at the surface of a WD. Is this larger or smaller than the temperature at which hydrogen is
1/2 ionized for photospheric densities appropriate to MS stars?
We know that the Saha equation can be written as
n p ne g p ge
=
nQ,ee/kTe f f
nH
gH
we know that in a gas of 1/2 ionized hydrogen ne = n p = nH and also g p = 1 ge = 2 gH = 2 and =
13.6 eV. The quantum density is given as
nQ,e =

3/2

2me kT
h2

thus we can write the Saha equation as


nH =

2me k
h2

3/2

3/2

Te f f e13.6/kTe f f

we know that for every hydrogen atom there are both a proton and electron, thus the hydrogen density
accounts for 1/3 the total density
n
nH = =
3

2me k
h2

3/2

3/2

Te f f e13.6/kTe f f

Plugging in all the constants yield


5 /T
ef f

3/2

7.43 105Te f f e1.5710


56

1 = 0

solving this numerically yieds a temperature of


Te f f 2.6 104K
which is a higher than the 1/2 ionization temperature for the solar photosphere T 1.3 104K. This is
due to the differences in the densities.
Problem # 4
Consider the late stages of evolution of a 25Msun star. Focus on the core which has a mass 1Msun
and a radius 108 cm. The stars photon luminosity is 3 105 Lsun . You might find some of the numbers
in Table 4.2 of Phillips useful.
a) Estimate the temperature at which cooling by neutrinos (which are optically thin and leave the core
directly) exceeds cooling by photons (which random walk out of the star). Use the expression for the
neutrino luminosity from class. At what stage of nuclear fusion (H, He, C, O, Ne, Si, ....) does neutrino
cooling become dominant?
We know that the luminosity given by the neutrinos is given by
L 1012 T93

Rc
2
10 Rsun

3

e11.9/T9 Lsun

and the luminosity of the photons is given by


L ph = 3 105 Lsun
we can find the temperatures at which this are approximately equal by setting these expressions equal to
each other and solving it numerically, i.e
L = L ph

1012 T93

Rc
2
10 Rsun

3

e11.9/T9 Lsun = 3 105 Lsun

this simplifies to
3.3 106 T93 e11.9/T9 1 = 0
solving this numerically yields
T9 .794 T 7.93 108K
Neutrino cooling becomes important after Helium fusion and before Carbon fusion due to the temperatures
given in table 4.2 of Phillips for these reactions.
b) If neutrino cooling were unimportant (and thus the photon luminosity determined the energy lost by
the star), estimate the time it would take the 1Msun core of the star to fuse from 20 Ne to 56 Fe. Compare
this to the true time of about 1.5 years (from Phillips Table) set by neutrino cooling. Assume that the
luminosity of the star is independent of time and that fusion of heavy elements releases 0.7 MeV per
nucleon ( 10 times less the fusion of H to He because the binding energies of heavy nuclei are closer to
each other.
We know that the time is given by
t=

NQ
E
=
L ph L ph

57

we know that the number of particles are given by


N=

Msun
mp

and the energy released is


Q 0.7 MeV/nucleon
thus we find that the time is
t=

Msun Q
3.6 104 yrs
m p L ph

We can see that neutrino cooling is very effective in killing a star. Due to the fact that the star would
exist for much longer if it were not for those pesky neutrinos.

Problem set 12

Problem # 1
The Energy needed to dissociate one 4 He nucleus into two neutrons and two protons is Q = 28.3 MeV.
Derive an expression relating the numbers of 4 He nuclei, neutrons and protons coexisting at a temperature
T in an equilibrium set up by the reactions
+4 He 2n + 2p
Calculate the temperature for 50 % dissociation when the density is 1012 kg m3 . [Note :This is a simple
example of nuclear statistical equilibrium (NSE) discussed in class, i.e., the balance of nuclei determined
when nuclear reactions go both ways at high temperatures (because photons have enough energy to photodisintegrate nuclei into their more basic constituents).] In addition to calculating the temperature for 50
% dissociation of He, also show explicitly that at high temperatures, NSE favors the nuclei being broken
apart (n and p in this case) while at low temperatures it favors nuclei being bound (He in this case).
We can use the Saha equation, which is given by
() + (4He) = (2n) + (2p)
which can also be written as






g p nQ,p 2
gn nQ,n 2
gHe nQ,He
2
2
2
= 2mn c kT ln
+ 2m p c kT ln
mHe c kT ln
nHe
nn
np

58

rearranging this equation yields


c2 (mHe 2mn 2m p ) = kT ln

gHe nQ,He
nHe

n2n n2p
(gn nQ,n )2 (g p nQ,p )2

!!

we are given the binding energy needed to dissociate a helium atom


c2 (mHe 2mn 2m p ) = 28.3 MeV
thus

n2n n2p
=
nHe

gHe
(gn g p )2

nQ,He
(n2Q,n n2Q,p

e28.4MeV/kT

we are also given


gHe = 1 gn = g p = 2
and the quantum concentration is defined as
nQ,A =

2mA kT
h2

3/2

and if we assume for simplicity


nQ,n = nQ,p =
it follows that

2m p kT
h2

3/2

nQ,He =

8m pkT
h2

3/2

n4Q,p 28.4MeV/kT
n2n n2p
=
e
nHe
nQ,He

inserting the expressions for the quantum density gives




n2n n2p
2m p kT 9/2 28.4MeV/kT
e
= 16(8)
nHe
h2

(7)

if we assume that this gas is 50% dissociated gives us


1
2
nn = n p = n nHe = n
5
5
for every helium nuclei there are two neutrons and two protons, giving a total of five particles. We also
know that
mHe + 2mn + 2m p 8

5
n=
m =
mp n =
m
5
5
8 mp
Using this along with plugging in all the constants into Equation 1 gives us
9/2

7.84 1021 T12 e0.328/T12 = 1 T12 =


solving this equation numerically yields a temperature of
T12 = 0.0109 T 1.09 1010 K
59

T
1012 K

To show explicitly that at high temperatures, NSE favors the nuclei being broken apart (n and p in this
case) while at low temperatures it favors nuclei being bound (He in this case) we must consider Equation
1 along with assuming
n = nHe + nn + n p nHe + 2n p
we can see that
nHe n 2n p
putting this into Equation 1 gives


n2n n2p
2m pkT 9/2 28.4MeV/kT
e
= 16(8)
n 2n p
h2

(8)

If we consider the case where T (very high temperatures) we can see that Equation 2 goes to infinity,
this only happens if the denominator is 0
1
n 2n p = nHe = 0 n = n p
2
thus there are no bound nucleus only protons and neutrons in equal numbers. If we now consider the case
where T 0 we can see that Equation 2 goes to 0 this can only happen in
n p = nn = 0 n = nHe
and thus it favors nuclei being bound.
Problem # 2
Compare the total energy released by a 25 Msun star during (a) its pre-main sequence evolution (KH
contraction), (b) its time on the MS, (c) its post-main-sequence-evolution, and (d) the supernova explosion
to form a neutron star.
To calculate the total energy released during the pre-main sequence we can assume can just calculate
the total gravitational energy relesed from contraction in the star. We know that the energy for a bound
system is given by the Virial theorem as
U
E
2
is given by
GM 2
U
E |Ei | |EF | |EF |
2
2RMS
where RMS is given by the main sequence radius which is defined as
RMS Rsun

M
Msun

6/7

since we know that the mass of this star is


M = 25Msun
we find the radius to be
R = 15.78Rsun

60

and the gravitational energy, which is the total energy is


G(25Msun )2
7.45 1049 ergs
2 15.78Rsun

E preMS =

To calculate the total energy of a 25Msun star during the main sequence we need to multiply total
luminosity by the total time that the star spends on the main sequence
E = LMStMS
where the luminosity is given by
L = Lsun

M
Msun

3.5

= 7.8 104Lsun

and the time a star spends o the main sequence is given as


tMS = 10

10

M
Msun

2.5

yrs = 3.2 106 yrs

this gives the total energy as


EMS 3.08 1052 ergs
we can compare this to the value derived by nuclear energetics, we can say that the total energy is given
by
M
E 0.1NQHe = 0.1 QHe
mp
where we know
M = 25Msun QHe 7 MeV/nucleon
Where QHe is the nuclear binding energy per nucleon of helium. We find the energy to be given as
E 3.33 1052 ergs
which is almost them.
To find the total energy during the post main sequence stage we need to consider the energy to be given
as

E = 0.5NQ

where Q is the difference in the binding energy of iron to the binding energy of helium and we assumed
that 50% of the star will undergo fusion. The binding energy per nucleon of iron and helium are QFe
8.78 MeV/nucleon and QHe 7 MeV/nucleon, thus
N=

25Msun
mp

Q 1.78 MeV/nucleon

and we find the total energy to be given as


E postMS 4.24 1052ergs
61

To find the total energy released during the supernova explosion can be estimated using the same
equation as part a) except now the final radius is given by the radius of te neutron star.
E

GM 2
2RNS

where the mass that we will consider will be the mass of the core
M 1.4Msun

R 10 km

thus we find the energy to be


ESN 2.59 1053ergs
this energy release is much greater than all other processes.
Problem # 3
Consider an ideal degenerate gas of electrons, protons and neutrons, and the equilibrium established
by the reactions
n p + e + e and e + p n + e
Assume equal numbers of electrons and protons and assume that the density is so high that all the degenerate particles are ultra-relativistic. Show that the number densities of the particles are in the ratio
ne : n p : nn = 1 : 1 : 8
Using the Saha equation, along with the knowledge that all of the particles are now reletavistic we find
(n) + (e) = (p) + (e)
and we know that the Fermi momentum is given by
pF =

3n
8

1/3

and the Fermi energy for relativistic particles is


F = pF c
thus we find


3ne
8

1/3

3n p
hc +
8

1/3

3nn
hc
8

1/3

hc = mn c2 m p c2 0

where we made the assumption that mn c2 m p c2 0. We are also told that


ne = n p
thus we find
1/3

2n p = nn

1
ne = n p = nn
8

Problem # 4

62

Assume that a hot, bloated neutron star emits thermal neutrino radiation from a surface of radius R at
an effective temperature equal to TE . Assume that three types of massless, or nearly massless, neutrinos,
e , , and their antiparticles, are emmited in equal numbers, in thermal equilibrium with zero chemical
potential. Show that the luminosity is given by
L =

21 4
TE 4R2
8

where is Stefans constant. Find an expression for the average energy for a neutrino in this radiation.
[Hint: Look back at Chapter 2 and reconsider Problem 2.5]
If we refer to Philipps problem 2.5 we find that the energy density of fermions is given by
7
uF = aT 4
8

(9)

and we know that the energy density of a photon is


u p = aT 4

(10)

the differences in these two expressions comes from solving the following two integrals
 3 Z
Z
kT
x2
1
N(p)d
p
=
8
dx
n=
x
V 0
hc
0 e 1
this is the number density of particles with momentum p and p + d p, the is to differentiate between
bosons and fermions. Since neutrinos are fermions. And the energy density is given as
 3 Z 3
Z
1
kT
x
u=
p N(p)d p = 8
kT
dx
x
V 0
hc
0 e 1
from Equationquations 3 and 4 we can see that the solution to the energy density for a fermion is given by
7
uF = aT 4
8
this was using the assumptions that the polarization of the fermion is 2, but we know that the polarization
of neutrinos is 1. We also need to take into account the 6 different species of neutrinos, thus for neutrinos
we find that the energy density is given as
u =

6 7 4 21 4
aT = aT
28
8

we also know that Stefans constant is defined as


=

ac
4

thus the energy density is now given as


21 4 4
T
8 c
but we know that the flux due to the neutrinos is given by
u =

F =

21 4
T
8
63

where the factor of 4/c was taking care of the fact the intensity radiated at a particular frequency is c/4
times the photon energy density at this frequency. We know that the luminosity of neutrinos is given by
L = F 4R2
and we just derived the flux for neutrinos, thus the luminosity is
L =

21 4
T 4R2
8 E

Problem # 5
In this problem we will calculate the properties of the neutrinos emitted by a newly formed neutron
star (a proto-NS). The neutron star is formed during a supernova explosion and its gravitational binding
energy ENS is released in the form of neutrinos on a timescale tKH , so that the neutrino luminosity of
the NS is L = ENS /tKH . Assume that the NS has a mass of 1.4Msun and approximate it as an n = 3/2
polytrope supported exclusively by neutron degeneracy pressure. The initial central temperature of the NS
is 1011 K.
a) Calculate the radius and central density c of the NS.

We know that the relationship between the radius and the mass of a NS is given by


M
RNS = 15 km
Msun

1/3

and since we know that M = 1.4Msun we find the radius to be approximately


RNS = 13.41 km
the mass density can be found by using the central density of a n = 3/2 polytrop which is given as
c =

3M
an
4R3

an = 5.99 n = 3/2 polytrope

thus the central density is


c 1.43

M
R3NS

given the mass and radius we find


c 1.65 1015 g/cm3
b) Show that the neutrinos are initially degenerate in the core of the NS. Use this fact to estimate the
typical energy E of a neutrino in the core of the NS. The neutrinos are relativistic.
we know that
E EF (e)
where
EF (e) =
we know that

3ne
8

1/3

1
1 c
ne nn
8
8 mp
64

hc

thus we find the Fermi energy to be


E EF (e)

3nn
64

1/3

hc 303 MeV

to show that they are degenerate we just need to show


E ET
where ET is the thermal energy, we find that for relativistic particles is
ET 3kT 25.8 MeV
thus we can see that
E ET
c) The cross section for neutrinos interacting with matter is
10

44

E
me c2

2

cm2

Estimate the optical depth = n R R/ of the NS to neutrinos, where is the neutrino mean free
path in the core of the NS.
Since we know that the optical depth is given by
= n RNS
where we have defined n to be the total number density
n

c
9.86 1038cm3 RNS 13.41 km me c2 0.508 MeV
mp

we find
4.71 106
and we know that the neutrino mean free path is given by
=

RNS
0.284 cm

d) The timescale tKH for the NS to radiate away its binding energy in neutrinos is the time for the
neutrinos to random walk out of the NS. Use your result from c) to estimate the time tKH and the neutrino
luminosity L of the NS.
we know that the time it takes for a neutrino to random walk out of the NS is given by
R2NS R2NS
=
t=
v
c
given the radius of the NS and the mean free path we find the time to be
t 210 s
65

we know that the luminosity is given as


E
t
using the neutrino energy in part b) and the time we find that the neutrino luminosity is
L =

L 9.19 1050 ergs/s


e) Look at Problem 4 Note Phillips hints at the back of the book for problems 6.3 and 2.5.
f ) Use your results from d) and problem 4 to calculate the effective temperature of the neutrino radiation (in K) and the average energy of a neutrino emitted by the newly formed NS (in MeV). For comparison
to the results you have calculated in this problem, the observed timescale of neutrino emission was 10 s
for SN 1987A and the typical neutrino energy was 20 MeV.
The effective temperature is given as
TE =

8L
21 4R2

1/4

1.48 1010 K

the energy of a neutrino can be calculated by knowing the energy density and the number density. The
energy is given by
u
E
n
where the number density is given by Philips equation 2.42 with a modification coming fro the fact that
we are dealing with fermions
6
8k3
n = bT 3 where b = 1.803 3 3 = 45.48K3 cm3
2
h c
thus we know that the energy of a neutrino is given by
E =

21 a
T 6.24 MeV
8 b

66