Dr.

Partha Ghosh Department of Physics, Bangabasi College Kolkata

Chapter-2A Stars and stellar systems

(a) Properties of stars (distance, brightness, size, mass, temperature, luminosity).
(b) Measurement of stellar parameters: distance parallax, Cepheid variables, nova and supernovae, red shift),
stellar spectra, spectral lines, the Hertzsprung-Russell diagram, luminosity and radius, binary system and mass
determination, scaling relation on the Main Sequence.
(c) Basic equation of stellar structure, hydrostatic equilibrium and the virial theorem, radiative and convective
energy transport inside stars, nuclear energy production. Equation of state, opacity, Derivation of the scaling
relations.

Chapter-2B Stellar Evolution & Remnants

(d) Formation and evolution of stars star formation, pre-main-sequence collapse (gravitational instability and
mass scales, collapse of spherical cloud, contraction onto the Main Sequence, Brown Dwarfs), evolution of
high-mass and low-mass stars (core and shell hydrogen burning, helium ignition), late stage evolution of stars,
evolution of Sun-like stars and solar system.
(e) End stages of stars white dwarfs (electron-degeneracy pressure, mass-radius relation), neutron stars (mass
limit of neutron stars, neutron stars observable as pulsars), black holes as end point of stellar evolution,
supernovae.
******************************************************************************************
2A
(a) Measurement of Stellar Parameters
(i) Distance to the nearest stars is measured by trigonometric parallax.
Motion of the earth around the sun produces apparent motion of stars
on the sky. Stars perpendicular to the plane of the ecliptic trace a circle
on the sky in course of a year; stars lying in the plane of the ecliptic
will trace a line that doubles back on itself, whereas others would trace
an ellipse. α
The angular size of the semi-major axis of the ellipse would be given
as α = d¤/d, where d¤ = 1A.U, the sun earth distance. Stellar distances
are also measured in parsecs, light years etc. d
Apart from the apparent motion of stars due to parallax, they have real
motion too, with respect to themselves and also with respect to the
sun. Superimposed upon the parallax motion is thus a linear proper ●
motion, producing curly/wavy trajectories that are observed from the
surface of the earth.

(ii) Stellar temperature and types: Stars are hotter at their cores and temperature gradually decreases outwards.
Stellar spectrum is determined by temperature of the photosphere (outermost surface), from which photons can
escape without being further scattered or absorbed. The scattering & absorption probability depend on
wavelength, and so depth of the photosphere can be wavelength dependant.
The photosphere base emits a Planck spectrum, which is somewhat modified by
the colour transparent gas above it. By examining the spectrum, one can
measure the colour temperature by matching with the Planck spectrum through Photosphere
Base
Wien’s law. However, in practice, the peak is often outside the range of
detectors we have and absorption processes also modify it a bit.

1|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

A more practical way is to find the ratio of the fluxes f1(λ1)/f2(λ2) for two wavelengths and find the blackbody
temperature that gives rise to such a ratio. Colour temperature can’t be found if measurements are in the R-J
side of the spectrum (long wavelength, low energy) as the ratio has the same shape for all wavelengths (~λ -4)
irrespective of temperature. In that case, we can only set a lower limit of temperature. We can associate another
temperature by reexamining the absorption features at discrete wavelengths. The atoms/molecules in the cooler,
less dense gas in the photosphere may absorb photons with energy equal to the quantum of transition of various
levels and that would be predominantly absent in the spectrum. When they de-excite, the emitted photons will
have random directions. The absorption lines primarily depend on the level of ionization and excitation of the
gas, reflecting mainly the temperature of the photosphere and its chemical composition.

For hydrogen, the Lyman series lies in UV, Balmer in visible, while Paschen, Brackett and Pfund lie in IR and
higher wavelength range. The atmosphere is transparent to visible, fewer IR, microwave and radiofrequencies.
Before stellar physics was well understood, stars were classified according to spectral types of absorption lines
appearing in their optical spectrum. The following mnemonic may be remembered for this purpose:

Oh Be A Fine Girl/Guy Kiss My Lips Thou

Brown Dwarfs
BLUE, HOTTER STARS RED, COLDER STARS

Stellar luminosity can be calculated as L(λ) = f(λ) × 4πd2, if the flux ‘f’ and distance ‘d’ is
known. Integrating over all wavelengths, the stellar bolometric luminosity can be obtained. ●
Again, if the stellar luminosity is already known along with either of its radius/temperature,
then the other can be obtained from the very useful relation L = 4πr2×σT4. The scheme is
useful for Cepheid stars in open clusters, which are short lived phases of some intermediate
stars during part of the helium-core-burning giant stage. Cepheids are sometimes pulsating;
their periodic variations in luminosity vary from 1 to 100 days. One can then calibrate the
period-luminosity relation for a standard candle. Once the relation is calibrated for nearby
cepheids, it can be used to determine luminosity or distance of cepheids of other galaxies
with measured periods.

Another method used to obtain distance is supernovae light echo. After some time Δt of explosion, a ring of
circumstellar material that is ejected by the nova progenitor star during previous stages in its evolution, become
photo-ionized by the flash of UV radiation from the explosion. The ring begins to shine giving emission lines
characteristic to the chemical composition of the gas, slowly fading away as the atoms recombine. The time
delay Δt, which is due to the light travel time from the supernova to the ring, is related to the distance D to the
supernova as Δt = R/c = (θ×D)/c, where θ is the observed angular radius of the ring. The distance can be
calculated from this easily.

Redshift is a term used to describe situations when an astronomical object that is observed to being moving
away from the observer, such that emission or absorption features in the object's spectrum are observed to have
shifted toward longer (red) wavelengths. The change in wavelength of the spectral features is due to the Doppler
Effect, the change in wavelength that results when a given object and an observer are in motion either toward or
away from each other. The radiation coming from a moving object is shifted in wavelength: z = v/c = Δλ / λ0.
When Hubble plotted the redshift vs. the distance of the galaxies, he found a surprising relation: more distant
galaxies are moving faster away from us, as v = Ho d. He used a “standard ruler assumption”, all galaxies of the
same type have same physical size, whatsoever their nature may be.
2|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

(iii) Binary systems and measurement of mass: Direct measurement of stellar mass is possible for Binary or multiple star
systems. A large fraction of the stars are found in binaries. This indicates that stars are formed in groups. The sun as a
single star is an oddity, as for these kinds of stars, the ratio of single: double: triple: quadruple systems are 45:46:8:1.
Binary stars have orbital separations in the range 3×109 m to 3×1015 m. From the determination of period and line of
sight velocities, one can measure the masses of binary stars. In these systems, one can have both normal stars; one of the
stars may be a neutron star or white dwarf etc. If they are close enough, mass transfer is possible which leads to accretion,
and which may completely change the evolution process of the star.
There are generally four classes of binary stars:
Visual binaries → Can be resolved individually by
telescope with centers separated by more than one
arc second, orbiting the common centre of mass.
Long-term observations can then be made to plot
the relative positions of the members of the
system. Over time this data is accumulated and
used to calculate the orbits of the stars. The
inclination of the orbits and their relative
orientation to our line of sight is very important.
Astrometric binaries → One member is faint, other
is found to rotate about a point. The brighter of the
two stars is the primary, and the fainter is the
secondary. If the two stars are of equal brightness,
it can sometimes happen that, when they become
very close to each other, they merge and cannot be
distinguished until their separation is sufficiently
great for them to be resolved again. It may then not
be obvious which of the two had been designated
the “primary”. Some stars, if observed repeatedly
over time, may show a periodic perturbation or
wobble in their proper motion. We can then infer
that this occurs due to the gravitational influence
of an hence before unseen companion. Binary
systems detected by such means are called
Astrometric binaries.
Spectroscopic-binaries→Though spectroscopically
unresolved, the observed absorption spectrum is
understood as the superposition of two stars, from
the large enough periodical Doppler shift produced
in it by their orbital velocities. We see only one
source of light. The spectrum shows Doppler shifts
which change from redshift to blueshift
periodically. This means that the star is moving in
a periodic orbit. When the star moves towards us,
we see a blueshift. When the star moves away
from us, we see a redshift. The period of time from
redshift to blueshift back to redshift again is the
orbital period.

3|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

Eclipsing binaries → The orbiting plane is so

inclined that the rotating pairs are seen to eclipse
each other periodically. This can occur between
two main sequence stars or between a main
sequence and a red giant etc. If the plane of the
orbit is along our line of sight, we can see that the
light from the system gets dimmer for a short time
as one star comes in front of the other, thus
obstructing our view. By measuring the duration of
the eclipse, we can find the size of each star and
hence mass. A primary eclipse occurs when the
hotter star is blocked by the colder star and the
reverse scenario gives rise to the secondary
eclipse. Partial eclipses may also occur under
certain conditions. One can even have two
conditions to exist simultaneously; for example, a
binary system can be eclipsing as well as visual.

For a circular orbit, M1R1=M2R2 gives the position of the C.M. Thus if
R1+R2 = a, we have R1 = M2R2/M1 = M2(a-R1)/M1 or R1 = M2a/(M1+M2),
R2 = M1a/(M1+M2). The equation of motion of the first mass in the
gravitational field of the second is given as M1ω2R1 = GM1M2/a2 or
M1ω2M2a/(M1+M2) = GM1M2/a2 or ω2 = G(M1+M2)/a3. For the sun-earth
system, since M¤ >> Mo, ω2 is approximately equal to GM¤/a3 or
M¤=4π2a3/T2G, where T is equal to one year. For most binary stars, the
plane of orbit would appear to be an elliptical, inclined to our line of sight
by an angle i. That means that the size of the orbit we observe is actually
smaller than the true size of the orbit. Generally, the angle ‘i’ is in the
range 0 to π/2; the positions are known as face-on & end-on respectively.
If the distance from the observer to the binary star system is ‘d’, and the
orbital radii of the two stars subtend angles θ 1 and θ2, their minor axes
would appear shortened by a factor of cos i, as θ 1cosi & θ2cosi
respectively.
Now θ1=R1/d, θ2=R2/d, and R1/R2 = θ1d/θ2d = M2/M1. If the distance a
and angular velocity from time period T is known, the total mass M 1+M2
can be obtained from Kepler’s law, ω2 = G(M1+M2)/a3, and combining
with the already obtained ratio, M1 and M2 can be determined
individually.
For spectroscopic binaries, as the distances R1, R2 and hence ‘a’ cannot be measured directly, we need to use
amplitudes of oscillations of the line of sight velocities from the observed Doppler shifts. Let us find out how
this can be done.
For a system having circular orbits, the speed of each star will be constant. If the plane of the orbit is face on (i
= π/2), then the measured radial velocities will produce sinusoidal velocity curves when plotted against the ratio
of observation time and period t/P. When the orbits are elliptical (eccentricity not equal to zero), the graph
become skewed. The two cases are illustrated as below.
4|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

In general, the elliptic orbital plane of the system would also be tilted relative to our line of sight, the speeds of
the planets are not constant and hence velocity curves become more complicated. Let us try to keep things
simple at first while starting. For two stars having period P in concentric circular orbits of radii R 1 and R2, we
have the velocities v1 = 2πR1/P, v2 = 2πR2/P; linear momentum conservation as M1v1 = M2v2, M1R1 = M2R2 and
finally as before, R1 = M2a/(M1+M2), R2 = M1a/(M1+M2). Also, in the relation G(M1+M2)P2 = 4π2a3, we can
3 3
use P=2π/ω and for the first star, we get GP2
M2
 4 2 3
R 
M2

P
R1 3. For orbits
M1  M 2 2 1
M1  M 2  2G
2

inclined at angle i, the observed velocities will be v1, obs  v1 sin i and v2, obs  v2 sin i . Also, since the ratio of
3
M 2 sin 3 i
observed velocities |v1,obs|/|v2,obs| = R1/R2 = M2/M1, we finally have f1 
P
v1 sin i 3. Here f1 is

M1  M 2  2G 2

known as the mass function for star 1. Similarly, one can have the mass function f2 for the second star.
Generally, there are three unknown quantities, the masses of stars and angle of inclination ‘i’ against two
equations. A third equation can be obtained from the duration of an eclipse or from the orbital brightness
variations due to tidal distortion of the observed star by its binary companion from the light curve, as they move
past one another. In any case, stellar masses are determined only to a factor of sin3 i, and for all possible values
of i and M1, the value of f1 gives the minimum possible mass of M2.
1/ 3
 P 
An important case arises when M2 << M1. Then we have M 2 sin i  
2/3
 v1, obs M 1 . This has been applied
 2G 
to study exoplanets outside our solar system. Let us consider a hypothetical spectroscopic binary with circular
orbits having inclination ‘i’. The line of sight velocity of star M1 would show sinusoidal variation as vr(t) =
v1sin(i)sin(ωt). For the given system, since ‘i’ is constant, the maximum value of this quantity is v1 sin(i), which
is directly measurable from Doppler data and one can get the value of mass function f1 instantly. The result has
units of mass, but is not the mass of either component. An exception occurs when M1 << M2 and i = π/2, where
one can have the value f1= M2 directly.
Eclipses can be used to determine radii of the stars. If the suffices‘s’
and ‘l’ stand for the smaller and larger star respectively, then from the
adjoining figure, we simply get rs = v(tb-ta)/2; rl = v(tc-ta)/2 = rs+v(tc-
tb)/2. Here v = vs+vl, the relative velocity of the two stars. Also, the
radius of a relative circular orbit is a = vP/2π, and therefore rs/a = (tb-
ta)π/P, rl/a = (tc-ta)π/P.
The ratio of effective surface temperatures of the two stars can be obtained by assuming them to be blackbodies.
It can be seen from the figure that the dip in the light curve is deeper when the smaller, hotter star passes behind
its companion. The radiative surface flux Fr=σTe4. Regardless of whether the smaller star is behind or in front of
the larger one, same total cross section is eclipsed. When both stars are equally visible, the amount of light
received can be taken as B0 = k(πrl2Frl+πrs2Frs), where k is a constant depending upon distance, intervening
medium and nature of detector. When the hotter, smaller star is passing behind the larger, cooler one, the
5|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

amount of light received is Bp=kπrl2Frl. On the other hand, when the smaller one is front of the larger one,
amount of light detected is Bs=k(πrl2-πrs2)Frl+kπrs2Frs. Since it is not possible generally to exactly determine
the value of the constant k, ratios are employed. Straightforward calculations thus directly yield to the result
4
B0  B p F T 
 rs   s  . In reality, however, binary systems are elliptical rather than circular, and we need two
B0  Bs Frl  Tl 
additional parameters, eccentricity of the orbit and orientation angle, which pose two limitations to our
calculations. These can be determined from other observed astronomical data.

(iv)The Hertzsprung Russel (HR) diagram: The logarithm of stellar

luminosity when plotted against decreasing values of effective surface
temperature TE, gives rise to a scatter plot as shown in the figure. About 80-
90% of the stars lie in a narrow diagonal strip called Main Sequence, which
corresponds to L ~ TE8. Since for black bodies, L=4πr*2σTE4, we immediately
infer that for these stars, r* ~ TE2. Our sun is a main sequence star, radii of
other main sequence stars are in a range to 0.25 to 25 times r ¤. There is a
concentration of points in the HR diagram for stars that are cool (i.e., red),
with higher luminosities than main sequence stars.
They are called Red Giants. In the lower left part, we have hot stars, from whitish to bluish, with lower
luminosities. They have radii of the order of the earth, and are called White Dwarfs.
From measurements of binary stars, it became clear that the main sequence is actually a mass sequence; with
highly massive stars with high luminosities and temperature spread across along-with lower mass stars with low
luminosities and temperature. Masses of the stars in the sequence range from 0.1 M ¤ to 100M¤. Red giants have
masses around 1 to 2 times that of the sun, while white dwarfs are never more massive than 1.4 times M ¤.
Stellar luminosities vary with mass in the range from L ~ M3 to L ~ M5.
Most stars spend their maximum lifetime at the same location of the main sequence, when H-burning takes
place at the stellar core. When most of the hydrogen gets synthesized to helium, a temporary increase in
luminosity occurs. Stars having initial masses < 8M¤, eventually shed off their outer layer and become white
dwarfs, the compact remnants of the original stellar core. They are devoid of nuclear radiations and slowly
radiate away heat. On the other hand, stars having initial masses > 8M¤, after passing through the red/blue
supergiant stages (appearing near the top edge of the HR diagram), undergo a runaway process of gravitational
core collapse ending in supernovae explosion. After that, the remnants become either neutron stars or black
holes. Neutron stars are hotter and more compact than white dwarfs, and also less luminous.
The HR diagram played a very important role in the understanding & development of modern astrophysics.

6|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

(v) HYDROSTATIC EQUILIBRIUM, VIRIAL THEOREMS & OTHER ASPECTS:

A star is a sphere of gas that is held together by self gravity and balanced against collapse by pressure gradient.
To understand the phenomenon, let us consider the free fall time; i.e, the time taken by the star to collapse to a
point if there is no outward pressure support.
The potential energy of a mass element dm at rest on the surface of a collapsing star of
●dm mass M(r0) and radius r0 is given as dU= -GM(r0 )dm/r0. When the star collapses to a
radius r from r0, assuming no stellar mass loss, we have, the loss in P.E is equal to a
M M(r0)  1 1  dr  2GM r0   r0
1/ 2

2
1  dr 
gain in K.E; dm   GM r0 dm   or      1 . Here we
2  dt   r r0  dt  r0 r 
 dr
have retained the negative sign as dr/dt is negative. Also, since dt  , we can integrate
 2GM r0   r0
1/ 2
 
   1
 ro  r 
time from 0 to τFF and radius from r0 to 0, changing the limit of the variable x = r/r 0 from 1 to 0, and using the
1/ 2
 r03 
1/ 2
 x 
1
 
standard result    dx  to get finally the free fall time  FF     . Putting in the values,
0
1 x  2  2GM r0  2
we find that for our sun, the value comes out to be 1800s; or without support of outward pressure, the sun would
have collapsed to a point within half an hour. At that leads us to think differently as follows. We now consider a
P+dP A cylindrical element of height dr and base area A, with a pressure difference dp between
AdP
dr the outer and inner faces. While gravity acts inwards, the force due to pressure balances
 GM r dm
 Adp  0, dm   r Adr and we
P
it by acting outwards. Under equilibrium,
r2
Gravity
dpr   GM r  r 
have finally, the Equation of Hydrodynamic Equilibrium as  . This equation indicates that
dr r2
pressure gradient is negative and decreases outwards. If we multiply both sides by 4πr3dr and integrate r from
GM
r* r * * r
dp GMdM
zero to the stellar radius r*,  4r dr    2 4r 3dr   
3
, where dM=4πr2drρ is the mass within a
0
dr 0
r 0
r
thin shell of thickness dr. As obvious, the R.H.S is the work done to construct the star, shell by shell, and is
nothing but the gravitational self energy Egr possessed by it.

The L.H.S can be integrated by parts to get  pr 4r 

r*
3 r*
o  3 4r 2 pr dr.
0

We define the surface of a star as where the pressure is zero, i.e., p(r*) = 0, whence the first term vanishes
r
1 *
immediately. The second term is three times average pressure multiplied with volume, as p   4r 2 pr dr ,
V 0
giving p   Egr 3V , known as the Virial Theorem of First Form, which states that the mean pressure of a star
is numerically equal to one third of its gravitational energy density.
If we consider a simplified picture of a star being composed of total ‘N’ classical, monoatomic, identical non-
relativistic ideal gas molecules, then the equation of state is pV = NkBT, the thermal energy is Eth = 3NkBT/2, the
pressure is p = 2u/3 = 2Eth/3V. Multiplying the last equation with 4πr2 and integrating over the volume of the
7|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

2 Eth 2 EthTotal
star, we get p  4r 2 dr 
3V 
4r 2
dr  pV  where EthTotal is the total thermal energy of the star. On
3
inspection, we immediately have the result, EthTotal = -Egr/2, the Virial Theorem of Second Form, stating that
when a star contracts and looses energy, its gravitational self energy becomes more negative and thermal energy
rises. It now becomes clear that heat capacity is negative as loss of energy results in rise of temperature. Also,
the Third Form of Virial Theorem gives ETot = EKTot + Egr = -EthTotal = Egr/2. This means that the total energy
of a star, composed of classical, non-relativistic, monoatomic ideal gas type molecules is negative, meaning that
it is bound. Since stars radiate away their energy, ETot becomes more negative and they are doomed to collapse
eventually; exception occurs when the stellar gas moves from classical to the quantum regime.
GM r 4r 2 dr  3GM*2
r*

For a fixed (ρ = constant) density profile, the gravitational self energy Egr     . The
0
r 5r*
Egr
mean pressure can be taken as p  ; the characteristic virial temperature can be defined from the following
3V*
3NkBTvir E GM*2 GM* Nm
relation:  EthTot  gr   , where we have assumed in total N non-relativistic ideal gas
2 2 2r* 2r*
GM *mH
molecules of mean mass m  (m p  me ) 2  mH 2. On substituting in the previous equation, k BTvir  .
6r*
For our sun, the virial temperature comes to about 4×10 6K. At this high temperature, nuclear reactions set in
and replenish the thermal energy that the star radiates away; halting the gravitational collapse temporarily. Since
the mass of stellar material contained inside a thin shell of radius r and thickness dr is dM(r) = 4πr2drρ(r), we
dM r 
have, in the stellar context, the Equation of Mass Continuity (conservation) as  4r 2  r .
dr
(vi) RADIATIVE & CONVECTIVE ENERGY TRANSPORT:
The pressure gradient p(r) of a star that supports it from gravitational collapse is produced by gradients of
density and temperature, namely, ρ(r) and T(r). In much of the stellar volume, the temperature gradient depends
on luminosity L(r) and radiative transport is important. Also, in other cases, convection plays an important role.
For the stellar remnants like white dwarfs, the process of conduction for heat transfer also takes place.
Photons travelling as a beam through a medium may be absorbed / scattered out of their path by interaction with
molecules, neutral/ionized atoms and also with electrons. If n be the number density of such targets and σ be
their effective cross section, then the number of such interactions on travelling a path length dx are given by
nσdx. The collision probability, i.e., no. of collisions on travelling unit path length is nσ, and mean free path l =
1 1
1/nσ. Since there are a variety of absorbers, we effectively have l   , with mass density ρ and
 ni i K
i
2 -1 2
opacity K. The Opacity has dimension of L M , units of m /Kg and depends on local density, temperature &
elemental abundance in the medium of travel. Actual mean free path of photons in stars is found to be 1mm. As
   
they are randomly oriented along various directions, we have displacement D  l1  l2  ....  lN ; the squared
 
      
displacement as D 2  l12  l22  ....  lN2  2 l1.l2  l1.l3  ... and mean squared displacement as D 2  Nl 2 , the
second term cancelling out due to randomness of orientations. The linear displacement covered in this random

8|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

 1/ 2
walk is thus D  D2  N l. Since a star has radius r *, the number of steps required is given as N=r*2/l2.
Each step requires a time l/c and thus the time of escape of the photon from the star is  RW  Nl c  r*2 lc.
Stars radiate like blackbodies, at least locally, to a very good approximation, as apparent
Δr from the small free paths, numerous scattering / absorptions, remission etc. However,
u
L(r) u+Δu there is anisotropy, as the net energy flow is outwards. As evident from the adjoining
figure, the rate of energy flow from a shell of radius r, thickness Δr is given by
 4r 2 ru  4r 2 ru  4r 2lcu
Lr     ,where Δu is the excess energy density. A more rigorous
 r 2 lc r
Lr  lc du
calculation yields a factor of 1/3rd, giving  . This is basically a diffusion equation, with L(r)/4πr2
4r 2
3 dr
being the energy flow per unit time per unit area, du/dr being the energy density gradient and –lc/3 being the
diffusion coefficient. In the limiting case, du/dr can be replaced by u/r* and we can have an estimate of stellar
luminosity as L*=4πr*2(lc/3)(u/r*) = 4πr*(lc/3)aT4.
Opacity is the quality that controls the flow of radiation through the star. For low opacity (large free path l), the
flow would be relatively unobstructed, hence luminosity would be high and vice versa. At every point in the
stellar interior, since the ambience is of a blackbody, the energy density u=aT4, giving du/dr = 4aT3×dT/dr or
 3Lr  du dT 1 3Lr 
dT/dr = (1/4aT3)×(du/dr). From the diffusion equation, we have  and thus   .
4r lc dr
2
dr 4aT 3
4r 2lc
dT  3Lr  r K r 
Putting in the value l=1/ρK, we get the Equation of Radiative Energy Transport as  .
dr 4r 2  4acT 3 r 
The stellar luminosity is produced by nuclear reactions with output energies that depend on density &
temperature locally and hence on radius ‘r’. If ε(r) be the power produced/unit mass of stellar material, we have
dLr 
dL = εdm = ε×4πr2drρ, giving the Equation of Energy Conservation as  4r 2  r  r .
dr

The convective energy transport mechanism takes place when stellar material continues to be displaced from its
equilibrium position without being restored back. This is very efficient in transporting heat energy.
T+dT, p+dp, ρ+dρ Let us understand the condition that is required for convection to be set in inside
the star. We consider a volume element as shown at a radius r, where
temperature, pressure and density are T, p and ρ respectively. Let it be displaced
dr
to a larger radius r+dr, where surrounding parameters have changed to T+dT,
p+dp & ρ+dρ respectively. Since the gas in the star obeys ρ ~ p/T, taking
d dp dT
T, p, ρ logarithmic derivative of both sides, we have   .
 p T
For simplicity, we assume that at the new location, the volume element expands adiabatically, with the entropy
change dS=0, till its parameters reach T+δT, p+δp and ρ+δρ. (We have introduced ‘δ’ to be different from‘d’.)
As per the equation of state for adiabatic process, we have p~ργ, which on logarithmic differentiation yields
 p  1 p
  or  . We note that both dρ & δρ are negative.
 p   p
The said element will continue to float up rather than falling back into equilibrium if its density after expansion
is still less than the surroundings, that is, ρ+δρ < ρ+dρ, or simply, δρ < dρ. Dividing both sides by ρ, we get
9|P ag e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

dρ/ρ < δρ/ρ. Substituting the condition in the previous equations obtained from logarithmic derivatives, we get
1 p dp dT dT   1 dp
  . Recalling again that the changes in pressure δp and dp are same, we get  . This
 p p T T  p
dT   1 T dp
can be divided by dr to get  . Since the radial temperature and pressure gradients are both
dr  p dr
negative, the condition for convection to set in is that, the temperature profile must fall fast enough with
dT   1 T dp
increasing radius,  .
dr  p dr
For a non-relativistic gas (ionized or excited molecules), as number of internal degrees of freedom (rotational
and/or vibrational) rises, γ becomes more and more small, and thus convection becomes possible even for
smaller values of |dT/dr|. Convection thus sets in cooler regions, as in outer layers of intermediate-mass main
sequence stars (like sun) and red giants; in larger ranges in radius of low mass stars. Once set in, convection
helps to mix materials at different radii and thus equilibrating temperature, lowering the absolute value of
temperature gradient |dT/dr| to a critical value. The energy transport equation, in this case, would be just
dT   1 T r  dp dT  3Lr  r K r 
 in contrast to the previous radiative energy transport equation,  .
dr  pr  dr dr 4r 2  4acT 3 r 

(vii) THE EQUATIONS OF STATE OF STELLAR STRUCTURE:

In our previous discussion, we had obtained four coupled first order differential equations which control the
stellar properties, as follows:
dpr   GM r  r 
(i) Equation for hydrostatic equilibrium defining pressure gradient:  .
dr r2
dM r 
(ii) Equation of mass continuity defining mass gradient:  4r 2  r .
dr
dT  3Lr  r K r 
(iii) Equation of radiative energy transport defining temperature gradient:  .
dr 4r 2  4acT 3 r 
dLr 
(iv) Equation of energy conservation defining the luminosity gradient:  4r 2  r  r .
dr
Let us assume four boundary conditions as M(r=0) = 0, M(r=r*) = M*, L(r=0) = 0, p(r=r*) = 0. We also need
to have three equations connecting pressure, opacity and energy production rate of the gas with density,
temperature and composition, as p = p(ρ,T,comp.), K = K(ρ,T,comp.) & ε = ε(ρ,T,comp.), which remain
coupled with the abovementioned four differential equations. The dependence on composition is through
relative elemental abundances and ionization states of each element of the gas. In astrophysics, what we do
commonly is that, we parameterize mass abundance of H, He and heavy elements (which are collectively called
metals), as X  H  , Y   He  , Z  metals .
The four boundary conditions ensure that the solution set is unique. It means that the properties and evolution of
an isolated star is fully determined by its initial mass and chemical abundances (Vogt-Russel Conjecture). Note
that stellar rotation and magnetic fields if any have been neglected in the calculations.
Observable parameters like surface temperature, radius and luminosity of a star can be determined from the
solutions of these equations. Under normal conditions, the equation of state of a classical, non-relativistic, ideal

10 | P a g e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

gas holds good here, as p = p(ρ, T, X, Y, Z). Let us consider three different types of particles having mass mi and
m n  m2 n2  m3n3 
number density ni, to get the mean mass as m  1 1  . Thus the pressure can be expressed as
n1  n2  n3 n
k BT
p  nk BT  . For completely ionized pure hydrogen, the mean mass m  0.5mH as electrons are absent.
m
H X Y Z
Now, nH   . Similarly for helium, we have nHe  and for metals, n A  , writing in terms of
mH mH 4mH AmH
hydrogen masses. Next we note that complete ionization of hydrogen gives two particles, one nucleus & one
electron. For completely ionized helium, we have one nucleus and two electrons, or three particles. For case of
metals, we have Z electrons and one nucleus, i.e., (Z+1) ~A/2 particles. For a star comprising of all of these, the
mean density is given by, assuming the sum of the density functions X+Y+Z=1 and the relative abundances as
A X Y A Z A   3Y 1  X  Y   Y
above, n  2nH  3nHe   nA  2 3    2X    (3 X   1).
2 mH 4mH 2 AmH mH  4 2  2mH 2
 m 2 m
Also since n  , we have  . For our sun, X=0.71, Y=0.27, Z=0.02, and  0.85. In
m mH 1  3 X  Y 2 mH
addition to the kinetic pressure, photons exert radiation pressure too, as prad = u/3=aT4/3. Thus, we have the
 aT 4
complete Equation of State for normal stars as p  k BT  . The equations are different for the stellar
m 3
remnants, viz., white dwarfs, neutron stars and black holes.

(viii) OPACITY: Let us now pay attention to one of the most important stellar parameter, the opacity factor K.
The expression for this quantity, at every radius, will depend on density, temperature and composition of the
star at that point. It has been already obtained that K=1/lρ, with the mean free path for scattering being equal to
l=1/nσ. For electron scattering, Ke.s=neσT/ρ, with the Thomson scattering cross section σT.
We find that the electron density for an ionized gas comprising of H, He and metals is given as
X 2Y A Z A   Y Z  X  
A
ne  nH  2nHe   nA     X      1  1  X  . Thus the
2 mH 4mH 2 AmH mH  2 2  mH  2  2mH
expression of opacity due to electron scattering is Ke.s   T 1  X  2mH . At low temperatures, not all the gas is
ionized, some electrons are still bound to the atoms. Apart from Thomson scattering, there are possibilities of
bound-bound, bound-free and free-free absorptions. In the former two cases, an atom/ion is excited to higher
level or ionized to a higher degree by photon absorption. The free-free absorption occurs when a free
electron/ion shares the photon’s momentum & energy.
When averaged over all wavelengths, Kramer’s law states that the mean opacity Kbf , ff   T 7 / 2 .

(ix) SCALING RELATIONS OF THE MAIN SEQUENCE: For the main sequence stars, the functional form
of dependence of luminosity on stellar mass and effective temperature are observed as L ~ Mα, L ~ TE8. If we
assume rough power laws as p(r) ~ rβ, M(r) ~ rγ etc., then from the already discussed equations of
hydrodynamic equilibrium, mass continuity and radiative transport, we have –

11 | P a g e
Dr. Partha Ghosh Department of Physics, Bangabasi College Kolkata

dpr   GM M dM r  dT r  3LK T 4r

  p  ,  4r 2
  M  r 3
 ,    L  . For moderately
dr r2 r dr dr 4r 2  4acT 3 K
massive stars, kinetic gas pressure dominates, and thus p ~ ρT ~Mρ/r gives T ~ M/r. The dominance of electron
scattering gives opacity K=const., with L ~ T4r/Kρ ~ M4r/r4Kρ ~ M4/KM ~ M3.
We consider a star being formed from mass M, which is contracting under its own gravity and getting heated up
eventually. Once the density & temperature of the core are high enough for onset of nuclear reactions,
contraction stops and equilibrium will be set-up. The power density generated by nuclear reactions depends on
temperature. Thus for a given initial mass, radius r will stop decreasing when a particular temperature is
reached. Since T ~ M/r in all main sequence stars, we have r ~ M. When the density of stars decreases as M-2,
more massive stars have low density and low massive stars have high density.
High densities in low mass stars indicate the dominant role of bound-free and free-free opacity: K ~ ρ/T7/2. At
constant temperatures, we have r ~ M, ρ ~ M-2, K ~ ρ ~M-2 and L ~ T4r/Kρ ~ r/ρ2 ~ M/M-4 ~ M5. This is indeed
observed in practice.
A lower gas density in high mass stars makes the radiation pressure dominating, with p~T4 and the main source
of opacity being electron scattering, rendering it to be constant. So, the luminosity L ~ T4r/Kρ ~ pr/Kρ ~ M. This
flattening of the luminosity mass relationship is observed in most massive stars.
In the main sequence, since the luminosity L ~ M3 has been observed for moderately massive stars, we consider
a representative variation of the form L ~ M4, as one progresses from low to higher masses. Since r ~ M, we
have σTE4 = L/4πr *2 ~ M4/M2 ~ M2 ~ L1/2 or L ~ TE8. So, the mass vs. luminosity & effective surface
temperature vs. luminosity relations of main sequence stars are simply the consequence of different sources of
opacity and pressure in them. They are also a result of the fact that the H-burning onset keeps the core
temperature of all main sequence stars in a narrow range.

(x) NUCLEAR ENERGY PRODUCTION: We all know that nuclear fusion reaction is the basic criteria for
energy production in stars. To understand it properly, we need to have the power density ε (ρ, T, X, Y, Z) to be
described in terms of nuclear binding. From the virial theorem, Egr ~ 2Eth and Eth ~ GM2/2r*, the Kelvin-
2
1 GM* 1
Helmholtz time scale gives  KH ~ the value of which for our sun is 1.6×107 years. But our sun has
2 r* L*
been shining with the present luminosity for a much larger period of time than this. Thus the only viable energy
source for the sun and other main sequence stars is understood to be fusion of hydrogen into helium. The first
few steps of the p-p chain is given as follows:
1. p  p  d  e    e , a weak interaction with time-scale of 108 years. The positron annihilates with an
electron, producing two photons. The neutrino escapes carrying some K.E; the system quickly
thermalize by means of matter-matter and matter-photon collisions.
2. d  p3He   , an electro-magnetic interaction
3. He 3He4He  p  p, with a time scale of 3×105 years. The three step process, releasing 26.73MeV
3

of energy per two reactions, is capable to produce the existing luminosity over the age of the solar
system.

12 | P a g e