Astronomy 112

Physics of Stars

http://www.ucolick.org/~woosley/

- Stan Woosley
 Why study stars?

• Most of luminous matter in universe

• Abodes of life and origin of the elements
• Applied physics laboratories
• Precursors to interesting high energy phenomena
• Particle physics laboratories
• Big explosions
A star is a gravitationally confined thermonuclear
reactor. All “true stars” obtain their energy from
nuclear fusion, though there are interesting variations
that are not nuclear powered because they are
not hot enough or have become “degenerate”.

Because stars radiate, they must evolve. Their energy

is obtained at the expense of changing their composition.
With time, these composition changes alter the stellar
structure.

So long as a star is “non-degenerate”, the loss of

energy from its surface requires that it gets hotter
in its interior (Virial theorem). This continues until
the star becomes degenerate, explodes, or becomes
a black hole.
From these simple principles the scope of much
of the course can already be discerned:

• Stellar structure resulting from an interaction

of gravity, pressure (EOS), and nuclear fusion.

• Radiation transport and mass loss

• Nuclear and particle physics

• Solving the coupled differential equations to

get the time history (usually done on a computer
nowadays)

• Examining end points of stellar evolution

• Comparing with observations

 Observations of Stars I:

• Age of the sun

• Stars – general properties
• Luminosity and brightness
Stellar magnitudes

• Parallax and distances

• Standard candles
(Glatzmaier and Krumholz 1; Prialnik 1.1 1.2; Pols 1
and Ay 12)
 The Sun
The star we can study in greatest detail is rather typical.

M Mass = 1.989 x 1033 gm; about 300,000 Earth masses

R Radius = 6.96 x 105 km; almost 100 Earth radii
Average density 1.41 gm/cm3 3M
Age = 4.56 x 109 years 4π R3
L Luminosity = 3.83 x 1033 erg/s
Central temperature = 15.7 million K
Central density = 162 g cm-3
Photospheric temperature 5772 K

Rotation period 24.47 days at the equator

slower near poles

Surface composition (by mass) 70.6% H

27.5% He, 1.9% C, N, O, Fe, Si, etc
(like the modern “universe”)

A typical star, but a little on the heavy side.

 Age of the earth and sun

While other techniques exist for showing that the

age of the sun and earth is around 4 – 5 billion years
(solar models, for example), the most reliable date
come from radioactive dating.

Heavy elements often have several isotopes. Their

nuclei have different numbers of neutrons, but the same
number of protons. When the solar system was born,
it was well mixed and isotopically very homogeneous.
Samples from many places, on and off the earth, show
that an isotopic variation of over 1/1000 is extremely
unusual unless the isotopes have been affected by
radioactive decay (or, for deuterium, evaporation).

Elemental abundances on the other hand are highly

variable due to chemistry.
 Age of the earth and sun
Consider uranium and lead. The solar system was
born with some heterogeneous distribution of lead
and uranium – chemically - but a homogeneous
distribution of isotopes. Two uranium isotopes decay to
lead, but all presently existing lead isotopes are stable.

In a primordial sample of lead with no uranium in it,

the lead isotope ratios would then be just what the solar
system was born with. In a sample that contains
uranium (in variable proportions) though, both the lead
and uranium isotopes will be different today than when
the solar system was born.

It is assumed that the sun, earth, moon, and meteorites

all have very similar ages. There are other techniques
and models that show the interval between solar and
earth formation was small.
 Radioactive decay

The decay of a radioactive isotope with abundance

n is described by by the well known equation
dn
=−n / τ where τ is the mean lifetime.
dt
The solution is
n = n0e −t /τ where n0 is the initial abundance.

If this species is decaying to another that has initial

abundance m0 then the abundance of m0 as a function
of time is given by
m = m0 + n0 (1− e −t /τ )
 Age of the earth and sun
τ 1/2
235
U→ 207
Pb τ 235 =1.015 Gy ("mean" life= )
ln 2
238
U→ 206
Pb τ 238 = 6.45 Gy
204
Pb not affected by radioactive decay

Observe 204Pb, 206Pb, 207Pb in a large set of samples -

oceanic sediment, meteorites, etc with a variable
content of U/Pb chemical ratios
Because isotopes are
−t /τ 238 not subject to chemical
206
Pbnow = 206Pb0 + 238U0 (1− e ) fractionation, f1, f2, f3, g1,
−t /τ 235 and g2 are all constants
207
Pbnow = 207Pb0 + 235U0 (1− e )
through the solar system
204
Pbnow = 204Pb0 when it was born
206
Pb0 = f1 Pb0 207
Pb0 = f2 Pb0 204
Pb0 = f3 Pb0
235
U0 = g1U0 238
U0 = g2U0
206
Pb = 206
Pb0 + 238
(
U0 1 − e
−t /τ 238
)
U (1 − e )
−t /τ 235
207
Pb = 207
Pb0 + 235
0

206
(
Pb = f1 Pb0 + g2U0 1 − e
−t /τ 238
)
+ g U (1 − e )
−t /τ 235
207
Pb = f2Pb0 1 0

204
Pb = f3Pb0

( )
207
Pb f2 Pb0 g1 U0 −t /τ 235
y(t − now)= 204
= + 1 − e
Pb f3 Pb0 f3 Pb

( )
206
Pb f1 Pb0 g2 U0 −t /τ
x(t = now)= 204 = + 1 − e 238
Pb f3 Pb0 f3 Pb0
 ( )
206
Pb f1 Pb0 g2 U0 −t /τ
x(t = now)= 204 = + 1 − e 238
Pb f3 Pb0 f3 Pb0
Pb f2 Pbo g1 U0
( )
207
−t /τ
y(t = now)= 204 = + 1 − e 235
Pb f3 Pb0 f3 Pb0
Measure x and y at time t in a lot of samples with variable U0 /Pb0
and plot. Use the x equation to get U0 /Pb0

⎛ f1 ⎞
⎜x− f ⎟
U0 ⎝ 3⎠
=
Pb0 g2
f3
(
−t /τ
1− e 238 )
U0
Substituting for in the equation for y
Pb0
⎛ f1 ⎞
x −
g1 ⎜ ⎟
y=
f2
+ ⎜
f3
(
⎟ 1− e −t /τ 235 )
f3 f3 ⎜ g2
⎜⎝ f
3
(
1− e
−t /τ 238
) ⎟
⎟⎠
Now take the derivative

dy g1 1− e
=
( −t /τ 235
)
(
dx g2 1− e −t /τ 238 )
A prediction is that this derivative at a given time (now)
should be constant for a wide variety of samples with
different uranium contents and histories.
Also we can measure
−t /τ 235
235
U⎞ 1 g1e
= =
238
U ⎟⎠ now
137.8 g2e −t /τ 238
can be used to get g1 / g2
dy
Measure and compute
dx
 dy 25
≈
dx 40

Dalrymple (1986)
http://www.talkorigins.org/faqs/dalrymple/scientific_age_earth.html
dy 25
= =
g1
1− e ( −t /τ
235

=
1 e ) (
−t /τ 238 1− e
−t /τ
235
)
dx 40 g2 1− e (
−t /τ 238
) (
137.8 e −t /τ 235 1− e −t /τ 238 )
25 (
(137.8) = 86 = t /τ
−1e
t /τ 235
)
40 (
e 238 − 1 )
t = 4.5 Gy
Best value
4.55 Gy
The sun as a star
 In a volume limited sample – counting all stars –
low luminosity stars are much more abundant

To 33 ly

The sun

Only a few are more luminous than the sun

Mass-Luminosity Relationship

Main sequence
stars only

d logL
≈ 3−4
d logM
Very roughly
Since luminosity ∝ M3 (actually variable)
and fuel supply = f M (f ~ 0.1 - 0.2)
M 1
Lifetime ∝ 3 ∝ 2
M M
for stars that are spectroscopically main sequence stars

So more massive stars are more luminous

but considerably shorter lived. For example, the
sun lives in its present form 10 Gy but a star of
15 solar masses only lives 11 My.
 12.5 ly
www.atlasoftheuniverse.com

circles
indicate
plane of
Milky Way
galaxy

The nearest (24) stars within 12.5 light years of the

earth. The closest star system – Alpha Centauri–
is about 7000 times the radius of Pluto s orbit.
270,000 times the radius of the Earth s orbit,
 Some specific nearby stars:
• The sun – a typical yellow dwarf star. Type “G2” with 8 planets

• Proxima Centauri – closest of the triplet of stars loosely known as alpha-Centauri

Proxima Centauri is a faint red star that orbits Alpha-Centauri A and B with a
period of about one million years. Proxima Centauri is 4.22 light years from the
Earth (now) and about 0.24 light years from Alpha-Centauri A and B.

• Alpha-Centauri A and B – a double star system with a period of about 80 years. Component
A is a near twin of the sun (Type G2). Component B is a little fainter and orange.
Alpha-Centauri A and B are 4.39 light years from the Earth.

• Barnards star – highest proper motion of all stars. 5.9 light years away. It moves
0.29 degrees per century. In another 8000 years Barnard s star will be the closest
star to us (3.8 ly in 11700 AD). M star, faint, red, about 11 Gyr old. No big planets.

• Lalande 21185 – One of the brightest “red dwarfs” (low mass main sequence stars)
in the sky but still requires binocular to see it. In 1996 a couple of Jupiter sized
planets were discovered here

• Epsilon Eridani – 10.5 light years away. Searched for life by radio searches in the 1960 s.
May have a Jupiter sized planet orbiting at a distance of 3.2 AU. Young star (1Gyr?). K2

• Sirius A,B – At a distance of 8.60 light years Sirius A is the brightest star in the sky.
Sirius B is a white dwarf
 250 light years Number for isotropic
distribution and constant density
n ∝ d3
About 250,000 stars lie
within 250 light years of
the Earth. Beyond this
distance it becomes difficult
to see all the stars in the plane
of the Milky Way
Galaxy because of the
presence of dust.

Only the 1500 most luminous

of these stars are plotted.
Most of these are visible
to the unaided eye.

Note the presence of the

Hyades cluster.

< 1500 stars are visible

to the unaided eye. More
Starting to see some preference for Galactic plane for
distances beyond this. often it s a few hundred.
 The Hyades Open cluster of stars (151 light years)

The bright red star Aldebaran

is not in the Hyades

This cluster of stars is only about 625 million years

old and is in the process of coming apart. Stars like this
are born together from a giant cloud of molecular gas, most of
which is blown away by the young stars. About 200 stars
are catalogued at http://en.wikipedia.org/wiki/List_of_stars_in_Hyades
 Star Clusters

Open Cluster Globular Cluster

 Globular Clusters

47 Tuc M13
Second brightest globular cluster
This globular cluster in
(behind Omega Cen). There are about
Hercules is 22,000 ly distant
200 globular clusters altogether. This
and contains 105 – 106 stars.
one is near the direction of the SMC in
Age ~ 12 to 14 billion years.
the sky and about 20,000 ly distant. Lots
It is about 150 light years across.
of red giants visible here.
5000 light year view – Galactic spiral arm structure
is becoming apparent. The sun is on the Orion Arm
a lesser arm of the Milky Way compared e.g., to the
Sagitarius Arm. There is also a lot of gas and dust.

Betelgeuse 650 ly; Orion 1350 ly

 The entire visible galaxy is about 80,000 light
years across. Note orbiting galaxy and
globular clusters

http://www.atlasoftheuniverse.com/galaxy.html
Andromeda (2.2 Mly away)
220,000 ly in diameter
(larger than Milky Way)
 Mass of the Milky Way

Sun at 28,000 ly orbits the

r center of mass with a speed
256 km s-1

v 2 GMMW
≈ 2
(not exact for a disk)
r r
( )
2
2
v r 2.56 ×10 7
(2.8 × 10 4 )(9.46 × 1017 )
M≈ =
G (6.67 × 10 −8 )
= 2.6 ×10 44 gm = 1.3 ×1011 M
Much more outside the sun's orbit in the form of
"dark matter"
History of the Milky Way
Traditional theory now
somewhat out of date

Rotating cloud collapses into a

disk-like structure

Quasispherical cloud collapses

some part forms stars as the
collapse continues (Population II)

Rest of matter forms centrifugally

supported disk, Population I
stars form
 Population I Population II

Low velocity High velocity perpendicular

perpendicular to disk to disk

Low and high mass Low mass

Young and old stars Old stars

Blue and red Red

e.g the sun e.g. globular clusters

 DETERMINING DISTANCES

For relatively nearby sources, one can

measure distances by surveying -
by measuring the very small angles that
a star s position is displaced relative to
very distant objects because of the
motion of the Earth around the sun. Prior
knowledge of the AU is essential here.

For more distant objects one uses either

standard candles that are calibrated from
nearby sources or a theoretical model.
Stellar Parallax
-Wexler
 2p

For simplicity assume a

star 90o above the
ecliptic
P
⎛ AU ⎞
d= ⎜ if p measured
d ⎝ p ⎟⎠ in radians

AU

1 radian = 360/2"
= 57.296…o
For small angles, p  1, measured in radians
Sin p ≈ p Cos p ≈ 1
AU Sin p
= Tan p = ≈p
d Cos p
But astronomers actually report the angle p in seconds of arc.
1 radian is 360o/2  = 57.296 and each degree is
3600 arc seconds. So 1 radian = 206265 arc seconds. 1
1 arc sec = radians
Thus for p measured in seconds of arc (call it p ), 206265
AU
d=
p (in radians)

206265 AU 1 AU seen from one

d=
p '' parsec away would subtend
an angle of 1 arc second
1parsec
d = p’’ = parallax angle measured in
p ''
seconds of arc
This defines the parsec, a common astronomical measure
of length. It is equal to 206,265 AU s or 3.0856 x 1018 cm.
It is also 3.26 light years.

A little thought will show that this also works for stars whose
position is inclined at any angle to the ecliptic. What p
measures then is the semi-major axis of the parallactic ellipse .
 Hipparcos Space Astrometry
Mission (1989 – 1993)

Catalogue of accurate distances

(1 milli arc s angular resolution)

118,218 stars

Total stars observed (Tycho 2

Catalogue; median astrometric
accuracy 7 mas)

2,539,913 stars

Aside: In AD 150 Ptolemy in his Almagest Including 99% of all stars

catalogued 1,022 of the brightest stars. brighter than 11th magnitude
 Gaia – the successor to Hipparcos – will “survey”
one billion stars
launched Dec 2013 http://blogs.esa.int/gaia/2014/01/
5 year mission

parallax accuracy 20 micro arc seconds for 15 magnitude

200 “ “ “ “ 20 “
7 “ “ “ “ 10 “

7 µ arc sec is a human hair at 1000 km

Distance to 20 million stars determined to 1%

200 10%
100 µarc s =10 −4 arc s
Note implications for a lower bound to
1
the age of the universe... d = −4 = 10 kpc
10
or 31,000 ly
To go beyond distances that can be
surveyed using) parallax, one needs
standard candles .

Even with Gaia, this will continue

to be the case
for most extragalactic objects
 LUMINOSITY AND FLUX

• Luminosity is the total power emitted by a star.

It is measured in ergs/sec. Usually we are speaking
of the luminosity of light, or electromagnetic radiation
of any wavelength. But one can also speak of neutrino
luminosities. A synonym for luminosity is radiant power.

• Flux is a measure of how bright an object appears.

Its value involves both the inherent luminosity
of a source and its distance.

L
ϕ=
4π d 2
If we had a standard candle , a bright
stellar source of known luminosity, LSC, we could
determine its distance from measuring its flux

LSC
φSC =
4π d 2

LSC
d=
4 π φSC
 Measuring Flux: Magnitudes:
• The eye is a logarithmic flux detector
• In astronomy we measure fluxes using magnitudes.
Historically, a magnitude was about a factor of two .

• Calibrated more precisely by William Herschel in the

late 18th century (see also Pogson (1856))

5 magnitudes is defined to be precisely a factor of 100

in flux. One magnitude thus corresponds to a change
in flux of (100)1/5 = 2.512, i.e. (2.512)5 = 100

• A sixth magnitude star is thus 100 times less bright

than a first magnitude star. Larger magnitude is
fainter.
The Pleiades Cluster
 Magnitudes, apparent and absolute
L
According to Herschel s definition, for fluxes 1 and 2: φ=
m2 − m1
4π d 2
φ1
=100 5
φ2
That is, a star 5 magnitudes
m2 − m1 φ
log(100) = log 1 brighter has a flux 100 times
5 φ2 greater.
2 φ
(m2 − m1 ) = log 1
5 φ2
φ1
m2 − m1 = 2.5log
φ2

So, if 1 > 2 , m2 > m1. Keep in mind that bigger m means fainter .

Apparent magnitude, m, is a measure of flux.

 Absolute Magnitude
Absolute magnitude, M, is the magnitude a star would have if located
at a certain distance – 10 pc. Since the distance is the same for
all cases, M is a measure of the star s luminosity.

From these definitions of m and M, we can derive a relation which is

L
essentially the equivalent of φ = ϕ ↔m
4π d 2
L↔M
Consider a star with luminosity L at two distances, d1 = its
real distance = d, and d2 =10 pc. At distance d the star s magnitude
is m1. At 10 pc the star s magnitude is m2 = M. From the previous page:

φ1
m2 − m1 = 2.5log
φ2
⎛ L / 4π d 2 ⎞
M − m = 2.5log ⎜ 2 ⎟
⎝ L / 4π (10) ⎠
 ⎛ 102 ⎞
M − m = 2.5log ⎜ 2 ⎟
⎝d ⎠
= 2.5(2.0 − 2.0 log d )
M measures the luminosity,
m, the brightness, and
M − m = 5.0 − 5.0 log d d is the distance in pc.

For example, the apparent magnitude of the sun is

-26.74. What is its absolute magnitude?
M = m+5−5log d( pc) Here this (pc) just means
that d is measured in parsecs
⎛ 1 ⎞
= − 26.71+ 5−5log ⎜
⎝ 206265 ⎟⎠
= − 21.71− 5(−5.31)
= 4.84
What would be the apparent magnitude of the sun
at 10 pc? At 1.35 pc (distance to -Centauri)?
4.84− m=5−5log d
m= 4.84− 5+5log(1.35)
= − 0.16+ 5log(1.35)= − 0.16+ 5(0.130)
= 0.49
 The 10 brightest stars

Star m M
Sun - -26.74 4.8
Sirius Alpha CMa -1.46 1.4
Canopus Alpha Car -0.72 -5.6
Rigil Kentaurus Alpha Cen (A+B) -0.27 4.4
Arcturus Alpha Boo -0.04 -0.2
Vega Alpha Lyr 0.03 0.6
Capella Alpha Aur 0.08 -0.5
Rigel Beta Ori 0.12 -7.0
Procyon Alpha Cmi 0.34 2.6
Achernar Alpha Eri 0.46 -2.7

m = 0 was historically defined by the star Vega,

though modern readjustments have changed
m(Vega) = 0.03.

Which stars are closer than 10 pc?

 The 10 brightest stars

Star dist(pc) m M
Sun - -26.74 4.8
Sirius Alpha CMa 2.6 -1.46 1.4
Canopus Alpha Car 95 -0.72 -5.6
Rigil Kentaurus Alpha Cen (A+B) 1.3 -0.27 4.4
Arcturus Alpha Boo 11 -0.04 -0.2
Vega Alpha Lyr 7.7 0.03 0.6
Capella Alpha Aur 13 0.08 -0.5
Rigel Beta Ori 260 0.12 -7.0
Procyon Alpha Cmi 3.5 0.34 2.6
Achernar Alpha Eri 43 0.46 -2.7

m = 0 was historically defined by the star Vega,

though modern readjustments have changed
m(Vega) = 0.03.

Which stars are closer than 10 pc?

http://stars.astro.illinois.edu/sow/bright.html
 A Complication:
The Bolometric Correction

Unless otherwise indicated, m is usually the

apparent visual magnitude.
 BOLOMETRIC MAGNITUDE OF THE SUN

Our eyes have evolved to be most sensitive to the

light emitted by the sun. Hence the bolometric correction
for the missed emission in the infrared and ultraviolet
is small for the sun.

The visual magnitude actually corresponds to the flux

measured with a certain filter on the telescope. There are
also blue magnitudes, red magnitudes, and others. We will
discuss this later. For the sun.

M bol = M V − BC = 4.84 − 0.09 = 4.75

(actually 4.755 in 2012)
A similar equation would characterize apparent bolometric
magnitudes, mbol
https://sites.google.com/site/mamajeksstarnotes/bc-scale
Standard Candles
Cepheid variables are large luminous stars with
regular variations in brightness. The variation ranges
from a few per cent to a factor of 5

At 900 light years as judged by Hipparchos

Delta Cephi waxes and wanes with a period of 5 days.
200 Cepheids had their distances measured by Hipparcos.

Polaris, the north star, is also a Cepheid. Its magnitude

varies from 2.0 to 2.1 ever 4 days.
The great merit of Cepheid variables for distance
determination is that there is a clear relation between
the period of the brightness variation and the
average luminosity of the star.

Cepheid variables are also very bright and can be

seen from far away. (They are not main sequence stars).

A complication though is that there are two populations

of Cepheids and they have different period luminosity
relations
IN TERMS OF SOLAR LUMINOSITIES

• There are two populations of Cepheids.

• Type one is the classical type, they are about 4 times
brighter than type 2 and have a high metallicity.
• Type two are older stars with a low metallicity.
 Cepheids
Periods of light variation are in the range 1 to 60 days
and luminosities are up to 40,000 solar luminosities

The surface temperatures are similar to the sun but

the star undergoes regular oscillations in size.

The radial velocity curve is almost a mirror image of the

light curve, i..e., the maximum expansion velocity occurs
at maximum light.

Light variation is in the range 0.5 to 2 magnitudes

and radial velocities at maximum range from 30 to 60 km/s
A Cepheid variable is actually largest
when its brightness is declining and smallest
when it is rising.
 Modern Cepheids

Variable Example where Period Mass Luminosity

(Lsun)
Type I -Cephei disk 1 – 60 d 3 – 10 300 – 40,000
Cepheids

Type II W-Virginis halo 1 - 60 d <1 1.5 mag less

Cepheids globular
than Type I
(W-Virginis stars) clusters

globular <1 d <1 ~100

RR-Lyrae RR-Lyrae clusters

Most stars pass through a Cepheid stage at one time

or another. However the phase is short lived and
only about 1/106 stars are Cepheids at any one time
Cepheid variables are not main sequence stars
 Why do Cepheids Pulsate?
• He+ + light ! He++
• atmosphere becomes more opaque
• radiation pressure pushes the atmosphere
out
• atmosphere cools, He++ ! He+
• pressure decreases, star contracts.

nb. This is an oversimplification. It’s not just opacity - TBD

M100 is 17 Mpc (55 Mly) from the earth
 • With available instrumentation, Cepheids can be
used to measure distances as far as 20 Mpc to
10 - 20% accuracy.

• This gets us as far as the Virgo cluster of galaxies

- a rich cluster with over 1000 galaxies.

M − m = 5 − 5 log(d)
Typical MV for the brightest Cepheids is ~ -5
ST can easly measure fluxes to m = 28
-5 - 28 = 5 - 5 log (d)
log(d) = 38/5 = 7.6
10 7.6 = 40 Mpc
Cepheids play a critical role in bridging distance measurements
in the Milky Way to other “nearby” galaxies
 Type Ia Supernovae
Type Ia supernova in a galaxy 7 billion light years
away (z = 0.5) - Garnavitch et al (1998)
 TYPE Ia SUPERNOVAE ARE
ALMOST STANDARD CANDLES
The width of the light
curve is correlated with
its peak luminosity.
Brighter = Broader

This relation, known as

the Philipp s relation
exists because both the
brightness and width are
correlated with the
amount of radioactivity
(56Ni) each supernova
makes (to be discussed).

Using this correlation,

much of the spread in the
observations can be
narrowed.
It is currently quite feasible to measure supernova
light curves down to a magnitude m = 22. If the
absolute magnitude of a typical Type Ia supernova
is M = -19.5, how far away can we use them as standard
candles for getting distance?
M − m = 5 − 5 log (d )
−19.5 − 22 = 5 − 5 log (d )
−41.5 − 5 = − 5log (d )
−46.5
log(d ) = = 9.3
−5
109.3 = 2 ×109 pc

So, two billion parsecs or about 6 billion light years,

but are the supernovae there (then) the same as the
ones we observe locally (today)?