Lecture 3

Atomic Spectroscopy
and Abundances
Glatzmaier and Krumholz 1
Pols 3.5.1
 C = Balmer alpha
The solar spectrum F = Balmer beta
f = Balmer gamma
B = oxygen
D = sodium
E = iron
H, K = singly ionized calcium
others = Fe, Mg, Na, etc.

Wollaton (1802) discovered dark lines in the solar spectrum. Fraunhaufer

rediscovered them (1817) and studied the systematics.

Note D, H, K stronger tha C, F, f but the sun is not made of Na and Ca,
discovery of He
in the sun
Lockyer 1868
 Quantum mechanics tells us:
Radiation in the form of a single quantum (photon) is emitted
(or absorbed) when an electron makes a transition from one energy
state to another in an atom. The energy of the photon is the difference
between the energies of the two states.
E.g., for hydrogen-like atoms

emission absorption
E m → E n + hν (or E n + hν → E m ) m > n
hc
hν = = Em − En
λ
1 Em − En 2π 2 Z 2 e4 me ⎛ 1 1 ⎞
= = ⎜⎝ 2 − ⎟
λ hc h 3c n m2 ⎠
1 ⎛ 1 1 ⎞
= 1.097 ×10 5 Z 2 ⎜ 2 − 2 ⎟ cm −1
λ mn ⎝n m ⎠
−1
911.6 ⎛ 1 1 ⎞
λ mn = ⎜ − ⎟ Angstroms
Z 2 ⎝ n2 m2 ⎠
(for atoms with only one electron)
 −1
911.6 A ⎛ 1 1 ⎞
λmn = ⎜ − ⎟
Z 2 ⎝ n2 m2 ⎠
E.g.,
m = 2, n = 1, Z = 1
−1 −1
⎛1 1 ⎞
o
⎛3⎞
λ = 911.6 A ⎜ 2 − 2 ⎟ = 911.6 ⎜ ⎟
⎝1 2 ⎠ ⎝4⎠
⎛4⎞ o
= 911.6 ⎜ ⎟ =1216 A
⎝3⎠ Lines that start or
end on n=1 are
m = 3, n =1, Z =1
called the Lyman
−1 −1
⎛1 1⎞ ⎛8⎞ series. All are
λ = 911.6 ⎜ 2 − 2 ⎟ = 911.6 ⎜ ⎟
⎝1 3 ⎠ ⎝9⎠ between 911.6 and
⎛9⎞ o 1216 A.
= 911.6 ⎜ ⎟ =1026 A
⎝8⎠
m = 3,n = 2, Z =1
⎛ 1 1⎞
−1
⎛ 1 1⎞
−1 Lines that start or
λ = 911.6 ⎜ 2 − 2 ⎟ = 911.6 ⎜ − ⎟ end on n=2 are
⎝2 3 ⎠ ⎝ 4 9⎠
−1
called the Balmer
⎛ 5⎞ ⎛ 36 ⎞ o
series. All are
= 911.6 ⎜ ⎟ = 911.6 ⎜ ⎟ = 6563A
⎝ 36 ⎠ ⎝ 5⎠ between 3646 (i.e.
4*911.6) and 6564 A.
 H α , β ,γ ,.....

Lyα , β ,γ ,...
Adjusting the energy of each state in hydrogen by adding 13.6 eV (so that the
ground state becomes zero), one gets a diagram where the energies of the
transitions can be read off easily.
 BALMER SERIES

5→ 2 4 →2 3→ 2

H " H " H "

Hydrogen emission line spectrum
Balmer series
Heavier atoms have more energy
levels and more complex spectra
Stars show absorption line spectra
Absorption Spectra:
provide majority of data for elemental abundances because:
• by far the largest number of elements can be observed
• low fractionation due to convection zone - still well mixed
• well understood - good models available

solar spectrum (Nigel Sharp, NOAO)

Ionization state of the absorbing gas

As the temperature in a gas is raised,

electrons will be removed from atoms by
collisions and interactions with light. The gas
becomes ionized.

The degree of ionization depends on the

atom considered and the density and
temperature. High density favors
recombination and high temperature favors
ionization.
Notation: Ionization stages

HI neutral hydrogen 1p 1e
H II ionized hydrogen 1p 0e

He I neutral helium 2p 2e
He II singly ionized helium 2p 1e
He III doubly ionized helium 2p 0e

CI neutral carbon 6p 6e
C II C+ 6p 5e
C III C++ 6p 4e
etc.
 The ionization energy is the energy required to remove a
single electron from a given ion. The excitation energy is the
energy required to excite an electron from the ground state to the
first excited state.

Excitation energy Ionization energy

Ion
(eV) (eV)

HI 10.2 13.6

He I 20.9 24.5

He II 40.8 54.4
rare Li I 1.8 5.4

Ne I 16.6 21.5

Na I 2.1 5.1

Mg I 2.7 7.6

Ca I 1.9 6.1

Li is He plus one proton, Na is Ne plus 1 proton, Ca is Ar plus 2 protons.

The noble gases have closed electron shells and are very stable.
 Some of the stronger lines in stars

The composition is not varying, just the temperature,

and to a lesser extent, the density
and now L, T, and Y but these are mostly brown dwarfs
Spectral Classes
Main Sequence Luminosities
and Lifetimes
 DISTINGUISHING MAIN SEQUENCE STARS
FROM RED GIANTS OF THE SAME COLOR
The surface gravity GM
g=
R2

of a star is clearly larger for a smaller radius (if M is constant)

To support itself against this higher gravity, a the stellar photosphere

must have a larger pressure. As we shall see later for an ideal gas
P =nkT
where n is the number density and T is the temperature. If two stars
have the same temperature, T, the one with the higher pressure
(smaller radius) will have the larger n, i.e., its atoms will be more
closely crowded together. This has two effects:

1) At a greater density (and the same T) a gas is less ionized

2) If the density is high, the electrons in one atom feel the

presence of other nearby nuclei. This makes their binding
energy less certain. This spreading of the energy level is
called Stark broadening
All 3 stars have the same temperature but,
• The supergiants have the narrowest absorption lines
• Small Main-Sequence stars have the broadest lines
• Giants are intermediate in line width and radius
• In 1943, Morgan & Keenan added the
Luminosity Class as a second classification
parameter:
– Ia = Bright Supergiants
– Ib = Supergiants
– II = Bright Giants
– III = Giants
– IV = Subgiants
– V = Main sequence

And so the sun is a G2-V star

 ABUNDANCES AND IONIZATION
http://www.spms.ntu.edu.sg/PAP/courseware/statmech.pdf
The Boltzman equation derived from MB statistics
describes excitation in a single atom
ns gs exp(− βε s ) 1
= β=
N ∑ gr exp(− βε r ) kT
r

from the Boltzmann equation one can also derive the

Saha equation which describes ionization
http://www.phy.ohiou.edu/~mboett/astro401_fall12/saha.pdf

Ni+1 g 2
( )
3/2
= i+1 2π mekT exp(− χ / kT )
Ni gi neh3

where gi and gi+1 are the partition functions for the two
ionization states of the species i and i+1 and χ is the
ionization potential of the species i (the difference in energy
between states i and i+1). 2 is because the spins of the
proton and electron can be alligned or counteralligned in the
neutral atom
 One often sees terms like (2π mekT) and 1/(neh3 )
Aside: when dealing with quantum statistics. As we
will discuss a bit more next week, there exists a
volume in momentum space, much like in real
space - 4π p 2dp (like 4π r 2dr in spherical 3D
V Δ 3 p ~ h3 coordinates). The ratio 8π p 2dp / neh3 is essentially how
1/ V ~ ne many electrons one can have in that volume (between
p and p +dp) according to the uncertainty principle
Δp 3 ~ h3ne (2 electrons per cell) in phase space times the
volume of one electron V = 1/ne

In summing over all the states in a continuous distribution

of momenta, one gets integrals like
∞ 3/2 ∞
8π 4π
3 ∫
neh 0
p 2 − p 2 /2mkT
e dp = (
neh 3
2mekT ) ∫ x e − xdx
0

where x = p 2 / (2mekT ). The integral is π / 2 so

( )
3/2
2 2π mekT
one gets
neh3
As an example, lets apply these equations to understand
the strength of Ca and H lines in the sun. In the spectrum
the lines of Ca II (Ca+) appear stronger than those of H.
Is the sun made of calcium?

The calculation proceeds in 2 steps and follows GK 1.

We take as given the sun’s photospheric temperature

(5780 K) and electron pressure (this will be discussed
more later) – 15 dyne cm-2.The latter will be used
to provide the electron density, ne, assuming ideal gas
pressure Pe = nekT.

Pe 15 −3
ne = = =1.9 × 1013
cm
kT (1.38 × 10 −16 )(5780)
First is hydrogen neutral or ionized?
n(H II) g 2
( )
3/2
= II 3
2 π m e
kT exp(− χ / kT )
n(H I) gI neh
gII is just the statistical weight of the proton - 1 if it is unbound
gI is the partiction function for the neutral H atom which
in the general case is a bit complicated because one
must assign statistical weights to all the bound states.
From QM each energy state n is 2n2 degenerate (i.e.,
there are n2 levels with different m and l having the
same energy. Further the binding energy of each state
1
is 13.6 eV (1- 2
) so
n
∞
( )
∞
⎛ −13.6eV ⎛ 1 ⎞⎞
∑g = ∑ 2n exp ⎜
− E i −E1 /kT
gI = e 2
⎜ 1− 2 ⎟ ⎟
n=1
n
n=1 ⎝ kT ⎝ n ⎠⎠
but since kT =8.622 ×10 −5 (5780 K)= 0.50 eV, all terms
are negligible except n = 1, the ground state for which
gI = 2. The electron and the proton spins may be alligned or
counter-alligned
 3/2
n(H II) 1 2 ⎛ 2π mekT ⎞
= e −13.6/0.50
n(H I) 2 1.9 ×1013 ⎜⎝ h 2 ⎟⎠
3/2
⎛ 2π mekT ⎞
⎜ h2 ⎟ = 2.41× 1015 T(K)3/2
⎝ ⎠
n(H II)
= 8 ×10 −5
n(H I)
The hydrogen is neutral to a high degree.
Note that this also implies that the number of
free electrons is very much less than that of the
ions and electronic pressure, in the solar
photosphere is a small fraction of the total
pressure.
What about the number of H in the n = 2
level that can absorb Balmer alpha?
The difference in the n = 1 and 2 energy levels
1
is 13.6 eV times (1 - )=10.2 eV. The population
4
factor is then given by the Boltzmann distribution
Nn=2 g2 2(22 ) −10.2/0.50
= exp(−10.2 / kT )= e
Nn=1 gI 2
= 5.5 × 10 −9
a) Less than one hundred millionth of the H atoms are in their
first excited state (many fewer in higher states)

b) The number rises rapidly with increasing temperature

so (Balmer) hydrogen lines should be stronger in stars more
massive than the sun - for awhile.
What about Ca? Experimentally GK say the partition
function for neutral Ca between 5000 and 6000 K
is 1.32 and for Ca II is 2.30. The ratio of singly
ionized Ca to neutral Ca is then
3/2
n(Ca II) 2g(CaII) ⎛ 2π mekT ⎞
= exp(−6.11 eV / kT )
n(Ca I) g(Ca I)ne ⎜⎝ h 2 ⎟⎠
2(2.3) −6.11/0.50
= 2.41×10 15
(5780)3/2
e
1.32(1.9 ×1013 )
= 950
Unlike H, Ca is overwhelmingly ionized. We could also
calculate the ratio of Ca III/Ca II but it is small.
The fraction of Ca that is in the ground state is the partition
function for the ground state, which, it turns out, is 2, divided
by the total partition function for Ca II which is 2.3, so 87%
is in the ground state.
Together these results imply that:

a) Hydrogen is neutral but only 5 x 10 −9 is in

the first excited state and capable of absorbing a
Balmer alpha photon
b) Most calcium is Ca II in its ground state

If the observed line strengths show 400 more Ca II absorbers

than Hα absorbers, the actual abundance ratio of Ca to H is
( ) = 400 ⎛ 5.5 × 10
N Ca −9
⎞
= 2.5 ×10 −6

N(H ) ⎜⎝ 0.87 ⎟⎠

As we shall see the actual modern abundance ratio is 2.1×10 −6

Oscilator strengths?
For the same electron pressure, at what temperature
will hydrogen be half ionized?
Ni+1 g 2
( )
3/2
= 1 = i+1 2π me
kT exp(− χ / kT )
Ni gi neh3
3/2
g 2 ⎛ 2π me ⎞
( )
5/2
= i+1 kT exp(− χ / kT )
gi Pe ⎜⎝ h 2 ⎟⎠
3/2
⎛ 2π me ⎞
⎜ h2 ⎟ = 1.49 × 1039 χ = 13.6eV
⎝ ⎠
kT in erg= 1.602× 10 −12 kT in eV
(kT)5/2 in erg= 3.25 ×10 −30 (kT )5/2 in eV

IF Pe = constant = 15 dyne cm−2

incorrect
(kT )
5/2

(
1 = 1.49 × 10 39
)(3.25 ×10 )
15
−30 eV
e
−13.6/kTev

x 5/2e −13.6/ x = 3.10 ×10 −9 x = 0.724 eV or T = 8400 K

But this underestimates Tconsiderably because Pe will
increase with temperature due to ionization. Note that
bigger Pe (or ne ) means less ionization
Done correctly, ´when H becomes mostly ionized.
ne = n(HII).
3/2
n (H II) 2gII ⎛ 2π mekT ⎞
2
= e −13.6/kT
n(H I) gI ⎜⎝ h 2 ⎟⎠
n(H I) + n(H II) = n(H) = ρ X H NA Let n(H II)=y n(H )
n(H I) = (1− y)n(H) X H = 0.7
Then
3/2
y n(H) ⎛ 2π mekT ⎞
2
= ⎜ 2 ⎟ e −13.6/kT
(1− y) ⎝ h ⎠
3/2
y 2
1 ⎛ 2π mekT ⎞ ρNAkT
= e −13.6/kT Pgas =
(1− y) ρNA X H ⎜⎝ h 2 ⎟⎠ µ
3/2
1 ⎛ 2π mekT ⎞
= (kT )e −13.6/kT
µPgas ⎜⎝ h 2 ⎟⎠
The left hand side is 1/2 if y = 1/2, so we have pretty much the
same equation we had before except with µPgas (µ ≈ 0.6) instead
of Pe
3/2
1 1 ⎛ 2π mekT ⎞
= (kT ) e −13.6/kT
2 µPgas ⎜⎝ h 2 ⎟⎠
As we shall show next time the pressure in the solar
photosphere times µ is a few × 10 4 and this declines as "g"
in heavier stars. Solving as before (and including the .7 for
Habundance and the 33% contribution of electron pressure)

x 5/2e −13.6/ x ≈ 2×10 −6

T≈ 12,000 K

and this is closer to correct, though not very different

because of the exponential. Actually Pgas declines at
higher masses so this is an overestimate.
http://spiff.rit.edu/classes/phys440/lectures/saha/saha.html

*40
http://csep10.phys.utk.edu/astr162/lect/stars/spectra.html
 (Part of) the solar spectrum
Over 100,000 lines are visible in the entire spectrum
70 Elements have been measured
In practice, to obtain abundances, one builds
a model “stellar atmosphere” in which heat flux,
gravity, and hydrostatic equilibrium (TBD) are
given and solves consistently for the ionization,
electron density, pressure, and line strengths.

One of many concerns is the assumption of

“local thermodynamic equilibrium”, that the
level populations are accurately represented
by the Boltzmann equation.

In non-LTE situations one must solve rate equations

to get the level populations.
 Complications for photospheric
abundance determinations
• Oscillator strengths:

Need to be measured in the laboratory - still not done with sufficient accuracy
for a number of elements.
• Line width

Depends on atomic properties but also thermal and turbulent

broadening. Need an atmospheric model.

• Line blending

• Ionization State

• Model for the solar atmosphere

Turbulent convection. Possible non-LTE effects.

3D models differ from 1 D models. See Asplund, Grevesse,
and Sauval (2007).
 Meteorites H. Schatz

Meteorites can provide accurate information on elemental abundances

in the presolar nebula. More precise than solar spectra data in some
cases. Principal source for isotopic information.
But some gases escape and cannot be determined this way
(for example hydrogen and noble gases)

Not all meteorites are suitable - most of them are fractionated

and do not provide representative solar abundance information.
Chondrites are meteorites that show little evidence for melting
and differentiation.

Classification of meteorites:
Group Subgroup Frequency
Stones Chondrites 86%
Achondrites 7%
Stony Irons 1.5%
Irons 5.5%

Carbonaceous chondrites are 4.6% of meteor falls.

 H Schatz
Use carbonaceous chondrites (~5% of falls)
Chondrites: Have Chondrules - small ~1mm size shperical inclusions in matrix
believed to have formed very early in the presolar nebula
accreted together and remained largely unchanged since then

Carbonaceous Chondrites have lots of organic compounds that indicate

very little heating (some were never heated above 50 degrees K)
 Some carbonaceous chondrites smell.
They contain volatile compounds that
slowly give off chemicals with a distinctive
organic aroma. Most types of carbonaceous
chondrites (and there are lots of types)
contain about 2% organic compounds,
and these are very important for understanding
how organic compounds might have formed
in the solar system. They even contain complex
compounds such as amino acids, the building
blocks of proteins.
There are various subclasses of carbonaceous chondrites.
The C-I s and C-M’s are general thought to be the most primitive
because they contain water and organic material. They are fine
grained and contain no chondrules or refractory
inclusions.
The CM meteorite Murchison, has over 70
extraterrestrial amino acids and other compounds
including carboxylic acids, hydroxy carboxylic
acids, sulphonic and phosphoric acids, aliphatic,
aromatic, and polar hydrocarbons, fullerenes, 1969 Australia
heterocycles, carbonyl compounds, alcohols, ~10 kg
amines, and amides.

Five CI chondrites have been observed to fall: Ivuna, Orgueil, Alais, Tonk,
and Revelstoke. Several others have been found by Japanese expeditions
in Antarctica. They are very fragile and subject to weathering. They do
not survive long on the earth’s surface after they fall.
The table on the following page summarizes (Lodders et al
(2009) view of) the current elemental abundances and their
uncertainties in the sun and in meteorites.

The Orgueil meteorite is especially popular for abundance

analyses. It is a very primitive (and rare) carbonaceous
chondrite that fell in France in 1864. Over 13 kg of material
was recovered.
Katharina Lodders, Astrophysical Journal, 591, 1220 (2003)
and 2009 update (will be posted)

Uncertain elements above CNO.

Indium, tungsten, to a lesser extent Tl, Au, Cl, Rb, Hf

Unseen in the sun

Arsenic, selenium, bromine, technetium (unstable),

iodine, cesium, tantalum, rhenium, mercury, bismuth,
promethium (unstable)
Scanning the table one notes:

a) H and H have escaped from the meteorites

b) Li is depleted in the sun, presumably by nuclear

reactions in the convection zone

c) CNO also seem to be depleted in the meteorites

d) The noble gases have been lost, Ne, Ar, etc

e) Agreement is pretty good for the rest – where the

element has been measured in both the sun and
meteorites
per 106 Si atoms
Lodders (2009) translated into mass fractions
h1 7.11E-01 si28 7.02E-04 ti47 2.34E-07 zn66 6.48E-07
h2 2.75E-05 si29 3.69E-05 ti48 2.37E-06 zn67 9.67E-08
he3 3.42E-05 si30 2.51E-05 ti49 1.78E-07 zn68 4.49E-07
he4 2.73E-01 p31 6.99E-06 ti50 1.74E-07 zn70 1.52E-08
li6 6.90E-10 s32 3.48E-04 v50 9.71E-10 ga69 4.12E-08
li7 9.80E-09 s33 2.83E-06 v51 3.95E-07 ga71 2.81E-08
be9 1.49E-10 s34 1.64E-05 cr50 7.72E-07 ge70 4.63E-08
b10 1.01E-09 s36 7.00E-08 cr52 1.54E-05 ge72 6.20E-08
b11 4.51E-09 cl35 3.72E-06 cr53 1.79E-06 ge73 1.75E-08
c12 2.32E-03 cl37 1.25E-06 cr54 4.54E-07 ge74 8.28E-08
c13 2.82E-05 ar36 7.67E-05 mn55 1.37E-05 ge76 1.76E-08
n14 8.05E-04 ar38 1.47E-05 fe54 7.27E-05 as75 1.24E-08
n15 3.17E-06 ar40 2.42E-08 fe56 1.18E-03 se74 1.20E-09
o16 6.83E-03 k39 3.71E-06 fe57 2.78E-05 se76 1.30E-08
o17 2.70E-06 k40 5.99E-09 fe58 3.76E-06 se77 1.07E-08
o18 1.54E-05 k41 2.81E-07 co59 3.76E-06 se78 3.40E-08
f19 4.15E-07 ca40 6.36E-05 ni58 5.26E-05 se80 7.27E-08
ne20 1.66E-03 ca42 4.45E-07 ni60 2.09E-05 se82 1.31E-08
ne21 4.18E-06 ca43 9.52E-08 ni61 9.26E-07 br79 1.16E-08
ne22 1.34E-04 ca44 1.50E-06 ni62 3.00E-06 br81 1.16E-08
na23 3.61E-05 ca46 3.01E-09 ni64 7.89E-07 Etc.
mg24 5.28E-04 ca48 1.47E-07 cu63 6.40E-07
mg25 6.97E-05 sc45 4.21E-08 cu65 2.94E-07
mg26 7.97E-05 ti46 2.55E-07 zn64 1.09E-06.
The solar abundance pattern
256 stable isotopes
80 stable elements
 Lodders et al (2009)

Off scale In, W, Tl, Au, Cl, Rb

Abundances outside the solar neighborhood ? H. Schatz

Abundances outside the solar system can be determined through:

• Stellar absorption spectra of other stars than the sun
• Interstellar absorption spectra
• Emission lines, H II regions
• Emission lines from Nebulae (Supernova remnants, Planetary nebulae, …)
• -ray detection from the decay of radioactive nuclei
• Cosmic Rays
• Presolar grains in meteorites
 but see also Najarro
et al (ApJ, 691, 1816
(2009)) who find solar
iron near the Galactic
center.

From Pedicelli et al. (A&A, 504, 81, (2009)) studied abundances

in Cepheid variables. Tabulated data from others for open clusters.

For entire region 5 – 17 kpc, Fe gradient is -0.051+- 0.004 dex/kpc

but it is ~3 times steeper in the inner galaxy. Spans a factor of 3 in Fe
abundance.
 Variation with metallicity
Kobayashi et al, ApJ, 653, 1145, (2006)

“ultra-iron
poor stars”

SN Ia
Integrated yield of 126 masses 11 - 100 M  (1200 SN models), with Z= 0,
Heger and Woosley (2008, ApJ 2010) compared with low Z observations
by Lai et al (ApJ, 681, 1524, (2008)). Odd-even effect due to sensitivity
of neutron excess to metallicity and secondary nature of the s-process.

28 metal poor stars in the Milky Way Galaxy

-4 < [Fe/H] < -2; 13 are < -.26