Ch 12

Experiment 6: Atomic
Spectroscopy

The formulation of quantum mechanics in the Schr¨odinger equation is beau-

tiful and elegant. Unfortunately, however, their are very few problems that can
be solved exactly (more or less) which correspond to physical systems that
actually exist.

The hydrogen atom, or more exactly, atoms with a single electron, are
probably the best example of a true, measurable, physical system in which the
beauty of quantum mechanics can be tested without resorting to ap-proximation
techniques and models. In fact, many of the early successes of quantum
mechanics came in the study of atomic spectroscopy, with hydrogen providing
some of the most crucial tests of the theory.

Some other solvable problems with the Schr¨odinger equation are the har-
monic oscillator, and the rigid rotor. Although ideal cases of these things are
hard to nd in nature, good approximations are provided by the study of diatomic
molecules.

The energy levels involved in atomic physics typically give rise to transi-
tions in the few eV range. Consequently, the photons that are emitted are in the
visible region, and our studies of light production and detection will prove to be
useful. In this experiment we will actually use high resolution

197
198 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

spectroscopy to study the photon wavelengths to high precision. For exam-ple,

the wavelengths are slightly di erent in the deuterium atom as opoosed to
hydrogen because of the larger nuclear mass. You will be able to resolve that di
erence.

The physics surrounding atomic and molecular spectroscopy is a large eld.

For a good general reference, I suggest

Introduction to the Structure of Matter,

John J. Brehm and William J. Mullin, John Wiley and Sons (1989),
Chapters 3, 7, 9, and 10

Quantum Physics, Robert Eisberg and Robert Resnick,

John Wiley and Sons, Second Edition (1985), Chapters 4
and 7

The experiments we will do are more or less standard in the undergraduate

laboratory, and there are several good books available which describe such
experiments. A few of these are

The Art of Experimental Physics,

Daryl W. Preson and Eric R. Dietz, John Wiley and Sons (1991)
Experiments 12 and 13

Physical Chemistry; Methods, Techniques, and Experiments

Rodney J. Sime, Saunders College Publishing (1990)
Experiments 30-33.

Experiments in Modern Physics,

Adrian C. Melissinos, Academic Press (1966)
Sections 2.1 through 2.4.

The book by Sime is particularly useful for us since the experiments use the
same spectrograph as we have in our laboratory.
12.1. ENERGY LEVELS OF THE HYDROGEN ATOM 199

12.1 Energy Levels of the Hydrogen Atom

We'll start with the Bohr model of one-electron atoms. For most of the
measurements you will make, this model correctly predicts the results, and in
fact gives the same answers as the Schr¨odinger equation. This approach is
actually worked out in detail by Brehm and Mullin.

We will derive the energy of the atom, and impose simple quantization rules
on the result. Assume for now that the nucleus of the atom is in ntely heavy
compared to the electron, and that the electron moves nonrelativisti-cally. We
will examine these assumptions soon, but they are in fact quite good in general.
The total mechanical energy of the electron is

E = K+V
2
1 1 Ze
2
= 2 mv − 4 "0 r (12.1)
Here m, v, and −e are the electron's mass, velocity, and electric charge, +Ze is
the charge on the nucleus, and r is the \orbital radius" of the electron. The
potential energy, of course, is just the Coulomb attractive potential between the
electron and the nucleus.

Before we get into quantization, let's work with this equation a bit. We can
relate the velocity v to the other variables by applying F = ma, where
F is the Coulomb force and a is the centripetal acceleration. That is
2
1 Ze = m v2
2
4 "0 r r
2
which implies that Ze
2 1 1
v
= m 4 "0 r (12.2)
If we plug this into Eq. 12.1 we get
2 2 2
1 1 Ze − 1 Ze = − 1 1 Ze (12.3)
E= 2 4 "0 r 4 "0 r 2 4 "0 r
Now we can impose quantization rules on the energy just by considering the
implications for the orbital radius r.
200 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

Bohr's quantization rules are simple and elegant. The electron is also a wave,
with wavelength h=p = h=mv. As the electron moves around in its orbit, the
path length must be such that the head of the wave \links up" with the tail and
the wave pattern keeps repeating itself. If this wasn't the case, then the wave
would interfere with itself as the electron orbited, and the wave would
disappear, and so would the electron.

Bohr's quantization condition is, therefore, that the circumference of the

orbit be an integral number n of wavelengths of the electron. That is, 2 r =
n(h=p) or
r=n h (12.4)
mv
where h h=2 . This actually has a deeper physical signi cance. The quantity rp is
the angular momentum l of the electron, so this equation in fact says that l = nh,
that is, the angular momentum is quantized. This is really just the beginning of a
very interesting story about angular momentum and quantum mechanics, but we
won't cover it here.

Combining Eq. 12.4 with Eq. 12.2 gives us

n2h2 1 1 Ze2
=

and therefore
1 = m 1 Ze2
r n2h2 4 "0
which is a quantization relation for r that does not depend on v. Finally, insert
this expression in Eq. 12.3 and get
1 1 m 1
2 2 2 2
E = − 2 4 "0 Ze n h 4 "0 Ze
2 4
mZ e 1 (12.5)
2 2 2
= − " 2(4 "0) h #n

These are the quantized energy levels of the hydrogen atom. That is, the electron
can only have energies described by this formula with n = 1; 2; 3; : : :. These
energy levels are plotted in Fig. 12.1. Notice in particular that the energy
levels get closer and closer together near the top of the potential well.
This pattern will be clearly apparent in this experiment.
12.1. ENERGY LEVELS OF THE HYDROGEN ATOM 201

s
s
E=0
s n=6, E=-0.38 eV
s
n=5, E=-0.54 eV
n=4, E=-0.85 eV
n=3, E=-1.5 eV
???? n=2, E=-3.4 eV
H H HγH

n=1, E=-13.6 eV

Figure 12.1: Energy levels of the hydrogren atom, and the transitions which
make up the Balmer series of visible wavelength lines. Transitions to the n = 1
state are also possible, and these lines are called the Lyman series, but the
wavelengths are all in the far ultraviolet.
202 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

For the hydrogen atom (Z = 1), the expression in brackets for Eq. 12.5 works
out to be 13.6 eV. This is the energy required to take an electron in the ground
state (n = 1) and separate it from the nucleus completely (E = 0). That is, this is
the binding energy of the hydrogen atom. For atoms with more electrons, you
imagine that the energy needed to separate the \outermost" electron from the
nucleus and the Z −1 remaining electrons is about the same value, since the
other electrons \shield" the outermost electron from all but unit of charge on the
nucleus. In fact, the energy needed to separate one electron from an atom (the
\ionization potential") is pretty close to 10 eV for most atoms.

So what's the experiment? We need to add one more ingredient, namely

transitions between the energy levels. If you prepare an atom in one of the
excited states (n > 1), then the electron will make transitions down to the lower
states. Each transition emits a photon of energy h which must equal the di erence
in energy between the initial and nal states. If the electron starts from the energy
level with n = ni and ends up with n = nf then
h= E E
nf − ni

We use Eq. 12.5 to rewrite this expression, but the result is usually expressed in
terms of the photon wavelength instead of the frequency = c= . You
nd
1 1 1
2 2 2
= Z R1 nf − ni ! (12.6)
where the Rydberg constant R1 is de ned as
4
me
R1 8"02ch3 (12.7)
The subscript \1" refers to the assumption that the nucleus is in nitely heavy.
The Rydberg constant is actually known quite accurately. It's value is

−1
R1 = 10973731:534 m
Note that the hydrogen binding energy is just hcR1 = 13:6 eV.
You will measure several of these wavelengths in this experiment. You
should note, however, that only a few of the many combinations are ac-cessible,
because you will be detecting essentially visible photons. (See Ta-ble. 11.1.) In
fact, all of the transitions you will observe correspond to nf = 2,
12.1. ENERGY LEVELS OF THE HYDROGEN ATOM 203

and the series of lines corresponding to ni = 3; 4; : : : is called the Balmer se-

ries.1 The longest wavelength line (the 3 ! 2 transition) is red, but all the others
are blue or violet. These transitions are also shown in Fig. 12.1.

12.1.1 Corrections

There are various corrections due to the simple Bohr formulas. You will in-
vestigate one of these in particular, namely the e ect of a nite nuclear mass, and
it can be evaluated using a straightforward extension of the simple rules. The
other corrections actually led to the breakdown of the Bohr formula, and the
development of quantum mechanics based on the Schr¨odinger equation. You
won't likely measure these e ects in this experiment, but we will at least mention
the physics here.

Finite Nuclear mass

The proton mass mp is much larger than the electron mass, in fact mp=m =
1836, and all the other nuclei are even heavier. Therefore, the \in nite nu-clear
mass" assumption is a pretty good one. On the other hand, optical spectroscopy
experiments can be very precise (look at all the signi cant g-ures on R1), so you
might think you could see the e ect of a nite nuclear mass in this experiment. In
fact, you can, and you will measure the dif-ference in spectra for hydrogen
2
(nuclear mass mp = 938:3 MeV/c ) and for deuterium (nuclear mass md =
2
1875:6 MeV/c ).

The nite nuclear mass means that instead of the electron revolving around
the nucleus, both the electron and the nucleus revolve about their common
center of mass. The quantity r still refers to the distance between the electron
and the nucleus, but the orbit radii are actually re = M r=(m + M) and rM =
m=(m + M) for the electron and nucleus respectively. Since the nuclear mass M
is much larger than the electron mass m, re r and rM =re << 1, which is the
essence of the in nite mass approximation.
1
The other series corresponding to other values for nf are named after the other fellows who
discovered the lines, but none of these lines are in the visible.
204 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

We now quantize the energy as before, but we need to write E including the
kinetic energy of the nucleus, and form the quantization condition using the
total angular momentum which includes the nuclear contribution. The details are
worked out in Brehm and Mullin, and the result is that Eq. 12.6 is unchanged
except that the Rydberg constant R1 is replaced by
R
RM = m 1 (12.8)
where M m=(M + m) is called the reduced mass. The key to measur-ing this e
ect is to be able to measure the small di erence in wavelengths corresponding to
Rp and Rd for hydrogen and deuterium.

Relativistic E ects

By the time the Bohr model was being developed in 1913, people were pretty
much convinced that Einstein's theory of special relativity was right. Trying to
incorporate relativity into Bohr's model became a problem, and eventually led to
its downfall. Let's rst estimate how big a problem you expect this to be.

Relativity should become important when the electron speed v approaches

the speed of light c, so let's evaluate v=c. Substituting Eq. 12.4 into Eq. 12.2,
you nd
m
v 1 1 2
= Ze =Z

c m 4 "0 nhc n
where
1 e2 1
4 "0 hc = 137:036 (12.9)
is called the ne structure constant. In other words, for the hydrogen atom, the
electron speed is always less than 1% of c, so you wouldn't expect rel-ativistic
corrections to be very much larger than this. Of course, in heavy (i.e. large Z)
one-electron atoms, the velocity can actually get quite large compared to c, and
you can't expect the Bohr formula to work well.

I want to take a moment to talk about . Just about any physicist you know
can tell you that \ is around 1/137". It is a very fundamental quantity that you
will encounter more and more as you study physics. At
12.1. ENERGY LEVELS OF THE HYDROGEN ATOM 205

the very least, it helps you remember some of the formulas we've derived. For
1 22 2
example, the binding energy of a one-electron atom is just 2 Z mc , and the
1 2
Rydberg constant R1 = 2 (mc=h).
It is straightforward to incorporate the e ect of special relativity into the one-
electron atom, once you've solved the Schr¨odinger equation for the problem
without relativity. This is called perturbation theory, and the pro-cedure is
outlined in Preston and Dietz, as well as other places. The result is that the
individual lines are \split" according to the angular momenta of the initial and
nal atomic states. This splitting is called ne structure, and
it amounts to around = 0:1 A in the Balmer series.

Spin-Orbit Splitting

Relativistic corrections are only one contribution to the ne structure split-tings.

The second contribution, which is about the same size as from rela-tivity, is due
to the spin-orbit interaction.

The electron has some internal angular momentum we call spin. This
internal angular momentum shows up as a magnetic dipole moment on the
electron. In other words, the electron is like a tiny bar magnet. In addition, there
is the electron \current" due to the electron orbiting about the nucleus. This
current sets up a magnetic eld all around the atom, and the electron spin interacts
with this magnetic eld. This is the spin-orbit interaction, and
~

it is essentially of the form ~

~
B where ~ is the spin dipole moment of the
electron, and B is the magnetic eld set up by the electron orbit.
~
Depending on the relative orientation of ~ and B, which is quantized
according to quantum mechanics, the magnitude and sign of the spin-orbit
interaction will be di erent for di erent angular momentum states. This leads to
the spin-orbit splitting. Once again, it is about the same size as
relativistic e ects, i.e. = 0:1 A or so.
206 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

Higher Order Corrections

As you look at the spectral lines with higher and higher resolution, you discover
more (and smaller) splittings. Each of these has physics associated with them,
and some of this physics is very profound.

One example is the so-called hyper ne splitting. (See Brehm and Mullin,
Sec. 8-12.) This is the interaction between the spin magnetic moments of the
electron and the proton, for the hydrogen atom, or for the nucleus in gen-eral
assuming it has a nonzero spin. The splittings caused by the hyper ne interaction
in hydrogen are on the order of 0.01 A. Transitions between hy-per ne levels are
in the radio frequency range, and a particular transition in hydrogen ( = 21 cm)
is famous to radio astronomers who use it to identify hydrogen in hard-to-see
regions of the galaxy.

One of the most profound e ects in atomic physics is the Lamb shift, named
for Willis Lamb who discovered it in 1947. Up until that time, a synthesis of
quantum mechanics and special relativity, written down by Paul Dirac, was able
to correctly predict all the structures observed in the hydro-
gen spectrum. The Lamb shift, an unexpected splitting at around the 0.01 A
level (relative to optical spectroscopy), was inconsistent with Dirac theory. It's
solution turned out to hinge on a new formulation that explicitly includes the
radiation eld along with electrons, called Quantum Electrodynamics.

12.2 Measurements

Your rst goal is to determine the Rydberg constant from the Balmer series in
hydrogen. You will do this by measuring the Balmer spectral lines from atomic
hydrogen, and tting the wavelengths to Eq. 12.6. Next, you will calculate the
deuteron to proton mass ratio by measuring the isotope shift between the lines of
atomic deuterium and hydrogen. You need to make precise measurements of the
wavelengths to do all this.

Light from the various elements is produced in a gas discharge tube.

Wavelength measurements are made in either one of two ways. We have a
12.2. MEASUREMENTS 207

Baird model SB-1 1.5 m grating spectrograph, which exposes a strip of 35 mm

lm to a dispersed spectrum. A computer controlled table scans the lm and
measures the grain density as a function of position. The second method uses the
Jarrell{Ash 1 m Czerny-Turner Scanning Spectrometer which scans through
wavelengths on command. Data is taken directly from a photomul-tiplier tube as
wavelengths pass over a slit as the grating is turned. Results from the two
techniques will be compared in Fig. 12.8. Note that the lm is darkened at a line
position, so less light passes through and the scanner gives a smaller signal when
crossing a line. The Jarrell{Ash, however, detects the light directly and the
electronics provides something proportional to the light intensity.

The light source works by placing a very high voltage across the ends of the
discharge tube, and you will get a nasty shock from it if you are not careful.
Always make sure the voltage is o when you change tubes. Be careful not to
move the source or any other part of the apparatus while you change discharge
tubes, since a small change in position can have a big e ect on where the image
ends up on the lm. Try never to look directly at the discharge tube when it is on,
because some of the tubes have rather intense UV light that you cannot see.

12.2.1 Procedure: Baird Spectrograph

A diagram of the Baird spectrograph is shown in Fig. 12.2, and an expanded

view of the lm window and slit assembly is shown in Fig. 12.3. The discharge
tube is controlled through a timer that lets you preset the exposure time. Light
from the discharge tube enters through the slit and is focussed onto the concave
di raction grating in the rear. The grating disperses light according to wavelength
and focusses it onto the lm holder adjacent to the entrance slit. You can adjust
the vertical position of the lm image by moving the position of the Hartmann
slide, and you can put six or so di erent exposures on the same piece of lm. It is
also possible to run a HeNe laser beam through the spectrograph for calibration
purposes. Furthermore, there is a \reference line" you can superimpose on the lm
by pressing a button near the lm holder.
208 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

Film Light dispersed by wavelength

holder

Grating
Light
source
Slit and shutter assembly

Figure 12.2: Diagram of the Baird SB-1 Grating Spectrograph.

Spectrograph housing

Clamps Slit Assembly

Film Holder
Shutter
lever Shutter

Hartmann slide

Figure 12.3: Diagram of the window and slit assembly.

12.2. MEASUREMENTS 209

The wavelength range is 370 nm to 740 nm, pretty much covering the visible
spectrum. The dispersion at the lm focus is around 15A/mm, but you will
determine this yourself by calibrating the spectrograph. You can set the slit
width to 10 m, 32 m, or 60 m, and to get the best resolution, you may need to
experiment with di erent widths. The shutter can be operated manually, and you
should keep it closed if you are not exposing the source, even if the lights are o .

Before exposing lm, you should play around with the roughed-up plexi-glas
sheet that ts in the lm holder. You can see the lines with some care, and you
make adjustments of the slit, Hartmann slide, source position, and so on without
having to actually develop lm.2 A good source tube to use is neon, which gives
you many bright lines mainly in the orange and red. If you use the mercury tube,
you will see a strong yellow doublet at 577 nm and 579 nm, and a relatively
bright green line at 546 nm.

The lm holder uses 10-inch strips of standard 35-mm photographic lm, cut
from a 100 foot roll. Kodak Tech-Pan is a good lm to use, and that is what the
laboratory should be stocked with. This lm is sensitive to all wavelengths, so the
room must be in complete darkness when you are handling the lm. With the
lights out, remove a length of lm from the bulk lm winder, and slide it into the
lm holder. The concavity of the lm should match the concavity of the holder.
Place the lm holder over the window, and secure it with clamps. At this point,
light from the discharge tubes may be exposed to the lm through the shutter.
Spend a lm strip or two making measurements with the di erent slit widths, and
varying the exposure time between a few seconds and several minutes. After you
develop the lm and look at the result, you will be better able to judge the slit
widths and exposure times for your \ nal" set of data. Hint: Cut the lm a bit
long so an inch or so hangs out past the blue end of the holder. It will be
fully exposed and a good tag of which end is which.

Remember that you can move the exposure vertically on the lm using the
various settings of the Hartmann slide. You likely want to include some or all of
the following discharge tube combinations in your exposures:

2
To quote the users manual for the spectrograph, the Fixed slit Assembly is a very delicate
component. Handle it carefully. If you would like a copy of the users manual, please ask me or
the TA.
210 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

Helium+HeNe Laser
Hydrogen+HeNe Laser
Mercury+HeNe Laser
Hydrogen+Helium
Hydrogen+Mercury
Hydrogen+Deuterium
Nitrogen+HeNe Laser

The helium and mercury lines will be used to calibrate the spectrograph, that is,
to convert position on the lm to wavelength. The HeNe laser serves
a similar purpose, giving you a single strong line at 6328 A, and it might be a
good idea to add this line to all your exposures. You might also want to press the
reference line button for a moment, to help you orient the lm after you've
developed it. The hydrogen lines will allow you to determine RH , and the
simultaneous exposure of hydrogen plus deuterium allows you to measure the
small di erence between these two isotopes, although you will need pretty good
resolution to cleanly separate the two. This is one of the things to aim for when
you practice in the beginning. Nitrogen will give you a very complicated but
interesting spectrum that you can analyze to learn about the diatomic nitrogen
molecule. (See Sec. 12.3.)

Developing the lm

The chemicals you need to develop the lm should all be premixed for you, but
speci c directions are available in any case. You don't want to get a lot of these
chemicals on your skin, so handle them with some care. Remember, all
developing must be done in total darkness. We use HC-110 to develop the Tech-
Pan lm, and it is a good idea to make sure you are using a new (i.e. less than a
year old) bottle of it when you are developing.

Set out four developing tays, each containing around a half gallon or so of
developer, stop bath, xer, and photoflo, respectively and in that order. Place the
lm in the developing tray and agitate for 8-10 minutes at room temperature. This
may be longer or shorter if the temperature is cooler or
12.2. MEASUREMENTS 211

warmer, but the darkroom stays at a pretty even temperature. Dip the lm in the
stop bath for 30 seconds to a minute, and then put the lm in the xer and agitate it
for another 8-10 minutes. Rinse the lm in photoflo and hang it to try. You can
take a quick look at it using the light table in the developing room.

Scanning the lm

There are three ways you can scan the lm to measure the line position. One is
simply to use a ruler, although the experimental uncertainty will be pretty large.
A second possibility is to use the traveling microscope setup, equipped with a
vernier scale to make very precise measurements. The third method uses a
computerized lm scanner which measures the grain density by shining light
through the lm into a microscope and onto a slit, behind which is photodiode.
The lm holder moves by computer control through a stepper motor, and the
photodiode is read out at each step. The program lets you scan in \high
resolution" (small steps) or \low resolution" (big steps).

The computerized scanner gives you a a digital record of the density all
along the lm. The data is merely a list of numbers corresponding to the density
at each step. It should be simple to identify the peak position, or plot the peak
shape using matlab. The width of the peak is a measure of the resolution of the
instrument unless there is some narrow structure (\splitting") underneath, and we
will try to get some physics out of that.

One disadvantage of the computerized scanner is that you cannot move the
lm vertically in a precise way, so it could spoil the calibration (see below) if you
are interested in more than just the dispersion. If, for example, the HeNe line is
placed on each exposure, that will make your life a lot easier.

Try to include the HeNe line, or some other reference, in the scan. It is
already hard to get all the Balmer lines in one scan, since you are limited by the
range of the scanner, but get as many as you can along with a reference line. It is
probably best to include the red line and the HeNe (or other reference) line, the
green line, and as many blue lines as you can.
212 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

0.28

0.27

0.26

PHOTODIODE VOLTAGE SIGNAL

n=7
0.25

0.24 HeNe

0.23 n=6

0.22 n=5

0.21 n=4

0.2 n=3

0.19

0.18
0 20 40 60 80 100 120 140 160 180 200
SCAN POSITION ALONG FILM (mm)

Figure 12.4: Result of a low resolution computerized scan of a hydro-gen+HeNe

exposure. Because of the limited travel of the scanner, two scans are
superimposed by matching the H line position.

3
A low resolution scan of a hydrogen plus HeNe exposure is shown in Fig.
12.4. The rst ve Balmer lines are clearly visible, as is the HeNe line used for
calibration. Instead of the HeNe line, one could include a mercury or helium
exposure so that calibration lines are scattered throughout the scan.

12.2.2 Procedure: Jarrell{Ash Spectrometer

Use the Jarrell{Ash scanning spectrometer to measure wavelengths just as you

do for the Baird spectrograph. That is, use the same discharge tubes in the same
or similar combinations. The Jarrell{Ash has an advantage that it is much higher
resolution, so you can see much more ne detail. It is, however, a bit trickier to
use.

Figure 12.5 shows a schematic of the spectrometer. The light source sits

3 Data taken by Marc Crudele, Class of 1996.

12.2. MEASUREMENTS 213

Light
Source
Top View Mirror

Insertable mirror PMT

Viewport

Control Panel
with
Wavelength
Side View Rotatable
Dial
Diffraction
Grating

Figure 12.5: Schematic of the Jarrell{Ash scanning spectrometer. The di raction

grating is rotated either with the dial on the control panel, or remotely through
the computer control. You can look at the spectral lines with your eye, by
inserting the flat mirror and looking at the light on the viewport.

in the front of the device, and light enters through an adjustable slit. You can
place one of a variety of wavelength lters in front of the slit, to block out light
from other sources and from di erent directions. Light travels to the back of the
spectrometer, reflects from a curved mirror which creates a parallel beam that
travels to the di raction grating. The grating disperses the light according to
wavelength, and sends it back to the mirror, which reflects it back to the exit slit
where it is detected by the photomultiplier tube. The tilt angle of the grating
determines the wavelength which passes through the slit in front of the PMT. By
rotating the grating, you scan over wavelengths.

This spectrograph operates in \second order". In other words, the light for a
particular wavelength is deflected through an angle 2 = 2 =d instead of the \ rst
order" angle 1 = =d. (There are technical reasons why you might want a
spectrometer to operate in second order as opposed to rst.)
214 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

This can lead to some confusion, though, because light of shorter wavelengths
can show through from higher orders, and there are plenty of ultraviolet lines
around. For example, suppose you are studying visible light of wavelength 0,
observed when the grating angle corresponds to 2 = 2 0 =d. Then, ultraviolet
0
light of wavelength = 2 0 =3 will come out at the same position, through third
order in the spectrometer.

Remove this confusion by using optical lters between the light source and
the input slit. There are two ultraviolet/blue lters available. The 345 nm lter
removes all light below this wavelength, so your spectrum is \protected" from
345 nm up to (3=2) 345 = 518 nm. Using the 455 nm lter you can observe from
455 nm to 683 nm. Using a combination of these two lters, you can therefore
cover the full visible spectrum, from violet through red.

You will take your data by letting the computer turn the grating and record
the signal level on the PMT. However, it is a good idea to rst set things up by
looking at the spectral lines by eye. You can do this by inserting the mirror near
the PMT (using the \plunger" that comes out the back) which diverts the light to
the side. You can then open up the shutter attachment to look into the
spectrometer through the ground glass viewport. Make sure the PMT is
turned o when you open up the shutter! With the lights o , you can see the
faint, but sharp, lines of the particular light source you are using. This will help
you identify them in the PMT signal.

If you next close the shutter and take the insertable mirror out of the beam,
then the light will again pass through the slits into the PMT. Turn the PMT on
using the Keithley Model 247 High Voltage power supply. Make sure the power
supply is set to zero volts and negative polarity, then switch it on, and turn the
voltage up to around 500V. The white power cord coming out of the back of the
spectrometer powers the ampli ers for the PMT and for the grating stepper motor
control. Plug it in. Turn the grating with the hand crank, and measure the voltage
signal out of the ampli er with a voltmeter. As the spectral line passes over the
slit, you will see a clear increase in the voltage signal. You should \see" the lines
this way, just as you could see them with the ground glass plate.

Now get ready to read the signal out with the computer, controlling the
grating at the same time. Turn on the computer in the back of the room,
12.2. MEASUREMENTS 215

to the left, between the wall and the spectrometer. Start labview from its le
folder in the program manager when the computer comes up. When a
registration information dialog box comes up, press \OK", and then press \OK"
again when another dialog box comes up. You will now see a blank \vi" (virtual
instrument) appear as a grey window. Close this window, and now a dialog box
comes up asking for one of three options, namely \new vi", \open vi", or \close
labview". Choose \open vi" and in the \ le open" window, double-click on
ALPHSCAN.LLB, and then on \visible spectrometer". The program is now
ready to go.

Adjust the slit width using the micrometer dial on the front of the spec-
trometer, and set the starting wavelength position using the hand-crank on the
side control panel. Enter the starting wavelength you set on the spec-trometer
into the \start wavelength" eld on the program, and enter the nal wavelength (on
which the scan will end) in the \ nish wavelength" eld. Click on the resolution
button to read either \high" resolution (about

0.04 A/step) or \low" resolution (about 1 A/step). (Low resolution is the best
choice when you're rst starting out, otherwise it will take a long time to see if
you've got things set up correctly.)

When you are ready to start the scan, press the arrow button in the top left
corner of the vi. The scan will continue to the \ nish" wavelength previously set,
and then prompt you for a le name to which you can save the data. If you want
to end the scan before the nish wavelength, press the large STOP button in the
window, not the small octagonal button next to where the start arrow button
used to be. If you press the small octagonal button instead of the large STOP
button, you will be unable to save the data you just took. Also, the program will
lock up the stepper motor so that you will be unable to move the hand-crank.

If for any reason you are unable to move the hand crank, dont force
it!!. Press the square black reset button on the alpha stepper board to
release the hand crank.

The data that you get will be an intensity as measured by some voltage
output from the ampli er circuit on the PMT, as a function of the nominal
wavelength set on the dial. The nominal wavelength is just a conve-nient scale
that is close to the true wavelength. You will have to
216 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

1.5
345 nm filter
1

0.5

0
3000 3500 4000 4500 5000 5500 6000 6500 7000

2.5

2
455 nm filter
1.5

0.5

0
3000 3500 4000 4500 5000 5500 6000 6500 7000
Wavelength dial value

Figure 12.6: Scans of the helium discharge tube using the Jarrell-Ash spec-
trometer and the two ultraviolet/blue lters. Note that the horizontal axis is the
nominal wavelength, and must be recalibrated.

determine the true wavelength by using your calibration spectra. If you pre-fer,
you can just plot the data as a function of \step" number, instead of using the
nominal wavelength, to keep yourself from getting confused.

Figure 12.6 shows two scans of the helium discharge tube using the Jarrell-
Ash spectrometer. One scan uses the 345 nm lter, while the other uses the
455 nm lter. Note that the horizontal axis is the wavelength in Aas read o the
dial, by the computer. The existence of lines well below 3450 in the scan
using the 345 nm lter is clear evidence that the wavelength scale needs to be
recalibrated!

12.2.3 Analysis

There are many levels of analysis you can perform, exploring not only the
physics of atomic spectra, but also the operation of the spectrograph. The rst
thing to realize is that the spectrograph is designed so that the line posi-
12.2. MEASUREMENTS 217

tion on the lm corresponds to the wavelength through a linear relationship, that

is,
= Ax + B
where x is the position. The quanitity A is called the \dispersion", and for the
SB-1 it is supposed to be around 15A/mm. For the Jarrell{Ash, it will depend on
the step size you chose in the program.

You might want to try something very simple rst. The relative spacing of the
Balmer series lines must be given by Eq. 12.6, with nf = 2. That means that you
can just roughly measure the distance between pairs of lines,
and check that the ratio of any two distances follow this equation. You can't

1 1

get any physics out of this because everything cancels except the 4−
2
ni

terms, but at least you can check that you can make the ni assignments correctly.

You can go a little further and check that you are getting around the right
value for RH by using the nominal value for the dispersion. These are things you
can do just as you get the data, but to do a more careful job, use scanned data to
get the line positions. In this case, you will want to use the helium and mercury
lines for calibration.

Precise values for the wavelengths using the helium and mercury discharge
tubes are listed in Table 12.1. The values are the \wavelengths in air" as
tabulated in the MIT Wavelength tables, by G.R Harrison, et.al. (1969). You can
nd this book on the main floor reference section of the library under REF
QC453.M36 1969. These tables also give the relative intensities of the lines as
produced in a discharge tube.

Use the helium and/or mercury data to determine A and B. Determine the
uncertainties in A and B from your t, and propagate these uncertainties through
to the physics quantities you derive. How well does A agree with the
manufacturers speci cation? Plot the deviations from the linear t to see if there is
4
any evidence for nonlinearity. An example of a helium calibration done on the
Baird spectrograph is shown in Fig. 12.7. Do you think it would be worthwhile
2
to include a term proportional to x in the calibration? How large you expect the
deviations to be for a calibration of the Jarrell{Ash
4
Data taken by Marc Crudele, Class of 1996.
218 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

Table 12.1: Some prominent lines for calibrating the spectrograph. Wave-
lengths are in A.
Helium Mercury Helium Mercury
4026.2 Violet Blue 4916.0
Violet 4046.6 4921.9 Green
Violet 4077.8 5015.7 Green
Violet 4339.2 Green 5460.8
Violet 4347.5 Yellow 5769.6
Violet 4358.4 Yellow 5790.7
4387.9 Violet 5875.6 Yellow
4471.5 Violet 6678.1 Red
4713.4 Blue Red 6291.3

0.25

0.2
FIT WAVELENGTH - ACTUAL WAVELENGTH (A)

0.15

0.1

0.05

-0.05

-0.1

-0.15

-0.2

-0.25
4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000
WAVELENGTH OF HELIUM CALIBRATION LINE (A)

Figure 12.7: Result of a calibration using a helium spectrum on the Baird

spectrograph. The helium wavelengths are tted to a straight line as a func-tion of
scanner position. The deviation of the tted wavelength from the actual
wavelength is plotted versus the wavelength.
12.2. MEASUREMENTS 219

spectrometer? It will be important to carefully determine the position of the

maximum voltage signal to determine the wavelength of the line.

The deviations from the tted wavelengths in Fig. 12.7 give a good in-dication
systematic uncertainty associated with the scan. For this data set,

the uncertainty is 0:2A, or 13 m. Note that there is a systematic un-certainty of

about this size just because of the number of signi cant gures given in Table
12.1.

Determining RH and RD

You can determine the Rydberg constants for hydrogen and deuterium from
your calibration and the measured line positions for the two isotopes, and using
Equations 12.6 and 12.8. Recall that nf = 2 for the Balmer transitions.

First try the brute force method. Pick one of the lines of either isotope and
determine its wavelength from the calibration. Then, make a good guess for the
value of ni for the line you picked. (What color was the line?) You then calculate
R from Eq. 12.6. Calculate the uncertainty by propagating the uncertainties in
your caibration constants to and then to R. Does the accepted value agree with
your measurements to within uncertainties? You might try this on more than one
line. The brute force method is the most susceptible to uncertainties. For
example, it will be hard to measure the calibration \o set" B very precisely
because di erent spectra will be on di erent exposures. (Your best bet is probably
to always include a HeNe laser exposure that you can refer to.)

A di erent technique eliminates the need for knowing the calibration o - set
because you can use the di erence of two wavelengths, since this di erence only
depends on the dispersion and not the o set. Use Eq. 12.6 to express i − j for any
pair of lines i; j in terms of R. Again, you should try this for a few di erent pairs
of lines. Is the result more precise? (It should be, since you no longer need to
include the uncertainty in the o set.) How is the agreement with the accepted
value?

Try both of these methods on both hydrogen and deuterium. Compare them
to the accepted value. You should come very close. Is this consistent
 220 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

within your experimental uncertainty?

Hydrogen/Deuterium isotope shift

The values for RH and RD di er because MH and MD are di erent, even though
both are much larger than the electron mass. You could determine an expression
for (RH − RD)=RH in terms of the masses, and compare your individual
measurements this way. However, it would be hard to be accurate because RH
and RD are so close, the di erence might be smaller than your combined
experimental uncertainty.

A better way is to use the hydrogen and deuterium exposure and measure the
splitting = H − D on any of the Balmer lines. You can determine very precisely
by exposing hydrogen and deuterium together on the same strip of lm, for the
Baird, or by using a deuterium tube with some hydrogen in it on the Jarrell{Ash,
and scanning the double line, as shown5 in Fig. 12.8. The excellent resolution of
the Jarrell{Ash is obvious! Since you already know H to high precision, you can
write

D RH
H = 1 − H = 1 − RD
= m MD − MH = 1 − MH =MD m (12.10)
MD m + MH 1 + m=MH MH
1m
3 10−4 2 MH

The splitting is very small indeed! Measure = H for some of the lines and
compare them to this relationship, including uncertainties. Determine an average
value with uncertainty for MD or MD =MH , using Eq. 12.10, assuming known
values for m and MH .

5 Baird data taken by Marc Crudele, Class of 1996. Jarrell{Ash data taken by Steve Irving
and Davienne Monbleau, Class of 1999.

12.2. MEASUREMENTS 0.28

0.27
 0.26

INTENSITY THROUGH FILM STRIP

0.25
221
0.24

0.23
6

0.22

5
0.21

PHOTOMULTIPLIER VOLTAGE SIGNAL

0.2
4
0.19
0.7 0.8 0.9 1 1.1 1.2 1.3
POSITION ALONG FILM STRIP (mm)
3

0
0 10 20 30 40 50 60 70 80
GRATING SCAN POSITION

Figure 12.8: Scan of a single Balmer transition line, with both hydrogen and
deuterium exposures. The wavelengths di er slightly because of the isotope shift.
On the left is a scan of a lm strip taken in the Baird spectrograph, exposing
hydrogen and deuterium on the same lm strip. On the right, a scan using the
Jarrell{Ash spectrometer using a deuterium tube with some hydrogen gas inside
it as well.

Line widths and splittings

You can also learn about the spectrograph, and some physics as well, by looking
at the widths of the spectral lines. First, you should try various things to make
the individual line widths as narrow as you can. For example, make the input slit
width very narrow. The placement of the discharge tube might also be
important.

You can quantify the line widths in several ways. Probably the easiest is to
locate the \full width at half maximum", that is, the distance between the sides of
the line at the points where it is half the maximum value. Plot the widths of the
hydrogen lines as a function of wavelength and see if there is a trend. You
should keep in mind the e ect of \doppler broadening" on the spectral lines. The
atoms which emit these photons are moving in lots of random directions with a
more or less thermal distribution of energies. If a photon of natural wavelength 0
is emitted from an atom moving at velocity component v towards or away from
the observer, then its doppler shifted wavelength is just = 0(1 v=c). The mean
2
square velocity hv i is given
222 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

to good approximation by (see Appendix B)

1 2 3
2 MAT OM hv i = 2 kT
where T is the temperature inside the discharge tube. Can doppler broaden-ing
explain what you see?

It is probably not possible to see the ne structure splitting in hydrogen

(around 0.14 A) with the Baird spectrograph, but it might be possible with the
6
Jarrell{Ash. In any case, you can measure the line shape. Try to estimate the
widest separation you could have for two lines buried inside the single line
shape. This would give you an upper limit for the ne structure splitting, and it
should be bigger than the expected value.

12.3 Advanced Topics

If you've exposed the nitrogen discharge tube, you likely see something quite di
erent that the other samples. Instead of seeing discrete lines, you should see a
series of \bands". That is, lots of lines very close together, but clustered in
regularly spaced groups. There should be two major groups like this, one at the
red end of the lm, and one at the violet end.

You are looking at the deexcitation spectrum of the N2 molecule, not the
N atom. This opens up an entirely new area of quantum mechanics, based on di
erent types of potential energies. A thorough treatment is quite involved, but is
discussed to some extent in Brehm and Mullin, and in Sime, but a very complete
writeup on what's going on is in Preston and Dietz. Here I will just give you
some basics, and show you some simple physics you can get out of it.

The nitrogen molecule is your basic diatomic molecule, that is, two nitro-gen
nuclei separated by some distance, with an electron cloud surrounding them.
This object can undergo two basic types of motion. One of these is
6
An undergraduate laboratory experiment which resolves nd structure in the H and D lines is
described in S. Pollack and E. Wong, Amer. Jour. Physics, 39(1971)1386. You might want to
use some of their tricks.
12.3. ADVANCED TOPICS 223

a vibration along the axis between the two nuclei, where the separation dis-tance
oscillates about some equilibrium value. The other motion is a rotation in space,
around the center of mass. Since the two nitrogen nuclei have the same mass,
the center of mass is just halfway between the two nuclei.

The equilibrium separation corresponds to the bottom of a potential en-ergy

well V (x) where x is the separation distance. Pretty much any potential well can
be approxmated by a harmonic oscillator potential, so long as the displacement
from equilibrium doesn't get too large. The spring constant k of this approximate
1 2 2
harmonic oscillator is given by k = 2 d V =dx , evaluated at the equilibrium
point. The solution to the Schr¨odinger equation for a harmonic oscillator leads
to equally spaced energy levels
1
E = nv + 2 h! Harmonic Oscillator (12.11)
q
where ! = k= and is the reduced mass of the system, i.e. MN =2 in the case of the N2
molecule. The value of nv must be a nonnegative integer.

The molecule also rotates, and the quantum mechanics of a so-called \rigid
rotor" is well de ned. It gets complicated because the molecule stretches as its
rotation speed increases, but we won't get into that here. The energy levels of the
rigid rotor, made of two masses separated by a distance R, are given by
h2
n (n + 1)Rigid Rotor
E = 2I r r (12.12)
2
where I = R is the rotational inertia, and nr is a nonnegative integer.

Refer to Brehm and Mullin, for example, and you will nd that the spac-ing of
energy levels given by Eq. 12.11 is much smaller than typical atomic energy
level spacings, that is, a few eV. What's more, the spacing given by Eq. 12.12 is
much smaller than those for the harmonic oscillator, Eq. 12.11.

Don't confuse the di erent types of spectroscopy used to study these en-ergy
levels. Brehm and Mullin, Sec.10-8, does a good job of showing the var-ious
ways these excitations are observed, only one of which you are equipped to do
with this setup. Transitions between individual vibrational or rota-tional states
within the same electronic state correspond to very low energy photons,
typically with wavelengths in the far infrared. Your apparatus,
224 CH 12. EXPERIMENT 6: ATOMIC SPECTROSCOPY

however, is only useful for spectroscopy of visible photons, so you make use of
a transition between electronic states which are modi ed by the presence of
vibrational and rotation states built on top of them.

So now let's get back to your piece of lm taken with the nitrogen dis-charge
tube. You are looking at the deexcitation of electronic energy levels of the
molecule, that is, states similar to those found in atoms. However, the big di
erence is that each of these states correspond to molecular con gura-tions that
can undergo their own vibrations and rotations. Therefore, each of these states
has a series of states built on top of them, corresponding to the nearly equally
spaced vibrational excitations, and each vibrational state has rotational states
built on top of it. There are some rules as to which kind of state can decay to
another in a lower electronic con guration, but the result is still very complicated
as you might imagine. This gets compounded when you realize that the potential
well is not really a simple harmonic oscillator, and the rotor is not really rigid.
Nevertheless, we can still get something out of it.

Look at the major group of bands near the violet end of the spectrum. This
corresponds to the photons emitted from a speci c pair of electronic energy
levels, and is called the \second positive series" of nitrogen. A diagram of the
bands are shown in Fig. 12.9, where it is compared to the mercury spectrum in
this region. The gure is taken from Sime. Con rm that your spectrum agrees with
this gure. The band spacing, roughly 5 nm according to Fig. 12.9, corresponds to
transitions from the more or less equally spaced vibrational levels in one
electronic state to those in the lower electronic state. Use this band spacing to
determine a rough estimate of the equivalent spring constant k from Eq. 12.11.
This lets you draw the shape of the potential well near the equilibrium point.
Your next goal is to try and estimate the separation distance of this equilibrium
point.

Look closely, using either the traveling microscope or the computerized

scanner, at the lm within the bands. You should see a dense collection of lines
which actually make up the band itself. These are transitions between the
rotation excitations built on top of the electronic+vibrational excitations.
Measure the approximate separation between these lines, and use the data to
estimate the value of I in Eq. 12.12. From this value of I, you can determine the
internuclear separation distance R for the nitrogen molecule. Does this
12.3. ADVANCED TOPICS 225

Figure 12.9: The second positive series in the nitrogen molecule, compared to
the mercury spectrum, taken with the Baird SB-1 spectrograph. Taken from
Sime.

agree with what you expect?

More sophisticated analyses are possible. See Sime or Preston and Dietz for
more details.