Measurement of the Sodium D Emission Lines Using a Michelson

Interferometer
R. Price
University of Arizona
12 March 2008

The Sodium D emission lines (commonly called the ‘Sodium Doublet’) are spectral
emission lines that have wavelengths within a few angstroms apart. The experiment
performed involved calibrating a Michelson interferometer and measuring the
wavelengths of both emission lines. This was accomplished by measuring both the
average and beat wavelengths of the interference pattern created. From the
experiment it was found that the emission lines were given as 603±37nm and
604±37nm for the D1 and D2 lines respectively.

1 Introduction interference patterns. Also as a consequence of

the proximity of the two lines, the interference
The purpose of the experiment was to measure patterns together will form an observable beat
the wavelength of the Sodium D emission lines. pattern. From the resulting beat pattern the
Sodium has two emission wavelengths that are difference in the wavelengths can be determined.
extremely close in wavelength and without
sensitive equipment cannot be distinguished. 1.1 Setup
These lines, designated the D2 and D1 Fraunhofer
For the experiment, a standard Michelson
lines, have wavelengths of 589.6 nm and 589.0nm
interferometer was used to create interference
respectively1. Using sensitive interferometry, the
patterns which allowed measurements of the
average wavelength of the sodium lines was
wavelength of the light sources used. Below in
determined as well as the difference between the
Figure 1 is a diagram of the experimental setup
two emission lines.
used:
When a sodium lamp is used in conjunction with
an interferometer, the wavelengths of the two
emission line will create interference patterns.
The distance between fringes within the
interference patterns can be measured to
calculate the wavelength of the emission source.
By measuring the path length difference that
causes the movement of the fringe pattern by one
cycle, the average wavelength of the source can
be determined. However, because of the relative
proximity between the sodium D lines, it becomes
more difficult to directly observe the two
Figure 1 - Experimental Setup

1
Standard D Line values from University Physics, 11
th In the experimental setup, a computer controlled
Edition motor was used to move the adjustable mirror
 12 March 2008

small distances on the order of a few nanometers. fact that the light is reflected and travels back
To measure the interference pattern created from through the same distance twice. Thus it can be
the interferometer, a photodetector was used to said l=2d where d is the distance the mirror is
collect data. When the photodetector was placed moved. Therefore:
in the path of the outgoing light, the interference
pattern would create a sequence of ‘high’ and ݉ߣ = 2݀ (2)
‘low’ recorded values. The harmonic pattern of So as the mirror is moved a distance d, m fringe
the recorded values from the photodetector patterns will move past a given point on the
indicated the passing fringes of the interference detector. The wavelength of the light source can
pattern. be determined if both the number of passed
The computer system used was capable of fringes and change and mirror distance are
controlling the position of the adjustable mirror measured. Consequentially, if a light source with
as well as recording the intensity of light upon a given wavelength is known, the interferometer
the photodetector. Thus the system was able to can be calibrated to find the distance the mirror
record the intensity of light as a function of has moved.
position, which was used to experimentally However the previous derivations are only
measure the wavelength of the sodium lines. appropriate for monochromatic light sources.
However because the units of distance on the Consider in the case of the Sodium D lines two
system correlated to ‘motor steps’, the separate wavelengths λ1, λ2 separated by a small
interferometer was calibrated to find the amount. The resulting interference pattern is the
relationship between the motor step length and superposition of the two separate fringe systems
nanometers the mirror moved. for each wavelength. Over certain distances of the
mirror position the two fringe systems can either
1.2 Theory
constructively or destructively interference with
The working theory behind the experiment is each other when the fringe patterns overlap.
that of wave interference. Consider the When the two maxima of the patterns are at the
interferometer described in Figure 1. To create same position, a set of dark fringes are produced
interference from a monochromatic light source, indicating that the optical path lengths for both
the following condition must hold: wavelengths are equal. Consider the case when
both overlap as described. If the mirror distance
݉ߣ = ݈ (1) is increased, the two patterns will overlap again
when:
Where l is the path length difference between the
two light sources. The interference is caused by 2݀ = ݉ߣଶ = (݉ + 1)ߣଵ (3)
path difference l in the optical path to one mirror
and the optical path to the second. In Equation 1, So from Equation 3 and solving for Δλ = λଶ − λଵ
m is the order number and λ is the wavelength of then one has a result of:
the light source. In the instance of the ஛భ
interferometer, m is a measurement of how many Δλ = λଶ − λଵ = (4)
௠
fringes apart the associated distance represents.
Note in the case of the Michelson interferometer, Using this result in combination with Equation 4
if the distance the mirror moves is d, then the it can be shown that:
total path length difference is 2d because of the

Page 2 of 5
 12 March 2008

Δλ =
஛మౝౣ
(5) the wavelength and difference of the sodium
ଶୢ
doublet was prepared. The setup was similar to
Where λgm is the geometric mean of the two that in Figure 1 except that a lens was used to
sodium spectral lines2. Using both Equations 2 focus outgoing light. The photodetector was
and 5, the two spectral lines can be determined placed at the focal length of the lens where the
experimentally by making the appropriate contrast in the fringes would be the largest,
measurements. creating clearly defined maxima and minima of
the interference pattern.
2 Experimental Methods
Once the experimental setup described produced
Before measuring the interference patterns of the an observable fringe pattern, the computer
sodium doublet, the interferometer setup used system was used to control the mirror motor to
was calibrated. As stated in Section 1.1, the vary the mirror distance and observe the
adjustable mirror on the interferometer was intensity pattern created by the passing fringes
connected to a computer controlled system that on the photodetector.
moved the mirror in units of ‘motor steps’. Using
To determine the averaged wavelength of the
Equation 2, the distance the mirror moves can be
sodium D lines, the same method used to
directly related to the wavelength of the light
calibrate the interferometer was used. The mirror
source and the number of passing fringes in the
was moved via the computer control over
interference pattern. To use the relationship, a
approximately 1000 steps of the motor. The
helium neon laser with a wavelength of
computer in conjunction with the photodetector
633±1nm was used as a light source in the setup
was able to create a plot of the intensity pattern.
in Figure 1. After fine tuning the placement and
From the plot, the number of step per the passing
angles of the mirrors to focus the interference
of ten fringes was recorded to get an average
pattern on a screen, the photodetector was
value for the number of steps per passing fringe.
placed in the pattern to measure the intensity of
Using Equation 2 and the calibration
the light. Next the motor attached to the
measurements, the average wavelength of the
adjustable mirror was activated so that the
sodium light was determined. To measure the
fringes within the interference pattern passed
difference in the two sodium D lines, the mirror
through ten cycles. The computer system
was moved over a large number of motor steps so
connected to the photodetector was able to plot
that intensity pattern wavelength and beat
the intensity of the light crossing the detector
pattern could be observed and measured by
versus the number of motor steps. Using
previous methods.
Equation 2 and the data collected from the
system, the distance between fringes in terms of 3 Calibration Data Set
steps could be related to the calculated distance
in nanometer. The process was repeated in order For the calibration data set, the data was taken to
to determine an average value of nanometers the determine the distance in nanometers of each
mirror moved per motor step. motor step on the mirror motor system. The data
below in Table 1 displays the values relating the
With the calibration data for the setup
steps per fringe to the nanometers per step.
determined, the experimental task of determining

2
C.J. Overback. “Selective Experiments in Science”, 1962.

Page 3 of 5
 12 March 2008

Trial Steps
per 10
Steps
per
Nanometers
per Step
Calculated
Error (nm)
4 Sodium Doublet Data
Fringes Fringes
(σ=±50) (σ=±5) In order to determine the average wavelength of
1 900 90 3.52 0.20 the sodium D lines, the interferometer mirror
2 800 80 3.96 0.25
was moved approximately the distance that
creates 10 passing fringes. The distance was then
3 880 88 3.60 0.20
4 870 87 3.64 0.21
5 825 82.5 3.84 0.23 used to calculate the wavelength of the incoming
6 845 84.5 3.75 0.22 light creating the interference pattern. The
7 865 86.5 3.66 0.21 results of the measurements are below in Table
8 850 85 3.72 0.22 2:
9 820 82 3.86 0.24
Average 851 85.1 3.73 0.22 Trial Steps per Steps Average Calculated
10 per Wavelength Error (nm)
σ 31.8 3.18 0.14 0.02
Fringes Fringe (nm)
Table 1 - Calibration Data Set (σ=±50) (σ=±5)

Note that to measure how many nanometers are 1 780 78 582 37.5

in one motor step, Equation 2 was used along

2 790 79 589 37.5
3 830 83 619 37.5
with dimensional analysis to provide the 4 825 82.5 615 37.5
following: 5 790 79 589 37.5
6 800 80 597 37.5
ଷଵ଺.ହ௡௠ ఒ
1 ‫= ݌݁ݐݏ‬ ௦௧௘௣௦ = ி
(6) 7 820 82 612 37.5
# ൗ௙௥௜௡௚௘
8 825 82.5 615 37.5
9 800 80 597 37.5
To find the error in the value, error propagation 10 830 83 619 37.5
techniques were used to result in the following Average 809 80.9 604 37.5
relationship: σ 19.0 1.90 14.15 0.01
Table 2 - Average Na Wavelength Data
ଵ ଶ ఒ ଶ
ߪ௫ଶ = ߪிଶ ቀிቁ + ߪఒଶ ቀிమ ቁ (7)
In order to calculate the wavelength of the
emission lines, Equation 8 was used to convert
Where λ=633±1nm, F is the number of steps per
the step count into nanometers. To calculate the
fringe and σx is the error in the nanometers per
error in the calculation of the wavelength, the
step calculation.
following relationship was derived using error
From the calibration data, it can be said that each propagation and the relationship in Equation 2:
step of the motor is approximately given as
ߪ௫ଶ = ߪிଶ (2ܰ)ଶ + ߪேଶ (2‫)ܨ‬ଶ (9)
݀ = (3.7 ± 0.2݊݉)(# ‫)ݏ݌݁ݐܵ ݂݋‬ (8)
Where F is the number of steps per fringe, N is
Where d is movement of the mirror in the conversion factor between steps to
nanometers. Note that in Table 1, the standard nanometers related in Equation 8 and σx is the
deviation of the distribution of number of steps error in the wavelength calculation. The values
per fringe (≈3) is less than the uncertainty of the for the trials were averaged together to give the
measurement (5 steps/fringe) thus the final measured wavelength and error. The
measurement uncertainty was used in the error measured value for the average of the sodium D
propagation. line is thus given as 604 ±37nm. Similar to the
calibration data, the measurement uncertainty of

Page 4 of 5
 12 March 2008

the data for the steps per fringe portion is much value error
larger than the variance between the trials. Steps per Beat 82500 2000
Therefore the measurement uncertainty was nm per Beat 307000 19000

used as the error in the measurement instead. ∆λ 0.59 0.08

Table 3 - Beat Pattern Summary
To find the beat pattern of the sodium lines, the
Furthermore, using the definitions of beat length
mirror motor was set to move over half a million
and the average wavelength:
steps so that the entire beat pattern of the
sodium doublet interference could be measured. λୠୣୟ୲ = λଶ − λଵ (11)
The collected data is displayed below in Figure 2:
஛మ ା஛భ
λୟ୴୥ = (12)
ଶ
600 Sodium Doublet
ADC Channel (Intensity)

500 The sodium D lines can be given exactly. From the

results λbeat = 0.59nm and λavg =604nm and
400 solving the linear system, the emission lines are
300 given as 603.7nm and 604.3nm for the D1 and D2
lines respectively. However considering error,
200
the measured values are more fairly given as
100 603±37nm and 604±37nm.

5 Results
0
Motor Steps
-100000 150000 400000
From the experiment, it was found that the
Figure 2 - Sodium D Line Beat Pattern emission lines for sodium were 603±37nm and
In order to find the beat pattern, ∆λ, the distance 604±37nm for the sodium D1 and D2 lines
between successive beats is required. From the respectively. Although these values differ from
data it is found that the average length of one of the recorded values of 589.6 nm and 589.0nm, it
the beat patterns over four beats is is interesting to note that the relative difference
approximately 82500±2000 steps. Using ∆λ for both the experimental and recorded values
Equations 8 to find the value and 9 for error it are very similar. Furthermore the error in ∆λ is
was found that the average beat length was less than one angstrom, which is interesting as
307000±19000nm the uncertainty in the average value is off by
approximately 40nm. However, the large amount
The value of the difference in the wavelengths of of error in the average wavelength measurement
the sodium D lines is given in Equation 5. Using comes from the measurement uncertainty of the
error propagation derivations, the error in the number of motor steps per fringe. As noted
wavelength difference is given as: earlier the actual deviation of the data points was
ଶ
less than the measurement uncertainty. If the
ఒ ଶ ఒమ
ߪ୼ଶ = ߪఒଶ ቀௗቁ + ߪௗଶ ቀଶௗమ ቁ (10) statistical variation was used as the error, the
variations in the average wavelength would be
Using Equations 5 and 10 together result in ∆λ = reduced to approximately 10 nanometers which
0.59±0.08nm. The following is a summary table of would make the value of 603nm for the average
the results: slightly out of range of the recorded 590 nm value.

Page 5 of 5