Lab Report on the Fabry-Perot-Etalon

Anton Haase, Michael Goerz

4. October 2005

GP II
Tutor: M. Fushitani

1 Introduction
Multi-Ray-Interference
The diffraction pattern of a single and double slit is a well-known interference
phenomenon, where the double slit produces more fringes than the single slit.
By increasing the number of slits (using a grid instead, for example), we get
a sharper pattern with small areas of constructive interference and large areas
of destructive interference. This is the result of multi-ray-interference behind a
regular structure of defracting elements. Huygens Principle is a simple geometric
explanation for the observation behind such a setup. The higher number of slits
leads to a multiplication of the wavelength difference, so that the condition for
destructive interference is reached at a lower angle difference (from the angle of
the related maximum) than behind a single slit. This means that the maxima
are sharper and the resolution in a spectral application is higher.

Fabry-Perot-Etalon
The Fabry-Perot-Etalon (FBE) is a optical resonator and a realization of a multi-
ray-interference device. It consists of two parallel, semipermeable mirrors with
an optical medium in between. Incoming light of a certain angle is partially
transmitted and reflected between the two surfaces as shown in Fig. 1. The
resulting phase difference between two rays can be calculated with a simple
geometric interpretation (vector addition).

δ = AC + CD − AB = 2 d cos(α) (1)

The phase difference of two rays must be δ = nλ to get constructive interference.

The values of n are called the order of interference.
The Fabry-Perot-Etalon realizes a very high amount of interacting rays and
interference orders. Therefore it is a highly accurate device, used to distinguish
small wavelength differences in spectroscopy, for example (high resolution).

Dispersion Range
The spectral range of light which can be analyzed by an optical device like the
Fabry-Perot-Etalon is limited. The difference between two bordering lines can

1
GP2 - FAP Haase, Goerz

Fig. 1: Fabry-Perot-Etalon

be the value of the interference order or a small wavelength difference. The

interference condition for this case is:

(z + 1)λ = z(λ + ∆λ) (2)

From this follows the dispersion range.

λ
∆λ = (3)
z
The high interference order of the FPE mentioned in the beginning, and the
resulting low dispersion range, makes it a perfect device to analyze nearly
monochromatic light.

Fabry-Perot-Spectrometer
The advantages of the FBE as a spectrometer have already been explained.
The spectrometer is realized with incoming divergent light and a lens to project
the interference pattern on an observation plane. The image then consists of
concentric circles. The interfence condition for this specific setup is
· ¸
2d r2
z= 1− 2 , (4)
λ 2f
where f is the focal length of the lens, r is the radius of the observed circle and
α was set to α = fr and cos(α) = 1 − 12 α2 , respectively.
The measurement of at least two circles would be enough to calculate the
distance between the two mirrors of the etalon. The equation comes from con-
dition (4) by combining it for the two values of r to
λf 2
d=i , (5)
ri2 + r02
where i is the number of circles counted from the innermost circle (i = 0).
Additionally, we get the relative weavelength difference by measuring the
radius of a certain order for two separate wavelengths:
λ ¡ 2 ¢
∆λ ≈ r − r02 (6)
2f 2
The accuracy of this calculation is determined by the accuracy of the radius
measurement. To use the high efficiency of the FBE, it is necessary to calibrate
the device. This can be done with light of a well known wavelength. The
wavelength then will be calculated directly from the interference condition.

2
GP2 - FAP Haase, Goerz

Resolution of the Fabry-Perot-Spectrometer

The resolution of the FPS can be calculated with the Equation of Airy. It
depends on the transmission and reflection coefficient of the mirrors.
· ¸2
I T 1
= 4R 2 (7)
I0 1 − R 1 + (1+R) 2 sin (φπ)

From this we get the full width at half maximum.

1−R
2∆z = √ (8)
π R

2 Assignments
1. Construct and adjust the setup.
2. Calculate the distance between the two planes of the FBE using the red
line (643.9 nm) of an cadmium lamp and determine the interference order.
3. Calculate the wavelength of the green and darkblue line of the cadmium
lamp.
4. Give an approximate value for the width of the maxima lines of the red
line and compare them to the theoretical expectation of the setup.

3 Analysis
Our measurements with the FPE where performed using a cadmium lamp as
light source.

3.1 Plane Distance and Interference Order

In the first setup, we examined the red line of the spectrum (643.9 nm) using a
red filter to have a clear image. We measured the inner and outer border of each
ring to get an average value for the real radius. The error was estimated from
the thickness of the circles. The results of this first calculation are presented in
Table 1.
order i left(av) / 10^-3* m right(av) / 10^-3* m radius / 10^-3* m error / 10^-3* m
0 8.2 6.12 1.04 0.09
1 8.85 5.56 1.65 0.11
2 9.27 5.04 2.12 0.07
3 9.61 4.69 2.46 0.08
4 9.94 4.39 2.78 0.08
5 10.22 4.12 3.05 0.08
6 10.47 3.87 3.3 0.09
7 10.71 3.61 3.55 0.08
8 10.93 3.37 3.78 0.07
9 11.13 3.18 3.98 0.07
10 11.35 2.98 4.19 0.08

Table 1: Radii of the Red Circles

Our intention was to determine the distance between the two planes in the
Fabry-Perot-Etalon and to calculate the interference order from this result. The

3
GP2 - FAP Haase, Goerz

graphical analysis of our data, by plotting ri2 − r02 over i, clearly demonstrates
the expected linear behavior.

1.8e-05

1.6e-05

1.4e-05

1.2e-05
Radius Difference / m2

1e-05

8e-06

6e-06

4e-06

2e-06

0
0 2 4 6 8 10
Observed Order i

Fig. 2: Graphical Analysis of the Radii of the Red Circles

We get the distance d by reading off the slope of the approximate line and
using Eq. (5). The result then is a distance of
d = (4.1 ± 0.2) · 10−3 m.
The relatively high error of the radii does not allow a very accurate deter-
mination of the interference order. However an approximate analysis following
Eq. 4 shows that the order difference between two neighboring radii was exactly
one, as expected. The calculated value for the interference order of the ring
labeled zero was
z = (13000 ± 2000)

3.2 Analysis of the Green Line

After the calibration of our setup in the first assignment, we changed the color
filter to analyze the interference pattern of the blue and green line of the cad-
mium spectra. The purpose of these measurements was to determine the wave-
length of both lines. The low quality of the color filters used, resulted into two
measurements for the green line, because we assumed the circles we measured
with the two filters to be different, which was mistaken. This additional mea-
surement will however be used to improve the calculated value. Table 2 shows
the data we retrieved using the green filter.
Again, a graphical analysis will be useful to determine the wavelength from
Eq. (5). The behavior is again highly linear as Fig. 3 demonstrates. From
the slope we get the required values to calculate the wavelength of the light
observed.
λgreen = (532 ± 20) · 10−9 m

4
GP2 - FAP Haase, Goerz

order i left(av) / 10^-3* m right(av) / 10^-3* m radius / 10^-3* m error / 10^-3* m

0 8.07 6.3 0.89 0.09
1 8.66 5.7 1.48 0.07
2 9.06 5.28 1.89 0.06
3 9.4 4.95 2.23 0.06
4 9.67 4.67 2.5 0.06
5 9.93 4.41 2.76 0.06
6 10.16 4.18 2.99 0.06
7 10.37 3.96 3.2 0.06
8 10.59 3.77 3.41 0.06
9 10.75 3.59 3.58 0.06
10 10.94 3.42 3.76 0.06

Table 2: Radii of the Green Circles

1.8e-05

1.6e-05

1.4e-05

1.2e-05
Radius Difference / m2

1e-05

8e-06

6e-06

4e-06

2e-06

0
0 2 4 6 8 10
Observed Order i

Fig. 3: Graphical Analysis of the Radii of the Green Circles

The results of the second measurement are presented in Table 3 and Fig. 4.
The value from this calculation is

λgreen 2 = (544 ± 20) · 10−9 m

The average value for both results then is

λgreen avg. = (538 ± 16) · 10−9 m.

This data is identical to a strong green line (537.9 nm) in the spectrum of the
cadmium lamp1 .

1 Source: National Physical Laboratory (UK)

5
GP2 - FAP Haase, Goerz

order i left(av) / 10^-3* m right(av) / 10^-3* m radius / 10^-3* m error / 10^-3* m

0 8.03 6.28 0.88 0.08
1 8.63 5.7 1.47 0.07
2 9.07 5.28 1.9 0.05
3 9.4 4.94 2.23 0.06
4 9.71 4.66 2.52 0.05
5 9.96 4.39 2.78 0.06
6 10.17 4.18 3 0.07
7 10.37 3.97 3.2 0.08
8 10.57 3.78 3.4 0.06
9 10.77 3.59 3.59 0.07
10 11.02 3.42 3.8 0.05

Table 3: Radii of the Green Circles (Second Measurement)

1.8e-05

1.6e-05

1.4e-05

1.2e-05
Radius Difference / m2

1e-05

8e-06

6e-06

4e-06

2e-06

0
0 2 4 6 8 10
Observed Order i

Fig. 4: Radii Differences of the Green Circles (Second Measurement)

6
GP2 - FAP Haase, Goerz

3.3 Analysis of the Blue Line

In a final measurement we improved the quality of our blue filter by putting
two of them together in front of the FBE. We were now able to observe the
darkblue line of the spectrum. However, the results should be analyzed critically,
because the interference pattern was more blurry and hard to observe. The radii
calculation is presented in Table 4.
order i left(av) / 10^-3* m right(av) / 10^-3* m radius / 10^-3* m error / 10^-3* m
0 7.43 6.06 0.68 0.21
1 8.05 5.53 1.26 0.12
2 8.47 5.12 1.68 0.09
3 8.85 4.79 2.03 0.07
4 9.09 4.53 2.28 0.07
5 9.34 4.29 2.53 0.07
6 9.58 4.04 2.77 0.07
7 9.77 3.83 2.97 0.05
8 9.99 3.63 3.18 0.05
9 10.18 3.45 3.37 0.05
10 10.35 3.25 3.55 0.06

Table 4: Radii of the Darkblue Circles

Again we performed a graphical analysis to calculate the wavelength from

the slope by using Eq.5.

1.8e-05

1.6e-05

1.4e-05

1.2e-05
Radius Difference / m2

1e-05

8e-06

6e-06

4e-06

2e-06

0
0 2 4 6 8 10
Observed Order i

Fig. 5: Radii Differences of the Blue Circles

The resulting wavelength from this plot is

λdarkblue = (473 ± 20) · 10−9 m.

This value is identical to the theoretical expectation as well.

3.4 Expected and Measured Linewidth of the Red Line

First of all, a theoretical value for the expected linewidth is hard to calculate.
The reason for this is the roughly determined interference order. However,

7
GP2 - FAP Haase, Goerz

we tried to compare our measured values to the theoretical expectation. The

linewidth we measured for a certain order was different from the width of other
orders of course. The most accurate value was measured for the zeroth order,
because any value for higher orders would come with a very high relative error
and would therefore be meaningless. The linewidth we calculated was

∆λ = (1.2 ± 0.6) · 10−12 m

The theoretical expectation assuming a reflectivity of 80% is

∆λ = (1.9 ± 0.3) · 10−12 m

The values are compatible within error. As already mentioned, a comparison

of both values can only be qualitative. The different reasons for this will be
explained in the conclusion below.

4 Conclusion
The Fabry-Perot-Spectrometer turned out to be a very accurate device to ana-
lyze light, as any other optical device we used in other experiments before. The
high interference order allows us to examine very small wavelength differences.
However, a careful calibration of the setup is necessary to obtain good results.
Obviously, we did this very successfully, because our data shows only a very
small deviation from the expected linear behavior. This applies to all measure-
ments, except for the last. However, the accuracy is influenced by the quality
of observation. This means that the interference pattern has to be projected
sharply into the observation plane. The focusing was very hard, because the
pattern itself was not sharp at all. The result of such systematical errors would
be wrong measurements of the radii. Another problem of the adjustment is to
get the pattern in the center of the observation plane, which is also necessary
to read off the correct radii.
If non-monochromatic light is used, additional problems with inadequate
color filters lead to multiple rings of nearly the same color in one interference
order, so that a correct measurement cannot be guaranteed.
At last, the measurement of the linewidth is very inaccurate, because of
the problems mentioned before. The rings do not have a clear border and
the reading error from the micrometer scale is too high. The comparison to a
theoretical expectation, which comes from the Fabry-Perot-Etalon only without
considering the lenses or the observer, is therefore only rough. However, in our
case, they were even compatible within (high) error, which confirms our data.