08Interferometer

[Version WEB 28.11.2012]

Optics and Photonics Technology Laboratory, Institute of Microengineering,
School of Engineering, cole Polytechnique Fdderale de Lausanne,
Rue A.-L. Breguet 2, 2000 Neuchtel, Switzerland
Phone: +41 32 718 3200
E-mail: [email protected]

Fax: +41 32 718 3201

Web : opt.epfl.ch

Contact person:
Dr. Toralf SCHARF,
Phone : +41 32 718 3286,
E-mail : [email protected]

Contents
1. Objective and overview
2. Safety Issues
3. Background
4. Setup and equipment
5. Summary of tasks of the experimental work

Appendices
A. Preliminary questions
B. Final questions
C. Supplementary information

1Objectiveandoverview
The practical work should introduce the following subjects to you:

Understand the operation principle of a Michelson interferometer

Evaluate conditions for fringe contrast and form under different conditions
Measure the (temporal) coherence properties of a laser diode
Measure a surface tilt with phase shifting interferometry

To get this done you need to read the reference document provided and you should be able to
answer the questions under Appendix A of this document.

2SafetyIssues
In this experiments laser sources and low power electrical equipment is
In this experiments laser sources and low power electrical equipment
are used. The laser is of class II.
Class II: low-power visible lasers that emit a radiant power not above 1
mW. The concept is that the human aversion reaction to bright light will
protect a person. Only limited controls are specified.
(http://www.osha.gov, Laser Hazards)
The laser is safe because the blink reflex of the eye will limit the
exposure to less than 0.25 seconds. It only applies to visible-light lasers
(400700 nm). Intentional suppression of the blink reflex could lead to eye injury. In our
experiment the laser sources are collimated and should be handled with care. A strongly
divergent beam will not be focussed on the eyes retina and represents often no danger.
Collimated beams will lead to small focus spots onto the retina special care is needed. Do
note stare into a collimated beam!
The electrical equipment used in the experiments is based on USB power (5V, 0.5 A, 2.5 W)
and not subjected to any particular security issues. Nevertheless you should not produce
short circuits on the printed circuit board (PCB) or to the computers USB connection to
avoid damage to the material. Make proper use of screwdrivers. Do not force any mechanical
parts.

3Background
Camera
In our experiment we use a pixilated camera. The pixels are arranged in a regular array that is
produced at very high precision. The camera chip has 1600 x 1200 pixels and is 4.536mm x
3.416mm wide and large respectively. Pixel pitch is 2.835 micron in both directions (square).
We can safely assume that there is no deviation of pixel position.
Interference and coherence
The addition of coherent light beams results in intensity modulation: the interference fringes.
The important term is coherent. Coherence has two aspects
Spatial coherence
Temporal coherence
3

Spatial coherence is linked to the physical size of the source. A monomode laser (one
wavelength, one polarization) for instance is a coherent source with the highest spatial
coherence but still has a finite coherence length. To see interferences, light has to be
superposed with path differences within the coherence length. If a spatially extended source is
used the superposition of coherent areas in the spatial domain has to be assured. This is
achieved by observing images of the source that overlap in the plane of observation.
Temporal coherence is linked to the spectral properties of light. Thin (limited) spectra have
higher coherence than wide spectra. It is possible to quantify the coherence by the coherence
length lc. There are different definitions based on the measurement of contrast of fringes in an
interferometer. We use the following formula:
lc =

1 c
1 2
=
n n

Eq. 1

The refractive index n is the refractive index of the surounding medium. The coherence length
depends on the bandwidth (optical frequency) , the speed of light c. It can be expressed in
wavelength and spectral width . We work mainly in air, therefore, we set n=1. The table
below gives indicative values of coherence lengths for the different sources we use
Source

Center
wavelength

Spectral width

Coherence
length

Halogen lamp
(visible)

550 nm

300 nm (400 700

nm)

1 m

LED

635 nm

20 nm

20 m

Monomode laser

635 nm

0.2 nm

2 mm

The coherence properties of the light emitted by different sources are determined by
measuring the contrast of fringes in an interferometer. In an interferometer the light
propagation is well managed and fringes of defined geometry appear.
Laser source
Interference is easily observed with a laser. As a laser source, a monomode laser diode
(model Opnext_Laserdiode-HL6354_EN) was chosen. The spectral properties are shown
below. The widths of a single peak is below 0.2 nm FWHM. If no special measures are taken
lasers have several modes. The spectral separation and spectral width of the modes are
determined by the geometry of the laser resonator (Fabry Perot effect!). The pumping
(electrically or optically) of the system gives its power generation properties.

Figure 1. Spectral characteristics of the monomode laser diode measured at different

resolutions. Left: as measured with a miniaturized grating spectrometer. Right: measured with
an optical spectrum analyzer in linear and logarithmic scale. The logarithmic scale (right side)
reveals that there are two peaks of different intensity and that the laser is not strictly
monomode (only one peak is allowed)
The laser diode has a very small active zone which leads to diffraction. Because the active
zone is not symmetric the emission profile is elliptical.

Figure 2. Appearance and emission profile of the laser diode. The emission profile is not
symmetric and leads to an elliptical shape in the far field.
For our experiments, it is important to note that the laser has at least two wavelengths. This
will lead to particular properties (beating) in space and can be detected with the Michelson
interferometer.
Interference and contrast
Interference is a wave phenomena. It appears when coherent waves are superposed. The
spatial and temporal coherence of the waves have to be considered. The result of intereference
phenomena are interference fringes in space. The mathematical description of interference
can be easily done for the one dimensional case. It can be extended to three dimensions. Here,
5

we repeat the description for the one dimensional case to define the fringe contrast and the
fringe period.
The amplitude of the electric field E has to be considered. For two waves of the same
frequency, the superposition gives:
Eq. 2
E = E1 + E 2
The irradiance I is the square modulus of the field (see for instance Hecht, Optics Addison
Wesley 1998 page 377 ff).
Eq. 3
I = I1 + I 2 + 2 I1I 2 cos
The irradiance values of the two waves I1 and I2 are given by the time average of the electric
field. The last term in the equation is the interference term. It is usually written as

= ( k 1 ir k 2 ir + 1 2 )

Eq. 4

One considers a position a space r (vector) and this is equation is a vectorial equation taking
into account all dimansions and the direction of propagation (trgough the vectors k1 and k2).
Innitial phase differeneces of interfering waves are put into the value 1 and 2. To get this
equations it is important to consider the time average of the measurement. This is because our
detector will not see the fluctuation at light frequencies. The phase difference governs the
intensity found in Eq. 3. The phase difference can take any value but the final term describing
the intensity is the cosine of that and hence varies between -1 and 1.

cos = cos ( k 1 ir k 2 ir + 1 2 )
Eq. 5
So the maximum value of intenisty and the minimum value of intensity for a given problem
can be caluclated. We will consider two examples.
Fringe contrast for monochromatic wave
An important parameter is the contrast of the interference fringes. The contrast is defined as
the difference of intensities divided by the sum of intensities.

Imax Imin
Eq. 6
Imax + Imin
We consider a one dimensional problem and wave travelling along the same direction at the
same wavelengths. The wave vector k1 and k2 are equal k1=k2 and only the initial phase delay
1 and 2 have to be condsidered. Here 1 and 2 represent optical path differences hence path
lengths of the travel to the observation points r at the same speed of light c. Equation 4 has
values from -1 to and 1 and a maximum and minimum intensity can be found. We caluculate
the fringe contrast and considers Eq. 1 which can be reformulated as
C=

C=

I1 + I 2 + 2 I1I 2 I1 I 2 + 2 I1I 2
I1 + I 2 + 2 I1I 2 + I1 + I 2 2 I1I 2

2 I1I 2
I1 + I 2

Eq. 7

If the intensities are equal the formula gives a contrast of 1. If the intensities differ then the
contrast is less than one.
6

Interference with several frequencies (beating)

Another very intersting case is when the wave vectors k1 and k2 are pointing in the same
direction but have different values because the wavelength is not the same. The general case
is rather complex and is the basis of Fourier transform spectroscopy. We consider here two
wavelengths only. We are interested in the contrast of fringes induced by this effect and we do
not consider additional phase delay. We set 1 - 2 to zero. We reformulate the phase delay
using a one dimensional form with scalar values of k, writing z for the position and taking 1 2=0.
cos = cos ( ( k1 k 2 ) z )
Eq. 8
The wave vectors are related to the wavelengths through k 1 = 2 1 and k 2 = 2 2 .One can
find
1
2 1
2 2
1

Eq. 9
k1 k 2 =

= 2
= 2
2 2
1 2

1 2
1 2

cos = cos 2 2 z

Eq.10

For a given position is space r we will find a particular value of and a certain intensity
corresponding to its cosine value. If we change z the intensity will change and will lead to a
full modulation over a distance that is given by the periodicity of the cosine function as a
2
multiple of its period which is . One full intensity variation is obtained when moved
z = 2 . Consider a concrete example. Assume we have a laser with a spectral width of
=0.2 nm at =635 nm. We get
2 ( 635nm )
z =
=
= 2mm

0.2nm
2

Eq. 11

It is possible to show that this behavior is not only found for the intensity but also for the
contrast of fringes in an interferometer.
Michelson interferometer
The Michelson interferometer is only one of a multitude of different interferometric
arrangements. The figure below shows its geometrical layout. It is a wavefront splitting
interferometer and it uses either a beam splitter cube or beam splitter plates to divide the
intensity of the beam in two paths. In the example below, the laser send light onto the beam
splitter and propagates further to the mirrors. The light is reflected back from the mirrors,
recombines at the beam splitter and reaches the detector. In our case, the detector is a camera
that allows to see the form of the fringes. Note that the same amount of light is send back
towards the source (laser) and might influences the source radiation conditions for some
particular cases.

Figure 3. Geometry of a Michelson interferometer with the source (laser), beam splitter cube,
two mirrors and a detector
When we illuminate the interferometer with a point source (monomode laser diode). The
radiation is not collimated. The figure below shows what will happen.

Figure 4. Point source illumination and the appearance of two virtueal sources behind the
mirrors of the interferometer.
The beam splitter divides the incoming light which is then reflected at the mirrors. If we were
looking inside the interferometer from the detector, we would see two point sources.
Depending on the position of the mirrors and the beamsplitter the virtual images of the
sources may be found at different positions. There are particular configurations that lead to
well defined fringe geometries. A few examples are given below
Point sources at the same depth but different positions

In the detector plane two waves with equal curvature arrive and lead to a set of parallel
fringes.
8

Point sources at the same position but at different depths

The curvature of the superposed spherical waves are different and circular fringes become
visible. Because the sources are approximately on axis (centered) with respect to the
observation plane the fringes are symmetric and appear as nice circles.
Point sources at arbitratry positions

In this general case the point sources are displaced in both directions. Curved fringes of
different spacing appear. The spacing of the fringes could be very narrow (!) and may be
difficult to detect.
Zero optical path difference ZPD

The most exclusive position appears when there is complete positional coincidence. In this
situation, the direct fringe intensity is observed. It might be bright or black. The
interferometer is now aligned. The source images are superimposed within a precision better
than one fringe that is half the wavelength (at 635 nm this means better than 300 nm). If the
intensity is very uniform the alignment is even better.

Optical path difference and fringe contrast

As discussed above the spectral characteristic of the laser (two wavelengths) will lead to a
beating effect. To make this visible the optical path on axis in the interferometer has to be
changed. The figure below shows the principle.

The images of fringes below show the contrast variation as a function of position x. The
contrast variation depends on the laser and its operation conditions and varies very much
between different sources of the same type.

Figure 5. Contrast versus measurement position.

As an example, we show in the figure above the measured contrast for the monomode laser.
The contrast shows two maxima at a distance of approximately 1.8 mm. Using Eq. 9, one can
calculate the spectral width (or distance of beating frequency). Here we have z=1.8mm and
=635nm
2 ( 635nm )
=
=
= 0.22nm
z
1.8mm
2

Eq. 12

Note that when moving the mirror the contrast and the fringe geometry are changing. A
sequence of frame is shown below.
10

Figure 6. Fringe appearance for movements of 2 mm between each image.

The fringes show different contrast and the curvature increases, a typical sign of depth
difference of the two sources in the interferometer. In the experiment to evaluate the
interference of sources with a certain spectral width and beating only the contrast is important.
The geometry of fringes is not considered.
Phase shifting interferometry
An important application of interferometry is the evaluation of the geometry of surfaces. To
do so the typical approach is to create fringes and analyze them. Phase shifting interferometry
is used to measure shapes of continuous surfaces as for instance lenses. We can only make a
very rudimentary approach here to show the principle of operation. For more details please
refer to the specialized literature. [P. Hariharan, Basics of Interferometry, Academic Pr Inc;
2nd Ed. 2006, ISBN 0123735890, Daniel Malacara, Optical Shop Testing, Wiley Series in
Pure and Applied Optics, 3rd Ed. 2007, ISBN-10: 0471484040)

Figure 7. Basics of phase shifting interferometry. (image from James C. Wyant, University of
Arizona,
http://www.optics.arizona.edu/jcwyant/Optics505%282000%29/ChapterNotes/Chapter09/pha
seshiftinginterferometry.pdf)
A practical implementation is shown in the figure above. One uses a minimum of three phase
shifted images of a given interferogram and can evaluate a phase as a multiple of the
wavelengths. This phase can be translated into a height profile. The difficulty is a nonambiguous evaluation of the profile after the phase has been determined from the Arctan
operation. The classical algorithm here uses 5 frames (Schwider-Hariharan algorithm). In our
example we will use three measurements that are shifted by 90 (/2). A typical measurement
sequence is below.
11

Figure 8. Phase shifted images of a series of fringes. The phase shift should be adjusted to 90
between the images to prevent artifacts. In the three images above the phase shift is not
perfect!
Three intensity images will be recorded, I1, I2 and I3. The phase difference between them
should be 90. In a real measurement system that has to be calibrated. After the images were
obtained the following formula applies to calculate the phase
I I
= arc tan 3 2
I1 I 2

Eq. 13

The phase contains information about the height. The difficult part is the unfolding of the
phase information because the phase has a certain periodicity. After unfolding the phase a
height profile can be extracted by multiplying the phase with the corresponding wavelengths.
The optical path is given by the equation

OPD =
unwrap ( )
2
Eq. 14
With this it is possible to find the surface profile of an object. To give an example we show
below the evaluation for a tilted surface. The interferograms of the figure above are used and
evaluated in MATLAB. The difficulty here is that the camera is too noisy to perform well.
Averaging of lines was used to obtain good fringe data. We start to show the three fringe
curves in the center of the image (averaged over 200 lines)
200
180

signal [counts]

160
140
120
100
80
60
40

200

400

600

800
1000
position [pixel]

1200

1400

1600

Figure 9. Phase shifted fringes from the original measurement averaged over 200 horizontal
lines.
12

NOTE: The full period is 360 and a 180 phase shift means here a shift by half fringe,
hence 90 a quarter fringe shift!
In the next step the phase value are calculated with Eq. 10 and the height profile is calculated
by using the wavelength of light.
2

2.5

1.5
2
1
1.5
height [micron]

signal [rad]

0.5
0
-0.5

0.5
-1
0
-1.5
-2

200

400

600

800
1000
position [pixel]

1200

1400

-0.5

1600

200

400

600

800
1000
position [pixel]

1200

1400

1600

Figure 10. Phase values after arctan operation (left) and height profile after phase unwrapping
and calibration.
In our case, the high profile is a simple line because we have only set a tilt on one of the
mirrors. The process works also for more sophisticated surfaces and is very versatile. There
are a lot of instruments using this technique for full field surface measurement.

13

Setupandequipmenttasksoftheexperimentalwork

4.1 Setup and alignment

A Michelson interferometer is a beam dividing device to interfere light. In our case a laser
beam is divided by a beam splitter and sent on two mirrors as shown below.

The mirrors reflect the light and a detector (CMOS camera) is used to visualize the result.
Because of the beam splitting, the detector sees light coming from two virtual sources. A laser
is a highly coherent source and allows to create interferences over large surface areas and
large movement distances of the movable mirror. In our case we use directly the laser diode
without collimation. The laser produces a diverging light cone that is sent to the
interferometer.
Materials
A CMOS camera (1600x1200 pixels, colour, pixel size 2.835 um) C600 from Logitech
Laserdiode, USB driven
Mechanical setup
Mechanical holders and setup

Figure 11: Top view of the Michelson interferometer

14

Figure 12: Schematic view of the mechanical setup; top view and perspective view with lens
cap that is ONLY used in pre-alignment procedure.
Detached pieces

Figure 13: Separate elements provided for the experimental setup.

15

The linear stage has to be mounted with 45 movement direction with respect to the source
and the camera. The camera is used without objective and IR filter to allow direct access to
the camera chip.

Camera is mounted in the long holder as seen below.

Figure 14: Close view of the CCD holder; a) front view, b) rear view. The camera is
mounted in the holder without its objective.

Figure 15: One mirror is mounted on the translation stage (side view).The other mirror is
already mounted on a tilting stage to align its angle and SHOULD NOT BE TOUCHED.

16

Figure 16 The second mirror is mounted on the six axis stage; left: side view, b) front view,
and SHOULD NOT BE TOUCHED.

Figure 17 The beam splitter is a 10 mm cube mounted on a cylinder and SHOULD NOT BE
TOUCHED.
CAUTION: Do not touch optical surfaces of the mirrors and the beam - splitter cube!
Consider the following details:
The laser source is fixed with its special holder.
Start by setting the linear movement stage to its central position (at about 6 mm).
Take care that the beam splitter is in the right position and sends light to both
mirrors!
Move the angular adjustment of the mirrors to set small angles at the beginning. Avoid
position of the alignment screws at the end of movement range.
Pre-alignment procedure
An interferometer is a high precision instrument and need carful alignment. It is recommend
to assure a good spatial alignment which is nothing else than to superpose the source images
on the detector. This is done by creating an image of the source in the detector plan. To do so,
use the lens cap.

17

Figure 18. Focalization (black arrow) and intensity adjustment with the lens cap.

Make sure that all components are well fixed.

Check camera position (height should correspond the beam splitter and laser).
Focus the laser on the surface of the detector.
Adjust intensity.
Check by naked eye where your focus spots are. You should see two spots that
correspond to images created by the two arms of the interferometer. You need to rotate
the beam splitter to adjust position.

Figure 19. Focus spots at the detector during pre-alignment procedure. The two spots have to
be superposed by adjusting the mirror position.
Identify the image from each mirror on the detector by turning the tilt alignment
screws for the two mirrors separately.
Superpose the images in the middle of the detector by adjusting the tilt of the mirror

18

Figure 20. Adjustment screws of the tilting stages to align the spots over each other.

Figure 21. Images of the spots as seen on the detector at different stage of the alignment
the two spots have to be superimposed as seen in the right image.
The pre-alignment procedure is done. You can remove the lens cap and look at the
interferogram directly.
4.2 Zero optical path difference ZPD and interferometer alignment
Make sure that all components are well fixed.
Check camera position (height should correspond the beam splitter and laser).
Set the camera exposure conditions to automatic for adjustment.

If the detector is saturated or needs more light increase or decrease the distance of the
laser to beam splitter respectively. (intensity follows a square law with distance)
19

 Move SLOWLY the linear motion table to find fringes on your detector. This might
look like shown below. The fringe distance can vary very much depending on initial
alignment.

The laser has an elliptical emission profile by turning the laser you can optimize
(minimize) the stray light (light that is reflected by components in the setup but not
used for the functionality).
An interferometer can be aligned to its zero optical path difference positions. In this case there
is only one fringe (either black or white) on the detector. To find this position the alignment
knobs of the mirrors have to be turned to increase the spacing between fringes. A sequence is
shown below when turning only one knob of the alignment mirrors.

During alignment fringe orientation will change and fringe distance will have a maximum at a
certain position. At this position you should use the second knob for the different alignment
direction of the mirror to continue. Usually it will give you a circular fringe picture like
shown below.

20

To correct this curvature you have to move the linear motion stage and eventually correct the
angular alignment until you find the position of zero optical path difference. At the OPD zero
position fringes are straight lines and spacing can be very large until one sees only one fringe.
SWITCH OFF the automatic exposure to get best results of fringe contrast.
See below for an example. The fringe intensity in the OPD zero position with one fringe could
be high or low. Touching the setup allows you to choose one or the other as shown.

To be done for the report: Find the zero OPD position and plot the intensity profile for the
low intensity state and the high intensity state. Take care that you use equal exposure
conditions for both states to show the contrast correctly.
4.3 Measurement of laser fringe contrast
The interferometer allows to measure the coherence properties of a laser. In the present case
we can assume that the laser has several emission lines. This will influence the fringe contrast
when studied out of zero optical path difference condition. More precisely it will vary the
contrast of the interference fringes. The task is to measure precisely the contrast of the
interference fringes as a function of the distance of the mirror.
To do so follow the procedure :

Find the zero optical path difference position.

Adjust the interferometer to get a number of fringes (between 10 and 20).
Record the distance on the translation stage.
Adjust the exposure and intensity setting to avoid saturation.
Move the translation stage and record images for different positions.
Evaluate the fringe contrast with Matlab.

Contrast is given as the difference of the maximum and minimum intensity over the sum of
maximum and minimum intensity.
I I
C = max min
Imax + Imin
Image at zero OPD position (A), after adjustment to have number of fringes without
saturation (B) and at a position x when moved with the translation stage (C)

21

An example Matlab routine is stored under the file name fringes_center_line_red.m

% Read image 1
I_1 = imread('picture 166.jpg');
% select a channel
Red_1 = I_1(:,:,1);
%plot the channel
figure
imagesc(Red_1,[0 255]);
xlabel('position [pixel]')
ylabel('position [pixel]')
% format conversion
Red_1 = double(Red_1);
% select a line of interest ROI (1 x 1600), here center line
CenterLine_Red = Red_1(600:600,1:1599);
%plot a single line
figure
plot(CenterLine_Red);
xlabel('position [pixel]')
ylabel('signal [counts]')
%average over several lines (vertical)
% select a region of interest ROI with several number of lines N (N x
% 1600), here 200 lines are averaged
ROI_Red_1 = Red_1((600-100):(600+100),1:1599);
%caculate the mean of the RIO obver several rows
N_Avg_Red_1 = mean (ROI_Red_1);
%plot the averaged line
figure
plot(N_Avg_Red_1);
xlabel('position [pixel]')
ylabel('signal [counts]')

22

The corresponding images and line plots in Matlab are given below.

averaged signal

averaged signal
250

200

200

200

150

100

50

signal [counts]

250

signal [counts]

signal [counts]

averaged signal
250

150

100

50

200

400

600

800
1000
position [pixel]

1200

1400

1600

150

100

50

200

400

600

800
1000
position [pixel]

1200

1400

1600

200

400

600

800
1000
position [pixel]

1200

1400

1600

Figure 22. Examples of the fringe contrast obtained at different positions of the linear stage.
The contrast changes between pictures varies between the lasers and might be difficult to
evaluate. The use of line averaging in the Matlab script help in reducing the noise. To apply
line averaging, you have to avoid saturation of the image and avoid deformation of the
fringes!! Do not use too many lines.
The curve is periodic. This model allows to judge the spectral width (coherence) of the
source.

Figure 23. Contrast as a function of measurement position for a real measurement. x-axis is
the measurement point number taken each 100 micron. y-axis is the contrast value. The
distance between two maxima is approximately 2 mm.

To be done for the report: Show three pictures with different contrast and plot their line
curve. Measure the contrast as a function of position of the translation stage (2 full cycles of
contrast variation, minimum 20 points). Calculate the spectral widths of the source using the
period of the variation. (Eq. 12, where z is the period found in your measurement!)
23

4.4 Phase shifting interferometry

Phase shifting interferometry can be used to measure surface geometries The simplest
geometry is a inclined surface at different inclination angles. We want to measure such an
angle by using the simplest three value phase shifting algorithm. To do so the interferometer
has first to be aligned and set up to create high contrast fringes.
Your should adjust the interferometer to find an interferogram like shown in the image below.

Figure 24 High contrast interferogram with vertically aligned fringes for easy averaging over
horizontal lines.
To shift the position of fringes one has to change the relative positions of the two arms. We
need to shift by maximum by 180 () which corresponds to 317 nm (wavelength 635nm) in
real space. This can be easily achieved by touching the board at a convenient position. One
possibility is shown below.

Figure 25. Touching position to shift the fringes. Pressing more or less strong allows to
control the fringe movement an position.
The quality of the measurement depends critically on the phase shifting values. We need 0,
90 and 180. It is recommended to take several images and choose the best one to make the
evaluation.
NOTE: A 180 phase shift means here a shift by half fringe!
24

Figure 26: Example of images that represent a series of fringes for different pressure accessed
on the board. The best candidates for evaluation have 90 phase shift. They are given in the
sequence below.
An example Matlab routine is stored under the file name phase_shifting_interferometry.m
% phase shifting interferometry uses three images with a defined phase
% step.The phase s
hould have 0, 90 and 180 shift. Interfereograms will be recombined.
% The calculation is based on subtracting interfereograms from each other.

% Read image 1 (0)

I_1 = imread('picture 164.jpg');
% Read image 2 (90)
I_2 = imread('picture 165.jpg');
% Read image 3 (180)
I_3 = imread('picture 166.jpg');
% select a channel
Red_1 = I_1(:,:,1);
Red_2 = I_2(:,:,1);

25

Red_3 = I_3(:,:,1);
%plot the channel
figure
imagesc(Red_1,[0 255]);
colormap(gray)
xlabel('position [pixel]')
ylabel('position [pixel]')
figure
imagesc(Red_2,[0 255]);
colormap(gray)
xlabel('position [pixel]')
ylabel('position [pixel]')
figure
imagesc(Red_3,[0 255]);
colormap(gray)
xlabel('position [pixel]')
ylabel('position [pixel]')
% format conversion
Red_1 = double(Red_1);
Red_2 = double(Red_2);
Red_3 = double(Red_3);
% select a line of
CenterLine_Red_1 =
CenterLine_Red_2 =
CenterLine_Red_3 =

interest ROI (1 x 1600), here center line

Red_1(600:600,1:1599);
Red_2(600:600,1:1599);
Red_3(600:600,1:1599);

%plot a single line

figure
plot(1:1:1599,CenterLine_Red_1,'g-',1:1:1599,CenterLine_Red_2,'r',1:1:1599,CenterLine_Red_3,'b-');
xlabel('position [pixel]')
ylabel('signal [counts]')
%average over several lines (vertical)
% select a region of interest ROI with several number of lines N (N x 1600)
ROI_Red_1 = Red_1((600-100):(600+100),1:1599);
ROI_Red_2 = Red_2((600-100):(600+100),1:1599);
ROI_Red_3 = Red_3((600-100):(600+100),1:1599);
%caculate the
N_Avg_Red_1 =
N_Avg_Red_2 =
N_Avg_Red_3 =

mean
mean
mean
mean

of the RIO obver several rows

(ROI_Red_1);
(ROI_Red_2);
(ROI_Red_3);

%plot the averaged line

figure
plot(1:1:1599,N_Avg_Red_1,'g-',1:1:1599,N_Avg_Red_2,'r',1:1:1599,N_Avg_Red_3,'b-');
xlabel('position [pixel]')
ylabel('signal [counts]')
%calculate the tangens of the phase shifting value (I3-I2)/(I1-I2)
Tan_Int = (N_Avg_Red_3-N_Avg_Red_2)./(N_Avg_Red_1-N_Avg_Red_2);
figure
plot(Tan_Int,'k-');
xlabel('position [pixel]')

26

ylabel('signal [counts]')
%creates the phase signa (arcus tangens) with values between -Pi/2 and Pi/2
phase= atan(Tan_Int);
figure
plot(phase);
xlabel('position [pixel]')
ylabel('signal [rad]')
%Calculation of the unwrapped phase and height profile
%(unwrapping within -Pi/Pi, multiply by 2 and devide the final phase by 2)
%Dividing by 2Pi and multiply by half wavelength of 635 nm
height_profile=unwrap(2*phase)/2/(2*3.14)*0.635/2;
figure
plot(height_profile);
% axis description
xlabel('position [pixel]')
ylabel('height [micron]')

The MATLAB program produces a complete evaluation as described below.

Recorded interferograms

Single and averaged line plots

27

Arctan values and phase within -Pi/2 and Pi/2

Height difference as a function of position.

To be done for the report: Plot the three phase images and demonstrate the different steps
for a complete evaluation. With the help of the detector size, calculate the tilt angle you have
realized with this measurement. (the tangent of the tilt angle is the maximum height divided
by detector width)
HINT:
Never touch the optical surfaces.
Each laser has different emission characteristics. It might be more or less difficult to
obtain fringes and measure the contrast. You might change the laser if needed.
The laser needs a certain stabilization time. The emission has therefore also a time
change and you should be careful when measuring the contrast.
The interferometer is extremely sensitive. So do not touch components ones aligned
and fix all mechanical parts.
The translation stage has to be used in one movement direction only when a precision
measurement should be done to avoid tilt of mirrors and measurement incertitude
because of backlash. Move only in one direction (forward or backward) when doing a
set of measurements.
28

5.Summaryoftasksoftheexperimentalwork
4.2 Zero optical path difference - ZPD
Find the zero OPD position and plot the intensity profile for the low intensity state and the
high intensity state. Take care that you use equal exposure conditions for both states to show
the contrast correctly.
4.3 Measurement of laser fringe contrast
Show three pictures with different contrast and plot their line curve. Measure the contrast as a
function of position of the translation stage (2 full cycles of contrast variation, minimum 20
points). Calculate the spectral widths of the source using the period of the variation. (Eq. 12,
where z is the period found in your measurement!).
4.4 Phase shifting interferometry
Plot the three phase images and demonstrate the different steps for a complete evaluation.
With the help of the detector size, calculate the tilt angle you have realized with this
measurement. (the tangent of the tilt angle is the maximum height divided by detector width)

29

APreliminaryquestions
Under which conditions interference appears?
Where do the shapes of the fringes come from?
What is so special about the zero optical path difference?

BFinalQuestions
In a single fringe situation and when the fringe is getting dark, explain where the
energy is found!
Explain the interdependence between coherence lengths and fringe contrast
qualitatively? (No calculations!)
What is the problem of phase shifting interferometry applications?

30