important 
Otherwise 
vhich pre- 
ting a few 
ew plates 
vay as the 
| a waiting 
led down, 
olates are 
ventional 
the plates 
1se hot air 
il-time or 
re placed 
rior to the 
d the ref- 
sandwich 
ct should 
If one of 
way from 
ld appear 
If the two 
ey are re- 
‘as a front 
ler during 
ircles will 
ircles will 
is studied 
on in the 
> exposed 
same test 
the object 
essary to 
oo large. 
| the way 
vo fringes, 
vich holo- 
f the two 
thing has 
h plate is 
lographic 
"lass base 
e) for one 
tion plate 
from one 
real time 
or undis- 
ce free or 
t to place 
te, it will 
hich will 
m of con- 
1 the back 
tioned in 
ite. Also 
real time 
fringes. Ifthe compensation plate in front of the back 
plate is not used, there will appear a number of con- 
centric interference circles again. The reason is that 
the back plate reconstructs the object as seen through 
the front plate during exposure. Therefore, the image 
of the object should, during real time hologram inter- 
ferometry, also be seen through a similar plate. These 
studies of real time fringes are very useful to find which 
plate was placed in the wrong position and to learn how 
to reposition the plates correctly. Most errors are, of 
course, made prior to exposure when the room has to be 
relatively dark. 
After that one or more of the checks described here 
have been made, it is time to combine the sandwich 
plates. The front plate from the second exposure 
should be placed in front of the back plate of the first 
exposure. After the object has been obscured, its image 
is studied using the reference beam for the reconstruc- 
tion. On the fixed reference surface and on undisplaced 
parts of the object, there should be no, or at most one 
or two, fringes. These fringes in the reconstructed 
object space can now be manipulated by tilting the plate 
holder as described in Ref. 5. 
If a permanent record of a certain sandwich combi- 
nation is wanted the two plates can be bonded together 
using, e.g., Cyanolite, which is a one-component cement 
based on cyanoacrylate. It is similar to Lock-tite and 
related to cements used for bonding strain gauges. 
When pressed into a thin film it hardens in less than 10 
sec. To bond the sandwich hologram together, the back 
plate is first placed in the holder, and two drips of the 
cement are applied close to two diagonal corners on the 
glass side of the front plate. This is then placed on the 
support pins and with a finger on the center gently 
pushed toward and pressed against the back plate in the 
plate holder for a few seconds. 
After the sandwich hologram is bonded it can be 
evaluated anywhere without the use of a stable plate 
holder. It is, however, important that it be illuminated 
by a reconstruction beam that has the same divergence 
as the reference beam during exposure. This is ac- 
complished most easily by arranging so that the distance 
of the point from which the reconstruction beam div- 
erges is the same as the distance was during exposure 
from spatial filter to hologram holder via the reference 
mirror. If those distances are not identical, interference 
rings and erroneous fringes will appear on the object. 
It is just as important that the wavelength of the re- 
constructions beam be identical to that used during 
exposure. Ifan argon laser is used for the exposure, a 
reconstruction with a He-Ne laser will produce fringes 
that again will have the form of concentric circles on 
undisturbed surfaces. 
During the reconstruction of a sandwich hologram. 
the image from the emulsion of the front plate has to 
pass through the glass base before reaching the emulsion 
of the back plate (Fig. 1 of Ref. 5). Thus it is distorted 
by variations in glass thickness of the front plate. The 
back plate, however, is reconstructed by a beam that has 
passed through the front plate, and therefore the two 
reconstructed images will have almost identical dis- 
tortions when they interfere. The smaller the angle 
between object and reference beam the better this 
compensation works. Therefore the risk of erroneous 
fringes caused by glass defects is minimized if the object 
occupies a small angle of view and is placed close to the 
reference mirror. Thus we have found a fifth reason for 
a large distance between object and hologram holder. 
When a good sandwich hologram is reconstructed 
correctly it should be possible to tilt it in such a way that 
not a single fringe is seen on the reference surface. Even 
if the eye is moved behind the hologram plate no fringe 
should appear on that surface. Sometimes, however, 
one or two false fringes appear which do not destroy the 
use of the hologram, but it is important that the eye or 
camera be kept close to the plate and in the same place 
during the observations. This test should be made 
prior to the evaluation, and it should be decided what 
part of the hologram to use. The rules for the evalua- 
tions of sandwich holograms have been presented in 
Ref. 5. It should be pointed out that the observation 
should be made from a point close to the plates and that 
the starting point from which the sandwich tilt is mea- 
sured is the angle at which no fringes are seen on the 
reference surface. The rules of tilt direction are true 
only if the first exposed back plate is placed behind a 
second exposed front plate. 
To save plates a hologram plate can be placed behind 
a compensation plate for the first exposure and another 
hologram plate in front of a compensation plate for the 
second exposure. One further advantage of this method 
is that the hologram image will improve slightly, because 
the back plate does not have to see the object through 
the emulsion of the front plate. Still another advantage 
is that even antihalo coated hologram plates can be 
used. 
There is, however, one main disadvantage in this 
method if many exposures are made with the hologram 
plates alternatively in front of, or behind the, compen- 
sation plate. In that case every combination cannot be 
used because the sandwich hologram must consist of one 
front and one back plate. 
V. Results of the First Experiment 
The vertical kneetype milling machine used in the 
first experiment was deformed by a simulated cutting 
force produced by a pneumatic membrane which was 
placed between tool and workpiece. A hologram plate 
with its emulsion forward was placed behind a com- 
pensation plate in the holder, and the first exposure was 
made with the machine at rest. The plate holder was 
emptied, and another hologram plate with its emulsion 
forward was placed in front of the compensation plate. 
Air pressure was applied to the membrane to simulate 
a low cutting force, and after a few minutes a second 
exposure was made. The two hologram plates were 
processed and put together to form a sandwich. 
The result is seen in Fig. 3. The top of the machine 
is the head, while the big part protruding to the right of 
the main body is the knee which is supported from the 
base by a heavy screw. On top of the knee is the table 
which has totally disappeared from the holographic 
image because it reached far outside the ellipsoid rep- 
resenting the limited coherence length (see Fig. 2). On 
September 1977 / Vol. 16, No. 9 / APPLIED OPTICS 2525