which illuminated the object. A string fixed on one end 
to the spatial filter and on the other end to the hologram 
plate holder was used to equalize the pathlengths of 
object and reference beams and to optimize the position 
of the reference mirror. 
On the floor to the right of the machine (Fig. 3) was 
placed a heavy (25-kg) piece of steel to be used as a fixed 
reference surface in relation to which the motions of the 
machine could be defined. As the forces deforming the 
machine were internal, there was little reason to believe 
that the floor would transmit any deformations from the 
machine to this surface. 
lll. Sandwich Plate Holder and Its Uses 
Close to the spatial filter at the laser we placed an- 
other heavy, about 1 m high, steel cylinder on top of 
which the plate holder was fixed. This specially de- 
signed sandwich hologram plate holder is seen in Fig. 
4. It is extremely stable and has no moving parts or 
adjustments. The hologram plate rests only by gravity 
because even weak springs tend to bend the thin glass 
plate. It leans slightly (5°) backward against three 
supports in the form of short pins of hard metal with 
semispherical ends. The plate is placed with its longer 
side resting on two 20-mm long hard metal pins with a 
diameter of 6 mm and positioned at different levels. 
Therefore, the plate slides down at an angle of 30° to the 
horizontal line until it comes to rest against a third pin. 
Eventually the plate rests by gravity on the three ball 
contacts at its back surface, on the two pins at its longer 
side, and on the single pin at its short side. 
With this plate holder we have found it easy to re- 
position hologram plates within one or two fringes for 
real time holography or sandwich holography. Without 
need of adjustments the same holder or an identical one 
is used to expose, reconstruct, and glue together the 
sandwich holograms. If many fringes are caused by 
reposition errors, it is usually found that a small piece 
of glass has broken away at one of the points where the 
edge of the plate made contact with one of the support 
pins. 
I do not agree with those who find it difficult to make 
sandwich holograms because of reposition problems and 
therefore have to invent alternate methods. Our 
students use the plate holder, mostly with the same 
result that I get; about eight times out of ten the errors 
will, without any adjustments, be less than two fringes. 
Even without any hologram holder, holding the plates 
by hand, it is with some patience possible to manipulate 
the two plates so that they form a hologram sandwich 
with zero fringe on the reference surface. 
When one starts using sandwich holography for the 
first time it might be useful to proceed as follows. First, 
one hologram plate is placed into the holder, its emul- 
sion, e.g., facing the object, its back side resting against 
the spherical contact points, and two of its edges resting 
against the long support pins. Thereafter another plate 
is placed in front of the first, also with the emulsion 
facing the object. This second plate is placed onto the 
long support pins and gently pushed with a finger at its 
center until their surfaces are pressed together, and the 
back plate rests against the contact points. (Neither 
2524 APPLIED OPTICS / Vol. 16, No. 9 / September 1977 
plate should have any antihalo coating.) It is important 
that the long support pins are of hard metal. Otherwise 
they might be scratched by the glass edges which pre- 
vents them from sliding smoothly. After waiting a few 
minutes the first exposure is made. 
The two plates are taken away, and two new plates 
are placed in the holder in exactly the same way as the 
first two. After deformation of the object and a waiting 
time of a few minutes until everything has settled down, 
the second exposure is made. The four plates are 
thereafter developed and fixed in the conventional 
way. 
IV. Checking for Reposition Errors 
Now it is time to check the result. After the plates 
have dried (in my experience one should not use hot air 
or alcohol to speed the drying process of real-time or 
sandwich holograms), the first two plates are placed 
back in the plate holder in the same way as prior to the 
first exposure. The object is obstructed, and the ref- 
erence beam is used for reconstruction of the sandwich 
hologram. The holographic image of the object should 
be seen fringe free or with at most one fringe. If one of 
the plates is pushed sideways with a finger, away from 
one of the support pins, straight fringes should appear 
and disappear as the plate slides back again. If the two 
plates by mistake are interchanged before they are re- 
positioned so that the plate which was exposed as a front 
plate (facing the object) is put behind the other during 
reconstruction, a number of dark concentric circles will 
appear covering the object image. Similar circles will 
also appear if the wrong (pseudoscopic) image is studied 
or if the plates are placed with their emulsion in the 
wrong direction. The two plates that were exposed 
after the deformation should pass through the same test 
program. 
If there exists no fixed reference surface in the object 
scene, this study of the fringes is almost necessary to 
make sure that reposition errors are not too large. 
When the combinations of the holograms in the way 
described here result in no more than one or two fringes, 
there is good reason to believe that the sandwich holo- 
gram will be successful. If, however, one of the two 
exposed pairs produces many fringes, something has 
gone wrong, and it can be useful to find which plate is 
erroneous. For this purpose we need a holographic 
plate with the emulsion removed so that its glass base 
can be used as a substitute (compensation plate) for one 
plate of a sandwich hologram. This compensation plate 
is placed in the plate holder. A front plate from one 
exposure is placed in front of it, and the real time 
fringes! are studied. The reference surface, or undis- 
placed surfaces on the object, should be fringe free or 
have at most one or two fringes. If we forget to place 
the compensation plate behind the front plate, it will 
be repositioned too far from the object, which will 
therefore appear covered by fringes in the form of con- 
centric circles. 
The same procedure should be repeated with the back 
plate of one exposure which should be repositioned in 
the plate holder behind the compensation plate. Also 
this time there should be no, or only a few, real time 
frix 
pla 
cen 
the 
the 
of 1 
fer 
stu 
pla 
to ! 
cou 
rel. 
ha 
pla 
she 
ex] 
is S 
tio: 
pai 
Or 
obj 
ho] 
nai 
usi 
ba: 
rel 
sec 
ple 
cer 
gla 
su] 
evi 
ho 
as 
Col 
of 
erg 
frc 
rir 
It 
CO] 
ex, 
rec 
th 
un 
th 
pa 
of 
by 
pa 
re 
toi