Technical Field The present invention relates generally to analysis of the microstructure present in large trans- parent or semi-transparent volumes or over large object surface areas and, more particularly, to the performance of such analysis through the use of opti¬ cal processing techniques.

Background Art As is well known in the art, an optical Four¬ ier transformation can be performed on a subject by illuminating the subject with a coherent light source and using a lens to collect the reflected, and/or dif¬ fracted and/or transmitted light containing optical information corresponding to the subject. The lens will define a Fourie ^' r transform of the orignal subject information in a plane located one focal length away from the lens at its focus. With this process, sub¬ ject information is redistributed in the Fourier transform plane to correspond to spatial frequency content. A second lens, located two focal lengths from .the first, images the information onto a display screen, recording medium, or the like.

The optical information, coming from the sub- jeσt may be manipulated for the purpose of enhancing detail, removing unwanted subject information, isolat¬ ing defects, or making precise topographic and optical path measurements . This is accomplished by blocking a portion of the subject information or changing its phase within the transform plane. These techniques are particularly useful where the subject consists of repetitive spatial frequency content such as is pre¬ sent in photomasks or wafers used in the production of microelectronic circuits, since the- optical Fourier transform will consist of an array of regularly spaced points of light whose distance from the optical axis is proportional to the spatial frequency.

- ^0M?I

One example of such a technique is disclosed in U. S. Patent No. 4,000,949 issued January 4, 1977 to Watkins. ^' Using the basic processing scheme out¬ lined above, a photomask is used as the subject. An optical spatial filter is placed within the Fourier transform plane for blocking all subject information other than that corresponding to nonperiodic defects in the mask. Since only defect information will pass through to the display screen, the number of defects , and their locations can be determined.

It can be easily seen, however, that aberra¬ tions within the optical components of an optical pro¬ cessor, particularly within lenses focusing the sub¬ ject information and/or the processed information, will be a significant detriment to use of the otpical processor with subjects having microscopic detail. On axis, for example, a focus error aberration can cause loss of spatial frequency information, as illustrated by way of example for a typical optical telescope system in Fig. 1.

A three-lens system 24 is shown, wherein lenses 1. and 1 ₂ are F/1.50, 50 mm diameter, 75 mm focal length lenses, and lens 1- is an F/1.50, 150 mm diameter, 225 mm focal length lens. Assuming the sum of the spherical aberration from lenses 1_ and 1 ₃ to be 2 mm, the focus of the rays 26 entering the system 24 is moved from a point 28 to a point 30, 77 mm away from lens 1, . This results in a smaller.

F/1.54 collection angle at lens 1..

Therefore , with a 225 mm focal length for lens I, , which would give an F/4. 5 collection angle for lens 1- and the system 24 as a whole if lens 1, collected F/1.50 , the lens can only collect: i__50 = ≤ __i y = (1 . 54) ( 4. 50) _ _/4

1. 54 Y ^Y 1. 50 * / ⁴ . t> ⁴ .

OMP m IP

Si ilarly, for 1 mm of special aberration, the focal point is moved to 76 mm away from lens 1, an :

(1.52) (4.50)

Y = 1.50 F/4.56 for the system 24.

To consider these effects directly, the optical transfer function (OTF) of the system may be looked to.

For a diffraction limited system:

where a(fx,fy) is the area of overlap of the pupil collecting spatial frequencies with the restricting pupil function.

The OTF with aberrations for two pupils given by:

*( ξ- ^Λ - n - Ififz )^d P (ξ ₊ ifA, η + ^λd i ^£ ₂ v)

is H ^* (f _χ ,f _y ) =

a{f ,f ) exp {ik [W (ξ +- ^λd i ^f x, n + ^λ _i_∑) ^{x y} 2 2

- W (ξ - ^λd i ^£ , η - ^λd i ^£ y)3 > dξdη 2 2

//dξdη a(o,o) where W is the aberration function. The Schwarz inequality,

//X Y dξdri ≤ ( // dξ dη) ( / dξdη)

can be used to show directly that f

^H x'fy)'no aberrations

This means that aberrations never increase the MTF (the modulus of the OTF) , but rather lower 5 the contrast of each spatial frequency component. Thus, the cutoff to the spatial frequency passed by the system will be effectively decreased and the effective F-number thereof will increase. This is especially important if there is a random background 10 to further decrease the contrast.

Off-axis, spatial frequency information can be lost even in an ideal aberration-free system directly due to vignetting, as shown in Figs. 2 and 3. Again, considering the sample three-lens optical 15 system 24, the limit of the system 24 is found by tracing rays from the edge of lens 1, . The angle of incidence of these rays at a point on the subject plane 34 gives the cone angle of spatial frequency information which can be collected from that point 20 on the subject.

In Fig. 2, using a thin lens ray trace, it can be seen that only an F/9 cone can be collected • by system 24 from a point 36 located 37.5 mm off the axis 37. In Fig. 3, a point 38 located 50 mm off 25 the axis 37 is considered. Again using a thin lens ray trace, it can be seen that only an F/13.5 cone can be collected due to vignetting.

In order to utilize optical processing techniques on objects having microscopic detail with ^• **- ^υ an optical system such as is used in Watkins , it is necessary to employ within the system opitical com¬ ponents of very high quality. Such components will then minimize aberrations within the system. ^" Addi¬ tionally, to avoid loss of information due to 35 vignetting, system apertures must be made as large

as possible. Unfortunately, large aperture optical components pf very high quality, particularly lenses, can be obtained only at extremely high cost, making optical processing for such objects far less practical.

Use of high quality lenses creates addi¬ tional complications other than cost where high- resolution records are desired, since as the resolu¬ tion of an imaging optic is increased, the depth of focus thereof decreases. Thus, the depth of focus available in the image plane is limited.

Full-field documentation with relatively high resolution can be accomplished through holo¬ graphy. By holographically recording a subject, and then reconstructing the holographic image with the conjugate to the reference beam, a three-dimensional real image of the object is created in space. This real image can then be examined microscopically as if it were the original illuminated object. An added advantage is that a holographic system can document not only reflecting surfaces but also the microstructure inside a thick, transparent object such as optical components. Since the holographic image has no substance, a microscope can focus through the image, even to the opposite side if desired, without the need for a long focus and high quality objective required when a solid object is in the way.

Examples of previous work in holographic microscopy are disclosed in Leith and ϋpatnieks,

"Microscopy by Wavefront Reconstruction," 55 J. Opt. Soc. Amer. 569 (1965); Toth and Collins, "Recon¬ struction of a Three-Dimensional Microscope Sample Using Holographic Techniques," 13 Appl. Phys. Letters 7 (1968); and Briones, Heflinger and

Wuerker, "Holographic Microscopy," 17 Appl. Optics

1944 (1978) . The records produced are large aper¬ ture wide field of view, large depth of field, three-dimensional images of both transmitting vol¬ umes and specularly or diffuse reflecting subjects. 5 Such work, however, has not as yet obtained both the necessary resolution and field of view for viewing and/or analyzing the microstructure of large object surface areas.

One technique available in holography, άes-

10 cribed in Toth and Collins, is known as reverse ray-tracing. When a hologram 10 is made of a sub¬ ject 12 through a lens 14 or other optical compon¬ ents, the image information 16 from the subject may be aberrated by the lens 14 (Fig. 4a) . If the holo-

15 gram 10 is repositioned accurately with respect to the lens 14, and the conjugate to the reference beam 18, i.e., the same wavefront as the reference beam but travelling in the opposite direction, is used to reconstruct the holographic image, the image rays 20

20 . will exactly retrace the path of the original sub¬ ject rays back through the optical system (Fig. 4b) . This is not the same as merely turning the lens 14 around, since the information about the lens aberration is stored in the hologram 10. Therefore,

25 the aberrations of the lens 14 will be completely compensated for upon reconstruction and the hologra¬ phic image 22 will be diffraction limited.

It can be seen that to produce a high resolution holographic image 22, it is not necessary

30 to use a high quality lens 14. Thus, it would appear promising to attempt to incorporate the holo¬ graphic reverse ray-trace technique into an optical processing system.

.What is needed, therefore, is a method and

^■• -5 apparatus for optical processing that utilizes holo¬ graphy, and specifically the reverse ray-tracing

technique. Such a method and apparatus would enable optical processing to be performed on both detailed three-dimension microscopic and macroscopic subjects with high resolution. Disclosure of Invention

The present invention provides a method and apparatus for performing optical processing on opti¬ cal information corresponding to a subject that com¬ bines known optical processing techniques with holo- graphy. By utilizing a reverse ray-trace hologra¬ phic system, aberrations introduced by the system optics are completely compensated for, producing diffraction-limited three dimensional images of a subject. A single large aperture lens is used to define a Fourier transform of the optical subject information as the hologram is recorded or the image reconstructed, and the subject information is manip¬ ulated with the transform plane.

A source for generating a coherent beam of light, preferably an argon ion laser operating sin¬ gle frequency, provides a beam to a beamsplitter that divides the beam into a subject beam and a reference beam. The subject beam is directed to the subject where it is reflected and/or diffracted thereby. The beam, which now contains optical information regarding the subject, .is passed through a lens and focused to define a Fourier transform, imaging the information in the vicinity of a photo-sensitive material, such as a photographic plate. The reference beam is simultaneously direc¬ ted onto the plate at a predetermined angle of incidence.

After processing, the hologram is replaced and illuminated with the reference beam which is redirected onto the plate in a direction conjugate to the original reference beam. Subject information

from the hologram will then exactly reverse ray trace back through the optical system following the same path but in the reversed direction as original subject rays during recording. A three-dimensional real image of the original subject will be formed at the original subject location. Because of the exact reverse ray-trace feature, optical distortions or aberrations introduced by optical elements ^' in the system will be completely eliminated thus providing a diffraction limited three-dimensional real image. The subject information is manipulated at the Fourier transform either during recording of the hologram or during image reconstruction. This may be performed, for example, by placing a blocking filter within the transform plane for blocking at least a portion of the .subject information. Alter¬ natively, a filter may be placed within the trans¬ form plane for altering the phase of at least one of the beams passing therethrough. Other processing techniques are also possible.

This invention offers several important advantages over known optical processing systems and methods.

First, the reverse ray-trace of the image information through the optical system provides an image that is free of aberrations introduced by optical elements.

Second, since such aberrations are elimi¬ nated, the optical quality of the imaging optic is not as critical as in known systems. This permits use of components of lesser quality, with very sig¬ nificant reductions in costs.

Third, the use of a single, large-aperture lens provides for a wide field-of-view, with little information loss due to vignetting.

Fourth, the high resolution obtained with this invention provides for microscopic evaluation of recorded images, and enhances results obtained through optical processing and interferrometric techniques.

Accordingly, it is an object of the present invention to provide a method and apparatus for the performance of optical processing to provide such a method and apparatus for utilizing holographic images free from aberrations due to components of the system; to provide full-field, high-resolution holograms capable of microscopic examination; and to provide such a method and apparatus that can perform a variety of optical processing and interferrometric techniques.

Other objects and advantages of the inven¬ tion will be apparent from the following descrip¬ tion, the accompanying drawings, and the appended claims. Brief Description of Drawings

Fig. 1 is a diagram of a hypothetical tele¬ scope system, showing the effect of focus error aberration;

Fig. 2 illustrates the hypothetical tele- scope system of Fig. 1, showing the effect of vignetting;

Fig. 3 illustrates the hypothetical tele¬ scope system of Fig. 1, again showing the effect of vignetting; Fig. 4a is a diagram illustrating generally the recording of a hologram, showing introduction of aberration into the image information;

Fig. 4b is a diagram illustrating generally the reconstruction of a holographic image, wherein reverse ray tracing eliminates aberration from the image information;

-10-

Fig. 5 is a diagram of the apparatus of the present invention;

Fig. 6 is a diagram of an alternative embodiment of a portion of the present invention for use when the subject is a transparent or semi-trans¬ parent volume;

Fig. 7 is a diagram of a further alternative embodiment of a portion of the present invention for use when the subject is a transparent or semi-trans- parent volume;

Fig. 8 is a diagram of a still further alternative embodiment of a portion of the present invention for use when the subject is a transparent or semitransparent volume; Fig. 9 is a diagram of an alternative embo¬ diment of a portion of the present invention for recording an image plane hologram; and

Fig. 10 is .a diagram showing image recon¬ struction with the embodiment of Fig. 9. Modes for Carrying Out the Invention

Referring now .to the drawings of the pre¬ ferred embodiments of the invention, and particular¬ ly to Fig. 5, the optical processing apparatus includes a laser 40 that provides a source for gene- rating a coherent beam of light 42.

In the preferred embodiment, an argon ion laser operating single frequency at a wavelength of 0.5145 micrometers is used, in particular, a Spectra Physics Model 166-09 Argon Laser. Any laser source, however, whose output wavelength is compatible with the recording medium will work, including but not restricted to helium-neon, krypton, ruby, neodymium YAG or glass, and metal vapor lasers.

Beam 42 is directed by mirrors Ml and M2 into the system- 44 and passes through a half-wave plate 46 which allows the angle of polarization to

be continuously adjusted before entering the polari¬ zing beamsplitter 48. It should be noted, however, that although polarizing beamsplitting techniques are used in the preferred embodiment to conserve energy and allow maximum flexibility, any beamspli - ting technique that divides the incoming beam into two components will work, such as metal films, grat¬ ing, dielectric films and holographic elements.

Beamsplitter 48 divides- the input beam 42 into a first or subject beam 50 and a second or reference beam 52 having an intensity ratio deter¬ mined by the angle of polarization of the incoming beam.

The subject beam 50 is then directed to a means 54 for expanding, filtering and collimating beam 50 including lenses 56 and 58 and a mask having a pinhole 60. It then passes through a second polarizing- beamsplitter 62 which serves as a selec¬ tive beam director. This is accomplished in the following manner: light entering this director (beamsplitter 62) passes through because of the polarization orientatioii established by beamsplitter 48. Beam 50 then passes through a quarter-wave plate 64 where it is converted from linear to cir- cular polarization. It then strikes a subject 66 and is reflected thereby back through the -quarter- wave plate 64 where it is converted back to linear polarization but orthogonally oriented. This causes the subject beam 50 to be reflected by the beam¬ splitter 62 towards a large aperture lens 68 which forms an image in the vicinity of a means for recor¬ ding the hologram. In the preferred embodiment, lens 68 is a symmetric doublet pair, imaging subject 66 on or near the recording means at one-to-one mag- nification to maximize the information collected by the lens 68. Other means for directing the subject

information to the recording means may be used, including placing the recording means close to the subject and eliminating the imaging system.

While the recording means is shown in the preferred embodiment as a photographic plate 70, it will be recognized that it may be any photo-sensi¬ tive material such as photographic films or plates, thermoplastic films, crystalline materials, or any other material capable of recording a hologram. Fig. 6 shows the preferred arrangement if the subject is a transmitting volume 72. Here a mirror M3 is placed behind the subject 72 which cau¬ ses the subject beam 50 to be reflected back through the quarter wave plate 64. This double pass scheme offers the advantage that anomalies of interest pre¬ sent in a subject volume 72 are illuminated from both the front and the back thus enhancing visibil¬ ity. When using the invention to make interferome- tric measurements of optical path changes or differ- ences, double passing the volume doubles the mea¬ surement sensitivity.

The variation shown in Fig. 7, or the more preferred Fig. 8, is used if the double passing or front illumination of a subject volume 72 is unde- sirable (e.g., in recording or optical processing of transparencies). Referring to Fig. 7, a half-wave plate 74 is placed in the subject beam 50 ahead of polarizing beamsplitter 62. This allows for adjust¬ ment of the subject beam polarization angle and is set to provide vertical or s polarization. Beam¬ splitter 62 now reflects the subject beam 50 toward mirror M4. This mirror directs the beam to mirror M5, where it is directed through the subject or sub¬ ject volume 72 and back into beamsplitter 62 an . thus into- lens 68 as described above.

IΠJIB

Similarly, in the preferred variation shown in F-ig. 8, the subject beam 50 is directed around the beamsplitter 62 by mirrors M6 and M7 where it is directed through the subject 72 and hal -wave plate 76 by mirror M8. It then passes through the system to the photographic plate 70 as previously described.

Referring back to Fig. 5, the reference beam 52 generated from beamsplitter 48 is directed by mirror M9 towards a means 78 for expanding, fil- tering and collimating beam 52 consisting of lenses 80 and 82, and a mask having a pinhole 84. Beam 52 is expanded and collimated and directed towards mir¬ ror M10. In the recording mode, mirrors M10 and Mil direct the reference beam 52 along reference path 86 to the photographic plate 70 where beam 52 combines with the subject beam 50 to expose plate 70, where¬ upon a hologram 88 is formed in plate 70.

After processing the hologram 88 and retur¬ ning it to position, mirror M10 is removed and mir- ror M12 directs the reference beam 52 towards the hologram 88 from the conjugate direction along con¬ jugate path 90. Subject information from hologram 88 will then exactly reverse ray-trace back through lens -68 and beamsplitter 62 following the same path but in the reversed direction as the original sub¬ ject beam 50 during recording. A three-dimensional image of the subject 66 will be formed at its origi¬ nal location. Because of the exact reverse ray-trace feature, optical distortions or aberra- tions introduced by optical elements such as lens

68, beamsplitter 62, or the like will be eliminated, providing a diffraction limited three-dimensional real image.

Iri the variation shown in Fig. 9, when the subject 66 or a portion thereof is flat (for exam¬ ple, an integrated circuit mask), the lens used is

color corrected (for example, achromat 92), and if care is taken to image the subject 66 on the surface of the plate 70, an image plane hologram is recor¬ ded. This image plane hologram can then be illumi- nated for reconstruction with a broadband source such as white light. This technique offers the advantage that incoherent source illumination will eliminate coherent noise in the reconstructed image. Reconstruction is illustrated in Fig. 10 where the hologram 88 is illuminated with a broad¬ band beam 94 along the conjugate path 90. Recon¬ struction of the image point will occur at the orig¬ inal subject position 96.

There may be background noise in the holo- graphic image which is not located at the subject plane. If the reference beam's angle of incidence is changed slightly during image reconstruction, this dislocated background noise moves across the field of view while the image is stationary. There- fore, much of this background noise originates from other optical components in the system-. Dithering the reference beam 52 while viewing or photographing the holographic image smoothes out the dislocated background and increases the ^" effective contrast. However, this is effective only for a single plane in the reconstructed image, because only a single- image plane of the subject can be exactly imaged in the plane of the hologram. Points in the recon¬ struction corresponding to points off the hologram are displaced with dithering of the reference beam 52 and will smear out.

In the preferred embodiment, hologram 88 is formed in an Agfa 8E56HD 10.2 cm x 12.7 cm (4" x 5") plate. The preferred developing processes, as experimentally determined, are either PAAP or HRP/Bromine Methanol. These were found to give

substantially better resul ts than standard D-19 process i ng . The lens 68 descr ibed here in collec ts approximately an F/3. 5 cone of light . Thus , the reconstr ucted real holographic image is dif f raction limited for the F/3.5 optical system:

dif f raction limi ted diameter = 2. 44 λ (F-number ) .

= 2. 44 ( .5145 ym) ( 3. 5) = 4.4 microns .

In experimental testing of the system, however, it appears that resolution of the holographic image better than four microns is attainable.

Because of the high resolution obtained with the presently disclosed system, documentation of the microstructure of a subject is readily facil¬ itated. The reconstructed image may be studied mi-croscopically at the original subject plane through an appropriate means therefor, such as microscope 98 shown in Fig. 5. Moreover, the observed information may be recorded, for example, by a camera (not shown) or other means, used in conjunction with microscope 98.

Because collimated coherent light is used to illuminate the subject in conjunction with a focusing element, an optical Fourier transform of subject information is formed in each of the embodi- ments of Figs. 5-9 between the hologram 88 and the lens 68 , one focal length from the focusing elemen , as seen for example at 91 in Fig. 6. This transform plane 91 can be used to manipulate subject informa¬ tion such as spatial frequency content, distribu- tion, and phase, by placing appropriate phase and blocking filters at this location. Such filters.may be of any of the presently known optical filters, examples of which are disclosed in U. S. Patent No.

4,000,949, issued January 4, 1977, to Watkins, and ϋ. S. Patent No. 4,299,443, issued November 10, 1981, to Minami et al. This serves to optically process subject information and can be performed prior to initial recording of the hologram or after the hologram has been recorded. In either case, the present inventio offers several advantages over standard optical processors.

First, the inverse Fourier transform is performed by the same optics that produced the original transform because of the reverse ray-trace feature of this process. Second, the optical quality of the imaging optic need only be sufficient to separate spatial frequencies at the Fourier transform plane relating to the desired image infor¬ mation. ^' This significantly reduces the cost and broadens the applications for optical processing _* In general, the Fourier transform is less sensitive to lens aberrations than the image. Third, the use of a single, large aperture lens provides for a wide field-of-view, with little information loss due to vignetting. Finally, the same apparatus may be used to produce effective blocking filters on photosensi¬ tive material such as photographic plates, which are accurately matched to the subject's spatial fre¬ quency locations in the Fourier transform plane. The optical processing may be performed through a number of techniques for a variety of reasons. For example, the blocking filters may be placed within the transform plane in order to par¬ tially or completely block all or a part of the subject information within the Fourier transform. This will enable the reconstructed image to .have certain details enhanced or suppressed, or a portion of the image to be removed entirely. Particular information may be isolated, for instance where the ^*

locations of defects within a particular subject are sought. While these advantages are available from optical processing generally, the high resolution achieved with the present invention enables such techniques to be carried out at microscopic levels and at reasonable cost. Additionally, since the filtering may be performed during image reconstruc¬ tion, the subject itself is required only long enough to record a hologram, allowing optical processing to be performed at a later time.

^• The optical processing may also be per¬ formed through a technique known generally as matched filtering. Rather than completely blocking all or a portion of the subject information as with blocking filters, all or a portion of the informa¬ tion may be diffracted (and/or reflected) out of the optical system at the transform plane. This may serve as a means of effectively blocking informa¬ tion, but can also be used to direct the deflected portion of the information into one or more other * systems where the information can be detected, observed, recorded, and/or further manipulated. Matched filtering may have any one of a number of practical applications, such as the separate isola- tion or detection of specific information pertaining to selected objects or features of a single object. This technique is accomplished using a selective refractive and/or diffractive medium, such as a hologram, as the matched filter. The subject can also be compared to an unperturbed reference wavefront in the following manner. A hologram 88 is made of the subject, pro¬ cessed, repositioned at its original location, and reilluminated with the reference beam 52 from the conjugate direction 90 via mirror M12. A second

reference beam is then combined with the real holo¬ graphic image at the original subject location via the path of original subject beam 50 and beam direc¬ tor 62. If the intensities are then made equal by adjusting half-wave plate 46, a fringe pattern can be observed. This fringe pattern will map the dif¬ ference in optical path between the surface figure- or optical density of the subject and the wavefront shape of the reference wavefront. Thus, the subject may easily be compared to an ideal flat or spherical surface.

If a plane reference wavefront is combined with the subject image wavefront but is shifted in phase by a quarter-wave relative to the beam used to reconstruct the holographic image, small changes in the subject's surface topography will be highlighted by interference fringes. This is accomplished using the technique outlined above for comparison to an unperturbed reference wavefront, but with a filter placed at the Fourier transform 91 of lens 68 so as to pass only the zero order component therethrough.. Either the reference beam 52 or the subject beam 50 is shifted in phase by means of a quarter-wave plate. The phase contrast fringe pattern is then viewed at the subject location.

A permanent record of the phase contrast image of the subject can be recorded by means of a double exposure hologram. A hologram of the subject is first recorded, and the filter is placed at the transform plane 91. Either the reference or subject beam 52 or 50 is shifted in phase by a quarter-wave, and a second hologram is recorded on the same holo¬ graphic plate. After processing and replacement to its original position, the hologram is then illu i- nated from the back or conjugate reference beam direction via mirror M12. Both the holographic

O

i age of the subject and the reference wavefront will then be reconstructed at the original subject position and can be viewed, for instance, with microscope 98. Real time holographic interferometric analysis of the subject may be accomplished by first recording a hologram with the apparatus arranged as in any of Figs. 5-9, depending on the subject type. The developed hologram 88 is then replaced in its original position, with the aid of a kinematic mount, and illuminated with the recording reference beam 52 with mirror M10 in place. Viewing would occur from the position indicated by arrow 100 (Fig. 5). If either the original subject 66 or 72, or some new subject to be compared, is simultaneously illuminated with light from subject beam 50, and made equal in intensity to the holographic image intensity by adjusting half-wave plate 46, then differences in subject optical paths will be observed as fringes appearing on or near the sub¬ ject. These optical path differences are referenced to the subject as originally recorded.

A diffraction limited interferogram can be reconstructed in the following manner. A double- exposed hologram is recorded by first recording a hologram of the subject in its reference state. Changes to the subject are then made or a different subject is inserted, and a second hologram is recor- ded on the same holographic plate. After processing and replacement to its original position, it is then illuminated by reference beam 52 from the back or conjugate direction 90 via mirror M12. Both holo¬ graphic .images will then reconstruct at the original subject position and can be viewed therefrom. Opti¬ cal path differences due to changes in the subject

occurring between exposures or differences between the two subjects will be observed as fringes on or near the reconstructed subject's image. Since this interferogram is inherently diffraction limited, due to the reverse ray-trace feature, extremely small fringe spacings can be accurately measured through a magnifying device, such as microscope 98.

While the method herein described, and the form of apparatus for carrying this method into effect, constitute preferred embodiments of this invention, it is to be understood that the invention is not limited to this precise method and form of apparatus, and that changes may be made in either without departing from the scope of the invention, which is defined in the appended claims. What is claimed is: