RECYCLABLE HOLOGRAPHIC STORAGE MEDIA

Joseph Bordogna, Scott A.Keneman,

Juan J. Amodei

Moore School of Electrical Engineering,University of

Pennsylvania, Philadelphia,USA

RCA Laboratories, Princeton, New Jersey, USA

Itek Corporation, Lexington, Massachusetts, USA

Performance parameters of state-of-the-art recyclable holographic recording madia are compared to develope tradeoffs for their use in holographic storage and imaging applications. Included are record energy, recoed time, erase time, diffraction efficiency, linearity, resolution, cycle lifetime, natural decay time, and a figure of merit. Materials and devices discussed include: magneto-optic, electroptic, and photochromic materials; metal films; ferroelectric-photoconductor, thermoplastic deformation, elastomer, liquid crystal-photoconductor-Pookel's effect devices.

The advent of the laser as a reliable commercial product has made possible the use of optical holography for information storage, processing, display, and image conversion system /1-3/. It is the purpose of this paper to discuss briefly and compare the materials /4/ and devices which researchers hope will store the holographic data and perform the real-time holographic imaging in such systems. In particular, an overview is given of the physical basis, fabrication, and operating parameters of available holographic storage media.

Recyclable media are characterised by their ability to be era-

sed and reused for sequentially storing different information patterns. Referring to such media as "holographic" implies, at least, that they are capable of supporting micrometer resolution. "Storage" implies an ability of the medium to keep the information interact after the hologram is recorded. In view of the embryonic state of development of light sensitive media for information applications this paper is not limited in scope only to media that satisfy the above criteria; materials and devices which satisfy some, but not all, of the requirements of ideal recyclable holographic storage media are discussed. Some media possess insufficient resolution; others are subject to decay. Some make no attempt at permanence but are useful for holographic imaging. Others fatigue after numerous operating cycles. We include materials subject to these deficiencies because it is likely that as technology advances and the understanding of the various media increases, some of these deficiencies may be removed and the realization of higher performance recyclable holographic storage media achieved.

The media discussed may be characterized as materials or devicies depending on their fabrication and use. The former possess a single constituent which is directly light sensitive. The latter are composite structures containing a light sensitive element but requiring more than light along to operate properly. The materials discussed include magneto-optic films, electro-optic crystals, photochromatic materials, and metal films. The devices include feroelectric-photoconductor devices, thermoplastic deformation devices, elastomer devices^*, photoconductor-liquid crystal devices, and photoconductor-Pockel.'s effect devices.

The media are compared on a number of operating parameters: record or write energy (quantity of energy per unit area required for recording-frequently referred to as "sensitivity"); re-

cord time (time required to record a hologram); erase time (time to clear storage medium for next hologram); diffraction efficiency (percentage of light energy in reconstructed image to incident energy of readout light); linearity (measure of attainable contrast or "gray scale"); resolution capability (number of resolvable lines per linear distance); cycle lifetime(number of times medium can be recorded and erased without deterioration of performance); and natural decay time (length of time hologram can be stored with no sustaining power). A comparison summary is presented in Table 1. Unfortunately, this table does not tell the complete story. Factors such as difficulty of fabrication, reproducibllity of parameters, and projected eventual cost are missing. Further research on these and other factors is in progress.

Curie point writing on manganese-bismuth films /5-10/ can be used for direct storage of digital information or for holographic storage or digital or pictorial information. Writting is achieved by switching the magnetieation of a semitransparent film (~ 300 to 700 nm thick) in response to the heating caused by local light absorption. Readout can be accomplished through the magneto-optic Faraday or Kerr effect which gives rise to rotation of the polarisation of the light transmitted or reflected. Because the writing mechanism relies on selectively heating the film above the Curie temperature in the region of higher light intensity, while leaving the regions of low intensity below the Curie temperature, this medium requires that writing be accomplished in times that are short compared to the thermal time constant of the film this time is normally of the order of 100 nanoseconds. This calls for high power pulses lasers, or focused CW lasers for high resolution writing, in order to supply the energy necessary to reach the Curie teaperatare in a sufficiently short

Table 1. Performance Parameters for Recyclable Holographic Storage Media.

time. Erasure of the pattern is accomplished by uniformly remagnetizing the film by means of an external magnetic field. The film deposition technique and other details of the operation and recording process are discussed at length by Mezrich and Cohen elewhere in this issue^*.

The most common magneto-optic recording material is manganese bismuth (MnBi). Its pertinent performance parameters are listed in Table 1. Other important parameters include: Faraday rotation > 5*10⁵ degrees/cm^-1, Kerr rotation ≈3°; absorption coefficient =3,5*10⁵ cm^-1; additional physical parameters are given by Mezrich and Cohen^*. Perhaps one of the most serious disad vantagee of MnBi is its performance deterioration under humidity conditions greater than 50%; however, this effect is not completely understood as yet /11/.

Other materials being investigated include gadolinium-iron-garnet (GdIG) /12/^* and europium oxide (EuO)/13/^*, EuO has a major disadvantage in that it operates only at liquid helium temperatures. All things considered, MnBi appears to be the best candidate for magneto-optic holographic storage at the present state of the art.

High efficiency volume phase holograms can be recorded in certain transparent electro-optic crystals. Examples are lithium niobate co-doped with iron (LiBbO₃ : < Fe), lithium tantalate (LiTaO₃), strontium-barium niobate [(Sr,Ba)Nb₂O₆, SBN, or Ba_0.75Sr_0.25Nb₂O₆], and doped barium - sodium niobate

(Ba₂NaNb₅O₁₅)/14-18/.

The recording process relies on the optical excitation and subsequent frift or diffusion of electrons which originate from localized centers in the crystal. This creates fields that modulate the index of refraction of the material and the result in a phase hologram whose efficiency could theoretically be as high as 100%. Proper recording procedure requires that the crystal optic axis be perpendicular to the fringe gratings of the hologram. High efficiency readout can be accomplished only with light polarized in the plane containing the optic axis.

In addition to the materials discussed above, several other electro-optic hosts have been tried but with only marginal success. These include barium titanate (BaTiO₃), bismuth oxide (Bi₁₂SiO₂₀), and bismuth germanate (Bi₄Ge₃O₁₂)/17,18/.

Color centers in transparent crystals are caused by the presence of impurities and imperfections which give rise to localized states that trap electrons or holes within the forbidden energy gap of the material /21/. Such states may show absorption at visible wavelengths where the pure crystal is transparent. In particular, the photochromies materials have the ability to switch colors under the influence of optical irradiation. This change is normally caused by the transfer of an optically excited electron from one type of color center to another with the absorption properties of both centers being changed accordingly. Light of one wavelength produces a given change in coloration, and the crystal returns to its original state when exposed to light of another wavelength.

The performance of inorganic photochromies can be compared to other media by referring to Table 1. They are characterited by the maximum absorption change induced by light, the energy required to produce a given change, the background absorption, and the storage time of the material. The inherent resolution of the materials is practically unlimited as illustrated by the fact that reflection holograms with fringe spacing of less than 100 nm have been successfully recorded by Amodei in SrTiO₃ : Pe, Mo.

The sensitivity of inorganic photochromic materials varies considerably from hundreds of J*cm^-2 to nJ*cm^-2 depending on the quantum efficiency of the process. The maximum storage time of the crystals, which is usually determined by the thermal activation energy of the centers, varies from minutes to months.

Examples of inorganic photochromic materials include silver halide photochromic glasses /20-23/, and inorganic photochromio crystals (e.g. SrTiO₃:Fe, Mo; BaF₂:La; Sodalite:Cl, electron beam colored NaCl; and fluorescein-boric acid glar organophosphor) /24-26/.

Successful cyclic recording and erasure of holographic infor-

nation on thin organic photochromic films has been reported by Milcaeliane and hia assooiates in the Soviet Union /28/, and N.Lescinsky and M.Miller in Czechoslovakia /28/. Performance data for these materials(essentially photochromic epiropyran dissolved in styrene polymer) are rather sketchy at this time, however. In the United States, Ross /29/ has investigated the use of thioindige dye. The efficiency of the material is temperature dependent, but its densities are higher than those of the inorganic photochroioic materials.

In summary, a clearly destinot advantage of photochromic materials is that there is no inherent resolution limit since absorption takes place on an atomic or molecular scale. Disadvantages include: sensitivity limited to narrow spectral regions; relatively thick samples (1-10 mm) required to obtain useful optical densities; and image deterioration due to both thermal decay and optical bleaching during readout.

Concentrated layer energy may be used to raise the temperature of a material to its vaporization point. This principle has been used to store evaporation holograms in thin metal films /30/. Films may initially be deposited on substrates using conventional technology. Then, during hologram recording, locations of constructive interference rise to higher temperatures than do points of destructive interference.

Holographic recording by vaporation relies on the fact that the energy absorbed by a thin film (7,5 - 20 nm) during a very short laser pulse does not have time to diffuse away. Exposure to such light pulses, therefore, causes the material to vaporise from each point of the surface in amounts that are nearly proportional to the integrated intensity of the light absorbed at that point. When a holographic intensity pattern is applied to a film in this manner, it leaves a film thickness profile which closely resembles the spatial intensity variations of the recording light. Such a pattern is considered an amplitude ho-

logram because, while the absorption is substantial, the film is not sufficiently thick to significantly affect the phase of the light traversing it. The hologram thus stored can be read out by reflection or transmission with comparable efficiencies. In tests with 7,5-20 nm thick bismuth (Bi) films vacuum deposited on glass substrates /30/, both transmission and reflection holograms have efficiencies of 6%, which is very close to the theoretical maximum for amplitude holograms.

The sensitivity of this direct holographic recording material depends to some extent on film thickness; for films approximately 10 nm thick, the energy required to record was experimentally /30/ found to be less than 50 mJ/cm², which compares favorably with other materials such as photochromies and magneto-optic thin films (see Table 1).

Another feature of the thin metal films is that they may be operated in a linear region because the energy required to reach boiling temperature is much lower than the vaporation energy. Also, since the recording technique is heat rather than light dependent, lasers of any wavelength can be used for recording. A chief disadvantage, of course, is the difficulty in recycling. It is possible to build a contained cell where the material is evaporated between two surfaces in very close proximity and is thus redistributed during the recording cycle. Such a device could then, in principle, be erased by exposure to a spatially uniform light pulse that redistributes the film evenly over the surface. Experiments by Amodei have shown that this approach is feasible, at least for a limited number of cycles (5 to 10), and it is conceivable that with proper de -sign, cycle lifetime could be extended considerably.

The use of a ferroelectric-photoconductor sandwich device for information storage is not a particularly new idea /31/. However, it is only recently that the suggestion was made to operate such a device with optical readout /32/. Early devices in-

corporated an optically controlled record operation but readout was electrical.

In many ferroelectric materials the two stable remanent polarisation (P_R) states may not be distinguished optically. Recently, two material configurations not posseasing this disadvantage have been suggested: (1) use of ferroelactric bithmuth titanate, Bi₄Ti₃O₁₂, in which non-180° switching of P_R may be performed /33/, and (2) "strain-biasing" of ceramic PLZT ferroelectrics to allow effectively non-180° switching of P_R /34,35/. With these ferroelectric materials, ferroelectric-photoconductor (PE-PC) devices capable of optical (with applied electric field) record and optical readout are possible.

The PE-PC device geometry ia shown schematically in Fig.la. The ferroelectric (FB) slab (either single crystal, ceramic, or thin film) is covered with a photoconductive (PC) layer, and the combination is sandwiched between transparent electrodes. External batteries or pulse generators are used to apply switching fields to the device. The operational characteristics of the PB-PC device are determined by the specific materials used. Two typical state-of-the art materials are Bi₄Ti₃O₁₂ and strain-biased PLZT.

Bismuth titanate, a member of crystal class m /33/, is prepared in single crystal platelets by flux growth /36/. It posseses four atable remanent polarisation states (having both a-axis and a-axis components) in a detwinned cryetal. By first poling /37/ P_Ra (in the plane of the crystal) uniformly, switching P_Rc between its two states leads to the non-180° switching described earlier. As shown in Fig.1b the orientations of the optical indicatrix for the two states differ by a rotation of 2a (roughly 50°) around the crystalline b(x₃)-axis. Since Bi₄Ti₃O₁₂ is biaxial, n₁₁≠n₂₂≠n₃₃. Two teehniques for observing for cohange in optical indicatix orientation (or equivalently, reading out a stored pattern in an FE-PC device) are possible: (1) tilting the crystal about the a-axis to achieve a difference in extinction directions for the two states /32, 38, 39,40/ (illustrated In Fig.la), and (2) tilting the crystal about the c-

Fig.1.FE-PC (a) Bi₄Ti₃O₁₂ device geometry showing WRITE and READ processes and (b) orientation of optical indicatrix for Bi₄Ti₃O₁₂ for the two states of c-axis remanent polarisation P_Rc (with fixed P_Ra).

axis to achieve a phase difference for light polarized perpendicular to b /39,41/. The reader may easily convince himself that light normal to the ab plane (i.e., parallel to the c -axis) will not be adequate to distinguish the two dates. Both techniques have been used to reconstruct holograms stored in FE-PC devices /38-41/.

PL2T (lanthanum-doped lead zirconate-lead titanate ceramic material) platelets /34/ coated with photoconductor layer and transparent electrodes, are held intension or compression causing ferroelectric domains to line up parallel and antiparallel to the strain axis (in the platelet) resulting in zero net remanent polarization. By switching the device with fields normal to the platelet a normal component of remanent polarization, allowing readout of a stored pattern.

Both bismuth titanate and PLZT FE-PC devices are included among the comparisons in Table 1. To date, holographic storage has only been reported in the former. Both devices possess their own unique advantages and disadvantages. Both possess an inherent gain mechanism, similar to that in photographic film. Incident light for recording does not directly cause the storage; rather, photoeleotrons in the photoconductor are "created" by the light causing an increase in conductivity and allowing the battery to switch the ferroelectric. The bismuth titanate device involves difficult fabrication and processing techniques because of its single-crystal nature, large-area devices are difficult to achieve. Furthermore, an optical degradation has been observed after numerous switching cycles with metal electrodes directly on the Bi₄Ti₃O₁₂ surfaces /42/. However, such degradation has not yet been observed in FE-PC devices with photoconductive layers /43/.

The PLZT device is simpler to fabricate because of its ceramic form. However, ferroelectric ceramics are notorious for their electrical and optical fatigue /44/, which would restrict their useful lifetime.

Perhaps the ultimate FE-PC for holographic applications would utilize thin films of Bi₄Ti₃O₁₂. Such films have recently

bean produced /45,46/. They might represent the optimum compromise between the diffioulties of single-crystal technologies and long term ceramic fatigue.

The technique of thermoplastic recording was first described by Glenn in 1959 /47,48/. Prior to its application to holography, this recording technique used an electron beam to scan a thermoplastic film (e.g., polystyrene, chlorinated polyphenyl, or alphamethylstyrene) and record video signals in the form of a pictorial image; the image was visible on the film immediately following recording. The technique offered very high resolution and recording bandwidth with the option of erasure if desired.

The making of phase holograms on thermoplastic film was first demonstrated by Urbach and Meier in 1966 /49/. As most recently described /50/, a phase hologram is recorded in a transparent thermoplastic film as a spatial variation of film thicknens corresponding to the light intansity variation of a holographic fringe pattern. Since the thermoplastic is usually a light-intensive resin, it is combined with a photoconduotor in a film structure which can respond to light. Although manufactoring a film structure with the photoconductor dispersed through the thermoplastic is possible, the usual structure consists of a layar of thermoplastic coated over a layer of photoconductor bacause it yields higher photosensitivity /50/. The complete structure consists of a glass substrate covered by a grounded transparent conductive film on top of which lie the photoconductor and thermoplastic film. The sequence of a complete hologram record-erase cycle is described elsewhere in this issue by Gredelle and Spohg^* whose extensive experiments with the

material staybelite, have resulted in improved readout efficiency and resolution.

It is important to note that, unlike most other hologram recording media, thermoplastic material responds only to a band at spatial frequencies /50,51/. Thus, there it an inherent tradeoff between afficiency and resolution which depends on iha angle between reference and object beams in system deign. Also, thermoplastic devices have relatively poor cycle lifetime although significant improvement in this parameter has been reported recently /52/.

In 1968 Chang /53/ investigated the physical parameters required of thermoplastic film for a system in which an electron beam was used for both recording and readout. More recently, Doyle and Glenn /54,55/ have described a real time optical processing system which records with an electron beam and reads out with viaible light. Their device, named "lumatron" makes use of a fixed reusable transparent thermoplastic surface mounted in a glass envelope. Surface modulation of the thermoplastic is accomplished by a high resolution (10 m) electro-statically focused (at 6,5 kV) and deflected electron beam. Readout is performed with a single lens schlieren optical system external to the vacuum but whose light path is coaxial with the recording electron beam and normal to the thermoplastic coated faceplate as shown in Fig.2. The thermoplastic is coated onto a transparent conductor on the faceplate which serves both as the final faceplate reference potential and as a heater. Erasure is accomplished in less than a second by supplying a pulse of heater current to the transparent conductor to themalty remove all deformations stored in the thermoplastic surface. In real time operation, the theimoplastic is kept at a temperature between the deformation and erase temperatures, and recording is continuous with the scanning electron beam. In this mode of operation the mechanical tine constant of the thermoplastic determines the rate at which the recorded pattern can be changed. According to Doyle and Glenn /54/, cycling can be accomplished at 1/30 second and gray scale is greater than 8 sha-

des. Thus, the luimatron certainly has potential for use in real time incoherent-to-coherant holographic image conversion systems which need such a device for their successful implementation /3/.

If the thermoplastic material in thermoplastic deformation devices discusses in the preceding section is replaced by an elastomer, another useful device for holographic imaging (but not permanent storage) is created. Such a device, the "ruticon"^*, has been investigated toy Sheridon /56/.

The ruticon device has a layered structure consisting of a conductive transparent substrate, a thin photoconductor, a thin defonnable elastomer layer, and a defortnable electrode.The deformable electrode can take three format (1) a conductive liquid, (2) a conductive gas, or (3) a thin optically opaque flexible metal; the separate devices resulting from these three different deformable electrodes are identified, respectively, as a ruticon, ruticon, and ruticon.The most interesting of these is the ruticon. In operation, a deelectric field (< 600 V) between the thin flexible metallic electrode and the transparent electrode establishes the initial uniform charge. Holographic information impinging on the device through the transparent substrate causes changes in the electric fields across the photoconductor and hence across the elastomer. These fields create mechanical forces which cause the elastomer, and consequently the thin natal layer, to deform and form a phase hologram.

In readout, light is reflected with high efficiency from the metal layer to reconstruct the original object wavefront. If the metal layer is non-transparent, readout light will not affect the conductivity modulation of the photoconductor layer. In some forms, the permanence of ruticon storage may be en-

hanced by utilizing an image locking mode. Erasure is accomplished by removing the do field. This is a distinct advantage over the lumatron where heating is required for erasure.

The ability to store optical images in liquid crystals /57/ has promted their use 'in both electrically- and photoconductively-controllable real-time imaging devices. Kieinie and Wolff /58/ have reported recyclable holographic storage in an electrically-controllable device consisting of mixed (90% Hematic, 10% cholesteric) liquid-crystal sandwich cells having matrix-shaped electrodes. Holographic recording was achieved as a result of dynamic scattering induced by application of a do or low frequency ac electric field across the electrodes. While initially the nematic ordering is intact and the nematic-cholesteric mixture is clear the dynamic scatlaring caused by the applied field emulsifies the cholesteric material. Because this colloidial-like suspension can influence the coherence of an incident laser light wavefront, holographic recording is feasible. Recent experiments /58/ indicate that this optical scattering state can persist for weeks without a sustaining field or can be erased by an ac field in 20-500 ms. A typical restoration field strength is 5 V. m^-1 (rms) at 3 kHz.

The matrix-shaped electrode structure required for the electrically-controllable device is complicated to construct and, worse, limits resolution capability. The implementation of the photoconductively-controlled liquid crystal device overcomes, in pronciple, both of these problems. Such a device is achieved simply by replacing the electrode matrix structure by a photocon - ductive layer. Of the photoconductive materials tested, zinc sulfide /59/ (ultraviolet recording) and cadmium sulfide /60/ (visible light recording) have given the best results thus far. Holographic recording has not yet been reported, however, because resolution of the device is on the order of only a few line-pairs per millimeter. In structure, the device consists of a layer of liquid crystal and a layer of photoconductor sandwiched bet-

ween two electrode layers of conductive material (e.g., tin oxide). The photoconductive layer is designed to have a "nonilluminated" resistance that is lower. In the nonilluminated condition, an applied dc voltage across the electrodes of the device lies alciost entirely across the photoconductive layer. Thus, the low voltage across the liquid crystal layer is not sufficient to produce scattering effects and the liquid crystal remains transparent. However, when a wavefront illuminates the photoconductor, its resistance is lowered causing more voltage to appear across the liquid crystal. This higher voltage produces sufficient current to produce dynamic scattering and allow image storage.

The photoconductor-liquid crystal devices permit image recording and readout at a rate controllable by the magnitude of the applied voltage and composition of the material used. The sensitivity appears to be better than that of other recyclable materials, being several orders of magnitude greater than that of photochromic films.

Thie device consists of a layer of photoconductive, electrooptic, single crystal material covered by insulating dielectric layers on one or both faces, and sandwiched between transparent electrodes (or one electrode and a simple ohmic contact). Experiments using highly resistive photoconductive ZnS and Bi₁₂SiO₂₀ as the crystal materials have been reported in the literature /61-63/. In operation, a dc voltage applied across the material is modulated by an incident optical wavefront in accordance with the photoconductive properties of the crystal, thereby storing the information in the wavefront structure. Since the material is simultaneously electrooptic, readout of the stored electrostatic pattern can be accomplished by the Pockel's effect; i.e., the pattern is reconverted to an optical wavefront by local phase ratardations resulting from the transmission of uniformly polarized light through the material. The combination of both photoconductive and electrooptic functions in a single material

has an advantage over a device using separate laminated photo-conductive and electro-optic films because of the fabrication problems inherent in interfacing a multilayared structure of this kind. On the other hand, the complexities involved fabricating a uniform single material which optimally exhibits both functions is actually more difficult than the interfacing problem at the present state of the art.

A comparison of performance parameters for recyclable holographic storage media is given in Table 1. In most cases we have listed the "best" state-of-the-art material in the storage medium column as an example although there may be potentially better materials being investigated. Detailed information on these may be found in the list of references accompanying this paper.

Although quantitative information for recyclable holographic storage media is in a rapid state of flux, we have attempted to offer a quantitative comparison in Table 1 by defining a "Figure of Merit". In deciding on a reasonable definition for this relative "tradeoff" value we have omitted record- and erase-time parameters because they differ by such great orders of of magnitude (e.g., picoseconds vs minutes). These particular parameters can be considered independently in any specific design application depending on the desirad recycling speed. The remaining quantitative parameters (resolution, readout efficiency, and write energy) listed in Table 1 thus yield the following definition:

Figure of merit =

The superior figure-of-merit values in Table 1 for the thermoplastic and elastomer devices are due primarily to their low write energy values. A look at the record time column, however, indicates that their record time values are several orders of magnitude higher than many media which have lower figure-of-merit values. Also, the thermoplastic and elastomer devices require more than light alone to operate properly, a fact which dulls the meaning of their high figure of merit in certain applications. In addition, the themoplastic device has poor cycle lifetime at present and the elastomer device is not useful for permanent storage.

Because of their low write-enargy values, the liquid crystal-photoconductor device and the photoconductor-Pockel'e effect device probably have figure-of-merit values equivalent to those of the thermoplastic and elastomer devices. However, since their diffraction efficiency values are not available we cannot compute a speciency figure of merit for them at present. Comparing these devices further we note that the photoconductor-Pockel'e effect device has exoellent sensitivity relative, to the other recyclable media while the liquid crystal-photoconductor device has poor resolution at present (although holographic recording has been reported in the more complicated electrically-controlled device /58/). On the other hand, the liquid crystal-photoconductor device is simple in structure, has low power consumption, and requires only low control voltages.

Generally speaking, the thermoplastic, liquid crystal-photo-conductor, and photoconductor-Pockel's effect devices can find significant application as incoherent-to-coherent image converters /3/. Most of the other materials in Table 1 find use in applications in which frequent readout and infrequent write-in are required.

Reviewing these other materials we see that magneto-optic materials are quite useful for two-dimensional storage and have the advantage of relative ease of fabrication compared to crystal de-

vices and the disadvantages of requiring accurate temperature control and pulsad operation. Electro-optic materials are useful for volume storage and therefore have high storage capacity which, in turn, implies that selective erasure is difficult. Also, being a single crystal material means that are relatively expensive compared to film media. One of their best advantages, however, is in nonoptically erasable applications because holographic information can be "fixed" in the material by a simple thermal process. In this respect, electro-optic materials are superior to photochromic materials, which, while three-dimensional storage materials in principle, cannot be used as such because infonnation stored in them cannot be fixed. Also, photochromic materials require two wavelengths for operation but have inherently unlimited resolution capability.

In the photoconductor devices, although the photoconductive material is not the key material for storagei it doss affect sensitivity both in terms of amount of light required and the wavelength which can be used, and it affeota readout efficiency since light is absorbed if reading out with the same wave-length as writing in. In this regard, the real beauty of the FEPG devices is their ability to retain information on read-out when light is absorbed) in comparison, information in the photoconductor-Pockel's effect device decays on readout.

In summary, Table 1 presents both quantitative and quntitative information on the state of the art of recyclable holographic recording media. The calculated figures of merit are intended only as an indication of possible potential performances since the state of the art in this area is so fluid at present a very careful and perhaps partly intuitive tradeoff must be made among all performance parameters for any projected application.

1. J.A.Rajchman. Appl. Phys., vol.41, p.1376, 1970.

2. R.A.Bartolini, W.J.Hannan, D.Karlsons, M.J.Lurie. Appl. Optics, vol.9, p.2283, 1970.

3. N.H.Farhat, W.R.Guard. Proc.IEEE, vol.58, p.1955, 1970.

4. J.J.Amodai. CRC Handbook of Lasers with Selected Data on Optical Technology, The Chemical Rubber Co., Cleveland, Ohio, 1971.

5. L.Mayor. J.Appl.Physics, vol.29, p.1003, 1958.

6. L.Mayer. J.Appl.Physics, vol.29, p.1454, 1958.

7. D.Chen, J.P.Ready, E.Bernal. J.Appl.Physics, vol.39, p.3916, 1968.

8. R.S.Mezrich. Appl.Phys.Letters, vol.40, p.132, 1969.

9. G.Y.Pan, K.Pennington, J.H.Greiner. J.Appl.Physics, vol.40, p.974, 1969.

10. M.J.Freiser. IBEB Trans.Magnetics, vol.MAG-4, p.152, 1958.

11. R.S.Mezrich. Private Communication.

12. N.Goldberg. IEEE Trans.Magnetics, vol.MAG-3, p.605, 1967.

13. D.I.Iehemev, O.W.Lewicki. IEEE Trans.Magnetics, vol.MAG-4, p.75, 1968.

14. F.S.Chen, T.T.LaMacchia, D.B.Fraser. Appl.Phys.Letters, vol.13, p.223, 1968.

15. J.B.Thaxter. Appl.Phys.Letters, vol.15, p.210, 1969.

16. J.J.Amodei, D.L.Staebler, A.W.Stephens. Appl.Phys.Letters, vol.18, p.507, 1971.

17. F.S.Chen. J.Appl.Physics, vol.40, p.3389, 1969.

18. J.J.Amodei, W.Philllps, D.L.Staabler. IEEE J.Quan.Elac., vol.QE-7, p.321, 1971.

19. J.J.Amodei, D.L.Staebler. Appl.Phys.Letters, vol.18, p.540, 1971.

20. G.Jackson. Optica Acta, vol.16, p.1, 1969.

21. S.Herman. Proc.Symp.Modern Optics, p.743. John Willey and Sons, New York, 1967.

22. J.P.Kirk. Appl.Optics, vol.5. P.1684, 1966.

23. R.J.Araujo. Recent Advances in Display Media. NASA Symposi-

um Proceedings, Cambridge, Mass., p.63, 1967.

24. D.R.Bosomworth, H.J.Gerritsen. Appl.Optica, vol.7, p.95, 1968.

25. A.S.Mackin. Appl.Optica, vol.9, p.1658, 1970.

26. T.A.Shankoff. Appl.Optics, vol.8, p.2282, 1969.

27. A.L.Mikaelian., A.P.Axenchikov, V.I.Bobrinev, E.H.Gulaniane, V.V.Shatun. IEEE J.Quantum Elec., vol.EQ-4, p.757, 1968; Soviet Physics-Doklady, vol.13, p.810, 1968.

28. M.Lescinsky, M.Miller. Optics Communications, vol.1, p.417, 1970.

29. D.L.Ross. Appl.Optics, vol.10, p.571, 1971.

30. J.J.Amodei, R.S.Mezrich. Appl.Phys.Letters, vol.15. p.45, 1969.

31. R.M.Schaffert. U.S.Patent 3, 146, 354, 1964.

32. S.E.Cummuns. Proc.IEEE vol.55, p.1536, 1967.

33. S.E.Cummuns, L.E.Cross. J.Appl.Physics, vol.39, p.2268, 1968.

34. J.R.Maldonado, A.L.Meitzler. Proc.IEEE, vol.59, p.368, 1971.

35. J.R.Maldonado, L.K.Anderson. IEEE Trans.Elec.Dev., vol. ED-18, p.774, 1971.

36. A.D.Morrison, F.A.Lewis, A.Miller. Ferroelectrics, vol.1, p.75, 1970.

37. M.M.Hopkins, A.Miller, Ferroelectrics, vol.1, p.37, 1970.

38. S.A.Keneman. G.W.Taylor, A.Miller, W.H.Fonger. Appl.Phys. Letters, vol.17, p.173, 1970.

39. S.E.Cummins, T.E.Luke. IEEE Trans.Elec.Dev., vol.ED-18, p.761, 1971.

40. S.A.Keneman, G.W.Taylor, A.Miller. Ferroelectrics, vol.1, p.227, 1970.

41. S.A.Keneman, A.Miller, G.W.Taylor. Appl.Optics, vol.9, p.2279, 1970.

42. G.W.Taylor, S.A.Keneman, A.Miller. Appl.Optics, vol.2, p.11, 1971.

43. S.A.Kenenan, A.Miller, G.W.Taylor. Ferroelectrics, vol.3, p.131, 1972.

44. W.C.Stewart, L.S.Cosentino. Ferroelectrics, vol.1, p.149, 1970.

45. W.J.Takel, N.P.Formigoni, M.H.Francombe. Appl.Phys.Letters, vol.15, p.256, 1969.

46. S.Y.Wu,W.J.Takei, M.H.Francombe, S.E.Cummins. Ferroelectrics, vol.3, 1971.

47. W.E.Glenn. J.Appl.Phys., vol.30, p.1870, 1959.

48. W.E.Glenn. J.SMPTE, vol.74, p.663, 1965.

49. J.C.Urbach, R.W.Meier. Appl.Optics, vol.5, p.666, 1966.

50. L.H.Lin, H.L.Beauchamp. Appl.Optice, vol.9, p.2088, 1970.

51. J.Urbach, R.W.Meier. Appl.Optics, vol.8, p.2269, 1969.

52. H.R.Andereon, Jr., E.A.Bartkus, J.A.Reynolds. IBM J.Res. Develop., p.140, 1971.

53. L.S.Chang. Photographic Science and Engineering, vol.12, p.238, 1968.

54. R.J.Doyle, W.E.Glenn, IEEE Conf.Record 1970 IEEE Conf. on Display Devices (70 C.55-ED), p.88, 1970.

55. R.J.Doyle, W.E.Glenn. IEEE Trans.Electron Devices, vol.ED-18, p.739, 1971.

56. N.K.Sheridon. IEEE International Electron Devices Meeting, Washington, D.C., 1970.

57. G.H.Heilmeier, J.E.Goldmacher. Proc.IEEE, vol.57, p.34,1969.

58. H.Kiemle, U.Wolff. 1970 Annual Meeting of Optical Society of America, 1970.

59. J.D.Margerum, J.Nimoy, S.Y.Wong. Appl.Phyc.Letters, vol.17, p.51, 1970.

60. G.Assouline, M.Hareng, E.Leiba. Proc.IEEE, vol.59, p.1355, 1971.

61. D.S.Oliver, P.Vohl, R.E.Aldrich, M.E.Behrndt, W.R.Buchan, R.C.Ellis, J.E.Genthe, J.R.Goff, S.L.Hou, G.McDaniel. Appl.Physics Letters, vol.17, p.416, 1970.

62. D.S.Oliver, W.R.Buchan. IEEE Trans.Elec.Dev., vol.ED-18, 1971.

63. S.L.Hou, D.S.Oliver. Appl.Phys.Letters, vol.18, p.325, 1971.