Holographic Data Storage Systems

LAMBERTUS HESSELINK, SERGEI S. ORLOV, AND MATTHEW C. BASHAW

Invited Paper

In this paper, we discuss fundamental issues underlying holo-

graphic data storage: grating formation, recording and readout of
thick and thin holograms, multiplexing techniques, signal-to-noise
ratio considerations, and readout techniques suitable for conven-
tional, phase conjugate, and associative search data retrieval. Next,
we consider holographic materials characteristics for digital data
storage, followed by a discussion on photorefractive media, fixing
techniques, and noise in photovoltaic and other media with a local
response. Subsequently, we discuss photopolymer materials, fol-
lowed by a discussion on system tradeoffs and a section on signal
processing and en/decoding techniques, succeeded by a discussion
on electronic implementations for control, signal encoding, and re-
covery. We proceed further by presenting significant demonstrations
of digital holographic systems. We close by discussing the outlook
for future holographic data storage systems and potential applica- Fig. 1. Optical roadmap. (Courtesy TDK, with permission).
tions for which holographic data storage systems would be partic-
ularly suited.
from and writing data onto the optical disk. Blue lasers
Keywords—Diffraction gratings, gratings, high density data
storage, holographic data storage, holographic recording mate- achieve spot sizes of a few hundred micrometers, providing
rials, holography, optical data storage, optical storage materials, approximately 25 GB of storage capacity per layer. Up to
photopolymer media, photorefractive media, volumetric data four layers—two on each side—increase capacity to near
storage. 100 GB per disk. To further improve capacity and transfer
rates, several options are available, including increasing the
I. INTRODUCTION NA beyond 0.85, reducing the wavelength below 400 nm,
or adding more layers. All these options present significant
A. Motivation obstacles. Additional improvements in NA require more
Optical data storage is a commercial success story. Each costly and complicated optical systems for an additional
year billions of recordable disks are sold worldwide, and in 40–50% gain in NA, leading to doubling in capacity. Shorter
almost every household with a computer there is a CD-ROM wavelength lasers are not commercially available and re-
or CD-recordable drive. The industry published roadmap quire special optical materials that are transparent below 400
shows future DVD products to reach capacities of near 100 nm. Increasing the number of layers also proves difficult
GB on a disk, and data transfer rates exceeding 300 Mb/s for a variety of reasons, the most significant one being
sometime in the latter half of the current decade, shown in manufacturability and complexity of implementation for the
Fig. 1. Such high capacities are obtained by using a very media, optical stylus, and the associated optomechanical
high numerical aperture (NA) optical stylus for reading data system for tracking and focusing. The industry, therefore,
has been researching new methods for extending the optical
Manuscript received March 7, 2003; revised March 9, 2004. This work data storage roadmap well beyond 100 GB per disk.
was primarily supported by the Defense Advanced Research Projects Two promising candidates are near-field recording and
Agency Industry/University PRISM and HDSS Programs and, subsequent holography. In near-field recording approaches, the NA of
to 2000, in part by the Japanese Science and Technology program.
L. Hesselink and S. S. Orlov are with the Solid State Photonics Lab, the optical stylus is made larger than one, resulting in very
Department of Electrical Engineering, Stanford University, Palo Alto, CA small spot sizes approaching dimensions less than 100 nm.
94305-4070 USA (e-mail: [email protected]). The focusing spot is near to the optical lens, requiring close
M. C. Bashaw is with Lockheed Martin Corporation, Integrated Systems
& Solutions, San Jose, CA 95134 USA. proximity between the optical head and disk, making remov-
Digital Object Identifier 10.1109/JPROC.2004.831212 ability of the media more difficult, as small contaminants

0018-9219/04$20.00 © 2004 IEEE

PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004 1231

can cause interference between the two. The advantages of
near-field recording are the extremely high areal densities
achievable, but at the expense of optomechanical complexity
and less robust removability, among others. The second
promising candidate technology is optical holography. In
optical holography, data are stored throughout the volume
of the recording medium, as opposed to on the surface. Data
are impressed onto an optical coherent beam using a spatial
light modulator (SLM) or page composer. The signal bearing
beam is interfered with a reference beam inside the recording
Fig. 2. Perpendicular (90 ) angle multiplexing architecture,
medium to produce an interference grating, representing a including the electronic control unit.
data page. Multiple gratings are superimposed by varying
the optical properties of the reference beam, a process re-
ferred to as multiplexing. Upon data retrieval or readout, a numbers of pixels and have fast frame rates, exceeding thou-
single reference beam is incident on the medium under the sands of frames per second for pages containing hundreds
same conditions as used for storage, producing a diffracted of thousands of bits. The sensitivity of the detector must be
beam representing the stored data page. The diffracted beam high, as superposition of holograms reduces the intensity of
is detected by a detector array, which allows extraction of the diffracted data pages as more and more pages are stored
the stored data bits from the measured intensity pattern. in the same volume. The larger the capacity, the smaller the
Data encoding and image processing techniques enhance readout signal, presenting a challenge to system performance
data robustness. Data pages contain large numbers of data optimization.
bits or pixels; practically up to 1 million b/page have been At present, it is uncertain which technology, near-field
demonstrated. As a whole page is either stored or recalled recording or holography, will be selected by the optical
by the reference beam, data transfer rates can be phenome- storage industry, sparking intense interest in both. According
nally high, exceeding 10 Gb/s. By superimposing multiple to IBM [1], holography, with its unique performance, may
data pages in the same volume, data storage capacity is find that applications in all-solid-state devices have ex-
very high as well. Although not quite physically correct, tremely short access times, effectively alleviating the large
we can think of the number of superimposed holograms I/O gap that exists between solid-state memory and hard
as the equivalent number of layers in a multilayer surface disk drive access times, which gives rise to latency that
recording system. Instead of the two or four layers used in reduces system performance. Memories could compete in
conventional optical recording, holographic recording can both the optical and magnetic storage markets. Holographic
support hundreds of superimposed holograms or “equivalent memories could also be employed in large data warehouses
layers” by employing the third dimension of storage media. where the extreme search and readout rates facilitate more
This leads to powerful features not available from surface efficient database management of large data repositories and
recording technologies, but also specific problems that need involving many users. Holography, it seems, is an attractive
to be addressed before holographic data storage can be a approach for future storage systems.
commercially viable option for the optical roadmap.
Among the unique advantages of holographic storage are B. Holographic Storage Architectures
extremely short access times, less than 50 s, extremely fast The holographic storage system architecture is largely
input and output rates, exceeding 10 Gb/s, as well as enor- determined by the type of recording medium. Broadly
mous search capabilities for finding unindexed information speaking, holographic data storage materials are divided
in databases at rates exceeding 100 Gb/s, far superior to cur- into two classes; systems based on thin (a few hundred
rent magnetic disk based storage systems. These advantages micrometers thick) photosensitive organic media and thick
result from storing data in a volume in the form of data pages (a few millimeters to centimeters), inorganic photorefractive
containing hundreds of thousands of bits. An all-solid-state crystals [2]. Thick, bulk crystals of photorefractive media
memory device could contain gigabytes of data, accessible in are ideal for recording geometries in which a reference and
microseconds, and allowing searches of image features and object beam are incident on the medium at right angles, as
recognition at unmatched rates. The powerful advantages of shown in Fig. 2. A typical photorefractive crystal used to
holographic storage have propelled decades of research and investigate this configuration is iron-doped lithium niobate.
development of this technology. Ironically, the challenges A laser beam is split into two beams, a reference beam and
facing commercialization of holographic storage technology an object beam. Data are imprinted on the object beam via
also result from the same desirable holographic features. As a page composer or SLM. The reference and object beams
many holograms are superimposed, media must exhibit ex- are focused and combined on the recording medium. The
cellent volumetric stability, low scatter, and high sensitivity medium is photosensitive and a phase hologram is recorded.
for recording, and must be easily manufacturable. Key op- By varying the reference beam angle, hundreds of holograms
tical components such as the page composer and the readout are superimposed in a single location of the medium. By
detector array must be of high performance, i.e., contain large scanning both beams over the medium, or by translating the

1232 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 the detector array. Shift multiplexing can be implemented
by using a spherical wave [11] or, alternatively, a random
or pseudorandom speckle pattern as a reference beam [12].
By modifying the phase front of the reference beam using a
random or pseudorandom phase mask to generate a speckle
pattern, higher shift selectivities can be achieved than in the
case of a simple spherical wave. Alignment requirements,
however, increase as well, as we need to maintain a constant
phase relationship between the signal and reference beam.
In photopolymer media, shrinkage is a considerable issue,
and tradeoffs are made to control shrinkage versus recording
sensitivity and total dynamic range of the medium. Pho-
topolymers, on the positive side, are more sensitive recording
Fig. 3. Shift multiplexing architecture using rotating media by at least one or two orders of magnitude than pho-
photopolymer disk medium. torefractive crystals. They derive their advantage from an
exothermal chemical chain reaction that “amplifies” one
medium relative to the optical beams, the total recording photon event into 100 or 1000 chain polymerization events,
volume is utilized. Upon illumination with a single reference whereas photorefractive media typically require one photon
beam, the corresponding data page is retrieved at the de- event for every photorefractive event that produces a local
tector array. Such systems excel in having ultrashort access index of refraction change in the medium. Consequently,
times—on the order of a few tens of microseconds—and sensitive photopolymer materials are write-once-read-many
extremely fast data retrieval rates exceeding 10 Gb/s, as (WORM) media. Attempts to make reversible polymer
well as all-solid-state operation with no moving parts. The media with high sensitivity have fallen short of the require-
media are rewritable and information can be fixed for tens ments for HDSSs [13], as they fundamentally require again
of years, if not centuries. Based on fundamental consider- one photon event for one index changing event. The price to
ations, storage capacity is typically tens of gigabytes, and pay for higher sensitivity of photopolymers is the change in
write times are slower than readout times by one to two media density, which leads to both an index change as well
orders of magnitude. Complete demonstration systems have as shrinkage caused by the polymerization reaction. The
been built by Stanford University, Stanford, CA, and Siros holy grail for developers of holographic data storage mate-
[3], IBM [4] and Rockwell, Thousand Oaks, CA [5], as rials developers is to devise a medium with high sensitivity
part of the photorefractive information storage materials while maintaining a low level of shrinkage. Researchers at
(PRISM) and holographic data storage systems (HDSS) Polaroid, Boston, MA, and Bell Laboratories, Murray Hill,
programs; while researchers at the California Institute of NJ, have taken different tacks on this problem, and excellent
Technology, Pasadena (Caltech) [6], and at Lucent [7] and system demonstrations based on cationic ring opening ma-
at other organizations around the world have subsequently terials (CROP) (Polaroid) [14], [15] and free-radial media
made system demonstrations as well. In particular, much (Lucent) have been made [16]. Data capacities exceeding
work has been carried out at Caltech on optical architectures 250 GB on a DVD-like disk have been measured by Aprilis,
based on LiNbO . Notably, the Psaltis group at Caltech and and data rates exceeding 10 Gb/s have been demonstrated at
the group at Northrop have demonstrated analog recordings Stanford University using Polaroid, and later Aprilis, CROP
of large numbers of holograms [8]. media [17], [18], To further increase data storage density
Polymer-based systems resemble conventional DVD sys- or capacity per disk, additional material improvements and
tems, having a rotating, removable thick (0.5–1 mm) holo- system optimization are needed.
graphic polymer disk, and a two-sided optical head, as shown System optimization, as usual, is a complex problem in-
in Fig. 3. An image bearing object and reference beam are volving a large number of tradeoffs. For HDSSs in particular,
incident on the recording medium. The medium is a photo- the tradeoff between capacity and transfer rate is different
sentive polymer, typically sandwiched between two parallel from other storage systems. Fundamentally, holographic
glass plates with a thickness of a few hundred micrometers. data storage is based on multiplexing many holograms in
Due to the small thickness of the medium, conventional an- the same volume of the recording medium. For media with a
gular and wavelength multiplexing do not allow hundreds of linear response, this implies that the dynamic range for each
holograms to be superimposed [2], because eventually inter- hologram is roughly equal to the total dynamic range of the
page crosstalk becomes too large. Upon illumination by light, medium divided by , the number of holograms. As the
a hologram is recorded through a chemical reaction that poly- diffraction efficiency of each hologram is proportional to the
merizes an initially gel-like medium. By rotating the disk un- squared index modulation, readout signal strength drops off
derneath both beams, a new hologram can be superimposed as . The larger the capacity of the device is, the smaller
at a slightly shifted location, partially overlapping previously the readout signal strength and the signal-to-noise ratio
recorded holograms (shift multiplexing [9], [10]). Data are (SNR) are. In turn, small SNR causes large raw bit-error
read out by illuminating the rotating disk with the reference rate (BER), which above a threshold of 10 to 10 cannot
beam, generating the stored data page and measuring it on be further lowered by error correction schemes. To boost

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1233

the SNR, detector integration times can be increased at the
expense of data transfer rates. High capacity systems are,
therefore, easiest demonstrated for small transfer rates, and
vice versa.

C. Brief History and State of the Art

In the early 1970s, photorefractive media were investi-
gated for use in holographic data storage devices because
of the potentially high storage capacity [19]–[21]. Theoret-
ical investigations indicated that a potential storage density
of 10 b/cm could be achieved [22], with data transfer rates Fig. 4. Holographic optics processes for: (a) recording and (b)
exceeding gigabits per second. At that time, however, the readout.
necessary input and output devices were not available to take
advantage of the high data bandwidths that page storage af-
fords. It was also difficult to fix information in the crystals, times better data storage capacity and transfer rates than the
resulting in data erasure upon readout. The full potential of first digital HDSS developed by Stanford in 1994 [24], using
holographic data storage was never realized. LiNbO as the recording medium.
In the mid-1980s, the Micro Computer Corporation Finally, in Section I, we have reviewed the motivation,
(MCC, Austin, TX) and Stanford University started another potential, and research activity on HDSSs. In Section II, we
effort to build a complete HDSS [23] based on the view that review the fundamentals of holographic optics, including
disk access latency—the time it takes to find information reciprocal space, also called -space, formalism. We com-
on a hard drive—was significantly impairing computer pare several basic multiplexing architectures to illustrate key
performance. This effort led to significant advances in properties of holographic optics and conclude with several
components and systems, but the key limiting element in more advance architectures that have been implemented
the system was the material. In 1994 and 1995, the Defense in various demonstrations during the last decade. In Sec-
Advanced Research Projects Agency (DARPA) initiated tion III, we review one of the unique features of HDSS,
two consortia, the PRISM and HDSS programs, to address namely, an extremely fast associative search capability. In
key holographic issues, led by Stanford University as Prin- Section IV, we review the basic material tradeoffs for an
cipal Investigator (PI) (L. Hesselink), and IBM Almaden as HDSS, including the tradeoffs among sensitivity, robustness
Co-PI (G. Sincerbox, later succeeded by H. Coufal)). The during readout, and permanency. In Section V, we review
PRISM activities were focused on fundamental recording specific details of photorefractive materials, typically crys-
material issues, while HDSS focused on research and de- tals, including techniques for nondestructive readout. In
velopment of new optical and electronic components and Section VI, we review specific details of photopolymer ma-
digital holographic storage systems. As a result of this terials, including sensitivity, shrinkage, and media stability.
team effort, Stanford University has developed, built, and In Section VII, we review various optical system architec-
tested a complete, very high transfer rate, and high capacity tures with a view toward capacity, transfer rates, and access
holographic disk storage system [14] including hardware times. In Section VIII, we review various signal processing
implemented holographic channel decoding electronics approaches, including data modulation and error correction.
for transfer rates exceeding 10 Gb/s, while providing high In Section IX, we review critical issues in the electronics
capacity storage ( 100 Gb per 6.5-in-diameter disk) in subsystem required to support an HDSS. In Section X, we
WORM photopolymer media. This system is, to the best of review various demonstrations completed in the last decade,
our knowledge, the only complete holographic system built including both page based and single bit holographic data
to date; it also still has record-setting performance. More storage. In Section XI, we comment on the outlook of future
recently, Aprilis has demonstrated data densities exceeding opportunities in this field. In Section XII, we summarize the
250 b/ m . Also, data search capabilities exceeding 20 Gb/s, overview presented here.
much faster than magnetic disk drives have been reported.
The achieved performance came about as a result of sig- II. HOLOGRAPHIC OPTICS
nificant improvements in the components, subsystems, and,
most importantly, the recording materials. The PRISM and A. Fundamental Concepts
HDSS programs delivered significantly new materials, most 1) Hologram Recording: Holograms are recorded using
noticeably polymer-based WORM disks that are two to three a reference beam and a signal beam comprising data. As
orders of magnitude more sensitive than the best inorganic shown in Fig. 4(a), the reference beam and signal beam
photorefractive read-write crystals, new laser sources, very interfere to generate a fringe pattern, which is transferred
large charge-coupled device (CCD) detector arrays and SLM to the medium. In HDSSs, the recorded hologram typically
arrays, high-powered red lasers, as well as new digital en- takes the form of an index or phase grating. As shown in
coding techniques and novel processes for system operation. Fig. 4(b), readout is accomplished by illuminating the holo-
By the end of 1999, the PRISM and HDSS programs demon- gram with the original reference beam , which diffracts
strated polymer-based systems having 100 000 to 1 million of the recorded hologram to generate an output signal beam

1234 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 5. Relationship between wave vectors of signal and reference
beam and grating vector.

. The electric field amplitude of the propagating reference

wave is represented as Fig. 6. Bragg mismatch illustrated upon rotation of the reference
beam wave vector.
(1)
wave vector provided the strongest diffraction efficiency
is the complex amplitude, is the wave vector, and when read out with the same reference beam. The original
is the spatial dimension. The electric field amplitude of the reference beam and the hologram are said to be Bragg
propagating signal wave is similarly represented as matched. When the wave vector of the readout reference
beam is substantially different, the diffraction efficiency is
(2) significantly less. The wave vector of the reference beam
can be changed by changing the angle of incidence or wave-
where is the complex amplitude and is the wave vector.
length of the reference beam, or both. Bragg selectivity,
The interfering signal and reference wave vectors generate a
therefore, allows multiple holograms to be recorded in an
grating with a grating wave vector given by
HDSS in the same volume of media, because only the holo-
(3) gram Bragg matched to the readout beam is reconstructed
with substantial diffraction efficiency; all of the others are
This relationship is illustrated in Fig. 5. The resulting holo- substantially attenuated.
gram is given by the following index grating: Bragg selectivity is illustrated in Fig. 6. Signal beam
wave vector and reference beam wave vector record
Re (4) a hologram grating having wave vector . In an isotropic
medium, wave vectors and have the same wavenumber
in which is the transverse direction of propagation and
and, therefore, reside on the Ewald sphere.
is a complex coefficient given by
Because the original signal and reference wave vector are
Bragg matched, the vector sum lies on the Ewald sphere
(5)
(8)
where is the proportionality constant that accounts for the
material response to illumination and is the total illumina- A second reference beam with wave vector diffracts off
tion. of a grating with wave vector ; the resulting vector sum
2) Diffraction Efficiency: The diffraction efficiency of a differs from the output signal by
recorded hologram is given by
(9)
(6) The second reference beam is, therefore, not Bragg
matched to the grating and is, therefore, substantially at-
where is the thickness of the medium, is the absorption tenuated. When the wavelength is the same and the angle
coefficient of the material, and and are the polariza- of the reference beam changes, as is the case for angle
tions of the reference and signal beams, respectively. multiplexing, the Bragg selectivity is approximated as
Note that the diffraction efficiency can approach 100% for
relatively strong holograms. For relative weak holograms, sinc (10)
with negligible absorption, the diffraction efficiency can be
represented as
where is projection of on the axis and sinc
. Thus, Bragg selectivity may be characterized
(7) by a minimum separation of grating vectors along the axis,
given by . The first null of the sinc func-
3) Bragg Selectivity: Holograms recorded in a material tion is obtained at this separation vector. The finite lateral
having a substantial thickness display a property called extent of the medium puts a similar constraint on the min-
Bragg selectivity. Bragg selectivity means that holograms imum separation of grating vectors in the and axis, so that
recorded with a particular, for example planar, reference the corresponding minimum separation vectors are given by

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1235

Fig. 7. Bragg degeneracy arrangement. Fig. 8. Fourier transform configuration for holographic data
storage. (a) Point source. (b) SLM.

and , respectively, 5) Material and Sample Sensitivity: Another important

for medium height and width , respectively. property of materials used in HDSSs is the recording sensi-
When multiple hologram gratings are recorded to increase tivity. The material sensitivity, typically measured near ,
the capacity of a storage medium, the Bragg selectivity re- can be determined using
sults in significant crosstalk between the holograms. As will
be shown in subsequent sections, holographic data storage ar-
chitectures are generally designed to minimize this crosstalk (14)
due to the finite size of the medium.
The crosstalk between gratings is generally geometry de- The material sensitivity depends on material parameters
pendent, and we need to consider the concept of degenerate and recording configuration and is generally reported in cen-
gratings, frequently called Bragg degeneracies. Fig. 7 illus- timeters per joule. The sample sensitivity can be determined
trates how two different pairs of signal and reference beams using
( , ) and ( , ) can write gratings having exactly the
same wave vector. In this case, any information stored using (15)
this wave vector is superimposed, leading to prohibitively
high crosstalk. The solution to this problem is to carefully de- where is the exposure time. The sample sensitivity depends
sign an architecture to manage grating overlap. Straightfor- on the material thickness and is generally reported in square
ward examples include angular multiplexing based on a set centimeters per joule. (The type of sensitivity referenced can
of reference beams having coplanar wave vectors and wave- be generally inferred from the units reported.)
length multiplexing with a reference beam oriented in a fixed Inorganic photorefractive materials such as lithium nio-
direction. These will be discussed in detail in subsequent sec- bate have sample sensitivity ranging from 0.01 to 0.1 cm J
tions. More sophisticated techniques have been proposed for and of about 1–10 for 1-cm-thick samples. Polaroid
reference beam sets as well as coded, nonplanar reference ULSH-500 photopolymers have sample sensitivity ranging
beams [25]–[27]. from 20 cm J and as high as 10 for 0.5-mm-thick
4) Dynamic Range: When multiple holograms are samples. Development of stronger, more sensitive, and more
recorded in a material, the dynamic range of the index stable holographic storage materials is an active area of re-
grating is divided up among the holograms. This dynamic search.
range is characterized by an , which is defined as
B. Fourier Transform Configuration

(11) A particularly advantageous configuration for holographic

data storage utilizes the Fourier transform properties of a
spherical lens. In the Fourier transform configuration, shown
in which is the total number of holograms and is the in Fig. 8, the signal path comprises a signal page generated by
diffraction efficiency of the th hologram. In an ideal mate- an SLM situated at the front focal plane of a lens, situated a
rial, the is independent of for a large number of mul- focal length in front of the lens. The holographic recording
tiplexed holograms. In this case, the is given by material is situated near the rear focal plane of the lens, but
not necessarily at the focal plane. Fig. 8(a) shows that point
(12) source becomes a plane wave on passing through the lens.
More generally, Fig. 8(b) shows an SLM having a spatial dis-
tribution given by , for position vector , which is trans-
The diffraction of each hologram is a system of multi-
formed into a spectrum of wave vectors according to the re-
plexed holograms is, therefore, given by
lationship

(13) (16)

1236 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 Equation (19) indicates that a material with an index of re-
fraction of 2.5, such as lithium niobate, has a channel density
of 2 10 resolvable grating vectors. In a practical appli-
cation, the number of degrees of freedom is reduced by a
factor to 10 to 10 due the limited NA of a signal beam
and limited angular excursion of the reference beam. In an
HDSS, the accessible data channels are the product of
the number of pages and the number of pixels per page .
We assume here a code rate of one, or, in other words, that
there is no extra overhead associated with adding error cor-
Fig. 9. Grating wave vector resolution in k -space. rection and channel coding bits. In practice this is not true,
and the code rate is smaller than one, typically around 0.75.
Thus, in this configuration, each pixel corresponds to a Once we have determined the number of data channels, we
plane wave in the holographic storage material. A recorded may then estimate total capacity through the Shannon law,
image of the signal page is imaged onto a detector array namely, by multiplying the total number of channels by the
using an appropriate lens. In the Fourier transform configura- capacity of a single channel
tion, the signal page is distributed throughout the holographic
storage material and is, therefore, robust against local defects SNR SNR (20)
in the material. This configuration is relatively insensitive to
displacements of the holographic storage medium, provided in which is the number of pages and is the number
it remains in the same orientation. of pixels per page, and SNR denotes signal-to-noise ratio.
For a white Gaussian noise model, given the constraint of
C. Grating Vectors as Data Channels constant laser power, and for constant total diffraction effi-
The data capacity can be estimated using a de- ciency, the SNR decreases with increasing number of pixels
grees-of-freedom model based on resolvable grating vectors per page SNR At the same time, due to a lim-
[26], [28]. The finite extent of a holographic storage medium ited dynamic range, the total diffraction efficiency decreases
generally limits the resolution of grating vectors. Fig. 9 il- as , leading to a corresponding behavior of the SNR
lustrated the resolution of grating vectors diagrammatically. SNR . Thus, there exists a tradeoff between the
In the direction of signal propagation, the grating vector SNR and the number of pages which can be superimposed.
resolution is given by ; in the directions As a result, given parameters of the system (laser power, op-
perpendicular to the direction of signal propagation, grating tical performance, electronic detection noise, media ,
vectors are given by and . and data rate), there is an optimum number of pages which
Here, , , and correspond to the medium length, width, need to be written in a single location of the volume holo-
and height, respectively. graphic memory for which the capacity is maximized
In an isotropic medium, the maximum wave vector length [29] (Fig. 10). For realistic parameters of storage in the 90
is given by , which defines a radius of a grating geometry lithium niobate system ( , , laser
vector sphere that bounds the space of allowable grating vec- power 1 W) the theoretical upper bound on the optimum
tors. The volume of this space is given by number of pages is estimated to be 2500, with total raw
capacity per storage location of 2.5 10 b [9].
It is worth noting that the capacity calculated above is
(17) merely an information theory upper bound and does not con-
sider the effects of SNR reduction due to nonideal imaging
Defining a data channel as a resolvable grating vector, the and interpixel and interpage crosstalk. The state-of-the-art
total number of accessible data channels is given by demonstration of digital hologram recording in the 90 ge-
so that ometry in lithium niobate [30] experimentally realized a ca-
pacity of 10 pixels per storage location (i.e., 1000 superim-
posed holograms of 1-megapixel images). The resulting raw
(18)
channel density of 350 b m (calculated as the ratio of
capacity and the area occupied by the holograms on the en-
where V is the volume of the holographic recording medium. trance face of the storage crystal) represented, however, only
This quantity allows two data channels in quadrature for 1.08 of the theoretical volumetric density limit of .
each grating orientation, which generally requires coherent
demodulation to separate the data channels. For one data
D. Multiplexing Techniques
channel in each grating direction, the total number of acces-
sible data channels is given by [26] and [28] 1) Angular Multiplexing: Angular multiplexing consists
of hologram formation in which each reference wave con-
sists of a plane wave of the same wavelength incident at a
(19) different angle. A reference beam selected from the set of

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1237

 the holographic medium. Each reference vector is limited to
a plane of incidence further situated so that the reference
beams and an included wave vector in the signal beam lie in
the same plane, as shown in Fig. 11. A separate hologram is
recorded with each reference wave, resulting in multiplexed
holographic data pages. A data page is read out by illumi-
nating the holographic medium with one of the reference
beams. The output data page is focused on the output de-
tector, and then demodulated by the appropriate electronics.
Throughout this paper, we are primarily interested in page-
oriented multiplexing, in which we organize a large quantity
of data recorded as pixels of an image. Each pixel stored by
each reference beam may be considered to be a separate data
channel. Angular selectivity is maximized for signal and ref-
erence wave vectors that are normally incident to perpendic-
ular faces of the holographic storage medium [8], [31]–[33].
For copolarized signal and reference beams, the diffraction
efficiency is expressed as

sinc (22)

where is the mean angle of the first and second reference

beam and of incidence of the signal beam. The minimum
N
Fig. 10. Capacity versus number of superimposed data pages separation angle is defined as the first extinction of the sinc
M
(courtesy of M. Neifeld, University of Arizona [29]). denotes
in this case the number of pixels per page, and represents the function, which occurs at
number of superimposed holograms.
(23)

In this expression, the separation angle refers to the angle

inside the material; refraction at the surface of materials
having a high index of refraction will substantially increase
the minimum separation angle.
For transmission holograms, in which the signal and ref-
erence beam are incident at the same side of the material, the
angle between the signal and reference beams is relatively
small, especially in media displaying a high index of refrac-
tion, and angular selectivity is moderate. Angular selectivity
is highest for the perpendicular geometry, in which the signal
and reference beam are incident at perpendicular surfaces of
Fig. 11. General angle multiplexing architecture. a holographic storage medium. The dependence of separa-
tion angle as a function of the angle between the signal and
reference beams is shown in Fig. 12. In the perpendicular
reference waves used to multiplex holograms at different an-
geometry, the minimum separation angle outside the holo-
gles is given by
graphic material is given by

(24)
(21)
in which is the index of the reference beam, is the wave In a holographic storage material such as lithium niobate
vector, does not depend on , is the angular frequency, having a thickness of 1 cm, the angular selectivity is 50 rad
is time, and is the amplitude of the plane wave. in free space for recording/readout wavelengths of 500 nm,
Fig. 11 shows a diagram for angular multiplexing. Data in the perpendicular geometry. Thus, 10 000 holograms can
are encoded on a signal beam using a signal amplitude SLM be accessed with an angular range of the reference beam of
(signal ALSM). The appropriate data pattern is loaded onto about 30 .
the signal ALSM through an appropriate electronic interface For moderate, nonoverlapping angular spectra of the
to a computer or peripheral system. The signal ALSM is sit- signal and reference beams, we can derive a simple expres-
uated at the front focal plane of a lens; the modulated signal sion to estimate the accessible data channels. A reference
propagates to the holographic medium in a Fourier transform wave illuminating the side face of the crystal, therefore, has
configuration. Reference beams are plane waves incident on a maximum number of degrees of freedom, or number of

1238 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 Fig. 13. Ewald sphere showing relationship between signal wave
vector  , reference wave vector  , and grating wave vector K .

Fig. 12. Angular selectivity as illustrated by separation angle

as a function of angle between the signal and reference beam,
Fig. 14. k -space representation of angle multiplexing.
in a holographic storage material such as lithium niobate
having a thickness of 1 cm, an index of refraction of 2.5, and
recording/readout wavelengths of 500 nm.
aration between the spectra of two holograms is not
constant at various positions in the spectra, so that the op-
grating vectors separated by the minimum separation vec- timum grating vector separation of a multiple of
tors, of . A signal wave illuminating a face has is not the same for all grating vectors in the spectrum. Thus,
a maximum number of degrees of freedom the null of the Bragg selectivity sinc function cannot be met
in the dimension and in the dimension. over the entire signal spectrum, resulting in crosstalk.
In this geometry, the number of accessible data channels This origin of crosstalk is illustrated in a different way in
is further estimated by taking into account the angular Fig. 15. A principal signal wave vector and a principal
bandwidths of the reference and signal beam to yield reference wave vector are incident in the perpendicular
geometry. For the principal wave vector, an angular separa-
NA (25) tion of is selected to correspond to the minimum separa-
tion grating vector projection on the axis . Thus, the
where NA is the numerical aperture of a signal incident at the peak readout angle of each hologram is situated at the null
surface normal and is the angular range of the reference angles of each of the other multiplexed holograms. For an-
beam, both in free space outside the material. For a wave- other signal wave vector , however, the peak readout angle
length of 500 nm, a signal NA of 0.25, and reference beam is not situated at the null angles of each of the other multi-
angular range of 30 , the number of data channels is about plexed holograms, resulting in crosstalk.
10 cm . Crosstalk is minimized when the signal and reference
Crosstalk: A practical HDSS generally displays devia- beams are incident on perpendicular surfaces of the medium,
tion from the grating vector separation rules identified above. for which the ratio between the crosstalk and signal, the
The result is crosstalk between multiplexed holograms. To crosstalk limited SNR, is given by [31]
understand the impact of crosstalk, we use a reciprocal space,
or -space, formalism [9], [28], [32], [34]–[36]. A signal SNR (26)
NA
beam comprising a data page and a reference beam generate
a hologram having spectrum of grating wave vectors. This in which is the number of holograms and NA is the nu-
spectrum is shown in the shaded arc in Fig. 13, which depicts merical aperture of the signal beam. When discrete pixels
a cross section of -space. In general, the spectrum of grating are stored in a Fourier transform arrangement, so that rows
vectors comprises a solid cross section of a sphere. Multiple of pixels span the dimension parallel to the reference wave
holograms are recorded by varying the incident angle of the vectors, the relative crosstalk takes on the form [32]
reference beam so that the wave vector rotates. The result is
a series of grating vector spectra indicated in Fig. 14. An im-
SNR (27)
portant observation in Fig. 14 is that the grating vector sep-

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1239

Fig. 15. Angle multiplexing properties for: (a) normally incident signal wave vector and (b)
oblique signal wave vector.

Fig. 16. Generic wavelength multiplexing architecture.

in which is the number of columns of pixels.

2) Wavelength Multiplexing: Wavelength multiplexing
consists of hologram formation in which each reference
wave consists of a plane wave of a different wavelength
incident at the same angle, such that

Fig. 17. Frequency selectivity as illustrated by separation angle

as a function of angle between the signal and reference beam, in
(28) a holographic medium such as LiNbO having a thickness of
where now and the corresponding angular frequency 1 cm, an index of refraction of 2.5, and recording and readout
vary but the orientation of is the same for all refer- wavelengths of 500 nm.
ence beams. Only the recording reference waves are Bragg
1matched, and any others result in Bragg-mismatched re- beam. For copolarized signal and reference beams, the
construction. diffraction efficiency is expressed as
Fig. 16 shows a diagram for wavelength multiplexing. As
with the angular multiplexing case, a Fourier transform con- sinc (29)
figuration is used. The signal and reference beams are inci-
dent on opposite sides of the holographic medium, known where is frequency difference between the recording and
as a reflection configuration. Wavelength selectivity is maxi- readout reference beam, is the velocity of light in free space,
mized for included signal and reference wave vectors corre- and is the angle between a planar signal and reference wave
sponding to counterpropagating directions in the holographic in the medium. For reflection holograms, the minimum sep-
storage medium [8], [31]–[33]. aration frequency is defined as the first extinction of the sinc
Fig. 17 shows the geometry in which the reference beam function, which occurs at
is incident normal to the surface of a medium and the signal (30)
beam is incident at an angle with respect to the reference

1240 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 18. k -space representation of wavelength multiplexing. Fig. 19. Generic HDSS using phase-encoded multiplexing.

For reflection holograms, the angle between the signal One suggested approach to address the issue of backscatter
and reference beam is generally substantial. Frequency is to use a perpendicular geometry, which has wavelength
selectivity is highest for the counterpropagating geometry, in selectivity comparable to the counterpropagating geometry,
which the signal and reference beam are incident at opposite as shown in Fig. 17 [38]; for moderate capacities, crosstalk
surfaces of a holographic storage medium and propagate in is not a factor.
opposite directions. The dependence of separation frequency 3) Phase-Encoded Multiplexing: Phase-encoded mul-
as a function of the angle between the signal and reference tiplexing uses a phase SLM to control the reference beam
beams is shown in Fig. 17. For counterpropagating signal and requires accurate control of the phase modulation and
and reference beams, the minimum separation frequency is efficient collection of phase patterns to minimize crosstalk.
An example architecture for phase-encoded multiplexing is
(31) shown in Fig. 19. Here, a Fourier transform arrangement
In a holographic storage material such as lithium niobate is used for both the signal and reference beam paths. A
having a thickness of 1 cm and an index of refraction of phase-encoded reference beam comprises a phase pattern
2.5, the frequency selectivity is 6 GHz for recording/readout established by an SLM. In this example, a one-di-
wavelengths of 500 nm. Thus, over 10 000 holograms can be mensional (1-D) SLM is situated at the front focal plane of
accessed with a wavelength range of the reference beam from the lens, so that the resulting reference beam incident on
500 to 560 nm. the material has a wave vector spectrum in one dimension.
Wavelength multiplexing generally displays crosstalk By using a set of orthonormal functions and an appropriate
between multiplexed holograms due to deviations from the optical configuration, multiple holograms may be superim-
wavelength selectivity rules identified above. The series of posed without significant crosstalk.
grating vector spectra for wavelength multiplexed holograms Multiple holograms are recorded by varying the phase
is indicated by bold line arcs in Fig. 18. As with angular codes of the reference beam. For each hologram, the result
multiplexing, the grating vector separation between is the grating vector spectrum indicated by the shaded
the spectra of two holograms is not constant at various region in Fig. 20. In contrast to angular multiplexing, the
positions in the spectra, so that the optimum grating vector grating vector spectra for each hologram overlap; what dis-
separation of a multiple of is not the same for all tinguished each hologram are the phase relationships among
grating vectors in the spectrum. Thus, the null of the Bragg the recorded holograms. Krile et al. [39], [40] and Morozov
selectivity sinc function cannot be met over the entire signal [41] have independently outlined this procedure for thin
spectrum, resulting in crosstalk. holograms, for which there is no Bragg selectivity. Because
Crosstalk is minimized when the signal and reference of its simplicity, we examine this procedure to illustrate
beams are counterpropagating, for which the ratio between basic properties of phase encoding. When volume effects
the crosstalk and signal, crosstalk limited SNR, is given by are taken into account, Bragg selectivity reduces crosstalk.
[37] In the general case, a two-dimensional (2-D) phase pattern
may be used. The transmittance of the medium with a linear
SNR (32) response is given by
NA
in which NA is the numerical aperture of the signal beam. (33)
Note that for this configuration, the crosstalk limited SNR
does not depend on the number of multiplexed holograms,
owing to the arrangement in reciprocal space. where and denote the Fourier trans-
In practical systems, crosstalk is not the limiting factor. forms of the signal and the coded reference ,
Rather, wavelength multiplexing is limited by stable tunable respectively, at the plane of the SLM, and is a proportion-
laser sources and strong backscatter in holographic media. ality constant dependent on the strength of the hologram.

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1241

 For the case illustrated in Fig. 19, the reference wave
would ideally be composed of point sources corresponding
to a pixel in the plane of the phase SLM. In the Fourier
transform geometry illustrated here, each point source corre-
sponds to discrete constituent plane waves after the Fourier
transform lens, such that

(39)
Fig. 20. k -space representation of phase-encoded multiplexing.
When a reference beam is modulated in phase in one dimen-
sion, as shown in Fig. 19, it consists of wave vectors spanning
Upon illumination with reference beam , light diffracted
a plane. We choose a spacing of plane wave components such
from the hologram is given by
that the discrete spectrum of wave vectors corresponds
(34) to the wave vectors optimized for angular multiplexing. For
proper selection of geometries, Bragg mismatch can be ne-
where is the readout reference beam. If a refer- glected. For this case, Krasnov has shown that the diffracted
ence and signal pair and are used for recording the th signal is given by [44]
hologram, then upon readout with we arrive at
(40)

in which the proportionality constant depends on the prop-

erties of the holographic material. For this special case,
(35) defines the inner product between and .
To minimize crosstalk, we require that the reference waves
be orthogonal. In a practical application of phase encoding,
The first summation gives the reconstructed wave of interest. we will typically have a configuration in which Bragg
If one hologram is stored, the diffracted beam that passes mismatch cannot be completely neglected; a more detailed
through the Fourier transforming lens to produce the field is analysis is required, but we can use the expressions obtained
[42] for above to place boundaries on the expected behavior
(36) of the system. Overall, we expect the crosstalk-limited
SNR of an orthogonal phase-encoded multiplexing system
at the detector plane, where the operators and denote to be comparable to the case of angular multiplexing, and
correlation and convolution, respectively, and is a propor- somewhat higher as we deviate from strict orthogonality.
tionality constant. A sharp autocorrelation peak is required For a 2-D reference beam, the presence of Bragg selec-
in order to reconstruct the signal . For many object pairs tivity reduces crosstalk due to Bragg mismatch, but crosstalk
this expression is generalized to due to Bragg degeneracy cannot be eliminated. Generally, we
desire a set of reference codes such that the autocorrelation
functions closely approximate a delta function centered at
(37)
zero, and that every pair of codes has a cross correlation that
is uniformly small and ideally zero. Then the reconstructed
If the cross correlation between two reference codes and signal will have minimal crosstalk.
is appreciable, then we have significant crosstalk between When an SLM or phase diffuser generating the reference
two images stored with these codes [42]. Krile et al. have beam in the Fourier transform geometry spans two dimen-
shown that this objective can be efficiently accomplished sions, significant crosstalk occurs due to Bragg degenera-
using Gold and Carter codes developed for spread spectrum cies. Generally, the phase can be assumed to be random for
communication applications [40]. each pixel, or wave vector address in the Fourier geometry,
Kral et al. [43] have shown that the lower limit for the where a random phase varies uniformly from 0 to For 2-D
crosstalk-limited SNR for holograms multiplexed in thin random phase codes, crosstalk is estimated as [27]
media is given by
SNR (41)
SNR (38)
where is the number of resolvable reference beam wave
The SNR can be improved using techniques outlined by vectors in the dimension
Kral et al., including illumination of a phase pattern with a
spherical reference beam. (42)

1242 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 The phase of each corresponding reference beam plane spectrum that will not generally have optimum Bragg selec-
wave component is individually determined. tivity properties. Rather, the cross-correlation formulation is
For the architecture in Fig. 19 having a 2-D reference a better model to describe the process, and in general a mixed
beam SLM, each pixel corresponds to a distinct reference approach is needed to develop a complete qualitative analysis
beam wave vector. These distinct reference beam wave [27]. In general, the shift multiplexing selectivity is equal
vectors may be referred to as degrees of freedom [27], [28]. to the autocorrelation length of the speckle field, which is
Usable SNR can, therefore, be obtained by having more NA in both the and directions.
degrees of freedom in the reference beam than number of The details of the crosstalk will depend on the details of the
holograms, or oversampling the reference beam. For ex- speckle pattern and the architecture of the HDSS. Crosstalk
ample, if a speckle pattern is defined by illuminating a phase decreases as the number of degrees of freedom associated
plate at the front focal plane of a lens, SNR generally im- with the speckle pattern increases.
proves as the number of grains increases, up to a saturation In this geometry, the number of accessible data channels
limit. can be approximately estimated as
Recent advances in phase-encoded multiplexing are
described in [45]. An important feature of phase-encoded NA (44)
multiplexing not afforded other multiplexing approaches is
encryption of recorded data using phase encoding. Ideally, where NA is the numerical aperture of a signal incident at
the capacity of phase-encoded multiplexing is comparable surface normal and the angles and are the minimum and
to that of angular multiplexing. Demonstrations and prac- maximum incident angles of the wave vector spectrum of the
tical implementations, however, have been limited by the speckle field of the reference beam, in free space outside the
suitability of available phase-encoding SLMs. material. The randomness of the speckle pattern generally
Correlation Multiplexing: For the holographic disk results in a number of accessible data channels significantly
geometry, shift multiplexing is the most suitable and natural less than this upper limit.
solution, since it does not require any beam steering or wave- While the details of correlation multiplexing are more
length change and is achieved by simply moving the medium complex than random phase-encoded multiplexing de-
with respect to the reference and object beams, as indicated scribed above, it is useful to compare the two to understand
in Fig. 2. Shift multiplexing can be implemented by using the crosstalk trends of correlation multiplexing. In general,
a spherical wave [46] or, alternatively, a random or, pseu- the details of the crosstalk will depend on the details of
dorandom speckle pattern as a reference beam [12]. Shift the speckle pattern and the architecture of the HDSS. We
multiplexing in the spherical reference beam case relies on can refer to (41) to estimate the upper bound of SNR. In
the fact that a spherical type beam contains multiple angular particular, (41) shows that crosstalk decreases as the number
components and, thus, a shift of the media or the reference of resolvable reference beam wave vectors in the dimen-
beam leads to conventional Bragg mismatch. Multiplexing sion associated with the speckle pattern increases. For
with a speckle beam is based on the spatial autocorrelation speckle multiplexing geometries, is given by
function of the beam used for the recording. In such case,
the diffraction efficiency of the hologram is determined by (45)
the degree of correlation between the speckle beam used for
the recording and the speckled beam presented upon readout By engineering the magnitude and the correlation function
(e.g., a spatially shifted version of the original beam). For of the phase mask, the decorrelation distance is modified. Ex-
random Gaussian speckle, the hologram shift selectivity is cellent densities can be achieved with this method, as shown
determined by the beam autocorrelation function [12] in Fig. 21.
With the random phase mask in the reference beam,
NA 100 holograms were superimposed with good diffraction
(43) efficiency, as shown in Fig. 22. These results indicate that
NA
densities of over 60 b m can be achieved with this tech-
where is the shift amount, , NA is the numer- nique.
ical aperture of the reference beam, and is the first-order Shown in Fig. 23 is an example of multiplexing multiple
Bessel function of the first kind. The speckle selectivity does holograms within virtually the same volume of the medium
not depend on the direction of the media shift (Bragg selec- using this technique.
tive or Bragg degenerate) and is independent of the media In the Stanford HDSS holographic disk system described
thickness, although the crosstalk buildup is usually much below, with the medium thickness of 200 m, the hologram
faster in the Bragg-degenerate direction (and in thin media). shift selectivity is approximately 1.0 m and the corre-
Thin-film-based materials are not typically suitable for an- sponding NA of the reference beam is 0.25.
gular multiplexing due to the difficulty in achieving the per-
pendicular geometry. E. Phase-Conjugate Readout
Qualitatively, speckle multiplexing can be modeled using An important approach for readout of holographic data
the formalism developed for phase-encoded multiplexing de- is to use a phase-conjugate reference wave. A hologram
scribed above. The speckle pattern will produce a plane wave is recorded in a typical manner, as shown in Fig. 24(a). In

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1243

Fig. 21. Spherical wave and random phase mask multiplexing results for 200-m-thick sample of
Polaroid photopolymer ULSH-500.

Fig. 22. Superposition of 100 holograms without implementation of scheduling.

Fig. 24. Holographic optics processes for: (a) recording and (b)
phase-conjugate readout.

Fig. 23. Holograms superimposed using speckle shift simple proof-of-concept demonstrations for materials dis-
multiplexing (sample out of 140). NA of the speckle reference is playing strong scattering [49]–[51]. More recently, digital
0.2; media thickness is 200 m.
HDSSs using phase-conjugated readout have been evaluated
at Caltech [9], [52] and IBM [1].
a phase-conjugate readout configuration, a counterpropa- Phase-conjugate readout, although being an extremely
gating phase-conjugate reference beam is incident on the elegant technique realizable with inexpensive optical com-
material, as shown in Fig. 24(b). This reference beam gener- ponents, puts severe limitations on the performance of the
ates a phase conjugate of the signal beam, which reverses the system in terms of media and channel evaluation capabilities,
distortions in the material. Crosstalk between multiplexed since direct imaging (without writing a hologram) is not pos-
holograms and noise caused by strong distortions in a holo- sible. Second, the phase-conjugation technique requires (in
graphic storage material can be reduced by readout with a most cases) the capability of generating a phase-conjugate
phase-conjugated reference wave. Such a scheme has been (or pseudophase-conjugate) replica of the reference beam to
suggested by several authors [47], [48] and implemented in be presented during readout, which is rather cumbersome

1244 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 25. Layout of the holographic search test bed. The diode-pumped YAG laser (DPSS 532) is
used to record/read/search holograms. The HeNe laser is used for alignment.

to implement within a shift multiplexing geometry using signal detection in order to increase the autocorrelation to
a speckle or phase coded reference beam. An alternative cross-correlation ratio.
approach employs storage of a phase-conjugated replica of Due to the lack of shift invariance, volume holographic
the reference beam itself using a buffer hologram [53]. optical correlators are most suited for relational database
searches in which each individual data field has a fixed
assigned position within the recorded hologram [54].
III. ASSOCIATIVE SEARCH
The holographic associative search testbed developed at
Volume HDSSs can be used to perform fast data searches Stanford has the capability to record and retrieve multiple
of the stored data pages. Because the readout of each holo- holograms recorded in a LiNbO crystal by angular multi-
gram multiplies two 2-D Fourier transforms (the transform plexing, as shown in Fig. 25. The correlation of a group of
of the search argument and the 2-D Fourier transform of the holograms with an arbitrary search key can be detected with
stored images), multiple 2-D correlations are computed opti- the correlation CCD camera. The control software allows the
cally. When multiple volume holograms are presented under hologram(s) corresponding to the most intense correlation
the SLM with the search argument presented on it, the recon- peaks to be selected.
structed reference beams are generated, with their respective Multiple holograms are recorded in the LiNbO crystal
amplitudes being proportional to the correlation strength be- using the 90 geometry by deflecting a galvanometer mirror
tween the search argument and the stored data pages. The in the reference beam path over different angles. The light
search speed is basically limited by the data recording den- source used initially was a diode pumped continuous-wave
sity and, for the system described below, can be as high as (CW) laser with nominal power of 400 mW and a wavelength
10–100 Gb/s. of nm. A twisted nematic liquid-crystal SLM (1024
Volumetric holograms exhibit substantial Bragg selec- 1024 pixels, pixel size m, video frame rate),
tivity with respect to small angular changes of the readout developed by IBM for the HDSS program was used as the
beams in the Bragg-selective direction. Thus, if the search page composer. The data readout CCD camera was manufac-
pattern (search argument) is shifted with respect to its tured by Dalsa, Waterloo, ON, Canada (1024 1024 pixels,
position in one of the stored images, the reconstructed pixel size 12 12 m). One-to-one imaging is accomplished
reference correlation signal will be substantially reduced with a high-quality imaging lens built by the Rochester Pho-
due to Bragg mismatch. In the Bragg-degenerate direction, tonics Corporation, Rochester, NY, for the HDSS consor-
the Bragg selectivity itself is a lot more forgiving, but the tium.
reconstructed reference signal will be angularly displaced in To read out the correlation field intensity measured, a
the Bragg-degenerate direction by an amount proportional Kodak Megaplus CCD camera is used, having 1008(H)
to the relative shift between the search argument and its (1018(V) pixels, of 9 9 m pixel size, and a center-to-center
position in the given stored image. The shifted reconstructed spacing of 9 m (20% fill ratio). We used a pair of F-theta
reference beams can be filtered optically (by a pinhole scan lenses, mm, scan angle 25 , spot diameter
and the correlation CCD pixel itself) before correlation 12 m (Fig. 25), to obtain a small, near diffraction limited

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1245

Fig. 26. Example of correlation field captured by Kodak CCD camera; 100 holograms are recorded.

Fig. 28. Query picture. 80% of the original photograph has been
erased.
Fig. 27. An example bitmap image from the demo database.
The bitmap image can be displayed on the SLM for recording and, after applying normalization to the correlation spot in-
or searching. A mask has been used to eliminate areas of poor
image quality from use. Each database field has a fixed position tensities, the processing software ranks all the correlation
in the bitmap image. The holographic image contains personal spots in the order of the correlation intensity values. An ex-
information of a person stored in the database. ample of search system operation is presented in Figs. 28 and
29.
correlation spot size for all holograms within the reference The performance of the volume holographic associative
beam angular range. To minimize crosstalk and optimize memory was evaluated in a number of search demonstra-
overall system performance, the spacing between correlation tions. When any one of the ten database fields was used as
spots is typically increased by decreasing the number of the query key, the expected records were consistently within
recorded holograms to 334—using three pixels per correla- the top ten of the 200 search results. With a larger query key
tion period. (for example, using the combination of the first name and last
Fig. 26 shows an example of the correlation field with the name field as the query key) better search results were ob-
scanning lens reference beam telescope (100 holograms were tained. Fig. 28 is a partial picture that we used as the query
recorded). key. Eighty percent of the photo in the query key has been
The experimental system was used to demonstrate fast par- erased, but it still returns the correct record. The retrieved
allel searches in a relational database. A relational database complete personal record (Fig. 29) is identical to the original
has data organizational structure such that the specific data is record.
located in a specific part of the stored holographic page, thus Each search operation includes the following: imprinting
eliminating the difficulties associated with holographic shift the search key on the object beam and illuminating stored
invariance. In the Stanford system, the demo database is a holograms; detection of the correlation signal, normalization
relational database that contains personal information about by hologram strength, ranking of the correlation intensity,
people. A typical data encoded page used in the demonstra- mapping of selected correlation spots to reference beam an-
tion is shown in Fig. 27. gles, address-based recall of the selected hologram, decoding
To demonstrate the parallel search ability of a volume of the retrieved data page, and finally, display of the search
holographic associative memory, 200 holograms were stored result.
in 90 -geometry in Fe : LiNbO crystal in the associative In the current system, the major time overhead is asso-
search testbed [55]. Each hologram represents one record ciated with mechanical components, signal processing, and
from the demo database. The reference beam angles were search argument display. The fundamental limit to the search
chosen in such a way that each reference beam was focused speed comes from the key generation and correlation signal
to a unique portion of the correlation camera. An exper- processing (ranking and normalization). With dedicated
imentally verified exposure schedule was used to achieve hardware and a fast SLM [such as a 1000-ft/s Texas Instru-
a uniform diffraction efficiency distribution among the ments deformable mirror device (DMD)], search speeds of
multiplexed holograms. The processing software converts 10 000 stored pages/s or more can be envisaged.
the user query inputs into a search key bitmap file by using At IBM, a similar associative search experiment was car-
appropriate modulation coding and encoding mask. Shown ried out (see Fig. 30) [1]. Researchers at IBM showed that
in Fig. 27 is an example search key bitmap file. if an unindexed conventional “retrieve-from-disk-and-com-
After the search keys is presented on the SLM, the image pare” software-based database is searched, the search is lim-
of the correlation spots is grabbed by the correlation CCD ited by the sustained hard-disk readout rate (taken to be

1246 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 29. The first ranked search result. While using pictures as the query key, the volume
holographic associative memory is in fact doing pattern recognition.

25 MB/s). For example, a search over 1 million 1-kB records recording light. Thermal fixing has been demonstrated al-
would take 40 s. In comparison, with off-the-shelf, video- lowing nondestructive readout for periods of years, but this
rate SLM and CCD technology, an appropriately designed process typically requires temperatures of at least 100 C
holographic system could search the same records in 30 ms, for fixing.
a 1200 improvement. Custom components could enable Photopolymer materials usually have much higher sen-
1000 or more parallel searches per second. sitivity due to chemical amplification effects and a high
dynamic range, but are rather limited in terms of thickness
IV. HOLOGRAPHIC MATERIAL CHARACTERISTICS FOR (1 mm or less). Other issues associated with polymers
DIGITAL DATA STORAGE are significantly increased scatter levels compared to
single-crystal photorefractive materials, and volumetric
Although different applications require materials of shrinkage upon photopolymerization. The latter effect can
varying properties, an ideal material would have a fast time be somewhat mitigated by use of high molecular weight
response (in microseconds), would be as as sensitive as pho- monomers and proper choice of the material chemistry. The
tographic film (in joules per square centimeter), would retain shrinkage and scatter limits the available usable thickness of
information over long periods ( ten years), would have a photopolymers.
large spectral sensitivity range extending into the infrared The typical minimum material parameters for optimized
region of the spectrum, and would be available in large holographic performance are listed below [56].
quantities and sizes with good optical quality, and would
cost little. Such a recording material does not exist today. Material:
In general, sensitive materials have small photorefractive — Thickness: mm.
response, and vice versa. For example, LiNbO is available — Sensitivity: cm J (or SL cm J).
in large size, of good optical quality, and at relatively low — Dynamic range: (or,
cost, and has a large index variation per absorbed photon. ).
Its response time, however, is rather slow, on the order of — Shrinkage: 0.05% (for 500 m).
milliseconds or longer, depending on the intensity of the — Scatter: 10 srad .

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1247

 Fig. 31 presents the summary of current holographic
storage materials properties (data collected at the Interna-
tional Workshop on Holographic Data Storage, sponsored
by NSIC, Nice, France, March 8–11, 1999).
The PRISM and HDSS team members have investigated
the properties of many available materials in the IBM PRISM
tester. The results are displayed in Fig. 32 The BER was mea-
sured for a given material as a function of background scatter,
as defined above. It is notable that LiNbO is still the best
material in terms of optical quality at a certain BER at all
page sizes. This is not surprising, as the polymer materials in-
trinsically scatter more light caused by the variation in index
of refraction of the components that constitute the polymer.
These components, such as the binder, the monomer and
the polymer are needed to create a hologram; therefore, a
tradeoff is required between sensitivity, diffraction efficiency
and scatter. For inorganic materials such as LiNbO the index
of refraction is induced by a space charge field and scatter (as
well as sensitivity) is considerably lower. Currently Aprilis
and InPhase are optimizing their materials for improved per-
formance. From this graph we also note that the page size has
a considerable effect on the achievable BER. For a page that
contains 1024 kb, the achievable BER is six orders of magni-
tude worse than for a 256-kb page, while reducing by only a
factor of four the capacity per page. To optimize total system
capacity, a careful tradeoff is needed between page size and
the multiplexed number of pages. For example, by setting a
threshold value for the minimum BER, we can determine the
optimal capacity by considering the number of pages that can
be superimposed using a certain error correction code (ECC),
as described in the section on en/deconding.

Fig. 30. The top two images show schematically the holographic
storage of a page of data, while in the bottom picture, the retrieval V. PHOTOREFRACTIVE MATERIALS
of the page address is achieved among all stored pages by using a
search key as input to the system. The page address is measured Photorefractive materials record holograms through pho-
with a linear detector array (Courtesy of IBM, with permission). toinduced charge-transport (diffusion, drift, and photovoltaic
effect), charge carrier redistribution between the deep traps
— Wavelength response: visible ( 532 nm or blue). and buildup of an internal space charge field. The resulting
index distribution is produced due to electrooptic (Pockels)
Media: effect (Fig. 34).
— Disk wedge 0.5 mrad. In general a variety of trapped and photoexcited states
— Format: 120 mm to 6.5-in-diameter disk. are possible, including electrons and holes. Typically, elec-
— Total wavefront error over 2 mm spot: . trons are trapped at impurity centers and photoexcited into
— Scatter (including substrate) 10 srad . the conduction band. This example is illustrated as follows.
The performance of an HDSS greatly depends on the Using the example of Fig. 4, intersecting signal and reference
quality and physical properties of the recording medium. beams generate an interference pattern with regions of high
The DARPA PRISM consortium has developed a preci- intensity and low intensity, which has the functional form
sion tester [57] at IBM for measuring the performance of
different holographic media using a common platform. (46)
Since 1996, many advanced holographic recording materials
were evaluated and compared using a common comparison in which is the modulation factor that quantifies the con-
framework. The most important parameters include image trast of the interference pattern.
quality (degree of image distortion), sensitivity, dynamic As shown in Fig. 33, photons excite electrons into the con-
range, fidelity (spatial resolution), stability, shrinkage, and duction band in higher concentrations within the regions of
available thickness. Based on the results published so far constructive interference than within the regions of destruc-
[58], the most promising photorefractive crystals remain tive interference. Electrons diffuse in the direction of the con-
Fe-doped lithium niobate, while Polaroid CROP materials centration gradient to the regions of destructive interference,
represent the better option in the photopolymer family. where they relax to empty states in the energy band gap. In

1248 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 31. Summary of holographic materials properties.

the presence of a photovoltaic current found in materials such interference pattern up to a phase shift, generating a volume
as lithium niobate, the effect is enhanced. hologram.
A graphical depiction of the relevant physical quantities In a holographic medium that does not display nonlinear
associated with the photorefractive effect is shown in Fig. 34 beam coupling, the diffraction efficiency is [60]
for a diffusion dominated photorefractive medium. A pe-
riodic intensity variation arises from the interference of (49)
the signal and reference beams. This pattern establishes a
coulombic charge density in the photorefractive material Detailed accounts of photorefractive transport can be
that records the intensity pattern; red indicates positive found in [61]–[66]. Lithium niobate is the most common
charge and blue indicates negative charge. At steady state, photorefractive material used in HDSS demonstrations.
a periodic potential associated with the charge density Other inorganic materials of interest include strontium
balances diffusion. The separation of electrons from fixed, barium niobate and barium titanate. Organic materials
positive ions establishes an electric field , typically called include reversible photorefractive polymers that may even-
a space-charge field, that replicates the functional form of tually be candidates for holographic storage systems [67].
the interference pattern, up to a phase factor Photorefractive polymer materials, however, require large
applied electric fields, and good sensitivity as well as long
(47) term stability of recorded holograms have not yet been
This expression is accurate only for small modulation depths, demonstrated in the same medium.
in which and is the normalized space-charge
A. Complementary Gratings for Hologram Fixing
field, representing the transport properties of the medium.
In materials displaying the linear electrooptic (Pockels) ef- In a simple photorefractive material with one photoexcita-
fect, for example, the space-charge field generates a periodic tion process and one species of trapped charge, light used to
index grating proportional to the electric field, such that record the hologram will typically erase the same hologram
upon readout. Thus, for permanent storage and multiple reads
(48) without data degradation, photorefractive materials require a
fixing process.
where is the index of refraction and is the effective Hologram fixing can be accomplished in a number of
Pockels coefficient, which in general depends on the Pockels ways. One method makes use of complementary charge
tensor, light polarization, and crystal orientation [59]. gratings residing in separate subsystems of charge transport
Therefore, the resulting index grating replicates the initial which are not sensitive to the readout light. The gratings are

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1249

Fig. 32. Optical quality of holographic storage materials (courtesy of R. M. Shelby, IBM Almaden
Research Center with permission [58].

Fig. 33. Photoexcitation of electrons in the presence of a periodic

intensity.

complementary in the sense that they have opposite polarity

and can screen each other. Fig. 35 illustrates a generic
process that uses complementary gratings to establish stable
holograms. Subsystem comprises a photoexcitable charge
species, such as a trapped electron, as shown in Fig. 35(a). A Fig. 34. Photorefractive effect in diffusion dominated materials.
coulombic charge density grating is written in subsystem
through the photorefractive effect. A second charge readout. For image bearing beams, an important issue is
density grating is established in subsystem using a the expansion of the crystal on heating. If holograms are
different process and may be accomplished by changing recorded at an elevated temperature and then cooled, the
a material environment parameter such as temperature, Bragg condition will be modified, and either the angle
electric field, or type of illumination, as shown in Fig. 35(b). of the reference beam or its wavelength or both must be
Optical, holographic, and system considerations put con- modified to obtain the greatest diffraction efficiency. This
straints on the use of complementary gratings for nonvolatile effect ultimately puts limits on the image field that may be

1250 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 35. Graphical depiction of the evolution of complementary gratings, illustrating charge density
grating  in subsystem A, charge density grating  in subsystem B , and the net grating  +  ,
for: (a) initial grating, (b) complementary gratings, and (c) exposed grating for readout.

Fig. 37. Arrhenius plot of the ionic hologram decay

Fig. 36. Mechanisms of holographic fixing via thermally assisted time(unscreened) in: (a) as-grown crystal (E =12 : eV, hydrogen
impurity); (b) and (c) samples with low hydrogen impurity content
=14
ionic drift.
and different Fe-doping (E : eV, lithium migration); and
([
(d) near-stoichiometric Li O] = 49 5%): Fe-doped dehydrated
crystal.
reconstructed with acceptable fidelity and will introduce ad-
ditional crosstalk because not all of the signal can be Bragg
matched simultaneously at the lower temperatures. It is a is Boltzmann constant. Mobile ions are thermally acti-
system issue whether simultaneous formation of holograms vated and move to screen out the electronic grating. The re-
at high temperature or formation a low temperature and sulting hologram becomes screened. However, at lower tem-
subsequent activation of mobile ions at high temperature peratures, the ions have significantly lower mobility and the
give optimal performance. grating lifetimes can be considerably longer than for an elec-
In lithium niobate, for example, subsystem may estab- tronic grating. Thus, at lower temperature, the ionic grating
lish an initial electronic charge density through the photore- will represent a backbone of the fixed hologram while, upon
fractive effect, and subsystem is established as a comple- light illumination, the electronic grating will redistribute it-
mentary charge density through mobile ion transport at ele- self, partially screening the fixed ionic grating. The ionic
vated temperature [68]–[71] (Fig. 36). The ionic conductivity grating also eventually decays at a much slower rate (from
usually obeys an Arrhenius-type dependence on the temper- several months to more than ten years) depending on the
ature residual conductivity at the storage temperature and the de-
gree of electronic screening.
Two alternative approaches exist for creation of fixed
(50) ionic gratings in lithium niobate. In the first case, electronic
gratings are formed at room temperature. Initial exposure
where is activation energy (approximately 1.2 eV for to the signal and reference beams generates an electronic
hydrogen impurity and 1.4 eV for interstitial lithium mi- grating which is followed by ionic compensation at higher
gration), is the density of conducting ions, is the elec- temperatures without illumination. At elevated temperatures
tron charge, denotes the permeability of free space, and (above approximately 100 C in LiNbO [72]), ions become

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1251

active and generate a complementary screening grating that
compensates the space charge arising from the electronic
grating. These paired electronic and ionic gratings consti-
tute complementary gratings and are derived from charge
patterns with opposite polarity. On cooling to room temper-
ature, the ions are significantly less mobile and represent
the ionic backbone of the fixed grating. Subsequent light Fig. 38. Two-photon gated recording requires a visible incoherent
exposure reveals the fixed grating by partially redistributing gating beam in addition to two IR recording beams. During readout,
only the IR reference beam is present. The light-induced IR
the electronic charges. In the presence of light exposure, sensitivity is about 100 times higher than without the gating beam,
the ionic grating is partially screened, but in the case of giving rise to prolonged readout.
LiNbO , strong photovoltaic currents significantly reduces
the screening. The actual degree of screening depends occurs at elevated temperatures with reading at room temper-
largely on the grating spacing, photorefractive impurity ature. For thick media, the diffraction efficiency, therefore,
concentrations (e.g., Fe ), and the reduction state of the reduces due to Bragg mismatch. This problem and the as-
material [66], [73]. sociated complications of recording at elevated temperatures
In the second approach, both electronic and ionic gratings can be overcome by using a two-photon recording method.
are formed simultaneously at an elevated temperature [68]. Two-photon gated recording techniques (Fig. 38) have
Simultaneous formation of electronic and ionic gratings typ- been investigated since the early days of research on holo-
ically results in stronger complementary gratings because of graphic memories, in particular at Bell Laboratories [79]
mutual screening during formation, in which the magnitudes [80] and the PRISM consortium [81]–[84].
of both gratings may be substantial, but their difference re- A signal and reference beam are incident at one (typical
mains relatively small [69], [70]. The typically used temper- near IR) wavelength, which is too long for photoexcitation,
ature range for this process is 150–180 C. and a uniform beam at a shorter wavelength provides the
Multiple fixed holograms based on ionic gratings have electron photoexcitation (and absorption at longer wave-
been demonstrated in LiNbO . [68], [74]–[76]. At least 500 length due to short-lived shallow traps). The gratings may
holograms may be stored using ionic gratings and for which then be probed nondestructively using the first wavelength.
diffraction efficiencies in the range of 2–25% have been re- This process has been identified in nominally pure Fe-doped,
ported [68], [75], [76]. As part of the PRISM consortium, and Cu-doped LiNbO [80], [85]–[87], in KTa Nb O
Siros Technologies first demonstrated fixed ionic gratings by [79], and in LiTaO [80]. This process is nonlinear with
recording over 500 pages of digital data in LiNbO [77]. Sub- intensity, and for optimal results, a resonant two-photon
sequently, the Psaltis group demonstrated fixing of thousands absorption process is desired. Nonresonant gratings require
of holograms [78]. energy fluxes on the order of joules per square centimeter
Holographic gratings based on charge redistribution in- and peak pulse intensities of gigawatts per square centimeter,
evitably decay due to ionic and electronic conduction. The whereas resonant gratings require the same energy fluxes
lifetimes of fixed ionic holograms are limited by the finite but peak pulses intensities of only 10 MW/cm [80]. Other
ionic conductivity at low (i.e., room) temperature [66], [73] work has focused on gated recording using rare earth doped
(Fig. 37). Ionic gratings are partially screened by trapped ferroelectrics [81], [82]
electrons upon readout (see Fig. 36), which decreases the re- The difficulty with the early work was the low sensitivity
sulting diffraction efficiency, but proportionally increases the of the medium and the high power of the lasers required to
grating lifetime. A significant increase in fixed ionic holo- record data in LiNbO . In the late 1990s, we have overcome
gram lifetime is realized in lithium niobate with low hy- this problem by modifying the composition of LiNbO
drogen impurity content. Fixed hologram lifetimes of about by making it more stoichiometric, adding Fe and other
two years at room temperature are projected in dehydrated dopants, combined with optimized oxidation and reduction
lithium niobate crystals. Due to partial electronic compen- postgrowth treatments [83], [84]. As a result, we have
sation, the actual hologram lifetime is prolonged but at the discovered an improvement of sensitivity of several orders
expense of reconstruction efficiency. Prolonged recording at of magnitude over previous results, making the medium
elevated temperature (in highly doped lithium niobate) usu- approximately equally sensitive as Fe-doped LiNbO in the
ally results in a nearly fully compensated gratings with little green. Typical sensitivity and index changes as a function of
or no electrooptic contribution, but with a sufficiently strong stoichiometry are shown in Fig. 39. It is interesting to note
index perturbation arising from the photochromic effect due that the sensitivity exhibits a threshold behavior near 49.5
to substantial modulation in the dopant concentrations. Such mol % Li O in the crystal.
gratings possess prolonged lifetimes (over 50 years or more) At a Li O concentration of 49.9%, the total accumulated
and do not require light for development. grating strength (the # or the sum of the induced refrac-
tive index changes of all superimposed holograms) in this
B. Two-Photon Recording material is very high, and we have measured values of the
10. This is very significant for data storage applica-
One of the difficulties encounted with thermally assisted tions, as it provides a large dynamic range for superimposing
ionic fixing for digital hologram recording is that recording holograms.

1252 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 39. Two-photon sensitivity and index change as a function of Fig. 41. Two-photon gated recording, nondestructive readout, and
stoichiometry in undoped LiNbO . erasure by gating beam in near stoichiometric lithium niobate.

Fig. 40. Mechanism of optically gated recording in lithium

niobate. Fig. 42. Schematic of the two-photon digital holographic storage
system. I: iris; L: imaging lens; RS: rotation stage.
The basic physical mechanism of two-photon gated
recording in LiNbO is shown in Fig. 40. Gating light ex- argon laser was used to illuminate a chrome-on-glass mask
cites electrons from deep traps (intrinsic or extrinsic) which containing digital data. Gating light from another argon
are temporarily trapped in the shallow levels (typically small laser at 476 nm is delivered to the crystal by an optical fiber.
polarons or shallow extrinsic dopants). The electrons from The mask pattern contained 512 512, 18 18 m square
shallow levels can be excited by red or near-IR recording random data bits (100% fill factor) and alignment crosshairs,
light, thus forming a holographic grating. When the gating but did not include any modulation codes or ECCs. An
light is switched off, the electronic population in shallow off-the-shelf multielement lens optimized for one-to-one
levels is transferred back to the deep traps, which do not imaging mapped the mask pixels onto the CCD pixels. The
absorb the longer wavelength recording light, thus making 18 18 m SLM pixels are centered on single 9 9 m
media insensitive, allowing nondestructive readout. square CCD pixels to reduce interpixel crosstalk and relax
The key in achieving efficient recording is in providing misalignment tolerances. The surrounding nonactive CCD
sufficiently long lifetimes ( 100 ms) for the excited elec- pixels act as a guard area that is equivalent to a smaller
trons in the shallow levels. This is achieved by using near-sto- fill factor, and no oversampling is used for postprocessing.
ichiometric lightly reduced lithium niobate with low concen- The crosshairs were used during the manual alignment for
tration of deep traps [83], [84]. Other approaches [88], [89] accurate adjustment of tilt and magnification. Closed-loop
rely of extrinsic dopants (e.g., an Fe, Mn pair) of optimum computer control of the CCD camera – position provided
concentration to provide pairs of deep and shallow levels fine alignment ( 1 mm) of the crosshairs before readout.
within the forbidden gap of lithium niobate, similar to in- A two-photon sensitive 1-mm-thick partially re-
trinsic defects described here. duced LiNbO crystal of near-stoichiometric Li O
The single photon sensitivity in the near IR is very small, mol composition [83] (doped with nonphotoactive
thus providing a means for prolonged readout, as shown in 0.2% Pr) was used in these experiments. The digital holo-
Fig. 41. When the gating beam is turned on again during grams of the mask were recorded at different wavelengths
readout, information is erased. in the transmission geometry by use of extraordinarily
The Stanford team [90] has implemented a digital polarized recording beams from a tunable laser and 100
recording system using two-photon illumination, as shown mW/cm of gating light (Fig. 42). The imperfect parallelism
in Fig. 42. A 1-W Ti : sapphire laser pumped by a multiline of the cut crystal caused distortions of the order of one

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1253

 Fig. 44. Transfer function of media with a local response.

the form of microdomains whose parameters (density and

average length) are correlated with the original electronic
space charge [100]. This approach is effective mostly for
the materials whose ferroelectric Curie temperature is close
Fig. 43. Intensity distribution of data bits contained in a 256-kb
hologram after the intensities are normalized to neighboring
to room temperature (e.g., Ba Sr Nb O , BaTiO ).
crosshairs. The high-intensity tail of the zeros and the low-intensity Using this technique, 1000 image bearing (but not dig-
tail of the ones are fitted to Gaussian functions. Their overlap is ital) holograms were angularly multiplexed and fixed in
then calculated to yield a best-threshold BER estimate.
Ce-doped SBN : 75 [101].
Another wavelength-dependent nondestructive readout
CCD pixel across the page, which were compensated for technique relies on the spectral response of a photorefrac-
by automated realignment of the camera on the crosshairs tive medium. Simple plane wave holograms written at one
while the hologram was scanned upon readout. For a stored wavelength may be read by Bragg-matched light of another
256-kb hologram, the estimated BER is found to be 8 10 wavelength, and if the medium is not photoactive at the
(Fig. 43), which is sufficient to guarantee a user BER of reading wavelength, repeated nondestructive readout is
10 when established modulation codes and ECCs are possible. For image bearing holograms, however, light at the
implemented. readout wavelength cannot in general be Bragg matched to
all the components of the signal. Reconstruction with spher-
C. Other Techniques for Nondestructive Readout ical wavefronts has shown improved results, and strategies
Several research results suggest that fixed hologram grat- to determine optimal reconstruction of wavefronts have been
ings may also be achieved using crystals with two species developed [102]–[105].
to establish and reveal a complementary grating with a life- In addition to fixing techniques which rely on the proper-
time longer than an initial principal grating; once a long-lived ties of materials, there are feedback techniques that refresh
complementary grating is established, shorter lived gratings and sustain holograms during their use. Using one technique,
in a second level may be exposed and erased without signif- holograms are periodically copied when their diffraction
icant erasure of the complementary grating. This technique efficiency falls below a certain level [106]. Holograms may
has been used in Bi TiO for real-time double-exposure also be circulated between two photorefractive crystals by
interferometry [91] and image subtraction [92]. For signif- transferring data from one to the other before the holograms
icantly long exposure, however, the complementary grating in the first are erased during readout, and then reversing
will decay, but can have a lifetime several orders of magni- the roles of the crystals [107]. Phased-locked sustainment
tude longer than the principal grating [93], [94]. Similar ap- may be achieved by generating a conjugate of the signal
plications have been demonstrated for semipermanent pho- and reference beams during readout, thereby refreshing the
tochromic gratings [95], [96]. KTa Nb O [97] has been hologram as it is read [108], [109]. By another technique,
shown to display fixed holograms for gratings established in phase-locked, binarized holograms may be dynamically
the centrosymmetric paraelectric phase at an elevated tem- written, read out, updated, and refreshed using either an
perature and then cooled through the phase transition to a optoelectronic or an all-optical feedback loop for spatial
ferroelectric phase. These may be due to the participation of thresholding in conjunction with a photorefractive crystal, as
two photoactive species during grating formation and the de- has been demonstrated in LiNbO [110] and BaTiO [111].
activation of one following the phase transition. Finally, extended readout may be achieved by illuminating a
Another mechanism to establish fixed hologram gratings storage material with a low-intensity reference beam during
related to complementary gratings relies on properties of readout and enhancing the output signal using an optical
phase transitions in ferroelectrics. Replication of photore- amplifier [112].
fractive space-charge density is achieved through sponta-
neous polarization modulation in strontium barium niobate D. Noise in Photorefractive and Other Media With a Local
(SBN : 75) [98] and barium titanate BaTiO [99]. Thus, Response
the hologram grating exists in the variations in the local Although LiNbO and other photorefractive media exhibit
spontaneous polarization of a noncentrosymmetic crystal in very low noise characteristics, as shown in the comparison

1254 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 45. The photovoltaic effect in single-crystal LiNbO produces large charge buildup resulting
in damage that severely distorts stored data.

Fig. 46. Periodic poling of LiNbO significantly reduces photovoltaic damage.

table in Fig. 31, in the absence of significant light expo- By alternating the direction of layers of ferroelectric
sure, noise builds up due to the dc sensitivity in media ex- domains, the dc damage effect is significantly reduced, as
hibiting a local response, as well as photographic film and shown in Fig. 46. Effectively, the transfer function has a
photopolymers. Here we discuss the case for LiNbO as it small value for low spatial frequencies, as the oppositely
has been well studied. In LiNbO , for example, as more and oriented domains prevent strong buildup of net charges
more holograms are recorded in the same location, a migra- [113].
tion of charges in the direction of the photovoltaic current
builds up local charge concentrations that lead to local index
VI. PHOTOPOLYMER MATERIALS
distortions in the medium. Aberrations in the readout and
the data beams then result in bit errors. After a few hundred Early photopolymers intended for holographic applica-
holograms are superimposed, this dc-noise effect can be- tions were developed primarily for display of single images
come significant and limits the further addition of holograms and were tolerant to a wide range of aberrations. For data
to the same location, thereby limiting the storage density. storage applications, however, more stringent requirements
This effect is frequently referred to as optical or photovoltaic are sought. Sensitivity, hologram shrinkage, and material
damage. Graphically the optical noise properties are shown stability are several of the most critical issues addressed in
in Fig. 44, where the sensitivity and index change are plotted HDSSs based on photopolymer media. Shrinkage is partic-
as a function of the grating vector, i.e., the transfer function. ularly important because it distorts the Bragg gratings that
In digital holographic storage, the usable dynamic range of record data holographically. Because most photopolymer
Fe : LiNbO as limited by optical damage is processes are not reversible, photopolymer systems are best
– (Siros Technologies, 1996). The solution to reducing suited for WORM applications.
this so-called photovoltaic damage is to reduce the dc re- In general, photopolymer storage media are complex sys-
sponse of the medium. One effective method in LiNbO is tems comprising several components and typically include
periodic poling of the ferroelectric domains. For example, the following: monomers and/or oligomers, a photoinitiation
in bulk LiNbO charges build up at the edges of the crystal subsystem, and a chemically inactive component. The pho-
when illuminated by a Gaussian beam, as shown in Fig. 45, toinitiation subsystem may further comprise a photosensi-
and this gives rise to distortions of the Air Force test target tizer that is specifically chosen to sensitize the polymer to
that is stored in the crystal (in this case a strong hologram is a specific light spectrum. Upon exposure to light, the pho-
recorded, which gives rise to similar effects as when super- toinitiation subsystem receives photons and triggers a reac-
imposing many weak holograms). tion that converts the monomers and/or oligomers to poly-

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1255

 Fig. 48. Generic mechanism of polymerization and reaction chain
propagation in Polaroid CROP photopolymer.

a gradient in the index of refraction due to three factors: a

change in the electronic structure of the monomer upon poly-
merization, called molar refraction changes; density changes
resulting from monomer diffusion into the regions of high
intensity; and direct refractive index changes that result due
Fig. 47. Graphical depiction of the evolution of the constituent to partial segregation of monomer and binder having signif-
photopolymer gratings, illustrating polymer density grating  , icantly different indices of refraction. Once an index grating
binder density grating  , and the net index of refraction n.
is established, any remaining monomer is polymerized to en-
sure that monomer and binder remain segregated.
mers. Polymerization generally results in local changes in Implementation of ULSH-500 in an HDSS is generally
the physical density, and the chemically inactive component, accomplished using the following procedure. Photopolymer
frequently called a binder, may diffuse in response to the media are deposited between two flat surfaces, typically
changes in density. Frequently, the binder has specific optical two glass substrates, which are separated by spacers. Pho-
properties that play a role in generating the index perturba- topolymer thickness is generally about 100–500 m; higher
tion responsible for the hologram; for example, the binder thicknesses can be achieved but generally require a decrease
may have a different index of refraction than the monomer. in sensitizer concentration, and, therefore, recording speed,
Photopolymer storage media may also comprise an internal to maintain usable optical density. Prior to polymerization,
polymer scaffold that serves to enhance the stability and re- the medium is a medium viscosity fluid. An initial preexpo-
duce the shrinkage of a material [16]. sure process, therefore, is typically required to establish a
We now focus on a specific example of a photopolymer photopolymer matrix, or scaffold, to provide the necessary
to illustrate the basic requirements for holographic data material stability prior to recording. The preexposure process
storage. Polaroid has developed a photopolymer system, may be accomplished by thermal treatment and/or flood ex-
ULSH-500, that has particularly good shrinkage properties posure to light. After preexposure the material is converted
[15]. This system uses specific monomers that polymerize into a gel-like state in which holograms can be recorded.
using a cationic ring-opening mechanism. The photoiniti- After the final hologram is recorded, a postexposure process
ation subsystem comprises a photosensitizer and an acid may be applied to polymerize any remaining monomer,
generator. The photosensitizer receives light used to record thereby improving the stability of the material. Single holo-
a hologram and transfers energy to an acid generator. The gram diffraction efficiencies can approach 100%; as with
resulting proton triggers the polymerization process, which inorganic materials, significantly less diffraction efficiency
proceeds as a monomer reacts with the acid to open a molec- per hologram is obtained for multiplexed holograms, as the
ular ring and bond to a neighboring polymer, generating an dynamic range is budgeted among the recorded holograms.
additional proton. This reaction proceeds until it reaches a Recording sensitivities can be about 3 cm/mJ for light at
termination impurity. A particular advantage of this pho- wavelength of 514.5 nm, and exposure fluencies of about 50
topolymer system is that opening a monomer ring partially mJ/cm2 can result in single-hologram diffraction efficiencies
compensates for shrinkage that typically occurs during the of about 80% for 100 m thick photopolymer media. The
polymerization process. high sensitivity of CROP photopolymer is merely because
During hologram recording, monomers are polymerized of the chain propagation effect (as shown in Figs. 48 and 49)
in the regions of high intensity, resulting in monomer gra- as well as the relatively high molecular weight of the stating
dients between the regions of high intensity and low inten- monomers [15].
sity. This is illustrated in Fig. 47, which shows the intensity A critical problem related to photopolymer media is
distribution and the resulting polymer density grating . shrinkage. As the medium is exposed to light, it changes
Monomers then diffuse into the regions of high intensity, pro- from a gel-like substance to a solid plastic. As a result, the
viding additional media for polymerization, and binder dif- density of the medium increases as the volume shrinks. As
fuses into the regions of low intensity, resulting in a binder a consequence, the grating spacing of previously recorded
density grating . The resulting polymer matrix displays data changes as more holograms are superimposed. This

1256 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 A variety of other photopolymer systems are under devel-
opment as candidate materials for HDSSs. Lucent, for ex-
ample, has developed a free-radical photopolymer system
that utilizes two independently polymerizable and compat-
ible chemical subsystems [115]. One subsystem defines a
polymer matrix scaffold in situ using low-index polymer pre-
cursors. The other subsystem records holograms using high-
index monomers, which remain intact after polymerization
of the first subsystem. The polymer scaffold enhances the
stability of the material and allows it to be shaped in a va-
riety of thick and flat formats. The scaffold is further de-
signed to avoid cross reactions with the monomer subsystem
Fig. 49. Mechanism of hologram recording in Polaroid CROP. that would reduce the refractive index contrast. The Lucent
PAG stands for photo acid generator.
team has reported implementing this type of photopolymer
system in a shift-multiplexing architecture to record about
leads to Bragg mismatch, as the grating wave vector is in- 4000 holograms in a configuration having a potential of about
creased due to shrinkage. As a result, the readout beam must 1 Gb of data, which is equivalent to about 50 Gbytes of user
be altered in order to satisfy the Bragg condition for efficient capacity in a 5.25-in disk [116]. These authors do not report
readout. In practice, either the readout wavelength or the any BER values, making evaluation of the storage capacity
readout beam angle is changed compared with the reference of these experiments somewhat difficult.
beam used for recording; usually both are changed [114].
For reconstruction of a plane wave, this can often be accom- VII. OPTICAL SYSTEM ARCHITECTURE ISSUES
plished, however, for data bearing object beams, all grating Design of HDSSs has been studied by many authors, from
vectors usually cannot be simultaneously Bragg matched, the early 1970s till recently. In early work [117], Vander
and as a result, only a portion of the data page is read out for Lugt, for example, examined the effects of optical constraints
a particular readout beam geometry. This, of course, leads to on the information storage capacity of block oriented page-
potentially large data errors and must be avoided. In practice, based memory systems. He did not take multiplexing of data
this can be achieved by optimizing the system for a certain into consideration in order to keep the SNR high and showed
amount of tolerable shrinkage in the medium. By trading that the packing density is limited by geometrical constraints
off recording sensitivity by preexposing the medium so it of the Fourier and imaging optics. The areal information den-
becomes closer to its final solid state, shrinkage is reduced, sity (raw) for holographic storage can be approximated as
but the available dynamic range is reduced commensu- a product of the individual hologram density (calculated in
rately as well, and fewer holograms can be recorded with [117] and [118]) times the number of superimposed holo-
lower sensitivity, requiring either longer exposure times grams
or higher recording powers. As CROP materials exhibit
lower shrinkage per unit volume for a given sensitivity than
free-radical media, they have practically proven to be the (51)
best performing media for holographic data storage. Their
optical properties are optimized for a particular holographic where is the size of the SLM, is the wavelength of light,
storage system by varying composition of the constitutive is the focal length of the imaging optics, and is a dimen-
photopolymer components. Overall system performance is sionless parameter which relates the actual hologram size to
optimized by carefully considering all system and media the Nyquist limit (typical values used in holographic storage
characteristics and determining the optimum capacity at a systems are within the range from 1.1 to 1.6, depending on
certain data transfer rate given limitation on cost and system the image quality requirements). The number of holograms
components. which can be superimposed is limited by a variety of pa-
A HDSS based on photopolymers is typically imple- rameters such as required transfer rate (i.e., photon budget),
mented using a rotating disk, as this provides a convenient material dynamic range (i.e., ), scatter, and information
format for using optomechanical systems developed for crosstalk. In order to achieve high storage density, it is, there-
other types of optical storage. Multiplexing techniques best fore, important to achieve high data density per hologram,
implemented in a disk-based architecture include rotational which according to (51) requires a high NA for the imaging
multiplexing (similar to shift multiplexing), which is based optics. The tradeoff between capacity and transfer rates is
on a reference beam with curved wavefronts [11], and estimated by computing the number of photons incident on
speckle multiplexing, which is based on a reference beam a detector element during page readout. From this calcula-
with a speckle pattern. Rotating the medium decorrelates the tion, we can estimate the maximum number of superimposed
reference beam with respect to the recorded hologram, thus pages readable with a certain SNR
enabling multiplexing by simply rotating the disk. This type QE
of photopolymer system has been implemented in HDSSs Number of pages
described in more detail below. (52)

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1257

Fig. 50. Tradeoff between capacity and transfer rate for the
parameters described in the text. The square dot denotes MO
recording at 2 Gb/in at 40 Mb/s. The round filled dot (just above
the “1000” marker) represents the Stanford/Siros Gb/s demo, while
the circle (above the endpoint of the solid line) denotes the planned
final HDSS demonstration performance.

where is a variable based on system and materials pa-

rameters, is the number of photons per watt per second,
QE is the quantum efficiency of the detector array, is
the optical efficiency of the readout system, is the fixing Fig. 51. Typical signal path in an HDSS.
efficiency, is the read time of the recalled data,
is the number of pixels in the data page, denotes the VIII. SIGNAL PROCESSING AND EN/DECODING
number of photoelectrons needed to get a 20-dB SNR, and
is the power of the readout beam. It is assumed that bit pat- In a data storage system, the goal of error correction
terns are balanced (although this does not have to be the case) coding, data modulation, and signal processing is to reduce
by making half the pixels in a page on and half off. From the BER to a sufficiently low level while achieving such
this equation, the tradeoff between capacity (the product of important figures of merit as high density and high data rate.
the stored number of pages, the code rate, and the number The distinct feature of digital holographic data storage is
of bits per page) and data transfer rate (the number of bits that this technology treats each individual pixel of the holo-
per page times the code rate divided by the readout time) is gram as a data channel and the entire image as set of data
determined in a photon-limited system. Optical damage and channels. This allows the employment of modulation coding
crosstalk limitations are not considered. In practice, however, and signal processing techniques to improve the system
optical damage often tends to limit capacity, but can be sig- performance in terms of total capacity at an acceptable error
nificantly reduced by recording at elevated temperatures [68] rate.
or by using periodically poled media. Equation (52) provides Coding and signal processing can involve several qualita-
an upper bound. tively distinct elements. The cycle of user data from input
As an example, assume that photoelec- to output can include error correction coding, interleaving,
trons to achieve a 20-dB SNR, the page size is modulation encoding, signal preprocessing, data storage in
, the number of photons per watt per second the holographic system, hologram retrieval, analog signal de-
is , the dynamic range is , the tection and binary digitization (using CCD or CMOS de-
camera frame rate is 1000 ft/s, i.e., ms, the de- tector array), signal postprocessing, channel decoding, and
tector quantum efficiency m, the laser power ECC decoding of deinterleaved binary data.
W, the fixing efficiency , and the op- A simple example of a signal path in an HDSS is shown
tical readout efficiency ; then the number of pages in Fig. 51. Input data are processed to generate data blocks
per stack is approximately 470, and the number of stacks for with error correction coding. The data are interleaved among
100-GB raw capacity is 2100. Thus, high data capacities and blocks to distribute burst errors among multiple blocks. Data
transfer rates require substantial spatial multiplexing. are subject to a modulation code that converts binary data into
However, capacities of several gigabytes could be specific modulation formats to be recorded in a holographic
achieved with no spatial multiplexing for data transfer rates material. On readout, data are demodulated and then deinter-
near a few megabytes per second. Using acoustooptic Bragg leaved. The error correction algorithm is applied to correct
cells, access times can be made less than 50 ms with no and/or identify error, generating the output data.
moving parts. At a transfer rate of 5 MB/s—typical for cur- The purpose of modulation coding is to simplify the detec-
rent magnetooptic drives—data density for HDSS is 50–70 tion of the analog signal values from the detector array. Mod-
times higher using a laser with two orders of magnitude more ulation codes generally restrict available patterns for format-
power. By varying the materials and systems parameters ting data to a subset that are less likely to be corrupted during
according to (52), the curve of Fig. 50 can be modified. In recording, storage, and/or readout. This restriction is typi-
particular, improvements in are effective in increasing cally achieved using constraints on the types of data patterns
storage capacity, an area of much current research. Further that may occur. A balanced constraint addresses the issue of
improvement is achieved by using signal processing and variation in signal strength across a data page; in a system
coding techniques. with binary channels, this issue is addressed by ensuring that

1258 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

a fixed number of OFF and ON pixels (zero and one, respec- raw BER have been made using differential encoding over
tively) occur in a block of a given size. A low pass constraint simple thresholding schemes [121]. Another advantage to
addresses intersymbol interference arising in high-frequency the differential encoding technique is that it can be used
data patterns, typically by eliminating the data patterns most when single pages reconstruct nonuniformly. This nonuni-
susceptible to intersymbol interference. formity can result from poor overlap of signal and reference
The simplest detection scheme is global threshold detec- beams in the crystal, from nonuniform illumination of the
tion, in which a specific threshold value is chosen: any CCD SLM, or from the introduction of interference fringes due
pixel with an intensity above the threshold is declared a one, to multiple reflections in the optical system. An important
while the rest are declared zero. disadvantage of this approach is a relatively high coding
If all holograms result in equal diffraction efficiencies, a overhead due to the constraint to use two pixels to encode
threshold intensity can be chosen to distinguish ON pixels a single data bit.
from OFF pixels. Note that we use OFF and ON when referring The code rate can be improved by increasing the array size
to SLM or CCD pixels and “zero” and “one” when referring and employing mathematically more sophisticated coding
to information bits. and decoding strategies [121]. More advanced balanced
For example, the probability of an error given that a zero is modulation codes provide local thresholding capability and
transmitted in a threshold detection system is given by [119], relatively high code rates (75% and higher). These codes
[120] assuming a model of additive white noise which is in- include balanced block codes [122] in which the fixed size
dependent from pixel to pixel pixel codeword, or data block, can contain only a specified
number of ON pixels. An example is the 6 : 8 code, in which
(53) a data pages is divided into 2 4 arrays, each containing
exactly four ON pixels and four OFF pixels. The number of
permutations for this constraint is 70, which can support 6 b.
where and is the average amplitude of OFF pixels at the Other modulation codes include parity threshold codes in
CCD array, is the standard deviation of the noise , is the which the number of ON and OFF pixels in a large uncoded
threshold value set to be the mean of the ON and OFF values, array is stored in a small coded (with highly redundant code)
or, in more general case of different statistics for ones and subarray, strip codes [122], and soft-decision iterative tur-
zeros, set to optimal value which minimizes the total BER. bocodes. In the case of block codes, the detection is based
The latter is achieved when the threshold is set to the point on a sorting or correlation technique, in which the detected
where the ones and zeros’ intensity distributions cross each block of pixels is compared to the set of valid code words
other. and is declared to be a specific code word based on max-
In the presence of variation in signal strength across a data imum likelihood criteria. Most advanced channel codes with
page, this simple scheme does not perform satisfactorily be- large coded array sizes also provide intrinsic error correc-
cause the optimal value of the threshold becomes dependent tion capability due to highly redundant coding, which allows
on the position within the page. A modulation code identified correction of a limited number of pixels directly during the
as an early alternative to the threshold approach in HDSSs channel decoding step. An example of a code with a large
is a balanced constraint known as differential decoding in coded array size is the 8 : 12 strip balanced array code [123].
which each zero and one of the data are encoded in form of Intersymbol interference can be introduced by a number
a group two adjacent pixels a and b, one ON, one OFF [24]. of factors, such as the optical bandpass of an HDSS, mis-
Upon detection and demodulation, if the intensity of pixel a registration of the imaged signal page and detector array,
is lowest, the data bit is declared a zero, and if the intensity crosstalk between pixels in an SLM, and electrical crosstalk
of pixel b is lowest, the data bit is declared a one. Viewed a between detected signal pixels. Optical misregistration can
different way, a zero–one sequence represents a data bit zero be addressed using sophisticated interpolation algorithms
and a one–zero sequence represents a data bit one. based on remapping the received data pages [124] or serial
The probability of an error in a differential detection correction processing techniques [125]. Other techniques
system with additive white noise is given by for reducing the negative effects of intersymbol interference
in page detection include Vitterbi algorithms [126], image
deconvolution upon detection [127], and equalization via
predistortion and inverse filtering during the recording [128].
Another approach is to use passband codes that eliminate
certain data patterns responsible for the highest incidence of
(54)
data error due to intersymbol interference, such an OFF pixel
surrounded by all ON nearest neighbors [122].
where represents the average amplitue of the ON pixels. The ECC adds redundancy to the data in order to pro-
Most notably, the error rate is independent of the choice vide protection from various noise sources. The ECC-en-
of the threshold value, thus making differential encoding coded data are then interleaved and passed on to a mod-
very robust and very easy to implement. This technique ulation encoder, then input to the SLM and stored in the
largely solves the problems associated with nonuniformity recording medium. On the retrieving side, the CCD returns
of holographic data pages. Significant improvements in pseudoanalog digitized data values, which then undergo a

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1259

Fig. 52. HDSS electronics operation.

postprocessing step (e.g., deconvolution in order to compen- Fig. 53. Image SNR distribution.
sate for the smearing of the optical signals due to image dis-
tortion and diffraction point spread function). The array of
the processed digital data is further passed to the modula- The decoder boards provide 6 : 8 channel decoding. A
tion decoder, which performs the inverse operation to modu- simple 6 : 8 code (i.e., 8 pixels correspond to 6 b of data,
lation encoding, then to the deinterleaver, and, finally, to the including ECC bits) was chosen for the HDSS holographic
ECC decoder. The purpose of data interleaving is in reduc- disk system because it provides a fairly high (75%) code rate
tion of the corrected BER (i.e., after the ECC encoding) by and has sufficiently low complexity to be implementable
making the digital data error rate before ECC encoding uni- in hardware. More advanced codes (e.g., 8 : 12 strip code)
form. Even with balanced modulation code, the BER may provide somewhat better performance [123] with respect
not be uniform across the page. Thus, by corrupting different to BER and storage capacity, but are significantly more
parts of the same ECC block in different locations in the page, complex for hardware implementation.
uniform BER distribution between different ECC blocks can The purpose of data interleaving is in reduction of the cor-
be achieved, thereby minimizing the total BER after the ECC rected (i.e., after ECC) BER. The SNR of the source images
decoding. is usually not uniform across the different image sections. By
As research in HDSSs progresses, more sophisticated ap- distributing data in different sections of the image and dein-
proaches, including algorithms implemented in one or a few terleaving the decoded data, one can achieve uniform BERs
chip sets, can be incorporated based on advances in data in all the ECC sequences before error correction operation,
storage and communications industries. thus reducing the error rate of the user data. Areas of high
BER SNR are usually excluded (masked) from the
IX. ELECTRONIC IMPLEMENTATION page. The masking operation is also performed by the demul-
tiplexer electronics. The deinterleaving and masking look-up
The purpose of the specialized processing electronics is tables are generated by software, based on the hologram SNR
to provide the holographic channel decoding, deinterleaving, distribution (such as the one shown in Fig. 53) and are down-
ECC decoding, and overall system synchronization, all at loaded to the electronics during the electronics initialization
high data rates (1 Gb/s). This cannot be accomplished (with step.
existing computers) in software due to both processing speed During hologram recording and readout, the synchroniza-
and timing limitations. The functionality of channel elec- tion electronics provide timed pulses to the SLM, the CCD
tronics is best described on the basis of the first ever hard- camera, and the laser according to the address table, which
ware implemented system developed by Siros Technologies is downloaded before the beginning of high-speed operation.
and Stanford for HDSS consortium. The HDSS decoding electronics are triggered automatically
The HDSS electronics are based on field-programmable by the CCD camera after each camera exposure.
gate array (FPGA) technology and can be reprogrammed
using software and hardware means. HDSS channel-de-
coding electronics consists of three main processing blocks: X. DEMONSTRATION PLATFORMS
holographic channel decoders, a demultiplexer, and a
Reed–Solomon ECC board. Basic processing steps of the A number of systems, including the first digital HDSS,
electronics are shown in Fig. 52. were designed, built, and tested at Stanford University and
Pixel deconvolution is performed on the raw 8-b pixel Siros Technologies for a variety of purposes. They were built
values. The 3 3 kernels are computed in software and under sponsorship of the Center of Nonlinear Optical Ma-
provide pixel corrections to reduce the effect of optical terials and the PRISM and HDSS consortia. The primary
point-spread function and distortions of the imaging system, goals were to gain insight into the underlying physics of the
thus increasing the SNR of the input image. recording mechanisms and into the system tradeoffs required

1260 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 54. Stanford/Siros demonstration platforms.

to achieve specified performance, and to integrate new com-

ponents into working demonstration platforms. The most sig-
nificant demonstrations are tabulated in Figs. 54 and 55 and
are briefly described in ensuing sections.

A. Stanford University All-Digital System Demonstration

(1994)
In 1994, we demonstrated the first fully digital HDSS [24]
by storing digital images and a short soundtrack. The purpose
of this experiment was to show that a fully functional system
could be implemented with off-the-shelf components. Fig. 55. Siros and Stanford University demonstration platforms in
Particular emphasis was placed on using digital signal historical order. The laser power for all experiments was 400 mW,
except for the April 1996, experiment which used a 250-mW laser.
processing techniques to overcome noise issues limiting ca-
pacity and transfer rates. A new differential coding scheme
was introduced that allowed us to significantly increase The noise-tolerant nature of digital storage makes it
capacity by adding extra bits for channel and ECC coding. It possible to overcome problems associated with the afore-
allowed us to evaluate the tradeoffs between storage system mentioned noise sources. However, prior to the publication
capacity, BER, and data transfer rate for several channel and of [24], no fully automated digital HDSS had been imple-
ECC coding schemes. mented and no comprehensive study of the bit-error-rate
The capacity is limited for the most part by the attain- performance of a system operating at reasonable data transfer
able diffraction efficiency of the stored holograms. As the rates had been carried out. Achievable BERs have been pro-
number of holograms stored in a single stack increases, the jected previously by sampling a small number of digital
diffraction efficiency falls as , mentioned above. As the information bits from a random sampling of 1000 holograms
strength of the diffracted signal decreases, the SNR decreases stored in a manually Bragg-tuned system [129]. Using
because the strength of noise due to scatter and CCD noise is various encoding techniques, we have implemented what
independent of . The total number of holograms that can be to the best of our knowledge was the first fully automated
stored is, thus, determined by the minimum acceptable SNR. system in which data are written and recalled in digital form.
A HDSS is also limited by page-to-page fluctuations in signal We have used this spatio-rotational multiplexing system to
strength, which can occur due to laser instability or, for ex- store digital representations of color images and compressed
ample, due to a nonoptimal choice of the recording schedule. video. Critical features of our implementation that allowed
Intrapage distortion can occur due to imperfections in the op- us to overcome previous obstacles include location of the
tical system or in the SLM or CCD array. These noise sources Fourier plane outside the crystal volume, use of cylindrical
limit the applicability of direct storage of images in analog lenses to implement spatial multiplexing, a novel differential
form. The majority of previous holographic storage imple- encoding technique used to increase error immunity, use of
mentations have involved direct storage and retrieval of pic- well-known Hamming error correction codes, and the dis-
torial information. For example, Mok has demonstrated an- tribution of consecutive information bits over multiple data
gular-multiplexed storage of as many as 5000 edge-enhanced pages in order to decrease the probability of burst errors.
analog images in a single crystal [8]. With these techniques, BERs of 10 have been achieved at

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1261

 To address these issues, we used the differential encoding
technique described above, in which the pixel sequence
OFF–ON represents a data bit zero and the pixel sequence
ON–OFF represents a data bit one. In addition to being insen-
sitive to page-to-page intensity fluctuations, the differential
encoding technique results in a lower probability of error
assuming a model of additive (in amplitude) white noise
which is independent from pixel to pixel as described above.
Our experiments showed for the first time that significant
improvements can be made using differential encoding over
simple thresholding schemes [24]. Another advantage to the
Fig. 56. Generic perpendicular (90 ) recording geometry.
differential encoding technique is that it can be used when
single pages reconstruct nonuniformly. This nonuniformity
readout rates of 6.3 10 pixels per second. We have used can result from poor overlap of signal and reference beams
the system to evaluate tradeoffs between BER and storage in the crystal, nonuniform illumination of the SLM, or the
capacity at a fixed data transfer rate. introduction of interference fringes due to multiple reflec-
A generic diagram of our system is shown in Fig. 56. tions in the optical system. When a substantial contribution
The storage medium is an Fe-doped LiNbO crystal cut to the variance is due to such large scale nonuniformities,
such that its axis was at 45 to the crystal surfaces. The we expect difference encoding to offer further improvement
crystal dimensions are 2 cm 1 cm 1 cm, of which only a over simple thresholding techniques.
0.1-cm portion was used for data storage. The crystal was Ideally, each pixel on the SLM would be used to represent
mounted on a computer-controlled rotation stage, which was a single bit of data. The liquid crystal array used in these
in turn mounted on a computer-controlled translation stage, experiments exhibited some interpixel crosstalk. In other
allowing the crystal to be moved in the vertical direction. The words, pixels adjacent to an ON (transmitting) pixel exhibit
SLM was a 480 440 pixel liquid crystal array taken from a significant amount of transmission. Also, the SLM and
an InFocus TVT-6000 video projector. It was addressed with the CCD array were manufactured with different horizontal
an analog video signal produced by a framegrabber board to vertical pixel pitch ratios. Thus, using a simple optical
in the computer. The camera was an intensified CCD array. system such as ours, it was impossible to image the SLM
Each stored hologram was read out in 1/30 s. The cylindrical onto the CCD array such that each SLM pixel is imaged
lenses in the reference beam path were used to collimate directly onto a CCD pixel. In practice, a multiple lens ar-
the reference to an area of approximately 10 mm 2 mm rangement could be used to achieve the proper anamorphic
at the face of the crystal. The combination of the translation imaging; however, this greatly increases system complexity.
stage and the collimation of the reference beam allowed The pitch mismatch resulted in a systematic geometrical
holograms to be stored in four different stacks. The signal source of errors if too fine a sampling grid is used. In order
beam occupied an area of approximately 1.5 mm 1.5 to circumvent these problems, we used a block of 8 8
mm on the front face of the crystal. The Fourier plane was pixels to represent 1 b. The differential encoding technique
located approximately 3 mm in front of the crystal. A filter increases the pixel-to-bit ratio to 8 16 pixels per informa-
was used to select only the central spot of this transform. The tion bit. The data is read out by sampling the CCD output
filter prevented erasure of previously recorded stacks during at one pixel for each 8 8 pixel block imaged onto it. The
writing and eased alignment tolerances at the CCD plane. By difference in intensity between adjacent samples is used to
recording in a Fresnel plane, rather than the Fourier plane, determine whether the information bit is a zero or a one.
the modulation depth was sufficiently uniform to eliminate We measured the raw BER of our system to be between
the need for a diffuser [130]. Because previously recorded 10 and 10 at video readout rates. In order to improve
holograms were erased in this demonstration as additional performance, we implemented a Hamming error correcting
holograms were recorded, an appropriate recording schedule code in which four check bits were added to each string of
had to be determined in order to store a large number of eight data bits. The Hamming code is capable of correcting a
pages with equal diffraction efficiency [2]. single bit error in the sequence of 12 b assuming that only one
If all holograms result in equal diffraction efficiencies, a bit error occurs in those 12. In order to reduce the potential of
threshold intensity can be chosen to distinguish ON pixels burst errors, each stored data page represented one bit plane
from OFF pixels. Note that we use OFF and ON when refer- in a two-byte sequence. Thus, each bit in the 12-b sequences
ring to SLM or CCD pixels and “zero” and “one” when re- was stored in a separate page in order to reduce the prob-
ferring to information bits. An incorrect measurement of the ability of multiple errors within a given sequence. With the
time constant can result in a schedule which leads to unequal error-correcting codes, we were able to achieve BERs of 10 ,
diffraction efficiencies for each of the stored holograms. A which is adequate for compressed video storage and more
10% error in the assumed time constant can result in more than sufficient for uncompressed image storage. We used our
than a factor of two variation in diffraction efficiency. In ad- system to store both digital color images and low-resolu-
dition, laser fluctuations or anomalous writing behavior can tion video data. The total capacity of the system was 1232
result in unequal diffraction efficiencies. pages with 1592 b/page, resulting in a raw capacity of 245

1262 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

kB. Taking into account the necessary error-correcting bits,
the total useful data capacity was 163 kB. The total pixel ca-
pacity of our system was 2.6 10 pixels, with a density of
3 10 pixels per cm . The transfer rate is 6.3 10 pixels
per second. The CCD camera outputs 1 B per pixel. The total
pixel capacity of our system indicates that our storage ca-
pacity does not represent a fundamental limit. Rather, the
primary limitation in determining the information storage ca-
pacity is the necessary oversampling on the SLM. With a
higher quality SLM designed in conjunction with the CCD
array, we later showed that the information capacity and data
Fig. 57. Electronic block diagram for the Siros real-time video
transfer rates can be enhanced by many orders of magnitude. readout system.

B. Siros Fully Automated Video Demonstration (1995)

The Stanford demonstration showed that digital encoding
techniques could be used to great advantage to improve raw
BER, and in combination with error correcting codes dig-
ital images could be retrieved with low error rates. The ca-
pacity of this demonstration, however, was rather limited.
The next step in the development of digital HDSSs accom-
plished storage of compressed high-resolution video data, re-
quiring substantially larger data storage capacity, and a good
BER of about 10 . Primarily due to the efforts of J. Heanue,
A. Daiber, R. Snyder, and J. Colvin at Siros Technologies, the
first video movie was stored and retrieved by recording ap-
proximately 5 MB of compressed video data . The optical ar-
rangement was similar to the Stanford demonstration, except Fig. 58. Siros Technologies compact holographic system.
that the SLM was a Texas Instruments DMD, with a resolu-
tion of 640 480 pixels. This device overcame all problems
effort, a fully electronic readout and control system was im-
of the previously used liquid crystal devices, and allowed us
plemented in 1996. We believe that this was the first demon-
to store the first digital video in a holographic system.
stration of a fully automated and electronically controlled
Both video imagery and sound were recorded in LiNbO
system. The block diagram of the electronic layout is shown
with the MPEG standard using the system of Fig. 56. Data
in Fig. 57. The electronic architecture is based on a virtual
were encoded using 6 : 8 channel codes and a Reed–Solomon
machine interface (VME) bus. The optical system architec-
15 : 13 error correction scheme that corrects one byte error
ture was the same, as shown in Fig. 56. Temperature fixing
per code word. This system was shown on the Discovery
in LiNbO was implemented for nonvolatile readout, and a
Channel in 1995 as part of the Next Step program.
total capacity of 5 MB of video data were stored and retrieved
In contrast to the Stanford demonstration, in this exper-
at video rates. Channel 6 : 8 coding, ECC Reed–Solomon
iment the information was permanently stored using ionic
13 : 15 coding, bit shuffling, and data warping were all im-
fixing. After recording of all data stacks, the crystal was
plemented in electronic hardware, as well as overall system
removed from the set up and placed in an oven. It was heated
control [3]
to approximately 120 C for 15 min, and then slowly cooled
A small compact demonstration system using LiNbO was
down to room temperature. The thermal fixing efficiency
built primarily by M. McDonald and L. Hesselink at Siros to
was rather high, approximately 50% (i.e., the diffraction
show that all components could be integrated into an approx-
efficiency after thermal fixing was approximately 50% of
imately 5.25-in form factor depicted in Fig. 58. This system
that before the fixing procedure) [77]. Data retrieval in this
includes both recording and readout optics and components.
experiment was implemented in software.
Associative data retrieval capabilities were added to
the Siros 90 demonstration platforms, and content based
C. Siros Fully Automated System With Electronic Readout searches were demonstrated at 20-Gb/s burst data rates by
at Video Rates (PRISM, 1996) R. McLeod in 1997.
All previous demonstrations were under software control
with offline data retrieval. Although it is well recognized that D. Stanford University and Siros Fully Electronic Data
software procedures can often be implemented in electronic Readout System Achieving 1 Gb/s (HDSS, 1999)
hardware, there are many practical issues that make such an In previous demonstrations at Stanford, Siros, and
implementation far from trivial. At Siros, primarily through IBM, the total system capacity was limited to tens of
the efforts of B. Okas and R. Snyder in cooperation with megabytes, with transfer rates on the order of a few
GTE, Mountain View, CA, as part of the PRISM consortium megabytes per second, far from being competitive with the

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1263

 E. Stanford University and Siros 1-Gb/s Holographic Disk
System Demonstrator (2000)
For thick media like LiNbO in the 90 geometry, capacity
is usually limited by media dynamic range and noise rather
than by multiplexing crosstalk. For thin media, such as pho-
topolymers, however, this is not the case, and the number of
superimposed holograms in a spatial location is largely de-
termined by the limited number of degrees of freedom avail-
able for multiplexing. Angular multiplexing in the transmis-
sion geometry does not allow sufficient data storage den-
sity, and other multiplexing techniques such as shift or peri-
strophic multiplexing are required. Unfortunately, the dis-
Fig. 59. Electronic block diagram for the 1 Gb/s electronics.
placements necessary for achieving low crosstalk in the shift
multiplexing technique are still too large and require the un-
then state-of-the-art optical and magnetic drives. The HDSS desirable stop-and-go architecture for the rotating disk when
consortium was formed to develop new components and the using a CW laser. To avoid this problem, a very sensitive, rel-
systems to demonstrate 1-Gb/s transfer rate and a capacity of atively thick storage medium is required, suitable for pulsed
125 GB on a 5.25-in disk. A reflective 1000 1000 element laser recording with nanosecond pulses. Developments in the
ferroelectric liquid crystal device was designed and built by DARPA-funded PRISM consortium have made great steps
IBM and Display Tech capable of recording 1000 frames forward toward achieving such a medium for recording in the
per second. A pixel matched detector array was developed green region of the spectrum using a few hundred milliwatts
by Kodak, having 1000 1000 pixels, and a frame rate of of power. To achieve recording during constant rotation of the
1000 ft/s using 8-b resolution. The output of this camera disk, a new multiplexing technique had to be implemented
was divided into 64 digital channels. A pixel matching very based on a phase-modulated reference beam [12]. In such an
low distortion Fresnel lens was designed and manufactured architecture, performance can be readily analyzed and com-
by Rochester Photonics to achieve less than 0.3 pixel el- pared with more conventional technologies. In the system at
ement distortion over the full 1 million pixels in the field Stanford University, densities of over 70 b m are reach-
of view. A novel random phase code reference beam was able for a total system capacity of 125 GB and a transfer rate
coaxially placed with the imaging lens. These components of 1 Gb/s using the electronics described above. The overall
were assembled in the optical set up of Fig. 1. In the first architecture for this system is shown in Fig. 60.
high-speed demonstration, a 90 LiNbO arrangement was A 1024 1024 pixel matched (12.8 m square) 1000 ft/s
implemented using the electronic architecture of Fig. 59. A Kodak C7 and IBM LC display are used as a detector and
6 : 8 channel code and a Reed–Solomon 136 : 156 ECC for page composer, respectively. Recording and readout is done
correcting up to ten byte errors in the sequence were used using a pulsed doubled Nd : YAG laser (532-nm wavelength).
for this demonstration. A rotating holographic photopolymer disk is mounted on a
A team at Stanford University and Siros consisting pri- precision air-bearing spindle. Angular addressing is acheived
marily of R. Okas, S. Orlov, E. Bjornson, X. Li, F. Dimov, by using a precision optical shaft encoder (16 384 counts
B. Phillips, R. Snyder, D. Kwan, Y. Takashima, and L. Hes- per revolution). Different radial positions are addressed via
selink developed the first implementation of a 1-Gb/s elec- moving the spindle with the mounted disk using a 25-nm-res-
tronic readout system. Six data files were recorded, five JPEG olution Newport PM600 translation stage. The optical layout
images and one counting sequence. The counting sequence for the HDSS WORM demonstration platform is shown in
was used so that the data could be easily inspected in nu- Fig. 61. A sophisticated hardware synchronization is imple-
merical format. Each source file was divided into two pieces, mented allowing addressing each individual radial location
giving a total of 12 data sequences, corresponding to the 12 with accuracy of better than 0.1 ms, thus allowing (if nec-
output byte streams. essary) incremental hologram recording with multiple pulses
Each recorded hologram stored data at a 55% efficiency of the laser.
due to 6 : 8 encoding, RS encoding, and masking. Thus, each The demonstration platform employs a custom-built (by
image of 1 048 576 pixels corresponds to 576 000 b. When Coastal Optical Systems, Inc., West Palm Beach, FL) short
this is divided into 12 streams, it gives 6 kB per image per focal length (17.1 mm) optical relay system [131] which al-
output stream. Thus, our source data files, each of approxi- lows 1-megapixel images to be relayed from the SLM to the
mately 48 kB (split into two 24-kB sections) required four CCD array with distortions of less than 1.5 m over the
images to store. The data were stored as four holograms and entire SLM format (13.1 mm 13.1 mm). The imaging NA
retrieved with an integration time of 600 s, consistent with for this system is rather high (0.36) while the outside area of
a 1-ms frame rate. Data passed through the entire system, the lens is used by the reference beam. (Fig. 61). The spot
and the 12 parallel byte streams were collected using the full size in the photopolymer media is 2 mm , which corre-
electronics and checked for accuracy. The raw error rate was sponds to a raw areal density per hologram of 0.5 b m .
8.7 10 . In the current implementation, the speckle correlation shift

1264 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 60. Overall system layout for the HDSS WORM demonstration platform.

The HDSS holographic disk channel decoding electronics

were designed for a sustained data rate of 1 Gb/s. In the first
demonstration (in October 1999), JPEG-encoded color im-
ages were stored on the holographic disk and retrieved at 1
Gb/s [14], [18]. The source data of 12 images were ECC
encoded and converted into holographic channel encoded
bitmaps, which were stored in a 200- m-thick Polaroid pho-
topolymer disk media. All the retrieved (at 1 Gb/s) data were
Fig. 61. Optical layout for the HDSS WORM demonstration captured by a logic analyzer and converted back into JPEG
platform. format files. Smaples of the original and retrieved images
are shown in Fig. 64. The storage density in the first experi-
ment was rather low (approximately 1 b m ) and individual
holographic page capacity was 65.3 kB (user).
In the later demonstration [132] (Fig. 65), 12 uncom-
pressed video streams were holographic channel encoded,
stored in the holographic disk, retrieved at 0.65 Gb/s, and
displayed in real time on three computer screens (four
streams per monitor). The total amount of data stored and
retrieved was approximately 1 GB (user) while only 10
of the area of a 6.5-in-diameter photopolymer disk was used.
The media thickness was 200 m.
Fig. 62. Photograph of the 1-Gb/s holographic disk WORM
demonstrator. F. Stanford Ultrahigh 10-Gb/s Optical Data Rate
Demonstration (2000)
In the holographic disk system at 1 Gb/s, the hologram
signal strength largely exceeds the camera noise. Thus, the
holograms can be read at much higher optical data rates pro-
vided that the holograms can be physically transported under
the optical readout head at sufficient and sustainable rates.
The latter depends on the recording density and the disk
linear velocity.
In later experiments, we achieved sustained optical data
rates as high as 10 Gb/s [17] by increasing the optical effi-
ciency of the optical delivery system. The storage density in
2
Fig. 63. Sample 1024 1024 data page (  2 10 , right) and
2
enlarged 48 48 portion of it (left). these experiments was approximately 10 b m . A sample
hologram read out at 10 Gb/s is shown in Fig. 66.
multiplexing technique was used for hologram multiplexing
(Fig. 62). G. Tamarack Multistore Jukebox
A sample multiplexed hologram recorded in rotating disk In the mid–1980s, MCC, in collaboration with Stanford
media and retrieved with a single laser pulse at 1000 ft/s is University, started a holographic data storage program in re-
shown in Fig. 63. The raw byte error rate is 2 10 , sponse to Japan’s Fifth-Generation Computer Program. The
which provides error-free reconstruction of data with a objective of the program was to address the issue of data
136 : 156 Reed–Solomon ECC. access latency caused by rotating disk drives. As processor

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1265

Fig. 64. Retrieved (left) and original (right) images.

Fig. 65. Twelve uncompressed video streams display.

speeds advance, the I/O gap—the latency of getting access containing 30 disks and a 5.25-in form factor drive. An
to data once a command for it has been issued—increases. excellent mechanical system was developed, but unfortu-
HDSSs could potentially solve this problem as they do not re- nately its holographic performance target of 30 GB was not
quire moving parts and, therefore, can provide much shorter demonstrated and thought to be too small to be attractive in
access times, on the order of microseconds. the mid–1990s. A lack of a suitable material ultimately lead
The MCC program culminated in the formation of Tama- to the demise of Tamarack. A photograph of the advanced
rack storage systems, a spinoff from MCC formed in 1992 Tamarack holographic optical head assemby is shown in
[23]. Tamarack targeted the development of a disk jukebox Fig. 67.

1266 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 J. Caltech and Holoplex Read-Only Memories and
Replication System
Digitial holographic disk systems based on polymer ma-
terials were also demonstrated by Holoplex, Pasadena, CA,
and at Caltech in the late 1990s [136]. The demonstration
system employed 100- m-thick Dupont photopolymer and
a combination of angular and peristrophic multiplexing.
The recording step was performed using a CW laser and
stationary disk. The reported density was approximately
40 b m while the channel decoding was performed by
software. The special emphasis of this work was on devel-
Fig. 66. Sample hologram reconstructed at 10 Gb/s. opment and demonstration of holographic disk replication
technology, where the data from a prerecorded master disk
was transferred to a replica disk via simultaneous readout of
H. IBM Holographic Material and System Testers holograms from the master and recording onto the replica.
Schematically, the setup is shown in Fig. 73 and a photo-
A research team at IBM Almaden, San Jose, CA, has built graph of a ten-beam replicator is shown in Fig. 74. Prior
three holographic digital data storage test platforms [133]. to demonstrating digital holographic systems, much pio-
The PRISM tester [57] shown in Fig. 68 was designed and neering work was carried out at Caltech relating to system
built for the purpose of having a stable platform for testing architecture issues, as described above.
holographic data storage materials. It was completed in 1996
and has provided an important means for testing a wide va- K. Optware Colinear Holography System
riety of materials ranging from photopolymers to inorganic Optware Corporation, Yokohama, Japan, is developing
photorefractive crystals. Some of the results of these mea- a colinear holographic system, as shown in Fig. 75. The
surements are shown in Fig. 32. Two additional demonstra- system involves a two-wavelength approach in which a red
tion systems were also built at IBM, the DEMON I and II in laser is used for servoing and a green laser for holographic
1997, shown in Figs. 69 and 70. storage. The green 532–nm laser is used for holographic
These platforms were used to study coding approaches reading and writing, while the red laser is used for tracking
[123] and to study the performance of holographic sys- and focusing. A dichhroic beam splitter combines the two
tems under various conditions. For example, IBM also beams onto the holographic medium. The Texas Instruments
stored videos similar to the earlier Siros demonstrations DMD is used as a page composer. The CMOS detector has
and achieved very high data storage densities in DEMON II a 3 3 oversampling ratio for each data pixel. The object
using LiNbO as a recording medium [30]. and reference beam are colinear at the medium, and both
Important optical approaches were developed and tested reflect from the lower surface of the recording medium. Po-
for achieving flattop intensity profiles for the image bearing larization optics are used to efficiently utilize available laser
signal beam [134] as well as methods for reducing the central power. The size of the holograms on the bottom surface of
spike in the object beam at the Fourier transform plane [133]. the photopolymer medium is approximately 200 m, which
Additionally, associative retrieval methods were developed is typically 400–500 m thick. Reported data densities are
which proved the superior performance of holographic-based approximately 40–50 b m .
searches over traditional digital computer searches, as de- The interesting aspect of this approach is the symmetrical
scribed above. Bragg selectivity in both the track and radial direction
affording close packing of holograms, which is not easily
achievable using an off-axis reference beam geometry. Since
I. Rockwell Ultrashort Access Time Testbed the object and reference beam pass through the medium
twice, signals are enhanced, but noise levels build up as well.
Ultrashort data access times were achieved by a team of re- The key research and development issue for this approach is
searchers at Rockwell International Research Center, Thou- to reduce noise through reduction of low-frequency media
sand Oaks, CA, as part of the HDSS program. Using acous- response and improvement in signal strength by increasing
tooptic beam deflectors for the reference beam, access times the thickness and reducing the inherent media scatter noise
of less than 50 s were demonstrated in LiNbO in the 90 while maintaining a low shrinkage rate. The reported media
configuration, as shown in Figs. 71 and 72 [5], [135]. Short for this development effort are produced by Aprilis, Inc.,
access times combined with associative data retrieval could Maynard, MA, and Inphase, Inc., Longmont, CO.
provide an interesting market opportunity, as other storage
devices based on rotating disks have access times of a few L. Aprilis High-Density CROP Disk Testbed
milliseconds and cannot contain as much data under the head. Aprilis, Inc., a spinoff company from Polaroid, has been
In particular, for searching video databases, such capability at the forefront of materials development based on CROPs.
could be valuable. Recently [137], significant advances have been reported

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1267

Fig. 67. Photograph of the Tamarack optical head system (with permission).

is shown in Fig. 76, where the histogram of bits is shown

for superposition of 155 holograms in Aprilis HMD CROP
media. The growth of the cumulative grating strength and
reduction in sensitivity as a function of total accumulated
exposure is shown in Fig. 77. In a recent experiment, as yet
unpublished, Aprilis has achieved 250-GB storage capacity
on a single disk, indicating that the Aprilis HMD CROP
material is currently the best available (based on published
data) medium because of its high sensitivity, large dynamic
range, and low shrinkage.

M. InPhase Technologies Demonstration Platform

InPhase Technologies has been developing photopolymer
Fig. 68. The primary features of the IBM PRISM holograhic media based on technology originally conceived and
materials test apparatus. The SLM is a chrome-on-glass mask, and practiced at Bell Laboratories—a two-chemistry medium
2
the detector array is a 1024 1024 of a large CCD camera. A composed of two independently polymerizable and com-
pair of precision rotation stages allow the reference beam to enter
the storage medium under test at any horizontal incidence angle patible chemical systems consisting of low-refractive index
(courtesy IBM, with permission). matrix precursors and high-refractive index polymerizable
monomers [138].
with improved materials that are capable of supporting The two-chemistry system allows in situ formation of the
digital data densities of 150 b m in 400- m-thick media, background matrix and optimization of various recording
corresponding to approximately 150 GB on a 5.25-in disk. chemistries to optimize dynamic range, sensitivity and
The raw BER for 262-kb data pages was approximately shrinkage. s of 6.7 have been reported, which compares
5.5 10 , which can be significantly reduced to better than favorably with LiNbO , which has a similar dynamic range,
10 with the use of error correction codes as described but lower sensitivity of one to two orders of magnitude. The
earlier. The dynamic range of the medium was large with an medium shrinks, however, by approximately 0.1%, reducing
of approximately 22. The media sensitivity is high with the usable dynamic range for digital storage as compared
a peak initial value of 6.75 cm/mJ reducing to 0.5 cm/mJ with LiNbO which is a solid and very dimensionally stable.
for attaining 87% of the available dynamic range or accu- Recent advances by InPhase indicate that shelf life can ex-
mulative grating strength. Approximately 155 holograms ceed years, and holograms retain their strength and fidelity
are recorded in the same spatial location using angular and for at least 1000 h of exposure. Media thickness can report-
peristrophic [9] multiplexing. Peristrophic multiplexing is edly also be increased to at least 1.5 mm, but no test data
a method for recording holograms in the same physical have been made publicly available. For thicker photopoly-
location by rotating the reference beam with respect to the mers, shrinkage is a serious problem, and its management
recording medium as well as by rotating the medium about requires tradeoffs between sensitivity, dynamic range, and
an axis normal to the surface of the medium. A typical result shrinkage. Theoretically, the or dynamic range should

1268 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 69. Salient features of the DEMON I holographic digital data storage engine. A five-element
zoom lens demagnifies the SLM image to an intermediate image plane, which is then imaged onto
the CCD detector array with a pair of lenses. This tester was designed for use with LiNbO media
(Courtesy IBM, with permission).

Fig. 70. Primary features of the DEMON II holographic digital data storage engine. This system
has demonstrated over 100 b=m (courtesy IBM, with permission).

increase quadratically with thickness, but as a result of these ometry is used, as shown. The reference beam is modulated
tradeoffs, in practice the useful dynamic range is often much in phase to produce a speckle pattern at the disk surface,
smaller. At higher media thicknesses, keeping optical distor- giving rise to small required displacement for multiplexing,
tions to a small fraction, typically 1/4 to 1/10 of the wave- as described above in the section on speckle multiplexing.
length of light, presents a challenge. InPhase indicates that a Using a 750- m-thick medium, a user bit density of
proprietary technique allows them to manage optical distor- 35 b m been reported. Media capable of higher densi-
tions efficiently during the manufacturing process. A com- ties are under development at InPhase, but no published
plete system demonstration incorporating media with thick- data describing media performance are avaialble. The re-
nesses exceeding 1 mm has not yet been reported, however. ported recording density of 45 b m was achieved in
The InPhase media has been tested in a demonstration plat- 750- m-thick media and correlation shift multiplexing at
form, indicated in Fig. 78, where a typical rotating disk ge- relatively low data rates (a few megabits per second) [116].

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1269

 multiplexing (i.e., recording holograms at different depth lo-
cations), total storage capacity is equal to
(55)
where denotes the number of layers, represents the
number of wavelength-multiplexed holograms, is the
number of angle-multiplexed holograms, and represents
the surface bit density. A capacity of 100 GB is, in principle,
achievable using two layers, ten wavelengths and 16 dif-
ferent angles for multiplexing a total of 160 holograms per
location, using a bit spot size of 1 m radius [141]. A widely
tunable laser is needed, as the wavelength selectivity is rel-
Fig. 71. Basic architecture as used by Rockwell for demonstrating
access times around 30 s (with permission). atively small, requiring tens of micrometers of wavelength
separation between multiplexed holograms. Such semicon-
ductor lasers do not currently exist, but, for example, dye
lasers can be scanned over a wide range of wavelengths.
The angular separation for good Bragg selectivity is also
quite large, as angular selectivity is at a minimum for the
reflection geometry, allowing only a few tens of holograms
to be multiplexed. In practice, storage densities not much
higher than a few gigabytes per disk have been obtained.
Siros Technologies investigated an alternative approach
using only a single hologram per location per layer, on mul-
tiple virtual layers in a homogeneous medium, and by using
confocal detection for readout. The layers are called virtual
because they are created during the recording process by
writing bits at particular depths. A schematic arrangement
is shown in Fig. 80.
Data are recorded as show in Fig. 79 using a green laser.
A second red laser is used for servoing and focusing. The
Fig. 72. Photograph of the Rockwell holographic fast access two laser systems are dynamically linked using a fast acting
storage demonstrator. optoelectronic servoing system. The servo layer is similar to
a grooved DVD recordable medium layer, which is used to
Issues related to improvement of material shrinkage and guide the writing beam during recording of the information.
media recording sensitivity are currently under investigation A split focusing and tracking detector is combined with a pin-
for this photopolymer material system. hole to achieve confocal detection allowing a layer spacing of
less than 10 m for light with a wavelength of 532 nm. This
arrangement allows a maximum of 10–15 layers to be packed
N. Single-Bit Holographic Storage
closely together without the need for spherical aberration
Storing single bits inside a holographic medium in layers correction using an NA 0.5 for the recording and readout
at various depths has been demonstrated in LiNbO and pho- beam. Embedded data in each layer allows fast in-plane and
topolymers. In analogy with the multilayer CD and DVD op- vertical seeks [142]. Experimental results for a nine virtual-
tical systems, bits are stored inside the medium on physical layer disk using Aprilis photopolymer media and a green
layers [139] or virtual layers [140] and read out with a fo- 532-nm laser are shown in Fig. 81. As signal strength is typ-
cused laser beam capable of addressing layers in depth. Data ically 10x smaller than for DVD systems, i.e., 1% versus
are stored by interfering two counterpropagating focused ob- 20%, respectively, high-gain low-noise amplifiers are used
ject and reference laser beams, as schematically shown in for data readout. Quantitative system performance measure-
Fig. 79, as reported by [141]. ments have not been reported, but initial results appear rather
In the recording region, small reflection holograms are promising [142].
formed, which can be read out using a single laser beam. Theoretical calculations indicate that a medium containing
Each hologram can represent a bit, or multiple bits can be 40–50 stratified layers might be possible using a red 658-nm
represented by elongated holographic marks in the direction laser and an NA 0.5. Shorter laser wavelengths allow
perpendicular to the optical axis to produce the equivalent of closer layer spacing and increased capacity. For example,
a DVD pit, where the pit length encodes data by measuring a blue laser having a wavelength of 405 nm and a focusing
readout signal transitions from low to high reflectivity, and lens of NA allows 25 channel b m to be recorded
vice versa. Multiple single-bit holograms can be superim- on BlueRay DVD disks. Assuming a reduced density of 15
posed to record multiple bits in the same physical location. channel b m , as holographic bits will not have the same
By combining wavelength, angle multiplexing, and depth sharp edges as data bits recorded on phase change media,

1270 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 73. Schematic of Holoplex repliactor for recording, replicating, and reading a hologram
(with permission).

least 150 channel b m or 150 GB per disk should be

achievable. Theoretical calculations show that for a signal
having a 40-dB SNR, 40–50 layers might be possible in
a single medium, giving rise to capacities of approaching 1
TB per disk. Data transfer rates can be increased by using
multiple beam readout, a technique developed by Zen Re-
search, Inc. Seven to ten bits can be read out simultaneously,
providing data transfer rates approaching 1 Gb/s.

XI. OUTLOOK

While there has been tremendous progress in research into

Fig. 74. Photograph of ten-beam Holoplex replicator (courtesy HDSSs over the last decade, there is still no mass-produced
Holoplex, with permission). or custom-built product available based on this technology.
The primary reason is that continued incremental advances
and allowing for lower signal strength of the holograms in magnetic and optical storage have met the needs for most
(around 1%) versus 20% for DVD signals, a density of at applications and have generated whole industries in which

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1271

 results in repeatable fabrication of media having the prop-
erties required for a system. Tens of tons of LiNbO , how-
ever, are annual produced at relatively low cost for signal
processing applications. At the same time, however, recent
demonstrations by Aprilis of their HMD product indicate that
capacities of over 250 GB per disk are feasible. This bench-
mark provides an important milestone toward a serious com-
mercialization effort. InPhase also has made good progress
in creating thick, high-quality optical media. Photopolymers
lend themselves well to disk-based architectures that could
leverage much of the mechanical technology used in cur-
rent disk drives and lend themselves well to mass produc-
tion. Although at present photopolymers have not yet ma-
tured to the point where they can consistently provide the
desired properties and reliability, a well-designed and exe-
cuted commercialization effort could provide the know-how
needed to achieve commercial quality media in a relatively
short period.
Materials used for holographic data storage are generally
specifically developed for this type of system. To date,
HDSSs have not been able to leverage the development
of materials for other purposes due to the tight material
requirements enforced by system requirements. Increas-
ingly, however, research into crystals and photopolymers
for photonic networking and interconnection has received
significant interest, and may result in materials that, with
appropriate modification, are suitable for holographic data
Fig. 75. Optical arangement of the Optware media tester (courtesy
Optware, with permission).
storage.
The various optical and electrical HDSS components used
component vendors are strongly synchronized to the needs of are adapted from components used in other industries. SLMs
the system integrators that provide the commercial products. are a significant driver in the display industry. The Texas In-
Although we continue to see advances in the components struments DMD is an extremely high-performance device ca-
and materials for HDSSs, the bar for market entry is raised pable of allowing data rates of 1 Gb/s with very high con-
each year as the incremental advances in traditional storage trast digital data pages. CCD and CMOS detector arrays have
systems continue. As far as mass production is concerned, numerous applications in image capture, including digital
until we begin to see a significant plateau in the advance- photography and manufacturing inspection. Size, data rate,
ment of the traditional technologies, they will continue to and performance of these detectors is close to what is re-
dominate. There may, however, be custom applications in quired for a holographic system. Advances in electronics in
which the high performance of HDSSs can be a technology the telecom and storage industries may also support HDSSs.
discriminator. One such example is associative data retrieval, Laser sources continue to improve incrementally and new
in particular for image-rich databases. Traditional search technology for tunable lasers developed for the telecom in-
approaches require some form of content indexing, which dustry may be applicable to holographic data storage. These
is difficult to do with pictures of objects that vary in size, may enable significant progress in research into wavelength
orientation and completeness of description throughout the multiplexing architectures. Additionally, the optical storage
database. To automatically find a person, from an incom- industry is developing high power blue lasers that would be
plete or fuzzy description, among thousands of people in excellent enablers for holographic storage. Once significant,
a 60-min video movie with better than 80% accuracy, for tangible commercial opportunities mature that current tech-
example, is currently beyond the capability of commercial nologies cannot economically capture and for which HDSSs
search engines. Holography is inherently suited for solving are genuine candidate solutions, we can expect significant
this problem, and it might provide the impetus for further synchronization of these various efforts to mature the tech-
commercialization efforts. The ultra short access times nology for mass production and custom development.
of a solid-state holographic system are also attractive for What can we then expect in the future? Several different
improving computer performance by reducing latency. scenarios come to mind, all taking advantage of the unique
A leading challenge in commercializing HDSSs is devel- properties of holographic systems, namely, the high data
oping suitable materials. Compact crystals generally require storage density, short access times, superior data transfer
sophisticated components and optical system engineering; rates, and ultrafast search capability. Schematically, four
these materials have not yet been reduced to a technology that potentiality different scenarios might play out for future

1272 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 76. Reconstructed 262-kb digital data page on CCD detector with 2/1 oversampling, direct
error map of page, histogram distribution of ones and zeros with application of local threshold
ontensities, and raw BER of the 60th of 155 colocationally multiplexed holograms, each recorded at
a fractional Fourier transform plane using peristrophic and planar–angular multiplexing in Aprilis
HMD CROP media with 400-m-thick material using an Nd : YAG laser at 532 nm (courtesy Aprilis,
with permission).

Fig. 77. (a) Growth of accumulative grating strength in Aprilis HMD CROP material. (b)
Reduction in sensitivity as function of recording fluence in Aprilis HMD CROP material (courtesy
Aprilis, with permission).

commercialization efforts [1], as shown in Fig. 82, pub- that make them close to suitable for commercialization.
lished by IBM. Capacities of over 250 GB per disk were demonstrated by
Photopolymer systems, either based on the CROP tech- Aprilis using their HDM media, and at Stanford University,
nology from Aprilis or the two-chemistry technology from transfer rates exceeding 10 Gb/s were measured in a fully
InPhase, appear destined to be potential candidates for DVD functional testbed using an earlier developed Aprilis CROP
follow-on products. Both materials exhibit characteristics photopolymer medium. There are, however, critical tests that

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1273

Fig. 78. Block diagram of the InPhase holographic data storage demonstrator (courtesy InPhase,
with permission).

storage systems require about the same energy on the de-

tector per bit, for writing and reading, 100 fJ/b for CCD
arrays. The strategic question that surfaces is: How large
should the page size really be? In CD or DVD systems, the
detected signal is from 50% to 20% of the incident power,
while in holographic storage this usually does not exceed
the range of 0.1% to 1%. On the other hand, single bits are
recorded in DVD systems having projected data rates of a
few hundred megabits per second; not too far from the 500
Mb/s required for follow-on systems. There is, therefore, an
opportunity to use holography with very small page sizes
(or even single-bit holograms), as has been researched by
Eichler in Germany and Siros Technologies in the United
Fig. 79. Single-bit hologram recording configuration (courtesy States, among others [141], [142], and earlier at IBM.
Eichler, with permission).
[143]. Small page sizes have significant advantages. The
SLM and detector arrays can be made cheap and small,
still must be performed. Even although media development but they require very fast switching times, approaching
has made significant progress over the past ten years, there is a few nanoseconds. Wavelength multiplexing and or
still no definitive test in which both density and transfer rates combined with spatial multiplexing of a few holograms
are measured in a working prototype, having performance would be sufficient to achieve data rates exceeding 1
characteristics that are commensurate with the optical data Gb/s and densities commensurate with capacities of 500
storage roadmap shown in Fig. 1, i.e., capacities exceeding GB on a 120–mm disk.
200 GB on a disk and transfer rates of 500 Mb/s or higher. As Small page size simplifies and makes cheaper the optical
holographic diffraction efficiency is inversely proportional design of high NA imaging optics, where a high NA is
to the square of the number of superimposed holograms, critical for achieving high density per hologram. The latter is
capacity and transfer rate are tradeoff parameters. Increasing important, since it results in higher diffraction efficiency of
one goes at the expense of the other one. Additionally, tol- recorded holograms due to their reduced number (perhaps
erances and systems issues require tradeoffs that lower the as few as 10–50) per multiplexed location. Higher bit
attainable capacity and transfer rates of the system further. It density per hologram means that fewer holograms have to
is, therefore, critical that in future system tests, total system be superimposed, which improves diffraction efficiency and
performance is measured, similar to what was accomplished SNR. The high NA approach was successfully exploited
in the Stanford HDSS consortium demonstration, as a pre- in the Stanford/HDSS system and most later designed
requisite to full scale commercialization. demonstration platforms described above. For small page
Another critical question in HDSS development is the sizes, Bragg degeneracy is less of an issue and tolerances
amount of laser power that will be available at a reasonable are increased, improving manufacturability. Additionally, a
cost. Currently all demonstration testbeds have used pow- small page size holographic system would afford a smoother
erful, expensive pulsed solid-state laser sources. Solid-state transition from the single-bit DVD system to a holographic
lasers are costly and require large amounts of electrical optical storage system. On the other hand, associative search
power. Fundamentally, the large required laser power is would lose its attractiveness, as the number of pixels under
caused by low diffraction efficiency and high data rates. the head will be small; therefore, such a holographic system
When normalized to photons per bit, almost all optical would not have the advantage of fast search capabilities

1274 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

Fig. 80. Dual laser system for recording and readout in green and servoing in red. The single-servo
layer is located below the virtual holographic data layers (courtesy Siros, with permission).

Fig. 81. Experimental results for reading nine layers of holographic single bit data recorded in
Aprilis CROP photopolymer media using an NA 0.5 lens and 532-nm laser light (courtesy Siros,
with permission).

over other forms of single bit storage, such as mag- tracking, as well as smaller lasers. These benefits are at the
netic or solid-state memories. Finally, paged based systems expense of reduced data transfer rate. Single-bit holograms
require massively parallel data processing for ECC and provide an evolutionary path to improvement in the perfor-
channel coding, potentially adding significantly to the cost mance of optical data storage systems, versus a revolutionary
of holographic systems. Encoding and decoding chips for (and much more challenging) improvement of larger size
Reed–Solomon and, more recently, turbocodes are com- holographic page based storage systems. One distinct advan-
mercially available and can be used in parallel to achieve tage of the latter approach is the enormously high-perfor-
the desired data transfer rates. Smaller page sizes require mance associative search capabilities which are not feasible
less image processing steps and, therefore, encoding and with holographic systems having small page sizes.
decoding of data is simplified. Finally, the issue of replication needs to be considered.
In short, when the page size becomes smaller, there is a The great commercial success of CD–ROM and DVD–ROM
higher tolerance to shrinkage, the tolerances for media flat- technology has been partially due to the low cost of replica-
ness and wedge are greatly reduced, the tolerances on high tion. CD– and DVD–ROM and recordable disks are stamped
NA imaging optics and the cost are reduced, and simpler out at a rate of one disk every few seconds at a cost that is
electronic control systems can be utilized for servoing and so low that customers use them as disposable media—very

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1275

Fig. 82. Four possible holographic storage commercialization scenarios: an all-solid-state memory
module having extremely short access times, two rotating disk geometries, with either erasable
or WORM type media, and a data warehouse with removable media. With its high volumetric
density, holographic storage has the potential to affect all types of data storage (courtesy IBM,
with permission).

much the same way as previous floppy disks were used, but croseconds, and the 90 configuration is ideally suited for
with 10 000 times improved capacity. The removability and associative retrieval at rates in excess of tens to hundreds
interchangeability of DVD media is a key to its success and of gigabits per second, and for direct data readout rates ex-
these attributes need to be preserved in future generation ceeding 1 Gb/s. Recording speeds are typically two orders
optical removable storage technology. Although page-based of magnitude slower. Fixing of information requires elevated
holographic data storage technology provides far-field op- temperatures, but this might be acceptable in server appli-
tical recording and readout, unfortunately it does not lend cations, for example, where data are recorded occasionally,
itself to simple copying procedures. As a result of multi- with readout by multiple users being the dominant use of the
plexing many holograms in the same location through either stored data. Cost of the medium could be brought down to
wavelength or angle multiplexing, a single-step replication $10–$20 per cubic centimeter 10 GB , produced in large
process is not practical, as all reference beam angles and volume. LiNbO , therefore, is still a very attractive medium
wavelengths must be simultaneously present to record data for holographic data storage for certain niche applications,
on the copied disk. So far, no good solution has been found such as for extremely rapid searches and in certain airborne
for this problem [144], and holographic system demon- and hostile environments where all-solid-state operation is
strations to date have been all of the WORM or rewritable required.
kind. Single-bit virtual multilayer holographic recording A jukebox approach, however, is not attractive as replace-
may have an advantage in this respect over large page based ment for DVD technology. A rotating disk architecture pro-
holographic recording, as replication techniques can be vides shorter access times than a jukebox and “looks and
envisioned. The single bits are recorded in the counterprop- feels” like a conventional disk drive. A significant effort has,
agating reflection geometry and a single, properly designed therefore, been expended to realize such a system with su-
reference beam can be used to create a copy of a master perior performance over current optical drives. The HDSS
hologram in a manner similar to techniques used for making demonstration, with 10 Gb/s as the highest data transfer rate
copies of display holograms [145]. of any storage system, was made with this approach. The
density achieved in this demonstration, however, was still
XII. SUMMARY too low. Due to material shrinkage and relatively high scatter
Research on HDSSs has progressed significantly since the levels, and limitations in shift multiplexing, the maximum
mid–1990s. New organic and inorganic materials have been density achieved so far (and published for verification) is 20
developed with much improved properties. LiNbO is still b/ m . This is about an order of magnitude too low for com-
the best quality optical medium for data storage as it does mercial viability of holographic technology. Unfortunately,
not shrink. Today, the best results in terms of density have for holographic data storage, density and transfer rate require
been demonstrated in this medium. Recording in LiNbO , a tradeoff of one against the other. As such, achieving either
however, is slow, and it is difficult to store large amounts high density or transfer rate in isolation is not sufficient, as
of data due to size limitations. A jukebox approach could both need to be demonstrated in the same experimental setup.
overcome this problem, but access times would be long, hun- The challenge for holographic data storage technology to be
dreds of milliseconds at best. For a single crystal, however, a viable commercial product, as we see it, is to achieve a ca-
access times to any stored data can be on the order of mi- pacity of exceeding 200 GB on a 5.25-in disk, with a transfer

1276 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

rate in excess of 500 Mb/s. The best photopolymer medium [14] S. S. Orlov, E. Bjornson, W. Phillips, Y. Takashima, X. Lee, L. Hes-
we have tested is made by Aprilis. With this medium, data selink, R. Okas, and R. Snyder, “High transfer rate (1 Gbit/sec) high
capacity holographic disk digital data storage system,” in Proc. Conf.
densities corresponding to 250 GB on a DVD disk have been Lasers and Electro-Optics (CLEO 2000), pp. 190–191.
demonstrated. The challenge is now to combine the high data [15] D. A. Waldman, H.-Y. S. Li, and M. G. Horner, “Volume shrinkage
storage capacity and the ultrahigh transfer rate in a single in slant fringe gratings of a cationic ring-opening volume hologram
recording material,” J. Imag. Sci. Technol., vol. 41, pp. 497–514,
testbed demonstration having a capacity of 250 GB and a 1997.
transfer rate exceeding 500 Mb/s. The outlook for such a [16] L. Dhar, K. Curtis, M. Tackitt, M. Schilling, S. Campbell, W. Wilson,
demonstration is very good and can reasonably be expected A. Hill, C. Boyd, N. Levinos, and A. Harris, “Holographic storage
of multiple high-capacity digital data pages in thick photopolymer
within the next few years. Achieving that goal would go a systems,” Opt. Lett., vol. 23, pp. 1710–1722, 1998.
long way toward establishing holographic data storage as a [17] S. S. Orlov, W. Phillips, E. Bjornson, L. Hesselink, and R. Okas,
viable follow-on technology to current DVD systems. “10 Gigabit/second sustained optical data transfer rate from a holo-
graphic disk digital data storage system,” presented at the OSA An-
nual Meeting 2000, Paper MK3.
[18] S. S. Orlov, “Volume holographic data storage,” Commun. ACM, vol.
ACKNOWLEDGMENT 43, pp. 46–55, 2000.
[19] L. d’Auria, J.-P. Huignard, C. Slezak, and E. Spitz, “Experimental
The authors would like to thank Dr. L. N. Durvasula of holographic read-write memory using 3-D storage,” Appl. Opt., vol.
DARPA for his pioneering view of HDSS and for his sup- 13, pp. 808–818, 1974.
[20] J. B. Thaxter and M. Kestigian, “Unique properties of SBN and their
port of the NSIC/DARPA/Industry HDSS/PRISM consortia. use in a layered optical memory,” Appl. Opt., vol. 13, pp. 913–924,
Without his support, many of the improvements in holo- 1974.
graphic data storage would not have happened. The authors [21] A. Mikaeliane et al., “Holographic bulk memories using lithium nio-
bate crystals for data recording,” in Optical Information Recording,
would also like to thank all the members of the two DARPA E. S. Barrekette et al., Eds. New York: Plenum, 1978, vol. 2, pp.
consortia for their contributions to the HDSS demonstration. 217–233.
The authors would also like to thank the Japanese Science [22] P. J. van Heerden, “Theory of optical information storage in solids,”
Appl. Opt., vol. 2, pp. 393–400, 1963.
and Technology Program. [23] S. Redfield, “Tamarack optical head holographic storage,” in Holo-
graphic Data Storage, H. Coufal, G. T. Sincerbox, and D. Psaltis,
Eds. Berlin, Germany: Springer-Verlag, 2000, pp. 343–357.
REFERENCES [24] J. F. Heanue, M. C. Bashaw, and L. Hesselink, “Volume holographic
storage and retrieval of digital data,” Science, vol. 265, pp. 749–752,
[1] J. Ashley, M.-P. Bernal, G. W. Burr, H. Coufal, H. Guenther, J. A. 1994.
Hoffnagle, C. M. Jefferson, B. Marcus, R. M. Macfarlane, and R. M. [25] D. Psaltis, D. Brady, and K. Wagner, “Adaptive optical networks
Shelby, “Holographic data storage,” IBM J. Res. Develop., vol. 44, using photorefractive crystals,” Appl. Opt., vol. 27, pp. 1752–1759,
pp. 341–368, 2000. 1988.
[2] L. Hesselink and M. C. Bashaw, “Optical memories implemented [26] P. Yeh, Introduction to Photorefractive Nonlinear Optics. New
with photorefractive media,” Opt. Quantum Electron., vol. 25, pp. York: Wiley, 1993.
S611–S661, 1993. [27] M. C. Bashaw, J. F. Heanue, and L. Hesselink, “Organization of
[3] A. J. Daiber, R. Snyder, J. Colvin, R. Okas, and L. Hesselink, “Fully data for monochromatic multiplex volume holography,” J. Opt. Soc.
functional digital video holographic storage system,” presented at Amer. A, vol. 13, pp. 2174–2186, 1996.
the Optical Soc. Amer. Annu. Meeting, Long Beach, CA, 1997, [28] C. X.-G. Gu, “Optical neural networks using volume holography,”
Paper ThR3. PhD dissertation, California Inst. Technol., Pasadena, 1990.
[4] R. M. Shelby, J. A. Hoffnagle, G. W. Burr, C. M. Jefferson, M.-P. [29] M. A. Neifeld and W. C. Chou, “Information theoretic limits to the
Bernal, H. Coufal, R. K. Grygier, H. Gunther, R. M. Macfarlane, capacity of volume holographic optical memory,” Appl. Opt., vol.
and G. T. Sincerbox, “Pixel-matched holographic data storage with 36, pp. 514–517, 1997.
megabit pages,” Opt. Lett., vol. 22, pp. 1509–1511, 1997. [30] G. W. Burr, C. M. Jefferson, H. Coufal, M. Jurich, T. A. Hoffnagle, R.
[5] J. H. Hong, I. McMichael, T. Y. Chang, W. Christian, and E. G. M. Macfarlane, and R. M. Shelby, “Volume holographic data storage
Paek, “Volume holographic memory systems: Techniques and archi- at an areal density of 250 gigapixels/in2,” Opt. Lett., vol. 26, pp.
tectures,” Opt. Eng., vol. 34, pp. 2193–2203, 1995. 444–446, 2001.
[6] D. Psaltis and G. W. Burr, “Volume holographic memory systems: [31] C. Gu, J. Hong, I. McMichael, R. Saxena, and F. Mok, “Cross-talk-
Techniques and architectures,” Computer, vol. 31, pp. 52–60, 1998. limited storage capacity of volume holographic memory,” J. Opt.
[7] K. Curtis, “Digital holographic data storage prototype,” in Conf. Rec. Soc. Amer. A, vol. 9, pp. 1978–1983, 1992.
2000 Optical Data Storage, pp. 164–166. [32] M. C. Bashaw, J. F. Heanue, A. Aharoni, J. F. Walkup, and L.
[8] F. H. Mok, “Angle-multiplexed storage of 5000 holograms in lithium Hesselink, “Crosstalk considerations for angular and phase-encoded
niobate,” Opt. Lett., vol. 18, pp. 915–917, 1993. multiplexing in volume holography,” J. Opt. Soc. Amer. B, vol. 11,
[9] G. Barbastathis and D. Psaltis, “Volume holographic multiplexing pp. 1820–1836, 1994.
methods,” in Holographic Data Storage, H. Coufal, D. Psaltis, and [33] E. N. Leith, A. Kozma, J. Upatnieks, J. Marks, and N. Massey,
G. T. Sincerbox, Eds. Berlin, Germany: Springer-Verlag, 2000, pp. “Holographic data storage in three-dimensional media,” Appl. Opt.,
21–62. vol. 5, pp. 1303–1311, 1966.
[10] L. Hesselink, “Digital holographic demonstration systems by [34] A. Yariv, G. Rakuljic, A. Yariv, and V. Leyva, “High resolution
Stanford University and Sires Technologies,” in Holographic Data volume holography using orthogonal data storage,” presented at the
Storage, H. Coufal, D. Psaltis, and G. T. Sincerbox, Eds. Berlin, Tech. Dig. Photorefractive Materials, Effects, and Devices 1991,
Germany: Springer-Verlag, 2000, pp. 383–397. Paper MD-3.
[11] D. Psaltis, M. Levene, A. Pu, G. Barbastathis, and K. Curtis, “Holo- [35] X. Yi, P. Yeh, C. Gu, and S. Campbell, “Crosstalk in volume holo-
graphic storage using shift multiplexing,” Opt. Lett., vol. 20, pp. graphic memory,” Proc. IEEE, vol. 87, pp. 1912–1930, Nov. 1999.
782–784, 1995. [36] C. Gu, P. Yeh, X. Yi, and J. Hong, “Fundamental noise sources
[12] A. M. Darskii and V. B. Markov, “Shift selectivity of holograms in volume holographic storage,” in Holographic Data Storage, H.
with a reference speckle wave,” Opt. Spectmsc. (USSR), vol. 65, pp. Coufal, D. Psaltis, and G. T. Sincerbox, Eds. Berlin, Germany:
392–395, 1988. Springer-Verlag, 2000, pp. 63–88.
[13] T. Bieringer, “Photoaddressable polymers,” in Holographic Data [37] K. Curtis, C. Gu, and D. Psaltis, “Cross talk in wavelength-multi-
Storage, H. Coufal, G. T. Sincerbox, and D. Psaltis, Eds. Berlin, plexed holographic memories,” Opt. Lett., vol. 18, pp. 1001–1003,
Germany: Springer-Verlag, 2000, pp. 209–230. 1993.

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1277

 [38] D. Lande, J. F. Heanue, M. C. Bashaw, and L. Hesselink, “Dig- [61] V. L. Vinetskii and N. V. Kukhtarev, “Theory of the conductivity
ital wavelength-multiplexed holographic data storage system,” Opt. induced by recording holographic gratings in nonmetallic crystals,”
Lett., vol. 21, pp. 1780–1782, 1996. Fiz. Tverd. Tela, vol. 16, pp. 3714–3716, 1974. English transl.: Sov.
[39] T. F. Krile, R. J. Marks II, J. F. Walkup, and M. O. Hagler, “Holo- Phys. Solid State, vol. 16, p. 2414, 1975.
graphic representations of space-variant systems using phase-coded [62] N. V. Kukhtarev, “Kinetics of hologram recording and erasure in
reference beams,” Appl Opt., vol. 16, pp. 3131–3135, 1977. electrooptic crystals,” Pis’ma Zh. Tekh. Fiz., vol. 2, pp. 1114–1119,
[40] T. F. Krile, M. O. Hagler, W. D. Redus, and J. F. Walkup, “Mul- 1976. English transl.: Sov. Tech. Phys. Lett., vol. 2, p. 438, 1976.
tiplex holography with chirp-modulated binary phase-coded refer- [63] N. V. Kukhtarev, V. B. Markov, S. G. Odulov, M. S. Soskin, and V.
ence-beam masks,” Appl. Opt., vol. 18, pp. 52–56, 1979. L. Vinetskii, “Holographic storage in electrooptic crystals. I. Steady
[41] V. N. Morozov, “Theory of holograms formed using coded reference state,” Ferroelectrics, vol. 22, pp. 949–960, 1979.
beams,” Kvantovaya Elektron., vol. 4, pp. 1694–1701, 1977. English [64] G. C. Valley, “Erase rates in photorefractive materials with two pho-
transl.: Sov. J. Quantum. Electron., vol. 7, pp. 961-964, 1977. toactive species,” Appl. Opt., vol. 22, pp. 3160–3164, 1983.
[42] D. Gabor, “Associative holographic memories,” IBM J. Res. De- [65] G. C. Valley and J. F. Lam, “Theory of photorefractive effects in
velop., pp. 156–159, Mar. 1969. electro-optic crystals,” in Photorefractive Materials and Their Ap-
[43] E. L. Kral, J. F. Walkup, and M. O. Hagler, “Correlation properties of plications I, P. Günter and J.-P. Huignard, Eds. Berlin, Germany:
random phase diffusers for multiplex holography,” Appl. Opt., vol. Springer-Verlag, 1988, pp. 75–98.
21, pp. 1281–1290, 1982. [66] A. Yariv, S. S. Orlov, and G. A. Rakuljic, “Holographic storage dy-
[44] A. E. Krasnov, “Thick-film phase holograms recorded by means namics in lithium niobate: theory and experiment,” J. Opt. Soc. Amer.
of coded reference waves,” Kvantovaya Elektron., vol. 4, pp. B, vol. 13, pp. 2513–2523, 1996.
2011–2013, 1977. English transl.: Sov. J. Quantum. Electron., vol. [67] B. Kippelen, “Overview of photorefractive polymers for holographic
7, pp. 1147-1148, 1977. data storage,” in Holographic Data Storage, H. Coufal, D. Psaltis,
[45] C. Denz, K. O. Miller, F. Visinka, and T. Tschudi, “Demonstration and G. T. Sincerbox, Eds. Berlin, Germany: Springer-Verlag,
platform for phase-coded multiplexing,” in Holographic Data 2000, pp. 159–169.
Storage, H. Coufal, D. Psaltis, and G. T. Sincerbox, Eds. Berlin, [68] D. L. Staebler, W. J. Burke, W. Phillips, and J. J. Amodei, “Multiple
Germany: Springer-Verlag, 2000, pp. 419–428. storage and erasure of fixed holograms in Fe-doped LiNbO ,” Appl.
[46] G. Barbastathis, M. Levene, and D. Psaltis, “Shift multiplexing with Phys. Lett., vol. 26, pp. 182–184, 1975.
spherical reference waves,” Appl. Opt., vol. 35, pp. 2403–2417, [69] M. Carrascosa and F. Agulló-Löpez, “Theoretical modeling of the
1996. fixing and developing of holographic gratings in LiNbO ,” J. Opt.
[47] L. Hesselink, “Photorefractive fibers for optical data storage and pro- Soc. Amer. B, vol. 7, pp. 2317–2322, 1990.
cessing,” Intl. J. Optoelectron., vol. 5, pp. 103–124, 1990. [70] G. Montemezzani, M. Zgonik, and P. Günter, “Photorefractive
[48] S. Wu, A. Mayers, S. Rajan, and F. T. S. Yu, “Use of photorefractive charge compensation at elevated temperatures and application to
fiber in optical interconnections and switching,” Appl. Opt., vol. 29,
KNbO ,” J. Opt. Soc. Amer. B, vol. 10, pp. 171–185, 1993.
pp. 1059–1061, 1990.
[71] A. Yariv, S. Orlov, G. Rakuljic, and V. Leyva, “Hologram fixing,
[49] F. Ito, K.-I. Kitayama, and H. Oguri, “Compensation of fiber holo-
readout, and storage dynamics in photorefractive materials,” Opt.
graphic image distortion caused by intrasignal photorefractive cou-
Lett., vol. 20, pp. 1334–1336, 1995.
pling by using a phase-conjugated mirror,” Opt. Lett., vol. 17, pp.
[72] J. J. Amodei and D. L. Staebler, “Holographic pattern fixing in
215–217, 1992.
electro-optic crystals,” Appl. Phys. Lett., vol. 18, pp. 540–542, 1971.
[50] M. C. Bashaw, A. Aharoni, and L. Hesselink, “Alleviation of
[73] S. S. Orlov and W. Phillips, “Hologram fixing and nonvolatile
image distortion due to striations in a photorefractive medium by
storage in photorefractive materials,” in Holographic Data Storage,
using a phase-conjugated reference wave,” Opt. Lett., vol. 17, pp.
1149–1151, 1992. H. Coufal, D. Psaltis, and G. T. Sincerbox, Eds. Berlin, Germany:
[51] A. Aharoni, M. C. Bashaw, and L. Hesselink, “Distortion-free mul- Springer-Verlag, 2000, pp. 127–148.
tiplexed holography in striated photorefractive media,” Appl. Opt., [74] H. Kurz, “Lithium niobate as a material for holographic information
vol. 32, pp. 1973–1982, 1993. storage,” Philips Tech. Rev., vol. 37, pp. 109–120, 1977.
[52] E. Chuang, W. Liu, J. P. Drolet, and D. Psaltis, “Holographic random [75] , “Photorefractive recording dyanmics and multiple storage of
access memory (HRAM),” Proc. IEEE, vol. 87, pp. 1931–1940, Nov. volume holograms in photorefractive LiNbO ,” Optica Acta, vol. 24,
1999. pp. 463–473, 1977.
[53] G. W. Burr and I. Leyva, “Multiplexed phase-conjugate holographic [76] W. J. Burke, D. L. Staebler, W. Phillips, and G. A. Alphonse,
data storage with a buffer hole gram,” Opt. Lett., vol. 25, pp. “Volume phase holographic storage in ferroelectric crystals,” Opt.
499–501, 2000. Eng., vol. 17, pp. 308–316, 1978.
[54] G. W. Burr, S. Kobras, H. Hanssen, and H. Coufal, “Content-address- [77] J. F. Heanue, M. C. Bashaw, A. J. Daiber, R. Snyder, and L. Hes-
able data storage using volume holograms,” Appl. Opt., vol. 38, pp. selink, “Thermal fixing for digital holographic data storage,” Opt.
6779–6784, 1999. Lett., vol. 21, pp. 1615–1617, 1996.
[55] X. Li, F. Dimov, W. Phillips, L. Hesselink, and R. McLeod, “Parallel [78] X. An, D. Psaltis, and G. W. Burr, “Thermal fixing of 10 000 holo-
associate search by use of volume holographic memory,” in Proc. grams in LiNbO ,” Appl. Opt., vol. 38, pp. 386–393, 1999.
29th Applied Imagery and Pattern Recognition Workshop (AIPR [79] D. von der Linde, A. M. Glass, and K. F. Rodgers, “High-sensitivity
2000), J. Aanstoos, Ed., pp. 78–83. optical recording in KTN by two-photon absorption,” Appl. Phys.
[56] S. S. Orlov, “Overview of holographic recording materials for major Lett., vol. 26, pp. 22–24, 1975.
system architectures in holographic data storage applications,” pre- [80] , “Optical storage using refractive index changes induced by
sented at the National Storage Industry Consortium Int. Workshop two-step excitation,” J. Appl. Phys., vol. 47, pp. 217–220, 1976.
Holographic Data Storage, Nice, France, 1999. [81] J. P. Wilde, “Spectroscopic characterization of photorefractive ma-
[57] M.-P. Bernal, H. Coufal, R. K. Grygier, J. A. Hoffnagle, C. M. Jef- terials for holographic storage applications,” in Proc. SPIE, Fluores-
ferson, R. M. Macfarlane, R. M. Shelby, G. T. Sincerbox, P. Wimmer, cence Detection Conf., 1996, pp. 82–92.
and G. Wittmann, “A precision tester for studies of holographic op- [82] Y. S. Bai, R. R. Neurgaonkar, and R. Kachru, “Resonant two-photon
tical storage materials and recording physics,” Appl. Opt., vol. 35, photorefractive gating in praeseodymium-doped strontium barium
pp. 2360–2374, 1996. niobate with cw lasers,” Opt. Lett., vol. 21, pp. 567–569, 1996.
[58] R. M. Shelby, “Media requirements for digital holographic data [83] S. S. Orlov, A. Akella, L. Hesselink, and R. R. Neurgaonkar, “High
storage,” in Holographic Data Storage, H. Coufal, D. Psaltis, and sensitivity two-photon nonvolatile recording in lithium niobate,” pre-
G. T. Sincerbox, Eds. Berlin, Germany: Springer-Verlag, 2000, sented at the Conf. Lasers and Electro-Optics (CLEO), 1997, Post-
pp. 101–112. deadline Paper CPD29.
[59] P. Günter and J.-P. Huignard, “Photorefractive effects and materials,” [84] L. Hesselink, S. S. Orlov, A. Liu, A. Akella, D. Lande, and R. R.
in Photorefractive Materials and Their Applications I, P. Günter and Neurgaonkar, “Photorefractive materials for nonvolatile volume
J.-P. Huignard, Eds. Berlin, Germany: Springer-Verlag, 1988, pp. holographic data storage,” Science, vol. 282, pp. 1089–1094, 1998.
7–73. [85] D. von der Linde, A. M. Glass, and K. F. Rodgers, “Multiphoton pho-
[60] H. W. Kogelnik, “Coupled wave theory for thick hologram gratings,” torefractive processes for optical storage in LiNbO ,” Appl. Phys.
Bell Syst. Tech. J., vol. 48, pp. 2909–2947, 1969. Lett., vol. 25, pp. 155–157, 1974.

1278 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004

 [86] D. von der Linde and A. M. Glass, “Photorefractive effects for re- [111] Y. Qiao and D. Psaltis, “Sampled dynamic holographic memory,”
versible holographic storage of information,” Appl. Phys., vol. 85, Opt. Lett., vol. 17, pp. 1376–1378, 1992.
pp. 85–100, 1975. [112] H. Rajbenbach, S. Bann, and J.-P. Huignard, “Long-term readout
[87] H. Vormann and E. Krätzig, “Two step excitation in LiTaO : Fe for of photorefractive memories by using a storage/amplification two-
optical data storage,” Solid State Commun., vol. 49, pp. 843–847, crystal configuration,” Opt. Lett., vol. 17, pp. 1712–1714, 1992.
1984. [113] S. Odoulov, T. Tarabrova, A. Shumelyuk, L. I. Naumova, and T.
[88] D. L. Staebler and W. Phillips, “Hologram storage in photochromic O. Chaplina, “Photorefractive response of bulk periodically poled
LiNbO ,” Appl. Phys. Lett., vol. 24, pp. 268–270, 1974. LiNbO :Y:Fe at high and low spatial frequencies,” Phys. Rev. Lett.,
[89] K. Buse, A. Adibi, and D. Psaltis, “Non-volatile holographic storage vol. 84, pp. 3294–3297, 2000.
in doubly doped lithium niobate crystals,” Nature, vol. 393, pp. [114] R. De Vré, J. F. Heanue, K. Gürkan, and L. Hesselink, “Transfer
665–668, 1998. functions based on Bragg detuning effects for image-bearing holo-
[90] D. Lande, S. S. Orlov, A. Akella, D. Lande, and R. R. Neurgaonkar, grams recorded in photorefractive crystals,” J. Opt. Soc. Amer. A,
“Digital holographic storage system incorporating optical fixing,” vol. 13, pp. 1331–1344, 1996.
Opt. Lett., vol. 22, pp. 1722–1724, 1997. [115] L. Dhar, M. G. Schnoes, H. E. Katz, A. Hale, M. L. Schilling, and
[91] D. Tontchev, S. Zhivkova, and M. Miteva, “Holographic interfero- A. L. Harris, “Photopolymers for digital holographic data storage,”
metric microscope on the basis of a Bi TiO crystal,” Appl. Opt., in Holographic Data Storage, H. Coufal, D. Psaltis, and G. T.
vol. 29, pp. 4753–4756, 1990. Sincerbox, Eds. Berlin, Germany: Springer-Verlag, 2000, pp.
[92] S. Zhivkova and M. Miteva, “Image subtraction using fixed holo- 199–208.
grams in photorefractive Bi SiO crystals,” Opt. Lett., vol. 16, pp. [116] K. Curtis, W. L. Wilson, M. C. Tackitt, A. J. Hill, and S. Camp-
750–751, 1991. bell, “High-density, high-performance data storage via volume
[93] M. C. Bashaw, T.-P. Ma, R. C. Barker, S. Mroczkowski, and R. holography: the Lucent Technologies hardware platform,” in Holo-
R. Dube, “Introduction, revelation, and evolution of complementary graphic Data Storage, H. Coufal, D. Psaltis, and G. T. Sincerbox,
holograms in photorefractive bismuth silicon oxide,” Phys. Rev. B, Eds. Berlin, Germany: Springer-Verlag, 2000, pp. 359–368.
vol. 42, pp. 5641–5648, 1990. [117] A. Vander Lugt, “Design relationships for holographic memories,”
[94] S. Zhivkova and M. Miteva, “Investigations of the characteristics Appl. Opt., vol. 12, pp. 1675–1685, 1973.
of fixed holograms in Bi TiO photorefractive crystals,” Opt. [118] A. L. Mikaelian, V. L. Bobrinev, S. M. Naumova, and L. Z.
Commun., vol. 86, pp. 449–453, 1991. Sokolova, “Design principles of holographic memory devices,”
[95] N. A. Vainos, S. L. Clapham, and R. W. Eason, “Multiplexed perma- IEEE. J. Quantum Electron., vol. QE-6, pp. 193–198, Apr. 1970.
nent and real time holographic recording in photorefractive BSO,” [119] W.-H. Lee, “Effect of film-grain noise on the performance of holo-
Appl. Opt., vol. 28, pp. 4381–4385, 1989. graphic memory,” J. Opt. Soc. Am., vol. 62, pp. 797–801, 1972.
[96] , “Applications of multiplexed real time and permanent holo- [120] J. W. Goodman, Statistical Optics. New York: Wiley, 1985.
graphic recording in photorefractive BSO,” Appl. Opt., vol. 28, pp.
[121] J. F. Heanue, M. C. Bashaw, and L. Hesselink, “Channel coding for
4386–4392, 1989.
digital holography,” J. Opt. Soc. Amer. A, vol. 12, pp. 2432–2439,
[97] V. Leyva, A. Agranat, and A. Yariv, “Fixing of a photorefractive
1995.
grating in KTa Nbx O by cooling through the ferroelectric phase
[122] B. Marcus, “Modulation codes for holographic recording,” in Holo-
transition,” Opt. Lett., vol. 16, pp. 554–556, 1991.
graphic Data Storage, H. Coufal, D. Psaltis, and G. T. Sincerbox,
[98] F. Micheron and G. Bismuth, “Field and time thresholds for the elec-
Eds. Berlin, Germany: Springer-Verlag, 2000, pp. 283–292.
trical fixation of holograms recorded in (Sr Ba )Nb O crys-
[123] G. W. Burr, J. Ashley, H. Coufal, R. Grygier, J. Hoffnagle, C. M. Jef-
tals,” Appl. Phys. Lett., vol. 23, pp. 71–72, 1973.
ferson, and B. Marcus, “Modulation coding for pixel-matched holo-
[99] , “Electrical control of fixation and erasure of holographic pat-
terns in ferroelectric materials,” Appl. Phys. Lett., vol. 20, pp. 79–81, graphic data storage,” Opt. Lett., vol. 22, pp. 639–641, 1997.
1972. [124] R. Snyder and A. Daiber, “Method for holographic data storage and
[100] S. Orlov, D. Psaltis, and R. R. Neurgaonkar, “Dynamic electronic retrieval,” U.S. Patent 6 064 586, May 16, 2000.
compensation of fixed gratings in photorefractive media,” Appl. [125] G. W. Burr and T. Weiss, “Compensation for pixel misregistration in
Phys. Lett., vol. 63, pp. 2466–2468, 1993. volume holographic data storage,” Opt. Lett., vol. 26, pp. 542–544,
[101] J. Ma, T. Chang, J. Hong, R. Neurgaonkar, G. Barbastathis, and D. 2001.
Psaltis, “Electrical fixing of 1000 angle-multiplexed holograms in [126] J. F. Heanue, K. Gurkan, and L. Hesselink, “Signal detection for
SBN:75,” Opt. Lett., vol. 22, pp. 1114–1116, 1997. page-access optical memories with intersymbol interference,” Appl.
[102] E. K. Gulanyan, I. R. Dorosh, V. D. Iskin, A. L. Mikaelyan, and Opt., vol. 35, pp. 2431–2438, 1996.
M. A. Maiorchuk, “Nondestructive readout of holograms in iron- [127] V. Vadde and B. V. K. V. Kumar, “Channel modeling and estimation
doped lithium niobate crystals,” Kvantovaya Elektron., vol. 6, pp. for intrapage equalization in pixel-matched volume holographic data
1100–2097, 1979. English transl.: Sov. J. Quantum Electron., vol. 9, storage,” Appl. Opt., vol. 38, pp. 4374–4386, 1999.
pp. 647-649, 1979. [128] G. W. Burr, H. Coufal, R. K. Grygier, J. A. Hoffnagle, and C. M.
[103] H.-C. Külich, “Transfer function for image formation of objects re- Jefferson, “Noise reduction of page-oriented data storage by inverse
constructed from volume holograms with different wavelengths,” filtering during recording,” Opt. Lett., vol. 23, pp. 289–291, 1998.
Appl. Opt., vol. 31, pp. 2461–2477, 1992. [129] F. Mok, D. Psaltis, and G. Burr, “Spatially- and angle-multiplexed
[104] D. Psaltis, F. H. Mok, and H.-Y. S. Li, “Nonvolatile storage in pho- holographic random access memory,” in Proc. SPIE, Photonic
torefractive crystals,” Opt. Lett., vol. 19, pp. 210–212, 1994. Neural Networks, vol. 1773, 1992, pp. 334–345.
[105] A. Aharoni, M. Jeganathan, M. C. Bashaw, and L. Hesselink, “Pro- [130] R. Brauer, U. Wojak, F. Wyrowski, and O. Bryngdahl, “Digital dif-
longed readout of photorefractive holograms by replay at a longer fusers for optical holography,” Opt. Lett., vol. 16, pp. 1427–1429,
wavelength,” presented at the Conf. Lasers and Electro-Optics, 1994, 1991.
Paper CTuJ4. [131] S. S. Orlov, R. Snyder, and M. C. Bashaw, “Optical relay for pixel-
[106] D. Brady, K. Hsu, and D. Psaltis, “Periodically refreshed mul- based holographic storage and retrieval,” , Aug. 22, 2000.
tiply exposed photorefractive holograms,” Opt. Lett., vol. 15, pp. [132] W. Phillips, S. S. Orlov, E. Bjornson, L. Hesselink, and R. Okas,
817–819, 1990. “Video demonstration of high data rate holographic disk data storage
[107] H. Sasaki, Y. Fainman, J. E. Ford, Y. Taketomi, and S. H. Lee, system,” presented at the OSA Annu. Meeting 2000, Paper MK2.
“Dynamic photorefractive optical memory,” Opt. Lett., vol. 16, pp. [133] C. M. Jefferson, G. W. Burr, and J. A. Hoffnagle, “IBM holographic
1874–1876, 1991. data storage test platforms,” in Holographic Data Storage, H.
[108] Y. Qiao, D. Psaltis, C. Gu, J. Hong, P. Yeh, and R. R. Neurgaonkar, Coufal, D. Psaltis, and G. T. Sincerbox, Eds. Berlin, Germany:
“Phase-locked sustainment of photorefractive holograms using Springer-Verlag, 2000, pp. 369–381.
phase conjugation,” J. Appl. Phys., vol. 70, pp. 4646–4648, 1991. [134] J. A. Hollhagle and C. M. Jefferson, “Design and performance of a
[109] D. Brady and D. Psaltis, “Control of volume holograms,” J. Opt. Soc. refractive optical system that converts a Gaussian to a flattop beam,”
Amer. A, vol. 9, pp. 1167–182, 1992. Appl. Opt., vol. 39, pp. 5488–5499, 2000.
[110] S. Boj, G. Pauliat, and G. Roosen, “Dynamic holographic memory [135] J. Ma, T. Chang, S. Choi, and J. Hong, “Ruggedized digital holo-
showing readout, refreshing, and updating capabilities,” Opt. Lett., graphic data storage with fast access,” Opt. Quantum Electron., vol.
vol. 17, pp. 438–440, 1992. 32, pp. 383–392, 2000.

HESSELINK et al.: HOLOGRAPHIC DATA STORAGE SYSTEMS 1279

[136] F. Mok, G. Zhou, and D. Psaltis, “Holographic read-only memory,” Sergei S. Orlov received the M.S. degree in
in Holographic Data Storage, H. Coufal, D. Psaltis, and G. T. physics and applied mathematics from Moscow
Sincerbox, Eds. Berlin, Germany: Springer-Verlag, 2000, pp. Institute of Physics and Technology in 1991
399–407. and the the M.S.E.E. and Ph.D. degrees in elec-
[137] D. A. Waldman, C. I. Butler, and D. H. Raguin, “CROP holographic trical engineering from the California Institute
storage media for optical data storage at greater than 100 bits/sq. of Technology, Pasadena, in 1993 and 1996,
micron,” in Proc. SPIE, Organic Holographic Materials and Appli- respectively.
cations, vol. 5216, 2003, pp. 10–25. From 1996 till 1999, he was a Senior Research
[138] M. Schnoes, B. Ihas, A. Hill, L. Dhar, D. Michaels, S. Set- Scientist at Optitek, Inc. and Siros Technologies,
thachayanon, G. Schomberger, and W. L. Wilson, “Holographic working on development of holographic data
data storage media for practical systems,” in Proc. SPIE, Practical storage systems (HDSSs) and advanced holo-
Holography XVII and Holographic Materials IX, vol. 5005, 2003, graphic recording materials within the Defense Advanced Research Projects
pp. 29–37. Agency HDSS and photorefractive information storage media (PRISM)
[139] S. Yagi, T. Imai, Y. Kurokawa, M. Endo, K. Ishihara, R. Yoshiyama, programs. He is currently a Senior Research Scientist in the Solid State
H. Kubo, and Y. Miwa, “Multilayered waveguide holographic ROM Photonics Lab of Electrical Engineering Department at Stanford University,
card,” Trans. Inst. Electron. Inf. Commun. Eng. C, vol. J84-C, pp. Palo Alto, CA. He has published over 30 scientific articles, contributed
635–643, 2001. several book chapters, and holds several patents in the fields of optical
[140] L. Hesselink, “Three dimensional recording (3DR) technology,” in design, holographic recording materials, and volumetric optical storage.
Tech. Dig. 2000 Optical Data Storage (ODS) Conf., pp. 149–151. His research interests include holographic data storage, ferroelectric and
[141] H. J. Eichler, P. Kuemmel, S. Orloc, and A. Wappelt, “High-density photorefractive materials, volumetric optical storage materials, optical
disk storage by multiplexed microholograms,” IEEE J. Select. Topics phase conjugation, and interaction of laser light with matter.
Quantum Electron., vol. 4, pp. 840–848, 1998. Dr. Orlov is a Member of the Optical Society of America. In 2001, he re-
[142] R. McLeod and L. Hesselink, Virtual Multilayer Holographic Data ceived a technical achievement award from National Storage Industry Con-
Storage, 2004, submitted for publication. sortium (INSIC) for “outstanding contributions to the development of holo-
[143] A. W. Bidwell and G. T. Sincerbox, “Holographic storage system,” graphic storage technology and to the NSIC HDSS and PRISM Research
IBM Tech. Disclosure Bull., vol. 12, pp. 1659–1660, 1970. programs.”
[144] F. Mok, G. Zhou, and D. Psaltis, “Holographic read-only memory,”
in Holographic Data Storage, H. Coufal, D. Psaltis, and G. T.
Sincerbox, Eds. Berlin, Germany: Springer-Verlag, 2000, pp.
399–407.
[145] S. A. Benton, “Hologram reconstructions with extened incoherent
sources,” J. Opt. Soc. Amer., vol. 59, pp. 1545–1546A, 1969.

Lambertus Hesselink received the Ph.D. degree

from the California Institute of Technology,
Pasadena, in 1977. Matthew C. Bashaw received the B.S. degree
He was the Principal Investigator of two large in physics from the University of Notre Dame,
$52 million Defense Advanced Research Projects South Bend, IN, in 1985 and the M.S., M.Phil.,
Agency/industry/university consortia on the de- and Ph.D. degrees in engineering and applied sci-
velopment of holographic data storage systems ence from Yale University, New Haven, CT, in
(HDSS) and photorefractive information storage 1987, 1988, and 1991, respectively.
media (PRISM) from 1995 to 2000. He founded He served as a Research Scientist at Stanford
Siros Technologies and Senvid, Inc., where he has University, where he led several research teams
held the positions of Chairman, Chief Executive focused on holography, nonlinear optics, optical
Officer (CEO), and Chief Technology Officer (CTO). He was also a Vis- signal processing, optical encryption, and other
iting Professor in China, Europe, and Japan. He is currently Professor of optical technologies. His team accomplishments
Electrical Engineering and, by courtesy, Applied Physics at Stanford Uni- include the first reported end-to-end digital holographic data storage
versity, Palo Alto, CA. He is considered a worldwide leader in the develop- system. He contributed significantly to the Stanford University Center
ment of ultrahigh-performance optical data storage systems and a pioneer for Nonlinear Optical Materials, the Photorefractive Information Storage
in Internet-assisted learning. He has authored over 350 papers in archival Materials Consortium (PRISM), and the Holographic Data Storage Systems
journals and presented over 250 keynote and invited lectures at scientific Consortium (HDSS), which included Defense Advanced Research Projects
meetings worldwide. He was Editor of Applied Optics and Applied Scien- Agency sponsorship and industry collaboration. After Stanford, he held key
tific Research, and he has over 50 patents and over 15 pending applications. positions at several startup companies developing leading-edge enabling
His current research interests span the areas of ultrahigh density optical data technologies for information system products. His roles included research
storage and nanophotonics. and development, technology transfer, product development, business
Dr. Hesselink is a Fellow of the Optical Society of America, a Fellow of development, and intellectual property strategy. In these capacities, he
the Society of Photo Instrumentation Engineers (SPIE), and a Member of the provided critical leadership in fiber optic networking, optical data storage,
Royal Dutch Academy of Arts and Sciences. Among other honors, he was a photonics technology, speech recognition tools, and online media distri-
Member of the Hubble Space Telescope Committee, a Member of the Scien- bution technology. He is currently a Senior Staff Engineer at Lockheed
tific Advisory Board of the U.S. Air Force, a Codesigner of the official astro- Martin Corporation. He is also a registered patent agent. He has authored
naut memorial at Space Port USA, a Fulbright Scholar, and a Stheeman Prize or coauthored over 100 papers and formal presentations, including 32
recipient. He has organized over 70 scientific meetings. He was an Editior peer-reviewed papers cited over 900 times. He holds nine patents and has
of the IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS. managed company patent portfolios as an intellectual property strategist.

1280 PROCEEDINGS OF THE IEEE, VOL. 92, NO. 8, AUGUST 2004