How often have you heard the expressions "it's in the mail," or "always a bridesmaid, never a bride," or "it will be ready next year"? I expect, if you're like me, more than a few times. I'm probably guilty of occasionally uttering them myself. This is especially true when talking to someone about holographic storage -- the product always seems to be "just around the corner." I realize there is an induction period for turning a technology into a product, but 40 years seems a little long!
It was shortly after the invention of the laser that holography became practical and the principles of holographic storage established. Holography is truly an elegant technology -- it allows an optical wavefront to be stored in such a way that it can be completely retrieved at a later time with all its properties intact. Amazing, really. In the case of information storage, the information-bearing wavefront comes from a page composer, such as a liquid crystal display. This wavefront is interfered with a second wavefront, called the reference wavefront, and this interference pattern is stored in a light-sensitive medium such as an inorganic crystal or polymer material. The stored pattern is the hologram. Illuminating the hologram with only the reference wavefront causes the other wavefront to be regenerated such that it can be imaged onto a detector array. Depending on the format of the medium, for example, by using different reference beam directions, many pages can be superimposed and independently recovered.
Over the decades, many critical issues have been resolved with novel and creative contributions made to the technology resolved with novel and creative contributions made to the technology. Finding a suitable recording material, however, has been an ongoing quest. This search is not over yet. We can specify the properties of the ideal material. They are:
- High optical quality (distortion) and low scatter (noise)
- Thickness (0.5-1.0 mm) to allow many holograms to be stored in one area and distinguished from each other
- High &Mac198;n, also so many holograms can be recorded in one area
- Low shrinkage
- High sensitivity so the recording can be fast
- Self-processing
- Erasable so the material can be reused
- If erasable, it must be fixable so the data is not volatile
- If not erasable, then it should be write-once
- Sensitive at laser diode wavelengths for packaging
- Long shelf-life
- Cheap
No one material has all these qualities; some are close, such as photopolymers. These may be the best choice for early holographic storage systems. The pace of development has accelerated in the past several years with the creation of two consortia and more recently, the formation of start-up companies to commercialize holographic storage.
In 1994, a consortium named PRISM (photorefractive information storage materials) was established that consisted of several companies and university research centers. This consortium, funded on a cost-sharing basis with DARPA, was to address key issues relating to the storage material. Of particular interest was to seek solutions to the volatility of holograms recorded in photorefractive materials. The same mechanism that permitted the recording of holograms in real time, electronic excitation and trapping, caused the holograms to be partially erased when a new hologram was added (superimposed) or when any hologram was read out. Because of this, the strength of the data page(s) would slowly decrease to the point that the reliability of data recovery would be compromised. What was needed was a way to "fix" the holograms to make them impervious to subsequent illumination.
Thermal and electrical fixing processes were investigated to convert the holograms to a less mobile and hence more stable species. These met with some success although thermal fixing (heating the crystal to about 120°C) caused some loss in signal strength and the long time-constants required for heating and cooling made the application of this method only practical for read-only storage. Electrical fixing was limited to small angles between the object and reference wavefronts and, as such, limited capacity.
A more elegant solution to controlling hologram volatility is to use two-color recording. One wavelength (color) is used to pump the electrons into a metastable state and a second wavelength is used to create the interference and record the hologram. Both wavelengths must be present in order to record -- neither one can do the job by itself. Reading the hologram requires illumination with only the recording wavelength and hence does no damage. Uniform illumination with both wavelengths allows erasure. While this process permits true read/write operation, the power densities required are still high (1 watt/cm2) and the recording speed too slow. Research continues to be performed on this important technique for controlling volatility.
To remove data volatility from the equation, one must give up erasability and consider developing a WORM (write-once read-many) recording material. Hologram formation still happens in real time (i.e., no post-processing is required), but the recording process is irreversible. Light-induced polymerization of a monomer is a candidate for this process. The list of desirable material attributes presented above is a challenge for polymer systems.
Shortly after the start of the PRISM program, it became obvious that another program was required to develop the enabling technologies necessary to build a complete holographic storage system and then to assemble these components into demonstration platforms. To do this, a new consortium was created between some of the PRISM participants and several new partners. It was funded in a similar cost-sharing manner with DARPA. Called HDSS (holographic data storage systems), this consortium developed a 1 megapixel spatial light modulator, a 1 megapixel detector array, laser, optics, multiplexing scheme, signal processing and coding techniques necessary to demonstrate a total capacity of 1 terabit, and a data transfer rate in excess of 1 gigabit/sec. Recording was done on a thick photopolymer disk developed by PRISM. Both programs successfully completed their research in late 2000.
Out of these consortia, emerged a spin-off company from one of the partners, Polaroid. Called Aprilis, Inc. (www.aprilisinc.com), this company is responsible for development of the photopolymer material under PRISM and used in the HDSS demonstration. They are continuing to refine the material and are planning on expanding their activity with the development of a holographic storage system.
Another center of activity was located at Lucent. They also developed a photopolymer recording material and used this material to study storage architectures and demonstrated the recording of holograms with an effective storage density of 45 channel bits/mm2 (29 Gigabits/in2). They estimate that 100 Gigabits/in2 can be achieved in this material (in the inorganic crystal lithium niobate, 225 Gigabits/in2 has been demonstrated). This activity spun-off a company called InPhase Technologies (www.inphase-technologies.com). There are several other companies and university research labs looking at various aspects of holographic storage but most of these activities are sub-critical in manpower and resources.
While there have been impressive demonstrations of the potential of holographic storage technology, it is not clear what premium the user will be willing to pay for enhanced performance. Certainly there are some niche markets where cost is less of an issue, but in general, the cost per megabyte will be the deciding factor. Even with a viable storage material, a holographic storage system has several costly components, such as the spatial modulator, detector array, laser, and optics, to name just a few. Unfortunately, the requirements placed on these components are such that one cannot take advantage of mass markets to reduce their cost. For example, one of the key performance attributes, data rate, requires a detector array that will operate far beyond video frame rates. Except for special-purpose streak cameras, I know of no other application that has this requirement and generates a mass market to influence their cost.
Because of these considerations, I doubt very much that we will ever see a high-capacity, high-performance holographic storage system attached to a personal computer. Perhaps at the server level, the capacity of the system can be sufficiently large to amortize the component cost and make this a cost-effective solution.
At the risk of using well-worn expressions: "the clock is running," "the window of opportunity is closing," and "the competition is keen." Do we really have the luxury of waiting "till next year?"
Glenn Sincerbox is a professor of optical sciences and director of the optical data storage center at the University of Arizona in Tucson, Arizona. He joined the faculty in 1996 after a 34-year career with IBM Research. He has published and presented extensively, holds 48 patents and is active in international conference management. He is a fellow of the Optical Society of America, a member of the SPIE and is currently treasurer of the International Commission for Optics. His interests include holographic and optical storage, displays and inspection technology. Until his retirement from IBM, he was co-PI for the two holographic storage consortia, PRISM and HDSS. The interested reader is directed to the recent book on Holographic Data Storage edited by Coufal, Psaltis and Sincerbox.
3255 Scott Blvd / Suite 2-102 / Santa Clara, CA / 95054-3013
tel 408.330.8100 / fax 408.492.1425 / www.idema.org
IDEMA is a registered trademark of the International Disk Drive Equipment and Materials Association.
This article first appeared in the September/October 2001 INSIGHT magazine. Reprinted with permission.
© Copyright 2001 IDEMA. All rights reserved.
|
