In the Forums...
Posted: June 26, 2000
Written by: Tuan "Solace" Nguyen
The object beam finally interacts with the reference beam inside a photosensitive crystal. The ensuing interference pattern -- the substance of the hologram -- gets stored as a web of varying optical characteristics inside this crystal. To read the data from the crystal, the reference beam merely illuminates the crystal. The stored interference pattern that was recorded with the two interacting beams diffracts the reference beam's light so that it reconstructs the checkerboard image of the light or dark pixels. In other words, the original beam will appear when the reference beams strikes the page in the crystal. The resulting image is then emitted upon a charge-coupled device (CCD) sensor array, and it instantly captures the entire digital page.
When reading out the data, the reference beam must strike the crystal at the same angle that's used in recording the page. This incident angle is extremely important and it can't vary by more than a fraction of a degree.
If the angle is off by even a very small amount, it will not reproduce the data beam and the data page image. This is obviously a bad thing if you are talking about data integrity and accuracy, but actually, it can be a good thing overall. Why? The apparent flaw in the recording process actually helps store more data and achieves higher transfer rates. It's how holographic storage achieves its high data densities. By changing either the angle of the reference beam or its frequency, you can write additional data pages in to the same volume of crystal.
However, as you keep recording more data pages slightly away from previous pages, the holograms will begin to appear dimmer and fogged up because their patterns must share the material's finite dynamic range and the data page is physically etched into the crystal. Eventually you will run out of space to store because the crystal has depleted all of its physical storage capacity, sort of like write once, read many media such as CD-R. If more data is continuously written, the data images become so dim that laser beam noise creeps into the read-out operation.
The dynamic range of the recording medium will determine how many pages it can hold with strong integrity. The PRISM project is currently examining limitations in a various photosensitive materials. Current tests are conducted using iron-doped lithium niobate, strontium barium niobate, or barium titanate crystals. IBM is also interested in using organic material to be the medium because of its dynamic ability to change data.
The PRISM project has stored up to 200 holograms composed of 37.5-KB data pages (640 by 480 bits resolution) into a crystal with less than 1 centimeter on each side, achieving a storage density of 48 MB per cubic cm. You must be thinking “48 MB!? We’ve been developing holographic technology to store that little?!” Well, yes and no. Yes because that is an achievement of the technology previously not possible, and no because soon the same size volume of crystal will hold up to 10GB. The incentive is a page data density of 10 GB per cubic cm, a bit smaller than a standard gambling die.