Holographic Storage

Do we still need Holographic Storage?

When I first wrote about holographic storage in the year 2000, magnetic disks were about to hit the superparamagnetic limit, and commercial holographic storage systems were expected to replace them by 2005. 2005 is over, magnetic disk data densities are still have a few years to go, and there are still no commercial holographic storage subsystems available. So is this a dead subject? No, not if the researchers are to be believed. However we are close to the end of 2008 and it still looks like it will take another few years before a good, commercially viable system is available.

The researchers expected that improvements in conventional hard disk magnetic recording will push the superparamagnetic limit into the 100 or 200-Gb per square inch, then that could be doubled again by using perpendicular recording instead of conventional recording. In conventional recording schemes, the 'magnets' on a disk essentially lie sideways on the disk, while with perpendicular recording, the magnetic poles point up out of the disk surface. This means that the magnetic bits can be packed tighter together. Seagate launched the perpendicular recording 7200 Barracuda series in 2007, with a maximum capacity of 1TB. HDS has also launched a perpendicular drive.

The next major breakthrough is expected to appear in early 2009, based on lithography. The Swedish manufacturer Obducat had developed a lithographic process to create 'nanoholes' in a sheet of aluminum oxide. These holes are then filled with a magnetic material so each hole can contain a single bit of data. The aluminum substrate acts like an insulator between each hole and so helps defer the impact of superparamagnetism. The holes are about 17nm in diameter and will allow an areal density of about 750Gb per square inch.

OK, so would it be possible to push the superparamagnetic limit to the Tb/in.2 range? Maybe, but this would present other engineering challenges, not least the requirement to float the read-write heads about 2 nanometers above the disks. This is molecular distances. It would be difficult, but not necessarily impossible. The other problem is that it is hard to drive the serial data channel at acceptable speeds from terabyte disks, so the performance/capacity ratio will be low.

Holographic Data Storage (HDS) remains a tantalising possible way to resolve these problems, as in theory it can store data by the terabyte, and transfer megabytes of data in a single operation. It is expected that early implementations will be used for fixed content storage, or to replace tape or flash storage systems. Data archives need to be retained for up to one hundred years, but disk and tape storage has a life of ten to twenty years. Holography is expected to have a life span of one hundred years or more. That makes holographic storage ideal for long term archives, so to answer the question in the title of this section, yes, we do still need holographic storage.

Another possibility is the holographic systems being developed by Colossal Storage. They use ferroelectric binary-state molecules that can be in one of two states. These states can be read or changed with ultra-violet light and so nano-optical holographic storage devices can be created. This is essentially an atomic switch with the potential to hold petabytes in a square inch.

How does Holographic Storage work?

In holographic data storage, an entire page of information is stored at once as an optical interference pattern within a thick, photosensitive optical material. This is done by intersecting two coherent laser beams within the storage material. The first, called the object beam, contains the information to be stored; the second, called the reference beam, is a simple light wave. When the two combine in an optical storage medium, they change the chemical or physical construction of that medium and so store the data.
If the storage medium is then illuminated with the reference beam again, the object data beam is produced. The combination of the two beams writes a complete 'page' of data into the crystal, and the single reference beam reads a complete page of data out of the crystal.

The photo sensitive material within which data is stored in a 1024*1024 bit array, called a page. Each element of the array is light-or-dark, or one-or-zero, so each page represents one Megabit of information. The entire page is processed in parallel, so an entire megabit is read in one operation.

As the hologram is three dimensional, several pages of information can be recorded in the same piece of material as long as they are distinguishable from one another. Two ways to do this are to change the angle between the object and reference wave or by changing the laser wavelength. Any particular data page can then be read out independently by illuminating the stored gratings with the reference wave that was used to store that page. In theory, a single crystal could store terabytes of information, and access time should be fast, because the laser beams can be moved rapidly without inertia, unlike the actuators in disk drives.

It should also be possible to find a particular page quickly too. If you combine a search pattern with an object beam and illuminate the crystal, all the reference beams that were used to store data will be produced. The reference beam with the highest intensity will be the one that most closely matches the search pattern.

There is an alternative architecture called the co-linear system. The reference beam and the data beam are combined together into a single 'light pencil', and that avoids the complex optics needed to maintain two light beams at a precise, separated angle. This is usually implemented by combining light from a green laser and a red laser.

HDS Components

A LASER which is split into two beams, a reference beam and an object beam. The interference pattern created by these two beams forms the hologram.

A Spatial Light Modulator (SLM) which is basically just a 1024 * 1024 array of light or dark squares. This array represents the data to be stored, and is usually implemented by a set of pixels on an LCD. An SLM can usually be refreshed at rates of about 1000 frames per second.

A Multiplexing Agent which is used to allow the laser beam to access different pages in the hologram. There are two types of multiplexing, shift multiplexing and angular multiplexing. Shift Multiplexing uses a rotating disk to vary the angle of the laser beam, and so access a different view of the hologram. Angular multiplexing uses mirrors to change the angle at which the laser strikes the crystal. SM is easier to implement, AM has lower intrinsic noise levels.

A Storage Medium

Three different storage media have been tested.
Fe-doped lithium niobate, now generally not considered sensitive enough for practical applications
Photo-addressable polymers or PAPs. The molecular structure of the material changes if they are irradiated, but the problem with PAPs is that their volume changes when data is changed.
A two chemistry polymer used by Inphase gets round the volume / state change issue
A glassy 'holomide' substance used by Polight

A Charge Coupled Device [CCD], an array of sensors which corresponds to the pixels on the SLM. The CCD is used to read the interference pattern from the reference beam, and so read the information from the hologram. The matrix construction of the CCD allows it to read all 1Mb of the data at once. Typical CCD dimensions are one square centimeter, and typical access rates are 1000 frames / second, or 1 Gigabit / second.

Some HDS components

The diagram is simplified (to fit with my artistic skills). The reference laser beam, and the multiplexing system are not shown.
As an example, when polarised laser light passes through a photo-addressable polymer (PAP) its chain-like molecules become aligned and stay like that even after the beam has been turned off. The holographic effect is created by shining two laser beams that are in phase onto the PAP. One of the beams, the data beam, falls first on an object which encodes the data, in this case a liquid-crystal display 'template'. This changes its phase. When the two beams meet on the polymer an interference pattern indicating the difference between their phases is etched into the substance. Then, by adjusting the angle of the beam slightly, an entirely new pattern can be recorded on the same substance without disrupting any of the information already recorded.
The PAP alignment can then be read by shining an unpolarised laser beam through the polymer. The beam picks up the pattern in the PAP, and it is then read by the CCD

Why is Holographic Storage not working yet?

The components needed for holography are generally available, and reasonably low-cost. The technology includes Liquid Crystal Display (LCD) and Charged-Couple Device (CCD) camera chips, both of which have been around for some time. So why are there no HDS systems in production yet? Well, they are still almost here. The TLA (Three Letter Acronym) of the moment is HVD or Holographic Versatile Disc.

Some of the reasons why commercial products are not available yet are -

• If the system uses a two beam approach then the reference beam and information beam are configured in two different optical axes. This makes the optical system complex.
• There are no agreed holographic recording media standards yet, which means no co-operation between manufacturers
• The industry has not decided yet if the holographic media should be compatible with other optical systems like CDs and DVDs.
• Holographic systems require an exact alignment of the recording plane and this makes bulk manufacture difficult.
• The production of a new holographic substrate requires substantial investments in new equipment, unless old technologies can be adapted.

Are there any Holographic Storage systems on the horizon?

The main problem with holographic storage is the expense required to tool up for a new technology, when the old technology, magnetic disks, still has growth potential. However, it appears that holography will no go away.

Plymouth University (UK) are working on opticube, a device that will provide a 1 terabyte of holographic storage, but whether or not this will get from an academic curiosity to a practical, marketable device remains to be seen.

InPhase Tapestry systems

InPhase has been shipping optical media since 2001, and in February 2004 extended its Tapestry brand of green-wavelength media to include blue wavelength media for use with 400 to 410 nm lasers developed for blue-laser DVD.
Although it is a small company, InPhase has managed to create a working holographic storage medium, and it sells that media on to major companies world wide to companies developing next-generation optical drive systems. InPhase media customers that have publicly stated they are investigating holographic storage include Pioneer, MEI, NHK, Sony, Thomson, Samsung, Daewoo, JVC, and Optware. Companies that have purchased testers include NHK and Sony.

InPhase ran a public demonstration of its 'tapestry' holographic technology in April 2005. InPhase plan to release their first generation drive in summer 2007. These will be write-once WORM discs designed for fixed content archiving with a capacity of about 300 GB. Re-writable discs with 1.6 TB capacities are planned for 2009. As well as enterprise class storage, InPhase are also considering small consumer targeted devices with capacities ranging from 2 GB on a postage stamp to 210 GB on a credit card, which could threaten Flash memory sales. They predict that the 200R will have a media shelf life of about 50 years, compared to 10 years for tape.

InPhase claim that the tapestry system has a recording density of 200 Gigabits / square inch and can read data at 20 MB/s. Contrast this with traditional magnetic disk that has a recording density of 120 Mbpsi and the newer perpendicular recording disks with a density of 240Mbpsi. High performance tapes can read data at 120 MB/s or more. The reason why the tapestry read speed is relatively slow, is because the DVD has to stop while it writes to each sector. This architecture needs to change so that the disk is spinning continuously and the data transfer becomes 'bitwise'.
Tapestry uses a twin polymer implementation for the storage medium. The recording medium polymer is dissolved inside a solid matrix polymer; This 2-chemistry combination is a 1.5 mm thick recording material that this is sandwiched between two plastic plates; there is no metallic layer such as used in DVD storage. Data is stored by crossing two separate laser beams inside the polymer, which records pages of data. Individual pages can hold approximately 1 megabit, and multiple pages are recorded by varying the angle of incidence and wavelength of the reference beam. 252 'pages' are collected together into one 'book', and fifteen books or 3780 pages can all the stored in the same piece of recording material.

Optware Collinear Holographic Versatile Disc (HVD)

The holographic systems described so far all use two laser beams to create an interference pattern in the storage medium, and it can be difficult to maintain the required precise alignment between these two beams. Optware changes this by combining two laser beams into a single co-linear light beam. A green laser beam is used to carry page data, and a red laser beam for tracking and controlling the signal. The two beams are combined by a dichroic mirror then focused by a lens onto the polymer resin recording medium. The data is then stored as an interference pattern in the holographic storage medium. A shiny metallic surface on the bottom of the disk then reflects that data back up to the laser to be read.

The reference beam is needed to read the data back, but the data cannot be read unless the exact reference pattern in used. This means that the data on an HVD can be encrypted very easily, and very securely, as each two dimensional page can have a different encryption key. The HVD could also protect from piracy, as it would be very easy to mass produce copies from a master disk, but the copied disks could not be used for further copying.
Optware plan to ship 200 GB capacity disks sometime in 2007, with a roadmap to increase capacity to 1TB then 3.9TB on a CD size disk. They will also produce a 30GB credit card size disk for consumer products, and a 'consumer' disk holding 100GB in 2008. The eventual data transfer rate is planned to exceed 1 Gbits/sec or 125MB/s, about three times the initial InPhase Tapestry transfer rate.

back to top

By entering and using this site, you accept the conditions and limitations of use