Etapa de planicación

The promise of using holographic techniques for fast, high-capacity data storage has been around for many years, but a commercial device has yet to be made available. However, extensive research into this area is ongoing. One of the most promising routes appears to be the use of photorefractive materials, of which lithium niobate is a popular choice. Holograms store data in a volume of material, not just an area, and many holograms may be recorded in the same volume, by recording at different angles or wavelengths. These properties make the prospect of holographic data storage very attractive, as much larger amounts of data can be stored in a given area than is currently possible. Estimates suggest storage densities of 50 times that of current optical media are possible, with a similar increases in transfer rates. The storage of holograms in lithium niobate is achieved by illumination of the crystal with a light pattern, which excites electrons from defects, such as impurities. Most lithium niobate crystals contain a small concentration of Fe impurities (Peterson et al., 1971) so electrons are easily excited from Fe2+_{to the conduction band, from}

where they migrate by diffusion, drift and the bulk-photovoltaic effect until they are captured by Fe3+ions in dark regions. The trapped charge creates space-charge fields, which affect the refractive index, through the electro-optic effect. Subsequent illumination of the hologram with with a homogeneous laser beam will result in the beam being diffracted by the space-charge fields and so the original optical pattern can be recreated.

To increase the extent of the photorefractive effect, lithium niobate crystals intended for use in holographic storage are purposely Fe-doped. Staebler and Phillips

(1974) produced an early report on the use of iron-doped crystals for storing phase holograms. They found that the once formed the hologram can be erased by exposure to a uniform light beam, which excites the trapped electrons and so allows them to redistribute uniformly. It was reported that only 12 mJ/cm2 of radiation at 488 nm was required to erase a hologram in reduced lightly-doped lithium niobate (0.05 % Fe), which was nearly three orders of magnitude less than had been previously observed.

Shah et al. (1976) studied eight Fe:LiNbO3 crystals with varying dopant concen-

trations, half of which had been annealed in oxygen, while the other half were unannealed, for use in holographic storage. The photorefractive sensitivity, defined as:

S=(∆n/∆E)Γ (2.25)

where Γ represents the ratio of the average beam intensity in the crystal to the incident intensity and ∆n is the refractive index change caused by the incident energy density, ∆E, was shown to be mainly sensitive to the concentration of Fe2+ _{ions. The crystal with the best photorefractive sensitivity was found to be}

unannealed with 0.05 % Fe concentration and with 20 - 25 % of the ions in the Fe2+ state.

Shah et al. also used IR spectra to identify hydrogen impurities, which are known to reduce the photorefractive sensitivity, (Smith et al., 1968). Hydrogen diffuses into the crystals during poling or annealing and combines with oxygen to form OH ions, which have a characteristic vibrational band in the near infrared at∼_2.86µm. The

motion of protons in lithium niobate crystals has been examined by Engelsberg et al. (1993), who found that there are two temperature dependent regimes of motion. At low temperatures (around room temperature) there appears to be a fast, fluctuating motion of the hydrogen ions between normal and interstitial sites, which suggests the intersitial sites are occupied for a substantial proportion of the time. This could be the reason for reduced photorefractive sensitivity and increased resistance to optical damage.

A problem with holographic data storage which still needs to be addressed is the durability of the data. Retrieval of the holograms can cause the charge fields to be reduced, due to the homogeneous light which is used to read the hologram. The space-charge fields also gradually decay even in the dark, giving a dark storage time, which is a measure of the time until the grating amplitude drops to 1/eof its initial value (Nee et al., 2000).

Holograms are often recorded using ultraviolet light, however, this can result in erasure of the holograms during reading. An alternative is to use infrared pulses, but Fe:LiNbO3suffers from poor sensitivity at these wavelengths. To overcome this,

Buse et al. (1995) proposed used a spatially homogenous pulse of green light prior to recording of holograms using infrared illumination, which sensitises the samples for recording and can also erase previous holograms. Reading the hologram with infrared light was found to be non-destructive.

A system using doubly-doped lithium niobate crystals has been recently proposed, by Buse et al. (1998), using iron and and manganese, which improves the stability of stored holograms. The doubly-doped crystal was found to be photochromic, i.e. the absorption of light of a particular wavelength affects the absorption characteristics of light of a second wavelength. In practice, this means that recording is carried out using a combination of red and ultraviolet light. The ultraviolet light, from a mercury lamp, uniformly illuminates the sample and the red light is from a laser source, and contains the data to be recorded in the hologram. For reading the hologram, only the red laser is required. The manganese ions can only be excited by ultraviolet light, because their energy levels are deeper, so the hologram is not degraded by repeated reading, and hence the data can be stored with greater reliability.

During the recording phase, the ultraviolet light excites electrons from deep Mn2+ levels, to the conduction band and the iron levels, forming Fe2+_{ions which can then}

absorb the modulated red light, which is also present, thus exciting more electrons, which diffuse then recombine with Fe3+ and Mn3+ ions to create space-charges fields and consequently a hologram is formed. The electrons will be trapped in Fe2+and Mn2+ions, and while ultraviolet light is still present, electrons will be still be able to move between iron and manganese. However, as red light only excites electrons from iron, the build up of charge in manganese levels will begin to mirror the spatial modulation of the red light.

Once the recording phase is complete the ultraviolet illumination is removed. The hologram can then be reconstructed using homogeneous red light. As before, this will quickly ‘wash-out’ the charge pattern stored in iron traps, but the charge stored in manganese will remain unaffected by the red light, as it is does not have enough energy to excite the electrons from the deeper energy levels. Consequently, the hologram will remain fixed in the crystal, allowing non-destructive read out. The hologram can be simply erased by uniformly illuminating the crystal with ultraviolet light.