Google may want to store every bit that you have ever flipped, but it faces the problem that current data storage technology uses a relatively low-density, 2-D approach. Of course, holographic data storage has been touted as the answer to this problem ever since, well, since the first hologram was demonstrated. Despite its potential, holographic data storage has failed to gain market share. This is because the current generation of optical and magnetic storage media are actually simple, robust, and just good enough to hold the competition at bay.
The upshot is that, until magnetic bits can no longer be shrunk and multilayer optical discs reach their limits, any new technology has to have all the good features of current data storage techniques and be better. A bunch of Aussies think they might have hit the sweet spot with a new multilayer optical storage medium that has the potential to store data at around 1.1Tb/cm3. A standard DVD clocks in at 51MB in a square centimeter in each of its layers.
At heart, the new medium looks rather like the old optical medium, consisting of multiple layers where data can be stored. However, instead of using a dye or a pit-island approach, the layers are filled with gold nanorods. The electrons in the nanorods will only respond strongly to an incident light if it has the right color and the electric field of the illuminating light lines up with the axis of the nanorod. When this occurs, the nanorod scatters light everywhere, glows like crazy, and heats up.
If hit with a sufficiently powerful laser, the rod will melt and change its shape, which also changes the color of the light it responds to. It is this step—exceeding the critical power density required to melt the nanorods—that is at the heart of this research.
The key feature here is that the nanorods' response depends on both the color and polarization of the light, which allows for multiple bits to be stored in exactly the same location. There are, of course, limitations to the technique. Only two polarizations should be used to minimize cross talk, and the colors must be separated such that no single nanorod will respond to more than a single illuminating color. Despite these limitations, this still clocks in at six to nine bits per location, and the researchers have already demonstrated that a ten layer medium can be written to.
Reading the information back is a little more complicated, but not by very much. The problem here is distinguishing bits written using the same color and polarization in different layers. To overcome this particular problem, the researchers looked at the nanorod's nonlinear optical response. They illuminated a layer with the same color and polarization used to write that bit, but instead of looking at the absorption or glow from the nanorods, they looked for a color that was much bluer. The bluer light comes from two-photon luminescence, where a nanorod absorbs two illuminating photons at the same time and emits one higher energy photon.
As with the write process, this requires a certain amount of power, so a low-power, sharply focused laser beam will only excite two-photon luminescence from a single layer. In keeping with their writing process, they demonstrated reading to ten layers deep.
So now for a few cautions: this was not demonstrated using a rotating disc, the read out process used a photomultipler tube rather than a photodiode, and the three colors are such that they cannot be obtained from a single diode laser. All this means that the read/write unit would require three laser diodes, a bulky vacuum tube, and a high-voltage power supply. The experiments were performed with a high power laser, but this is not too important, because the power densities used are readily achievable with low-power diode lasers.
Looking into the crystal ball, this is likely to be akin to the original CD technology, which was really a write once, read forever medium. However, unlike the original CD technology, writers should be relatively affordable, and, unlike (re-)writable CDs, these discs are likely to be reliable permanent storage.