Rust In Peace

Published Date
01 - Mar - 2006
| Last Updated
01 - Mar - 2006
 
Rust In Peace

Storage researchers have their work cut out all nice and neat: "all" that's needed is a way to make 1s and 0s. A Yes and a No; a hole and a bump; a groove and a ridge. This means a lot of freedom to experiment-there are a lot of materials and mechanisms researchers can tinker around with, with the goal of finding what materials can be in two states (or be brought into two states), and how to go about making those state changes. Given a decent lab, you could probably think up a storage mechanism yourself-with the caveat that what's needed is higher densities!

When you look at what's being researched and actually think about our current storage devices (and those in the past, such as punch cards), you'll see that they're primitive. They almost seem stupid. Holes in paper-how straightforward can you get? The Hard Disk, too, is based on an extremely simple, almost simplistic principle-different magnetic orientations on a layer of iron oxide.

It turns out we can now do much better. From a fairly lofty perspective, therefore, the hard disk is dead. This century belongs to nanotech, the science of the extremely small. And small is, of course, good news when it comes to storage, because it means higher densities. In what follows, we discuss several storage techniques in various stages of research and implementation, and with very different theoretical caps-from the one CD per square centimetre achieved using a type of rubber stamp, to the beyond-the-imagination one crore CDs on a square inch that might be achieved with the help of carbon nanotubes. Some of these technologies you'll see in the next couple of years; some you'll see sometime in your lifetime. Some will die because of competing technologies, some will tame the world's information. But they all demonstrate that innovation is alive, and kicking as hell!

Holographic Storage
Holographic Storage has been talked about for a long time; indeed, the Holographic Versatile Disc (HVD) is being eagerly awaited by people around the world. This technology uses lasers to record data in the volume of the medium, rather than on the surface. The idea, surprisingly, is not new, but it's only now that the technology seems to be getting up to speed.

How Does It Work?
A laser beam is split in two, the reference beam and the signal beam (called so because it carries the data). A device called a spatial light modulator (SLM) translates 1s and 0s into an optical pattern of light and dark pixels, as in the figure above. These pixels are arranged in an array (or "page") of about a million bits.





The signal beam and the reference beam intersect in the storage medium, which is light-sensitive. And at the point of intersection, a hologram is formed because of a chemical reaction in the medium, and gets recorded there. (A hologram is the interference pattern that results when two light waves meet; for more on interference, visit
www.physicsclassroom.com). For reading the data, only the reference beam is used: it defelcts off the hologram, and a detector picks up the data pages in parallel. The 1s and 0s of the original data can be read from the data pages.

By varying the angle or wavelength of the reference beam, or by slightly changing the placement of the media, lots of holograms can be stored in the volume of the medium.

What Work Is On?
Optware Corporation of Japan have already come out with their HVD: the HVD holds 1 TB (a terabyte), and is the same size as a regular optical disc. Enterprise versions were planned for this year: the estimated costs were something like $20,000 (Rs 9 lakh) for players and $100 (Rs 4,500) for discs.

In the meanwhile, InPhase Technologies, Optware's main competitor, is coming out with products of its own. In partnership with Maxell, InPhase has already come up with a 300 GB disk, with an 800 GB disc expected in 2008. And if the 1 TB of HVD weren't enough, Fuji Photo Film USA has demonstrated a type of HVD with a claimed capacity of 3.9 TB!

What Promise Does It Hold?
Besides high storage densities, holographic  storage means fast access times, because there are no actuators as in hard disks-laser beams can be focused around much more rapidly.

Holographic storage can kill off the hard disk within the next 10 years, but that's again speculation-some technologies step out quickly and graciously, but some are pretty stubborn!

Near-Field Optical Recording Using A Solid Immersion Lens

Think of a CD or DVD; while recording, the lens focuses the laser onto a tiny spot on the medium. This spot is tinier for DVD than for CD, and is even tinier in Blu-ray, for example. Near-field optical recording (NFOR) refers to the extremely sharp focusing of a laser beam, which means an extremely small distance between the lens and the recording medium. NFOR using a solid immersion lens (SIL) would be the child of Blu-ray and HD-DVD, and therefore, the grandchild of the DVD.



How Does It Work?
The density of the data that can be achieved on a disc is roughly proportional to the square of the numerical aperture (NA) of the lens, and inversely proportional to the wavelength of the laser (refer The Battle Of The Blue, Digit December 2005). The NA of a lens dictates how sharply it can focus the beam falling on it. The NA of a SIL is made very high, and the achievable data densities are therefore that much higher.

In NFOR using a SIL (refer figure alongside), the laser is very sharply focused: it converges at a point within the lens, instead of on the medium. The air gap between the lens and the medium is just about 25 nm! The photons "tunnel" through the air gap onto the surface of the medium (See Jargon Buster).

What's Being Done?
About half a year ago, Philips researchers reported "significant progress" in developing NFOR. Up to 150 GB of data on a dual-layer disc would be possible, they said, although they also said the technology was several years away from commercialisation.

What Promise Does It Hold?
NFOR using SILs seems in no way to us a hard disk killer, but it's the natural progression from blue-laser systems. NFOR is therefore a potential "Blu-ray killer." The promise NFOR holds all seems to hinge on how much comes in in terms of research dollars. Will the HVD have already taken over by the time NFOR takes off?

MEMS-Based Storage.
MEMS (Micro-Electro-Mechanical Systems) is, according to memsnet.org, "the integration of mechanical elements, sensors, actuators, and electronics on a common silicon substrate through microfabrication technology." The mechanical elements referred to here, range in size from a few micrometers to a millimetre. Actuators are just devices that convert an electronic signal to a physical action-for example, the device in the hard disk that is responsible for positioning the head precisely. In fact, we can take the example of a MEMS-based storage system to better explain what MEMS are.

How Does It Work?
Different MEMS storage systems work differently, but we can describe the concept. Take a look at the figure alongside. This isn't a working system, but just an example of a general MEMS storage system. The data "sled" at the top can move in all three directions; it is spring-mounted over the probe tip array, an array of mechanical tips that do the reading and writing (we aren't getting into the details here). There's an actuator on each side of the data sled, and it moves the sled in response to electric currents. Now when the first bit is written, the sled and the tip array are aligned, and then the sled moves along one axis while the tips do their work-writing a 1 or a 0. Note that the sled doesn't rotate; it slides. Also note that everything in this arrangement is mechanical and electronic.

What's Being Done?
At CeBit 2005 in Hannover, Germany, IBM showed off a MEMS-based storage device that it said could achieve densities in the range of 1 TB per square inch. The device is called the Millipede, because of the thousands of probe tips. The tips are of silicon, and the data substrate is a material called "plexiglass." To write a bit of data, a tip is heated to 400 degrees C. When it "pokes" the plexiglass, it softens it and makes a dent there. To read data, the tips are heated to 300 degrees C and pulled across the surface of the plexiglass. When it falls into a dent, the tip cools down because more surface area comes in contact with the (cooler) plexiglass. The temperature drop reduces its resistance, which can be measured. Finally, to erase a bit, a hot tip is passed over the dent, making it pop back up.



What Promise Does It Hold?
Plenty! MEMS-based storage devices such as Millipede could well be hard disk killers, depending on the research dollars spent. Seek times are lower and more stable than those of hard disks.

In the range of 1 to 10 GB, MEMS-based storage has the lowest cost per byte compared to non-volatile memory and hard disks. Data transfer rates can reach 1 gigabyte per second. Also, MEMS-based devices are smaller and use considerably less power.

Jargon Buster 
Spatial Light Modulator (SLM)
An SLM consists of an array of optical elements-pixels-in which each pixel acts independently as an optical "valve" to adjust or modulate light intensity.
Solid Immersion Lens
A type of lens that has a high refractive index, hemispherical in shape, optimised for precision.
Polymers
These are natural or synthetic plastic-like structures where two or more like molecules are joined to form a more complex molecular structure.

Photolithography
This is a process used in semiconductor device fabrication to transfer a pattern from a "photomask" to the surface of a wafer or substrate. Here, a chip, typically of silicon, is coated with a chemical called a photoresist. Flashing a pattern of light and dark onto the photoresist causes it to harden in the areas exposed to light. The parts not exposed to light stay soft and are etched away. (A photomask is a high-precision plate containing microscopic images of electronic circuits.)

Electron beam lithography (EBL)
As opposed to photolithography, in EBL, electron beams are used to create the patterns directly on-chip.

Scanning Tunnelling Microscope
The STM allows scientists to visualise regions of high electron density and hence infer the position of individual atoms.

Atomic Force Microscope (AFM)
This kind of microscope works by scanning a semiconductor tip over a surface. The tip is positioned at the end of a cantilever beam. As the tip is repelled by or attracted to the surface, the beam deflects. The magnitude of the deflection is captured by a laser that reflects at an oblique angle from the very end of the cantilever. A plot of the laser deflection versus tip position on the sample surface provides the resolution of the topography of the surface.

Tunnelling Current
Tunnelling is a quantum mechanical effect. A tunnelling current occurs when electrons move through a barrier that they "shouldn't" be able to move though. In the quantum mechanical world, electrons have wave-like properties. These waves don't end abruptly at a wall or barrier, but taper off quite quickly. If the barrier is thin enough, given enough electrons, some will move through and appear on the other side. When an electron moves though the barrier in this fashion, it is called tunnelling.

Cantilever
Simply speaking, a cantilever is a projecting structure supported only at one end, like a diving board.

Molecular Switches
Since we're talking miniature, let's go all the way: how about storing a bit in a single molecule? Short of using a single atom-which is more difficult -this is about as small as it gets. (Later, we do talk, though, about single atoms!) Like we've said, it's all about 1s and 0s; so what one needs is a molecule with two stable states, and a means of switching the molecule between these states.

How Does It Work?
A molecule currently being researched is "rotaxane." Now this may sound way out: the rotaxane molecule has a thread-like section, with a ring structure that moves around it. The ring doesn't fall off because of the dumbbell-type bulbs at the ends of the thread. Ways have been found to move the ring from one end of the thread to the other, both of which are stable states for the molecule. The reading procedure is complex, and cannot be discussed in this space.



What Is Being Done?
Scientists successfully demonstrated, in January 2003, that they could store information using rotaxane molecules. They reportedly achieved storage densities in the range of 1 to 10 gigabits per square inch.

What Promise Does It Hold?
There's currently more research going on in terms of using molecular switches as logic gates and such. But researchers are already envisaging fully-molecular computers, where the storage mechanisms, too, would be molecular. This is looking way into the future, but fully-molecular computers are an important idea-they represent the smallest you can get!

Atomic Storage Using STMs And AFMs
Take a look at the figure alongside. Notice the "IBM"? It's not a photograph, but it depicts what researchers have etched at the atomic scale: each of the little hills is an individual xenon atom! The thing was produced using an Scanning Tunnelling Microscope (STM) operating a few degrees above absolute zero.

How It Works
An STM has the ability to give a view of surfaces at the atomic scale, and researchers have envisioned the application of the technique to achieve ultra-high-density storage. The STM has an ultra-sharp tip placed extremely close to the substrate being written onto. A voltage applied between the tip and the substrate gives rise to a tunnelling current. The tunnel current depends on the separation between the tip and the substrate. As the tip is moved over the surface, the tunnel current is monitored, and the position of the tip is changed such that the current is constant-this way, the topology of the surface can be mapped out. The beauty of the STM is that it can be used not only to map a surface, but also to modify it.



There are difficulties with the STM approach-one is the problem of maintaining the distance between the tip and the surface at the angstrom level (an angstrom is 0.1 nm). To overcome these difficulties, researchers are concentrating more on the Atomic Force Microscope (AFM). Here, the tip rests on a cantilever spring. This allows for two things: first, the tip can actually touch the surface, because of the "bounce" enabled by the spring. Second, by monitoring and controlling the spring, extremely small forces can be sensed as well as applied.

What's Being Done
The letters "IBM" have been etched on a surface using an STM, as mentioned above. Researchers are playing around with the idea of using an AFM in contact with a hard disk-like surface to etch data pits. It will take a long time for this to materialise, but remember that we're talking about a "hard disk" that writes individual atoms! Disk storage just cannot get any denser than that now, can it?

What Promise Does It Hold?
The use of STMs, AFMs, and similar devices are almost the ultimate application of nanotechnology to data storage. The potential storage capacities are enormous-for example, the etched letters in the figure represent a storage density of 1 million gigabits per square inch! That's 2 lakh CDs on a square inch!

Rubber Stamps
SIn 2001, two Harvard University researchers designed a high-capacity data storage device based on the concept of a rubber stamp transferring patterns-1s and 0s, of course-onto a plastic film. The system could hold about 650 MB (one CD) of data on a square centimetre.

How Does It Work?
The method works on the idea of using electric charge-rather than magnetic orientations-to store bits. First, a pattern of 1s and 0s were etched on a piece of semiconductor. This piece was a mould for a liquid polymer, and solidifying the polymer resulted in the rubber stamp (see figure alongside). The 5-mm-thick stamp was then coated with an 80-nm gold film.

To write the pattern, the researchers placed the stamp on a silicon wafer covered with an 80-nm film of a substance called polymethyl methacrylate (PMMA), a polymer that can hold electric charge. They then applied an electric pulse to the gold plate. The current passed right through the PMMA film to the silicon wafer, and what resulted was an electric charge in the areas of the PMMA film that were in contact with the stamp. Where there's an electric charge, it's a 1; where it's not, it's a 0.


1.The rubber stamp with the gold film 2. The stamp in close contact with a PMMA-coated surface 3. The PMMA holds the stamped electric charge

What Promise Does It Hold?
The system is low-cost, which is its main advantage. But this seems to us an exotic piece of research, and its immediate or long-term value is not clear. What is also not clear is how data would be transferred between two systems: the rubber stamp was created using photolithography and electron beam lithography, so how would data patterns be created on an everyday basis? Still, the research shows that it is possible to use electric charge storage as an alternative to magnetic storage; and that it can lead to high data densities. So OK, here's yet another nanotech hard disk alternative!

Multiwalled Carbon Nanotubes
Carbon nanotubes have been much researched. They are folded sheets of carbon atoms, as in the picture alongside, where several nanotubes are shown in different colours. Multi-walled nanotubes are concentric tubes that hold together as a structure, as in the picture shown below. They're "nanotubes," so, of course, they are measured on the nanoscale-a nanotube can be smaller than a nanometre in diameter. It turns out that nanotubes can be used to write data: the tube is treated like a needle to make fine changes in a medium.


A bunch of single-walled carbon nanotubes. This isn't a photograph, but represents nanotube structure

What's Being Done
Researchers from IBM Research in Zurich, Switzerland, the Japanese Nanotechnology Research Institute, and Osaka Prefecture University in Japan, have demonstrated a storage density of 250 gigabits per square inch, using the tips of multiwalled nanotubes to write bits onto a film of a certain polymer.


A double-walled carbon nanotube is one tube inside another

The nanotube tip works something like a probe, pressing 1s onto the polymer surface; obviously, the absence of a 1 means a 0.


A multi-walled nanotube can be used like a pen to write data onto a polymer

What Promise Does It Hold?
The Swiss and Japanese researchers said that using nanotube tips in practical devices could not be possible until around 2008. 250 gigabits per square inch is high, but not astronomical by any means; what is astronomical is the 50 million gigabits per square inch that is being envisaged in a different scheme! That's one crore CDs on a square inch! Here, nanotube tips are used to place hydrogen atoms on a diamond or silicon surface.

The state of the art certainly does not match up to what is possible in theory, but you never know when a nanotech breakthrough just happens!

How Much, And When?
If you've read this far, you probably want to know exactly when you're going to be able to upgrade! The graph above indicates storage densities (on a logarithmic scale) and timelines. Though the hard disk seems pretty mighty in this graph, there are other considerations-for example, they've pretty much reached their theoretical limit; there's only so much you can pack onto a single platter; and access times are shorter with, for example, MEMS-based and holographic storage.

With so many labs doing so much work, standardisation is going to be a huge issue. Also, the technologies will directly compete with each other-so even just two years down the line, the scene is likely to be quite different. In any case, it's almost time to look lovingly at your hard disk and pay your last respects!


Team DigitTeam Digit  teamdigit@digit.in

All of us are better than one of us.