Optical storage – a brief history
has already briefly covered the early days of digital storage and the development of the first hard drives by IBM in the 1950s. Optical storage as we know it today required the culmination of several technologies that came into fruition at the start of the 1980s. One of the most important steps toward the creation of modern optical storage was the invention of the laser. In 1960, after an all-out race between scientists for its development, Theodore Maiman tested the first laser, based on a synthetic ruby crystal, at Hughes Research Laboratories in Malibu, California. Five years later, James Russell, a former employee of General Electric, joined the Pacific Northwest National Laboratory of Battelle Memorial Institute in Richland, Washington, where he developed the first prototypes of digital optical-recording systems, which many today consider ahead of their time.
Almost 15 years have passed, and in 1979 Philips and Sony set up a joint task force of engineers to design the new digital audio disc. Seemingly unaware of the work done by Russell, they drafted the basic standard for what became known as the compact disc, or CD. The standard was based on a previously existing technology called the LaserDisc, introduced to the market in 1978. Four years after the LaserDisc, Philips and Sony released to the Asian market the first CD player. The initial launch price was high—between $700 and $1000 for Sony’s CDP-101, the first commercial CD player. Initially CDs were used only to store audio, but in 1985 Philips and Sony began using CDs as data storage for the newly developed personal computer. The CD quickly became known as the CD-ROM.
Several years after the introduction of the CD, two consortiums started pushing a new standard for a higher-capacity optical media: the MultiMedia Compact Disc, backed by Philips and Sony, and the Super Density disc, supported by Toshiba, Time Warner, Matsushita Electric, Hitachi, Mitsubishi Electric, Pioneer, Thomson, and JVC. With a clear memory of the 1980s’ VHS–Betamax video-format wars, IBM’s president, Lou Gerstner, led an effort to unite the two camps behind a single standard. The attempt was a success: the MultiMedia Compact Disc was abandoned in favor of the SuperDensity Disc, becoming the digital videodisc or versatile disc, today known simply as DVD. The new format was supposed to hold 5GBs of data, but Philips and Sony insisted on a different version of encoding, reducing the disc density to a more familiar 4.7GBs—still more than seven times that of a CD. It took several years for the new format to mature, but in late 1996 Toshiba, Matsushita, Sony, and several other companies released the first DVD players in Japan, reaching the U.S. several months later. The first players cost more than $700 on their release, and the first computer DVD recorder, the Pioneer DVR-S10, hit the market in late 1997 with a whopping list price of $16,995.
As years passed by, DVD technology improved. While the DVD standard called for a 4.7GB disc, the first available media in the mid- to late-1990s held but 3.95GBs. DVD media continued to be developed and included DVD-RAM, DVD RW, dual-side DVD, and DVD dual layer (DVD DL), the latter having up to 8.5GB. Although the initial standard war was prevented in the mid-1990s, there was the threat of eruption a few years later with the introduction of plus/minus DVD media. However, unlike VHS and Betamax, the plus/minus variants were sufficiently compatible with each other to thwart a disturbing effect on consumers.
Like the CD of the 1980s that introduced a revolution in audio, the DVD of the 1990s radically changed the video world, dramatically increasing the amount of data users could save and back up from their computers. In recent years, a demand for even higher-capacity media capable of supporting high-definition (HD) video content led to the development of two new optical media technologies: Blu-ray Disc (BD) and the High-Definition DVD (HD-DVD). Unlike CDs that use an infrared laser (780nm) and DVDs that use red laser (650 nm), both BD and HD-DVD use blue-violet laser (405 nm), which allows for a much higher density and capacity. HD-DVD has 15GBs on a single layer and 30GBs on a dual layer disc (HD-DVD RAM will have 20GB), while BD has 25GBs on a single layer and 50GBs on a dual layer. Although many attempts have been made to unite BD and HD-DVD into a single format, the rivalry is still ongoing to date, slowing the adoption rate and therefore increasing prices which are currently too high for most users. LG, Samsung, and other manufacturers are already promoting combo-players capable of supporting both HD-DVD and BD, but it seems the new formats still have a long way to go before becoming acceptable widespread standards.
While the movie and video industries might be more or less satisfied with the current capacity of BD and HD-DVD media, the ever-growing computer industry continuously demands more storage for the IT market and increasingly for the consumer market. For that purpose, several companies around the world have been working on new optical-storage solutions able to hold hundreds of gigabytes of data on a single medium. The Israeli company Mempile is currently working on such a technology.
How optical media works
CDs, DVDs, and the current HD media have a number of physical similarities. Each uses a 1.2mm (4/100 inch) piece of clear polycarbonate plastic with microscopic bumps arranged as a single, continuous, spiral track of data. Optical media requires a player composed of a fast-spinning drive motor for spinning the media, a laser (infrared, red, or blue, depending on the player and media), and a tracking mechanism that moves the laser beam to follow the spiral track (with a resolution in the scale of submicrons).
The function of the player is to focus the laser on the track of bumps. In a CD, the laser beam passes through the polycarbonate layer of the media and is reflected off the aluminum layer and hits an opto-electronic device that detects changes in light. Since the bumps in the media reflect light differently than the rest of the layer, the opto-electronic sensor detects that change in reflectivity and translates this difference into digital information (zeros and ones).
The main difference between the current types of optical media is the size of the data track, bumps and the wavelength of the laser used for reading the data. In a CD, each track is about 1.6 microns wide and each pit has a depth of about 0.11 micron and a minimal length of about 0.834 micron. A DVD shrinks this almost by half with a 0.74 micron track and a pit length of 0.4 (interestingly, the pit depth of a DVD is a bit deeper at 0.12 micron). As we already mentioned, the 650nm wavelength (down from 780nm of a CD) allows the DVD to read this extra information. HD media does even better with a pit length of about 0.2 for HD-DVD and 0.15 for Blu-ray.
Another important technical change is the size of the laser spot which had to be reduced to read the ever smaller pits in the media. The original CD had a spot size of about 1.6 microns, which shrunk to 1.1 microns on a DVD—and even further in HD (0.62 micron in HD-DVD and 0.48 micron in Blu-ray). Besides the size of the data and reading apparatus, the data’s location inside the media also changed over time. While the original CD had only one layer located in the innermost part of the 1.2mm thick polycarbonate plastic (fairly close to the label), in a DVD (and HD-DVD) the data surface is located in the middle of the media. Blu-ray however is very different in this respect, locating its data surface on the opposite side of the label. Although each media type has a different location for its data surface, the overall volume taken up by the data inside the media is very small and the majority of space could be considered wasted.
So far we discussed read-only memory (ROM) media. However CD-R/RW, DVD-R/RW and similar HD media are in widespread use. Unlike ROM media, which is made in one go by commercial pressing machines, R/RW media use laser light to record the data onto the disc. In “write once” media (R) the burner turns the laser writer on and off according to the way the ones and zeros should appear on the disc. Some describe the operation of the laser as darkening the material to encode a zero and leaving it translucent to encode a one, although a more accurate description might be to say that the laser changes the volume of the disc in a specific location (“filling” the pit). A rewritable (RW) media is more complicated as it is based on phase-change technology. The phase-change element is a chemical compound made out of silver, antimony, tellurium, and indium (other compounds exist as well, including organic dyes). When the compound is heated above its melting temperature (about 600 degrees Celsius, or 1,112 degrees Fahrenheit), it becomes a liquid; at its crystallization temperature (about 200 degrees Celsius, or 392 degrees Fahrenheit) it turns into a solid. The crystalline form has less volume, so it leaves the pits “empty” while the noncrystalline form has a larger volume, so the pits are full. When the pits are full, there is constructive interference between reflection from the pits and their surrounding which means more light is reflected. When the pit is empty, there is a destructive interference, which means the light is reflected in a lower quantity.
In blank media, all of the material in the writable area is in the crystalline form, so light will shine through this layer to the reflective metal above and bounce back to the light sensor. In order to write information on the disc, the burner uses its write laser, which is more powerful and can heat the compound to its melting temperature. The melted spots have the same function as the bumps on conventional optical media. Nonreflective areas on the RW media indicate a zero, while areas which remain reflective indicate a one (here as well, things are a bit more complex in practice, and data on a disc is encoded as a series of lines having different lengths similar to Morse code called Run Length Limited or RLL for short).
Since the days of the first CDs, optical media increased its capacity by 75 times (from 650 MB in a CD to 50 GB in dual-layer Blu-ray media). However the demand for more storage space continues, and with ever larger hard drives now reaching capacities of 1TB and beyond, an appropriate next-generation optical media is in the making.
Mempile’s TeraDisc technology
For years Ortal Alpert tried to stay ahead of the game buying the latest hard drives and optical drives to store his ever growing library of data. In the mid 1990s, Alpert came up with a novel idea for storing data, and he decided to start his own company. Almost ten years later, Alpert’s dream lead to the creation of a new optical technology, one with the potential to hold 20 times more data than the best existing optical technology.
In late April 2007 the TFOT team visited the offices of Mempile, 20km Northwest of Jerusalem, Israel. Mempile, the company created by Alpert and a few of his colleagues in 2000, is now in advanced stages of developing its revolutionary optical technology, which we had a chance to see. When we first set our eyes on the see-through yellowish disc we were a bit surprised. Was the choice of color the idea of the PR department looking to draw attention to the new media, we inquired? The answer we received took us straight into the heart of Mempile’s technology and made us realize that looks could very well be deceiving.
In a DVD or HD optical media, there are either one or two layers of data. Adding more layers using existing technology would be expensive — but more importantly, it would have to get around a very basic problem: it’s difficult to read information embedded deep inside this kind of media. The current semireflective layers used to store data on CD/DVD/HD-DVD/BD reduce the amount of light that reaches the deep layers, making the amount of signal reflected from each layer smaller, after a few layers the amount of light reflected becomes so small and so noisy that reading the data becomes nearly impossible.
Overcoming this basic limitation of existing optical media is the goal Mempile set for itself, and the way to achieve it is by completely changing that way optical media works — starting from the material of which it is made. Mempile developed a special variant of the polymer polymethyl methacrylate (PMMA) known as ePMMA. After several years of trial and error, Mempile was able to develop this unique polymer, which it claims is almost entirely transparent to the specific wavelength of the laser used by its recorder/player. The yellowish color of the media is thus not a publicity stunt but the result of the special properties of the material used by Mempile.
Using ePMMA, Mempile was able to create a media with about 200 virtual (i.e., created by the laser) layers, five microns apart, each containing approximately 5 GB of data. Although current prototypes are still in the 600–800GB per media range, Mempile is convinced that further optimization will enable it to reach its goal of 1 TB per 1.2mm disc in the very near future.
But using specially designed polymers is just half the story. In order to make a media which could actually store all this data and effectively retrieve it, the old method of reading and writing on optical media had to be abandoned. Instead of the pits and flat surfaces representing zeros and ones, Mempile chose to implement a photochemical process, which happens when an ePMMA molecule is precisely illuminated by a red laser of a specific a wavelength.
In order to be able to precisely illuminate a specific molecule inside the disc, Mempile uses what is known as nonlinear optics. In linear optics the amount of light which is absorbed by an object is directly proportional to the amount of light used, in nonlinear optics the amount of light absorbed does not stand in direct proportion to the amount used — instead, a small decrease in the amount of light used will result in a dramatic decrease in the amount of light absorbed. The process that Mempile uses to write and read data is called two-photon absorption and is nonlinear in nature. When the laser beam is focused to a small radius on the disc, it is very easy for the photons to excite the ePMMA molecules (chromophores), but when the radius of the beam increases even slightly, it becomes very improbable for two photons to be absorbed by a chromophore, so no writing or reading can occur. Nonlinear optics is required in this case because in a 200-layer disc, linear optics would cause some of the light to be absorbed by the layers above the intended one resulting in errors and loss of signal.
In order to read data Mempile uses laser at a specific power which excites the chromophore in a particular layer of the disc. In order to record data, a stronger light is used which creates a different chemical reaction in the molecule. Mempile told TFOT that its technology can also be adapted to perform RW in the future, but market demand for such a product does not seem to be huge.
According to Mempile their product should be very reliable, and different simulations and acceleration tests showed data lifetime of about 50 years. Although Mempile is currently planning to launch their first product using red laser (which is a more mature technology), moving to blue laser further down the road will possibly allow the technology to achieve up to 5 TB of data per disc.
There are currently several other companies developing next-generation optical storage technologies. TDK recently announced a 200GB Blu-ray disc, which seems to be getting closer to the limit of Blu-ray media technology. A different path was taken by InPhase, which TFOT covered in 2006. InPhase uses holographic technology to record data on a special media currently containing about 300 GB. InPhase is working on increasing the capacity of its media and hopes to reach 1.6 TB by early next decade. The current main market for InPhase’s technology is professional users who are willing to pay extra for a fast and large backup storage system. Mempile is looking toward both the professional market and the consumer market and hopes to launch its first product early in the next decade.
Although this might seem like a long time to wait, there are some good reasons behind this decision. Besides the fact that Mempile developed an entirely new technology which is inherently different than that used by conventional CD/DVD/HD media, and hence bound to take longer to develop, the current market doesn’t seem ripe for such a revolution. In a time when 25/50GB media are still just a small percentage of the consumer market, bringing in 1 TB media doesn’t make sense from the point of view of most manufacturers. For that reason we shall probably see Mempile’s technology on the market just after HD media becomes mainstream.
However, when this transformation occurs, we shall reach a whole new stage in data storage. The invention of the CD-ROM made the question of storing documents (and to some extent images) irrelevant, as one disc could store more documents than most people write in their entire lifetime. The DVD allowed for the first time saving full movies (without the need for excessive compression). Only with the recent introduction of HD media did it become possible for higher-resolution movies to be saved on one disc. When Mempile’s technology reaches the market, it will make storing all major data types irrelevant. A single TeraDisc will be able to store over 250,000 high resolution, high quality pictures or MP3s, over 115 DVD-quality movies, and about 40 HD movies — not to mention an unimaginable number of documents. Mempile also sees its technology being used as a network-based backup technology, allowing users to save data from a variety of devices, including desktops, laptops, and digital video recorders (DVRs).
Although many people find it hard to imagine the need for such space on a single disc, it is not inconceivable that by the time Mempile’s technology reaches the market, even higher-resolution video formats will start to appear, requiring hundreds of Gigabytes per hour, on entirely new display technologies, such as holographic displays, which could require even more storage space.
TFOT recently interviewed Ortal Alpert, Mempile’s CTO, and Dr. Beth Erez, Mempile’s Chief Marketing Officer, to learn more about the company’s technology and future plans.
Q: When and how was Mempile created?
A: Mempile was founded in 2000 by myself (Alpert), Moshe Kelner and a few other people. First funding was received from Amiram Sivan; Millennium Materials was the first VC to invest in 2001. I had a crude version of this idea from 1997. Our first versions of the chromophore were not much different than the one we actually use today.
Q: Was there a Eureka! moment when reaching the idea or during the development stage?
A: There was one point where I got the idea, and started thinking about what could be done with it, although my friends remember it as being one specific day. However, optical storage is a multidisciplinary and fascinating field — since then, there were countless moments where we realized there is some way to solve a big problem that is much more elegant than others. Many of the employees of the company come with such good ideas from time to time.
Q: How does the TeraDisc work and how is it different from existing optical technology?
A: The TeraDisc is made of a material which is highly responsive to two-photon writing and reading. This allows us to write anywhere in that we can focus a red laser onto the disc, e.g. multiple layers. However, many other properties of the material have to be optimized to allow this to work properly. Especially the written points, and written layers have to remain transparent after writing, without which it would be very difficult for the reading process to see the 200th layer through 199 written, nontransparent layers.
When a red laser is focused to a small spot inside the TeraDisc, we can choose if we probe the state of this material (reading , low power) or alter it (writing at higher power). This is very similar to the way a regular CDR works, except for the fact that this is now done in 3D.
Q: What is nonlinear optics, how do you use it and how is it different than conventional optics used in CDs/DVDs/BRs, etc.?
A: Let’s start with linear optics — this is the normal way most optical phenomena behave. If you shine light on the wall, you get about 20% of the light back; if you shine it on a mirror, you would get 95% back, if the light passes through your sunglasses, about 10% of the light passes. Physicists call this linear phenomena because if the amount of light used is doubled, so will the amount of light that is reflected or passed. So if one draws a graph of the amount of light shined on the wall versus the amount of light reflected from it, this graph would look like a straight line (hence linear).
A simple example of a nonlinear phenomenon would be the weight of a steel ball compared to its radius — when the radius is about 5 cm you can pick it up and throw it away with one hand. If you increased the radius twice, you can probably pick it up with two hands, but you cannot throw it away very far, because now this is a rather large ball weighing about 33 Kg, if you further increase the radius by a factor of two, you will get a ball weighing 261 Kg, which means you need a forklift to pick it up.
The process that Mempile uses to write and read data behaves much in the same way. It is called two-photon absorption, and behaves according to the radius of the beam to the minus-fourth power (sometimes even to the eighth power). Which means that when the beam is focused to a small radius, it is very easy for the photons to excite the chromophores, but when the radius of the beam increases even slightly, it becomes very improbable for two photons to be absorbed by a chromophore — so no writing or reading can occur.
Q: Why did you choose red laser and not blue, and do you consider using blue laser in the future?
A: There are two answers for that:
1.Red lasers are established technology, so red lasers are easier to come by at the needed powers and beam quality than blue ones. We chose 650nm wavelength as it allows for large enough capacity combined with a mature chemical and laser technology.
2. A terabyte capacity is enough to enter the market. I hope we will have a market for blue materials with an expected capacity of ~5 TB when we finish developing it. This is a very similar path to the path DVD took evolving into blue-laser-based devices (HD-DVD and Blu-ray) and takes similar transition times.
Q: How are you able to direct two photons to a specific molecule in the disc so precisely without effecting nearby molecules?
A: We do not try to do that, all we do is focus a laser beam onto a small spot, which is the same as a DVD does. Physics and chemistry take care of the rest. In that focused spot there is a huge number of photons and a huge number of chromophores, so a lot of interactions happen.
Q: How reliable will the TeraDisc be?
A: We are very confident about data lifetime of 50 years, and similar shelf life. Our grandchildren will probably not be very surprised if the data is readable after 200 years, assuming they will still be able to find a working drive.
Q: What do you see as your advantage over other high-capacity storage technologies (TDK’s 200GB BR disc, Inphase, etc.)?
A: Two-photon technology is the beginning of a few generations of consumer optical storage. Mempile is the best in the field. No other technology I am aware of is capable of delivering a terabyte consumer optical disc.
Q: What do you predict will be the main applications of the TeraDisc?
A: Archiving — in consumer and enterprise markets where rich media content is growing exponentially. Home storage needs are growing exponentially and we are beginning to see 1TB hard-disk drives entering the home networking market. There are no solutions for archiving personal content other than low-capacity optical media. The TeraDisc fills this void. In enterprise markets, compliancy requirements are increasing, compounded by high-resolution content being produced. Healthcare, government, video surveillance, etc., are all searching for low-cost solutions that will provide high data reliability over increasingly longer periods of time for rich media content. Mempile believes that libraries of TeraDiscs will meet these archival needs.
Q: What will the price of the new technology be (both media and drives) and when will it be available?
A: We are planning, with our partners, to have a prototype in 18 months, and a drive and disc intended for the consumer market about a year or so after that. The first prototype will be aimed at compliancy markets with urgent needs where the form factor of the drive is less relevant. We would expect pricing to be around $3,000 for the initial drive and around $30-50 per 600GB disc. Prices will drastically decrease as the form factor of the drive decreases in size and as volumes of drive and disc manufacture increase.
Due to the overwhelming worldwide reaction to TFOT’s “Mempile – Terabyte on a CD” article and the numerous questions posted both here and on many other websites the company decided to grant us an additional Q&A published here for the first time and answering some of the questions raised by the readers.