c't 3/98, page 18 - Organic Mass Memory

The news is explosive: Evidently, for the cost of a few cents, a Norwegian company can produce a memory module with a capacity of up to 170,000 gigabytes, which could fit on a bank card.

Various newspapers and magazines have reported the achievements of Oslo-based Opticom, a company which conceivably could upset the entire industry with their mammoth memory made of polymers. Polymers are the stuff that panty hose and plastic bags are made of. The first series product of so-called organic memory should be on the market this coming year.

If these speculations are valid, the computer world is on the brink of a revolution. The Norwegians are promising a price which is far below that of established products. Hard drives, CD-ROMs, tapes, as well as flash memory and RAM/ROM would be replaced by this new technology. The entire memory branch could eventually have to call it quits.

And there's more: Non-volatile (mass) memory, sitting on a flexible substrata, as well as displays and polymer based logic, which are also being researched, would lead to complete change of paradigm. The radical jettison of components which use a relatively large amount of electricity and produce large amounts of heat would lead to completely new concepts in design, such as paper-thin computers, which can be rolled up to fit in a jacket pocket. Today's silicon industry would be obsolete.

However, the computer industry has been reacting in an extremely restrained manner to the sensational reports. We asked leading manufacturers from the hard drive industry if they were concerned about their livelihoods in view of Opticom's super memory. The answers: Either they didn't know who Opticom is or no comment.

In contrast, renowned scientists know about the potential of polymer electronics. According to Professor Dr. Christoph Bräuchle of the University of Munich, a known capacity for optical applications of polymer, would in principle support the possibility of producing non-volatile memory from light emitting polymers and proteins. In response to reports that they are ready for production he responded: "I can not imagine that they are already that far." Professor Dr. Josef Friedrichs, from the physics department Weihenstephan at the Technical University of Munich, also considers the production of organic memory to be scientifically possible. However, he didn't get quite so caught up in the sensational news. He claims to have already seen similar memory in the laboratory of a large Japanese manufacturer ten years ago. He said that he couldn't comment about the production-technical aspect of organic memory. Research groups have been working on polymer electronics for years. Above all, many university institutes and large companies have been researching organic light-emitting diodes. This has been verified by numerous scientific publications and talks, as well as by the plethora of patent-holders in the fields of polymer LED and polymer production. Names like AT&T, Motorola, Philips, Sharp, Toshiba and Xerox can be found on the list. A number of company research pages on the Web provide additional references to the promising field of polymer technology.

Display manufacturers seem to be registering the greatest interest. By using polymers, light-absorbing polarizers with background lamps in LCD's would become obsolete. Thereby simplifying the entire production process. Professor Dr. Alan Heeger has already presented prototypes of seven segment displays with light-emitting polymers. However, high-resolution polymer displays as a real alternative to monitors are still only a reality in the research labs of the big display manufacturers.

Sandwich

The principle of organic memory is as simple as it is brilliant. A polymer film which is contacted by a passive matrix, emits light onto the memory medium - a protein film. The light causes the proteins to switch between two stable states. The states can be distinguished above all by those colors which they absorb and those which they let pass. Once they have been changed, the states remain stable even without light. The data is then read with less intense light, that doesn't change the memory content. In one state the proteins absorb more light, in the other less. Another polymer layer, also matrix-regulated, acts as a photo-detector and measures the light which has been diffracted by the proteins.

The building blocks of the Opticom polymer memory: A matrix addresses the light emitting polymers. The light "writes" on the proteins in the middle of the sandwich at a cross-point. The lower polymer layer absorbs light thus reading the memory content.

A single matrix element of the Opticom memory is supposed to have a dimesnion of less than 100 nm. The entire layer is 350 nm thick. This is ten to hundred times smaller than the common size of microchips. Thus, the usual lithography procedures could not be sued in the production process. Dimensions that small could actually be achieved if the matrix's strip conductors are made of (conductive) polymers. The polymer chains, which are only a few nanometers thick, but quite long, line themselves up under certain conditions, thus serving as one of the matrix lines. The second polymer layer could also possibly be structured by exposure to UV light.

The catch with organic memory is the connection of this matrix. Every single strip conductor of the matrix must be connected to and powered by a transistor. The dimensions of modern transistors in 0.25 æm technology pose an obstacle of a few micrometers to the measurements of Opticom's dream memory with 100 nm line intervals. In addition, the mini strip conductors have to contact the (relatively speaking) giant transistor connectors in a confined area.

It is hardly conceivable that a small Norwegian company could be huge steps ahead of the semiconductor giants in the area of lithographic procedures. Therefore it would only remain for the driver logic to cahnge to another material: the use of polymer transistors. We asked Hans Gude Gudesen, director of the Opticom research team, how the measurement discrepancy could be explained. The company co-founder basically confirmed the use of organic transistors and provided several technical explanations. However, when it came down to making definite statements concerning the production procedure, he referred to the necessary secrecy surrounding the subject. According to him, Opticom is the creator of this memory technology. However, the company can not manufacture it and is thus required by its partner firms to maintain silence.

David Versus Goliath

With only 25 employees, the Scandinavian firm really is a small operation. Three renowned scientists in the area of polymer research build Opticom's technical advisory board: Professor Dr. Alan Heeger of the University of California, Santa Barbara, who played a leading role in the discovery of light emitting polymers, Professor Dr. Robert Birge of Syracuse University, who did pioneering work in the area of protein-based optoelectronics and Professor Dr. Olle Inganäs of the University of Linköping, a leader in the area of miniature, polymer-based LED's. In addition to this huge amount of polymer knowledge, research is above all propelled by cooperation with around 100 scientists and 30 institutes all over the world, according to Gudesen.

The company's intellectual property can only be protected by scattering individual fields of research, claims Gudesen. Opticom as of yet holds no patents in this area. The Norwegians have recently filed several patent specifications and are now waiting for the blessing of the respective patent agencies. They are hoping for patent approval by mid-year. The company has to protect themselves from technology theft. Therefore the publication of technical details is probably hung up on the patent process as much as on the non-disclosure agreements.

The rumors that Opticom would bring a manufactured product on the market by the end of 1998 was enough to make the young company's share prices take a temporary climb. Opticom's management owns over 34 percent of the shares. Around 19 percent are owned by Gudesen alone. The Eidos company, European manufacturer of entertainment software, holds another 15 percent. The two companies are connected through Opticom's Managing Director Robert Keith, who is also the director of Eidos. Therefore, around 50 percent of the shares belong alltogether to Opticom management. This gives rise to the suspicion that the memory sensation is really only a share sensation.

Cooperation

However, there are many signs which speak out against such nasty insinuations. In October 1997 Opticom signed a research agreement with the renowned Bell Labs (AT&T/Lucent Technology R&D). This joint development agreement is aimed at uniting both companies' know-how in the area of polymer circuits leading to the development of integrated circuits for organic memory. Bell Labs' researchers have been working on so-called plastic transistors for quite some time. They claim to be able to produce working circuits with polymer transistors. Their structures are printed on flexible substrats using a process similar to silk-screening, as they are employed in thick film circuits. This could actually be the solution to Opticom's problem of the connection of the memory matrix. In spite of this, the question of whether the technology is developed enough for mass production remains. Gude Gudesen declares that this is the case: The first protein memory will be manufactured in the coming year. It can be concluded from the research director's statements that the product in mention could possibly be a non-volatile RAM with a memory capacity of a few gigabytes. However, Gudesen wouldn't reveal the exact nature of the first product or the details of technical production.

The research director considers it theoretically possible that organic memory could be made of a thousand or more layers where only 9000 transistors would be needed to connect to 1 MByte. Due to parallel layers the ultimate memory would have a capacity of 170,000 GByte. According to Gudesen, the access time to a word would be fewer than 5 ns. Additionally, all layers of the super memory could be read in parallel. However, contradictions concerning access time appear between the spoken and written word. In contrast to Gudesen's statement, all documents available to us (diverse articles, stock analyses, bank analyses) mention an access time of 50 ns. It remains unclear whether this was a misunderstanding, if the time could actually be ten times better or if Gudesen was simply exaggerating.

Only One Layer

For the moment Opticom is not thinking in terabytes, but the company is baking much smaller potatoes. Gudesen says that two prototypes are ready. These are made of just a single layer. There is neither information available about the capacity of the single-layer memory nor about details of further stratification. A visit to the lab couldn't be permitted, allegedly due to the necessary secrecy.

One layer memory will certainly demonstrate the memory technology in principle. However, the path from lab prototype to mass production is long and rocky. A number of other companies are also working on three-dimensional protein memory. They use bacteriorhodopsin, for example, which allows addressing with laser beams. Organically based holographic memory systems are also popular among scientists. IBM is developing this kind of optical polymer memory.

Thus the question remains unanswered: Will Opticom be ready to go into production in the next year or will the competition beat them to it? (jr)

Top of page

Protein Memory

The bacteriorhodopsin protein is one of the most promising organic memory materials. Seven helix-shaped polymers form a membrane structure, which contains a molecule known as the retinal chromophor. The chromophor absorbs light of a certain color and is therefore able to switch to another stable state in addition to its original state. Only blue light can change the molecule back to its original state.

Translation by Stephanie Ford

Additional Translations from c't are available
Last changed by Jürgen Kuri, October 27, 1999

Seitenanfang