Merlin/ZeevHed - i must have missed this before, or its new to the ISON web page...it addresses some of the technical & economic questions we've addressed. even if this is the Atari computer of Silicon chip options, we could still make a lot of money on this stock if it gets some media PR...imagine the headline: ISON corp. - lands first contract for Isotopically Pure Si28 with Advanced Mirco Devices - annual contract value estimated at $97MM. ISON stock jumps to $42 on the announcement and at 2.5x next years sales is still cheap compared to the market.
Isotopically Pure Silicon:
Economic and Technical Drivers
Silicon -- The Semiconductor Building Block
The semiconductor industry had product sales of over $125 billion in 1997 and due to its size is a growth engine for the world as well as the U.S. economy. Nearly all of these products were manufactured from a single raw material ? silicon. Decades of experience and uncounted billions of investment have been expended developing manufacturing processes for silicon-based semiconductors. A higher performance alternative to silicon, gallium arsenide, has been generally unsuccessful in replacing silicon due to the necessity of investing billions of dollars to change the manufacturing infrastructure world-wide to accommodate this new material. This example clearly shows that any materials innovation in the semiconductor industry must build on, not replace, the vast silicon experience and manufacturing base.
Two drivers of the semiconductor industry are performance improvement and cost reduction. This has historically been achieved by shrinking the size of chips while at the same time putting more devices (e.g. transistors) on individual chips. Gordon Moore, a founder of Intel, recognized many years ago that the number of transistors on a chip double every 18 months. This relationship was named "Moore?s Law". Many predictions have been made that this rate of progress will necessarily decrease, but it has not. Engineers and scientists have so far found ways to go around apparent roadblocks to enable continued progress.
Too Much Heat -- "No Known Solution"
Device size shrinkage and higher operating speeds lead to increasing power density. Higher power causes higher operating temperatures which must be kept to a reasonable levels. To combat the problem so far, microprocessors are "thinned", i.e. the chip is reduced in thickness to facilitate removal of heat, while heat sinks and on-board fans are used to keep the operating temperature within limits. The Semiconductor Industry Association (SIA) in its 1997 "technology roadmap" has recognized this as a critical problem which has no adequate solution. A quote from that publication is provided in Exhibit A. As it indicates, heat sinks and on-board fans will not be practical when power reaches 110 to 120 watts. Closed-loop cooling systems are now being offered to increase performance of existing designs, but they cost hundreds of dollars each, more than the microprocessor itself! This is not a practical solution. Exhibit B from the same publication shows that the industry expects to reach that power level in just three years, in 2001. It also shows that the industry has found "No Known Solution" to this problem.
A New Type of Silicon is "The Solution"
Isonics believes the solution to this major industry problem is isotopically pure silicon.
Isotopically Pure Silicon
Isotopically pure silicon has better thermal conductivity than natural silicon which means that heat can be removed more effectively. A major microprocessor manufacturer has modeled isotopically pure silicon wafers and has told Isonics that the peak temperature of their "advanced 1 GHz microprocessor" was reduced by 35øC. Such a substantial reduction would make heat sinks and on-board fans practical far beyond 2001, thus removing the major industry roadblock.
The cost of isotopically pure silicon is "reasonable". Isonics believes it can profitably provide isotopically pure silicon bulk wafers for less than $10 per microprocessor, possibly much less. This compares favorably with $10 to $30 for a fan and heat sink and very favorably compared to a closed-loop cooling system costing several hundred dollars.
Finally, isotopically pure silicon is silicon. It is chemically the same and can be used in existing manufacturing processes without any changes. Isotopically pure silicon is "plug compatible" with silicon wafers being used today. Isonics expects to supply isotopically pure silicon to a major wafer manufacturer which will produce wafers using the manufacturer's most advanced processes. This partnership will result in the world's best silicon wafer.
Silicon exists in nature as three stable, non-radioactive, isotopes: Si-28 (92%), Si-29 (5%) and Si-30 (3%). By purification at the sub-atomic level, it is possible to remove essentially all of the Si-29 and Si-30 leaving isotopically pure Si-28, which has a more perfect crystal structure. This more perfect crystal structure exhibits reduced phonon-phonon and phonon-electron interactions (less lattice scattering) which increases certain transport properties, such as thermal conductivity. It has been demonstrated in the laboratory that isotopically pure Si-28 has 60% better room temperature thermal conductivity than natural silicon with its three isotopes. Improvements in other physical and electrical properties are theoretically predicted but have not yet been measured.
As semiconductor manufacturers design chips at higher densities and smaller line widths to increase speed and reduce costs, heat build-up has become a critical problem. Isotopically pure silicon is virtually the same as the material that is almost universally used today to manufacture semiconductors, but with one important advantage ? isotopically pure silicon has superior ability to transfer heat. This improved property should result in the near term in lower reject rates and a higher number of high performance chips. Longer term, isotopically pure silicon should enable the continuance of Moore?s Law far into the next century when its superior heat transfer capabilities becomes mandatory as deep sub-micron design rules are integrated into microprocessor manufacturing lines.
Patents
In 1995 Isonics signed an option agreement with Yale University for an exclusive license to U.S. Patent 5,144,408 covering the manufacture of semiconductor devices on isotopically pure silicon. In July 1997, Isonics exercised its option by providing the world?s first isotopically pure silicon wafers suitable for semiconductor fabrication. The exclusive license to this key patent provides Isonics the ability to direct the commercial development of isotopically pure silicon, essentially world-wide, since the patent provides a significant barrier to entry to potential competitors.
Exhibit A
"... major alternate innovations and solutions will be needed."
"The high-performance market sector [for micro processors] has experienced an exploding increase in power over the different generations. Air cooling is still the only option to keep costs within bounds. With chip power at 60 - 70 Watts, internal heat spots are of concern. Significant engineering development will be needed for power increases at each technology generation. It is expected that at approximately 110 - 120 Watts per chip, the heat sink sizes will become intolerable, and major alternate innovations and solutions will be needed. Capabilities equivalent to closed-loop cooled systems, which are acceptable to the end-use customers, will be needed." |
