Monday, October 8, 2007

Dead end to silicon technology and way out

It’s true that silicon industry has reached a dead end as they are facing problems in shortening the size of individual component in the electronic chip. When electronic chips emerged in markets around the world some 40 years ago, they replaced the bulky transistor tubes, which present generation may not even recall now. Since then, efforts have been made to increase the number of components on a single chip and it even reached upto 169 million in Pentium 4 processor. But further integrating more components on same area proves to be problematic now.
There are many factors which lead to the problem of integrating smaller sized components to a chip endlessly. One of the major problems is charge leakage. When two components come too close to each other, then electron can “tunnel” from one to other: resulting in unexpected erroneous functioning of the circuit designed. This starts to happen when one makes components in nanometer (a billionth of a meter) regime. Quantum phenomena like “tunneling effect” makes electrons to tunnel through the barrier if it is very thin and hence charge leaks out of the components to another unexpectedly. In such a case, behavior of the overall device cannot be controlled properly.
Another problem is the fabrication technology of components within nanometer size range. Continuous increase in density of components within an area of chip, has forced electronic engineers to fabricate smaller and smaller components. But with each step of reducing the size, fabrication technology becomes more complex and moreover it becomes difficult to keep up with rise in overall cost of fabrication. Generally, lithography techniques are employed for fabricating components of ultra small circuit in which electromagnetic radiations of particular wavelengths are exposed to particular parts of the material on the chip, which induces component formation. The limit of this method originates from the wavelength of electromagnetic radiation, as components cannot be smaller than that. This yields limiting problem as scientists have reduced the wavelengths up to Ultraviolet regime and now finding it difficult to further reduce it and still get same quality material as before. Moreover it should be mentioned that change in wavelength mostly require employing new materials and hence inducing burden of extra cost. These problems shows that silicon industry is facing a dead end now and it dearly needs new materials, fabricating technologies and new concepts for electronic circuits.
Apart from usual problems of lithography limit and charge leakage, one major problem is heat dissipation technologies. Each component generates heats when current flows through it. More components means more heat generation and this heat has to be taken out of the system before it melts the chip. Present heat sink materials and fan blowing technologies cannot handle this problem if one tries to go beyond a present number of components per square area.
All of these problems compel chip engineers to review the concepts of conventional electronics and adopt entirely new concepts. This has lead to many inventions since past few years. With the advent of nanotechnological research, semiconductor industry has learned that they can no longer rely on silicon as base material. There are other materials which perform better results in nanometer regime. Also, new fabrication technologies like self-assembly leads to cheaper and reliable fabrication within nanometer. New materials like a forest of carbon nanotubes act as better heat sinks. Nanowires and Nanotubes outperform most of conventional architectures for speed and load of charge transport. By tuning the sizes and shapes of nanomaterial, it can be made to act as an insulator, metal or semiconductor. This gives additional advantage of using only one material to fabricate all components, which is a great convenience for chip fabrication technologists. Conductive polymers open a whole new area of flexible electronics. Integrating magnetism and optics to electronics yields news devices based on magneto-electro-optical devices which present multifuctionality and multiple control mechanisms using magnetic fields, electricity and light.
Apart from improving the present semiconductor material and technology, scientists have been able to demonstrate some very different approaches to computing. One of them is called Spintronics where instead of charge; spin of electron shall be used to represent bits 0 and 1. Spin flipping consumes less energy and offers great switching speed which ultimately results in dual advantage in the form of lesser power consumption and higher computation speeds. Another technology is called Quantum Cellular Automata which provides the advantage of wireless communication between nanometer sized islands for computation. Many other architectures are also expected to open up soon. Now when humans face a dead end at silicon semiconductor technology, then they shall have to divert towards new materials and technologies.

No comments:

Popular Posts