Speed factors lead to possible usage of III-V semiconductors for microprocessor construction

Sep 10, 2008 13:28 GMT  ·  By

In the ever-increasing need for speed, the semiconductor industry prospects gallium arsenide semiconductors as potential replacements for silicon ones.

The year 1974 recorded the birth of the first microprocessor for general usage, the Intel 8080, which was able to operate with about 500.000 instructions per second. Based on the semiconductor industry's efforts in order to reduce their size, the best of its relatives from today execute 100.000 times more than that. Don't exhaust your minds: that means 50.000.000.000 (fifty billion) instructions.

The most important part of a microprocessor is comprised of MOSFETs (metal-oxide-semiconductor field-effect transistors). They're its basic building blocks that behave as small switches, and they are built by the billions on a silicon surface through the usage of photolithography. They used to get smaller and faster each year, as more of them fitted into a chip, so their cost remained roughly the same in the long term.

But the shrinking process seems to have come to a halt, since further decreasing the current MOSFET size (a couple of dozens of nanometers) proves to be extremely difficult in terms of fundamental physics. The best approach is still considered to be, as it was almost 50 years ago, when MOSFETs were invented, the usage of compound semiconductors such as gallium arsenide. This would render the transistors able to switch on and off a lot quicker than the actual silicon ones. But the constant failures gave birth to a Silicon Valley joke: "gallium arsenide is the technology of the future, and it will always be."

But Mr. Peide D. Ye, an associate professor of electrical engineering at Purdue University's Birck Nanotechnology Center in West Lafayette, Indiana, claims that he and his colleagues, together with other scientists in similar institutions, have made some advances that might soon allow the gallium arsenide transistors to be used on a large scale. Of course, this is only the key feature of the technology, which would provide 3 or 4 times faster microprocessors. But many other parameters have to be optimized, such as gate leakage, current-carrying capacity, or low-voltage operating (in order to reduce heating). In the meanwhile, the expected approach is for chip manufacturers to use compound semiconductors only when needed, in conjunction with silicon.

Gallium arsenide is formed by components coming from the IIIrd and Vth columns of the periodic table of elements, hence the name of III-V semiconductor. Other less studied compounds of the same groups are indium phosphide and gallium nitride. The devices built on the gallium arsenide technology are much more expensive, since the raw materials cost 10 times more than those used for silicon-based products. Its far better results are used for laser diodes, highly-efficient solar cells, and HEMT (the high-electron-mobility transistor), which is used for radars, communication systems or cellphones.

But, besides costs, Mr. Ye offers other reasons in order for us not to be too eager to see the gallium arsenide technology implemented: "One reason for keeping as much silicon as possible around is that it has considerably better physical properties for making the large wafers used in semiconductor manufacturing. Also, silicon is cheap and environmentally friendly, whereas gallium arsenide is expensive and, because it contains arsenic, potentially quite toxic."