 |
SAN JOSE, Calif. — The marriage of electronics with nanotechnology will be a long and rocky one, according to a number of speakers at Nanotech Planet Spring 2002, a two-day conference wrapping up today in San Jose.
 |
null
What this marriage will mean, in the short term, is technological advances that will help make traditional computer chips even smaller. In the longer term it will mean simple memory chips and logic devices at the nanoscale and, perhaps decades ahead, even a nanocomputer that will assemble itself and have tiny transistors (the on and off switches that form the foundation of a modern computers) only a few nanometers wide.
 |
null
These handheld devices will have the computational power of every computer in the world today, said R. Stanley Williams, a fellow at Hewlett-Packard Laboratories and a speaker at the convention.
null
Without nanotechnology, such a handheld marvel isn’t likely. Moore’s Law has to be in effect for 50 more years (every 18 months doubling computing power) for such a computer to be possible, Williams said. Thanks to some fundamental laws of physics, that is not likely to happen. Moore’s Law, in fact, will likely end in 10 to 15 years. Even if it doesn’t, there is another huge barrier: economics. The cost of building a new chip fabrication plant could reach $50 billion by 2010.
null
Enter nanotechnology, which could potentially pack large amounts of computational power in very small spaces.
null
HP, IBM and others are hard at work building nanotransistors. Recently, IBM built a double gate transistor that has a fin-like structure 20 nanometers thick.
null
In 15 or 20 years, this molecular circuit will be just as important a milestone as IBM’s other contributions to the field of nanotechnology, said Thomas Theis, director of physical sciences at IBM Research and the convention’s opening keynote speaker. Other IBM contributions include invention of the scanning tunneling and atomic force microscopes in the 1980s.
null
To build nanotransistors, IBM must also develop new manufacturing processes since the traditional lithographic method used to make computer transistors can’t effectively operate at this scale. Because “construction tools” could never be made small enough, nanoelectronic devices must “self-assemble,” or build themselves. This is how nature operates, said Theis.
null
“If we tried to write a digital file to describe a human being, it would be enormous.” Nature, on the other hand, builds mostly through analog means, which is why the human DNA contains only 40 megabytes of digital information, he said.
null
“There is no reason why we can’t eventually become very good at self-assembly,” Theis said. “This will enable more efficient and vastly less expensive manufacturing processes.”
null
At Hewlett-Packard, Williams and other researchers have spent “a lot of time trying to reinvent the computer.” HP researchers have built a new kind of transistor using the self-assembly method. Their transistor has parallel wires connected by a quantum switch. The wires, only 6 atoms wide and 4 atoms high, are made by placing a material on a foundation of silicon. When the material is heated, “the wires just form themselves, fast and inexpensively.” They also end up being perfectly straight and parallel, Williams said.
null
While these breakthroughs are important, they are just the first steps in what will be a long and difficult journey. The field of nanotechnology is where microelectronics was in the 1960s just after the integrated circuit was invented, Williams said.
null
While anything approaching a nanocomputer is decades away, in the shorter term, perhaps five years out, the first nanoelectronic products will begin to hit the market. “They will just add some interesting function to a device, like a toy or camera,” Williams said.
null
“The first ubiquitous (nanoelectronic devices) will be sensors integrated on silicon chips,” Theis said.
null
In 10-plus years, nanotechnology will begin to replace logic and memory devices, and in 20 to 50 years, we will build more complex systems, Theis said. “Microelectronics will become nanoelectronics.”
null
None of this is likely to happen (at least in this time frame and in the United States) unless government and private industry funds nanotechnology in the same way they funded the transistor and the integrated circuit. Thanks to the space race, in a five-year period in the 1960s the government poured in 1 percent of the nation’s gross national product, said Williams.
null
If this type of funding occurs and technology innovations continue apace, the possibilities are mind-boggling. In the 1950s, the ENIAC was the world’s most powerful computer. It could do 1,000 additions a second and was used to calculate the trajectory of artillery shells.
null
Today, a tiny handheld device has the capacity of 1,000 ENIACs. It is 100 million times smaller and a 100 million times more power efficient. “That is the most incredible advance humans have ever seen,” Williams said.
null
The nanocomputer could well be the second chapter to this remarkable story. Yet, said Williams, just because we can envision a technology doesn’t mean we will ever create it. As Williams said, we have known about the speed of light for about 100 years, but that doesn’t mean we will ever travel that fast.