World's fastest 3D chip manufacturing

The method of making new 3D computer chips, using materials that are carbon nanotubes, can give processing speeds up to 1,000 times faster than computer chips.

Successfully developed the world's fastest 3D chip

3D design allows scientists to increase data storage capacity by interlacing memory and microprocessors in a narrow space, Max Shulaker, a doctoral student in electrical engineering at the University. Stanford, California said.

According to Live Science, in a presentation on Sept. 10 at a technology forum of the US Army's Advanced Defense Research Agency, Shulaker said reducing the distance between components would significantly reduce machine time. Information processing properties.

The boom of computer technology in the last 50 years has been a huge part of making smaller and smaller transistors. According to Moore's Law , introduced in 1965 by lead researcher Gordon E. Moore, the number of transistors on a silicon chip will double after two years.

As expected, transistors were getting smaller and smaller, with the smallest part being 5 nanometers in size and 7 nanometer functional parts, much smaller than the 100,000-nanometer human hair width.

However, the size reduction is only limited. Too small, the quantum effects of particles will affect their function. Therefore, it is very likely that Moore's law will end within 10 years, according to experts. Later computers may not be able to speed up the processing anymore.

Therefore, scientists must take into account the optimization of the data transmission process to speed up the computer. Today's computer after processing information will transmit that information in the form of electric current through wires connecting the processor and hard drive. The resistance of the wires will affect this speed.

Picture 1 of World's fastest 3D chip manufacturing
3D chip with a calculation speed 1,000 times faster than the current fastest speed.(Photo: Max Shulaker).

While waiting for the information to move, the processor is in a "idle" state , does not perform any work, although it must still use energy.

"You are wasting a huge amount of energy," Shulaker said. Solving this problem is quite difficult, because it is impossible to place the CPU (central processor) and memory on the same silicon substrate. The temperature of the silicon base will reach more than 1,000 degrees Celsius, exceeding the melting temperature of many metals used in hard drives.

To overcome this difficulty, Shulaker and his advisers proposed using a different material, carbon nanotubes (CNT) . CNT has the same properties as traditional semiconductor materials.

"CNT will become a better transistor, making information move faster and consume less energy , " Shulaker said. However, because the CNTs are irregularly formed, they do not conduct electricity well, researchers have to develop methods to form CNT in narrow grooves so that they align with each other.

By this method, about 99.5% of CNT will be aligned. Small numbers arranged in a different direction are randomly processed by drilling several holes on the chip in certain locations.

Another problem is that when most CNTs act like semiconductors, there are still a few metal-like expressions, and there is no way to predict the properties of these tubes. They may damage a chip.

Shulaker and colleagues treated by "turning off" all semiconductor CNTs, and giving a huge current running through the remaining metal CNTs. The current will destroy all these CNTs, just as if there is a large current flowing through the fuse.

Group of Shulaker in 2013 designed a CNT transistor computer. However, it is quite cumbersome and slow, as well as a small number of transistors.

They have now created a multi-layer memory and transistor arrangement system, connected by small wires. This 3D design has cut the time it takes for information to travel between transistors and memory, and it is possible to create chips with a computational speed 1,000 times faster than today's fastest. The team has applied this new architecture in sensing panels to detect everything from infrared light to special chemicals in the environment.

The next step is to expand the system further, creating larger and more complex chips.