Transistors Twenty Years Later

Source: Adapted from IEEE.

2047 is the 100th anniversary of the invention of the transistor. What will the transistor look like then? Will they even become critical computing elements today? IEEE Spectrum asked experts around the world for their predictions.

Transistors are expected to be more diverse than they are today, according to one expert. Just as processors evolved from CPUs to include GPUs, network processors, AI accelerators, and other specialized computing chips, transistors will evolve to serve a variety of purposes. “Device technology will become application-specific, just as computing architecture has become application-specific,” said Philip Wong, IEEE Fellow, vice president of enterprise research at TSMC, and professor of electrical engineering at Stanford University.

Suman Datta, an IEEE Fellow, professor of electrical and computer science at Georgia Tech, and director of the multi-university nanotechnology institute, says that despite the variety, the basic working principle—the field effect that turns transistors on and off—is likely to remain the same. Center rises. “The minimum critical dimension of the device could be 1 nanometer or less, enabling a device density of 10 trillion per square centimeter,” said Tsu-Jae King Liu, an IEEE fellow, Intel board member, and dean of UC Berkeley’s School of Engineering.

Experts seem to agree that transistors in 2047 will require new materials and possibly a stacked or 3D architecture to expand on planned complementary field-effect transistors (CFETs, or 3D stacked CMOS). Transistor channels that are now parallel to the silicon plane may need to become vertical to continue increasing density, Datta said.

AMD senior fellow Richard Schultz said the main goal in developing these new devices will be power consumption. “The focus will be on reducing power consumption and the need for advanced cooling solutions,” he said. “There needs to be a focus on devices that operate at lower voltages.”

Will transistors still be at the heart of most computing in 25 years?

It’s hard to imagine a world where computing wasn’t done with transistors, but, of course, vacuum tubes were once the digital switch of choice. Startup funding for quantum computing that doesn’t directly rely on transistors will hit $1.4 billion in 2021, according to McKinsey & Company.

But electronics experts say quantum computing won’t be advancing fast enough to challenge transistors by 2047. “The transistor will remain the most important computing element,” said Sayeef Salahuddin, IEEE Fellow and professor of electrical engineering and computer science at UC Berkeley. “Currently, even with an ideal quantum computer, the potential application areas seem rather limited compared to classical computers.”

Sri Samavedam, senior vice president of CMOS technology at Imec, a European chip research and development center, agrees. “Transistors will remain very important computing elements for most general-purpose computing applications,” Samavedam said. “One cannot ignore the efficiencies achieved by decades of continuous optimization of transistors.”

Has the 2047 transistor been invented yet?

Twenty-five years is a long time, but in the world of semiconductor R&D, it’s not that long. “In this industry, it usually takes about 20 years from [demonstrating a concept] to introducing manufacturing,” Samavedam said. “It’s safe to assume that a transistor or switch architecture for 2047 has already been demonstrated on a lab scale,” even if the materials involved won’t be exactly the same. King Liu, who demonstrated modern FinFETs with colleagues at Berkeley some 25 years ago, agrees.

But the idea that the transistor of 2047 is already in a lab somewhere is not universally shared. Salahuddin also thinks it hasn’t been invented yet. “But just like with FinFETs in the 1990s, it is possible to make reasonable predictions about the geometry of future transistors,” he said.

AMD’s Schultz says you can glimpse such structures in proposed 3D stacked devices made of 2D semiconductors or carbon-based semiconductors. “Device materials that have not yet been invented may also be within this time frame,” he added.

Will silicon still be the active part of most transistors in 2047?

The heart of most devices, the transistor channel region, will still be silicon, or possibly silicon germanium (where progress has been made) or germanium, experts say. But in 2047, many chips may use semiconductors that are considered exotic today. These may include oxide semiconductors, such as indium gallium zinc oxide; two-dimensional semiconductors, such as metal dichalcogenide tungsten disulfide; and one-dimensional semiconductors, such as carbon nanotubes. Or even “other technologies that have yet to be invented,” says Imec’s Samavedam.

Gabriel Loh, an AMD senior fellow and IEEE fellow, pointed out that silicon-based chips may be integrated in the same package with chips that rely on newer materials, just as processor manufacturers integrate chips using different silicon manufacturing technologies into the same package today.

Which semiconductor material is at the heart of a device might not even be a central question in 2047. “The choice of channel material basically depends on which material is most compatible with the many other materials that make up the rest of the device,” Salahuddin said. We know a lot about integrating materials with silicon.

Where will transistors be ubiquitous by 2047 that aren’t there yet?

everywhere. No, seriously. Experts do expect some level of intelligence and perception to permeate every aspect of our lives. This means devices will be attached to and implanted in our bodies; embedded in all kinds of infrastructure, including roads, walls and houses; woven into our clothes; stuck to our food; every step in the supply chain; and doing many other things where no one has yet thought of.

Transistors will be “ubiquitous, requiring computing, command and control, communications, data collection, storage and analysis, intelligence, sensing and actuation, interaction with humans, or a portal into the world of virtual and mixed reality,” Stanford’s Wong said. concluded.

About Cheersonic

Cheersonic is the leading developer and manufacturer of ultrasonic coating systems for applying precise, thin film coatings to protect, strengthen or smooth surfaces on parts and components for the microelectronics/electronics, alternative energy, medical and industrial markets, including specialized glass applications in construction and automotive.

Our coating solutions are environmentally-friendly, efficient and highly reliable, and enable dramatic reductions in overspray, savings in raw material, water and energy usage and provide improved process repeatability, transfer efficiency, high uniformity and reduced emissions.

Chinese Website: Cheersonic Provides Professional Coating Solutions