CONNECT WITH US

North American startup Zinite pursues the real 3D chip with thin-film transistor innovation

Misha Lu, DIGITIMES Asia, Taipei 0

Credit: Zinite

Zinite is seeking to shake up the semiconductor industry through innovations in transistor architecture to achieve more chip performance density. Founded by Gem Shoute, Douglas Barlage, and Kenneth Cadien, and seeking to meet the demands of HPC, AI, and AGI by accelerating 3D chip adoption, Zinite's core technology is a thin-film transistor (TFT) with CMOS-level drive current and a large temperature process window, including channel material processing stability at 400C.

Talking to DIGITIMES Asia, Zinite founder and CEO Gem Shoute observed that TFT technology has been explored for a very long time because of its potential to allow transistors to be stacked on top of each other and this capability is well known to improve performance density. However, these advantages have not yet been fully realized. Even today, TFTs have been facing challenges with performance, and remains not fully compatible with CMOS processes that dominates global semiconductor manufacturing.

Against this backdrop, Zinite was born as its founding team initially sought to answer three fundamental questions facing TFT development, namely: how to get the best materials for TFT; how to bring TFT performance close to or even surpassing silicon-based transistors: and if the right physics is being employed in understanding the problem. "It turned out that the last question that we asked was the most important," remarked Gem.

The Zinite high-performance TFT is different in operation than a traditional thin film transistor. Zinite's thin-film transistors operate more similarly to fully depleted n-type SOI (silicon-on-Insulator) devices, and can be fully synthesized with temperatures at or less than 400 °C, making them integrable with back-end (BEOL) and middle of line (MOL) CMOS processes.

"We had to start with a clean sheet of paper," said the Zinite founder, "We couldn't make the same assumptions or use the same engineering shortcuts that worked for silicon-based transistors for the TFT we imagined." Having been trained by some of the greatest physicists in the semiconductor industry, Zinite's founding members were able to look at the problem of TFT development differently and find the right combination to make it all work. "At some point, all the pieces aligned: from materials to process to architecture and beyond for the device," Gem noted. Zinite was subsequently founded when the team realized that it had come up with something special.

When the founding team was thinking about the transistor that eventually became Zinite's core piece of technology, it wanted to build a thin-film transistor that could be as versatile and mainstream as silicon. In a sense, a suitable TFT would enable true 3D chips.

Silicon chips are built two- dimensional layers, and silicon transistors are built at the lowest layers in a chip, even for FinFET and GAAFET. "Because they're at the bottom-most layers of the chip, it is important that any subsequent processing does not affect the performance of the underlying layers," Gem pointed out, adding that anything processed on top of the conventional CMOS transistors is thus restricted to a certain thermal budget, and simultaneously must show stability at 400 degrees Celsius. Meeting these criteria makes it challenging to build useful transistors atop the CMOS layer.

Speaking of differences from current advanced packaging approaches, the Zinite CEO observed that performance is not just about density, but about cost as well. "When it comes to advanced packaging, the cost hasn't scaled for a long time. And it does not scale as fast as an integrated technology," noted Gem. "That is one of the many advantages that Zinite has, the cost advantage. Another is the density advantage."

Nevertheless, Gem emphasized that Zinite is not here to displace proven CMOS technologies, but to work synergistically with all of them to meet computational demands of AI, HPC and other possible future drivers of demands.

Meanwhile, Zinite's priority at the current stage is to put its technologies in customer lines. "Our next immediate phase starts with transferring the process into customer lines, proceed to Risk Production, then convert to high volume production. We're going to start out in market segments where we already have collaborations," indicated Gem. A participant of Computex Taipei, Zinite is seeking to further establish partnerships with flagship manufacturers to integrate Zinite's technology.

"We want to be very selective and targeted with our early partners and early adopters. Our end goal is to make Zinite's early adopters the market leaders for their segments - to make them winners."