Making chips on plastic for less than 1 cent each

Imagine that the objects around you are full of intelligence. A bandage, a banana peel, a bottle, etc. are all intelligent. At present, this kind of scene can only appear in science fiction movies. You may be wondering why all this has not been realized with the rapid development of technology today. This is because humans have not yet produced cheap processors.

The number of IoT devices worldwide is growing by billions every year. It may seem like a huge number, but in reality the potential in this area is much greater, and fairly expensive silicon chips are holding it back. The solution might be to introduce plastic chips that are many times cheaper.

Some research institutions have made various attempts before. For example, in 2021, Arm launched the PlasticArm M0 new plastic chip prototype, which can print circuits directly on paper, plastic or fabric. This chip does not use silicon as a substrate. Instead, it uses a plastic processor core, which is a project that Arm has been researching for nearly ten years, but even so Arm's research cannot meet the standards.

The problem, say engineers at the University of Illinois at Urbana-Champaign and British chipmaker PragmatIC Semiconductor, is that even the simplest industry-standard microcontrollers are too complex to be mass-produced on plastic.

At the International Symposium on Computer Architecture later this month, a research team from the University of Illinois at Urbana-Champaign will demonstrate a simple but functional plastic Processors that can be manufactured for less than a penny. The team designed both 4-bit and 8-bit processors. However, further details of this study have not been made public.

Team leader Rakesh Kumar said, "About 81% of the 4-bit processors can work, which is enough to break the 1-cent threshold."

Rakesh Kumar

Kumar said that flexible electronics have been a segment of the market for decades, and the processor made by the team is made using the flexible thin-film semiconductor indium gallium zinc oxide (IGZO), which Thin film semiconductors can be built on plastic and continue to work even when bent within a radius of a few millimeters. But while a solid manufacturing process is a prerequisite, it’s the design that makes the difference.

Source: https://technewsspace.com/scientists-have-developed-penny-plastic-flexicore-chips-they-promise-to-revolutionize -the-internet-of-things/

Why not silicon?

You may be wondering, why can’t silicon processors be made super cheap and have flexible computing performance? Kumar’s analysis concluded that this was impossible. Silicon is expensive and inflexible compared to plastic, and if you make a plastic chip small enough, it can continue to work within the range of bends. There are two reasons why silicon fails: One is that although the area of ​​the circuit can be made very small, you still need to leave a relatively large space at the edge of the chip to cut the chip out of the wafer. For a typical microcontroller, there is more space at the edge of the chip than the area containing the circuitry. What's more, you also need more space to install enough I/O pads so that data and power can enter the chip. As a result, blank silicon wafers will be wasted.

Rather than adapt an existing microcontroller architecture to plastic, Kumar's team created a design called Flexicore from scratch. Since the scrap rate increases with the number of logic elements. Knowing this, they came up with an alternative design aimed at minimizing the number of doors required. They use 4-bit and 8-bit logic instead of 16-bit or 32-bit logic. Like separating the memory where instructions are stored from the memory where data is stored. But this comes with reducing the number and complexity of instructions the processor can execute.

The team further simplified the design of the processor, designing it to execute instructions in a single clock cycle rather than the multistep pipeline of today's CPUs. They then implemented the logic for these instructions by reusing parts, which further reduced the gate count. "Overall, we were able to simplify the design of FlexiCores by tailoring them to the needs of flexible applications, which tend to be computationally simple," said Kumar's student Nathaniel Bleier.

Through the above design, the team achieved a 4-bit FlexiCore chip with 5.6 mm^2, consisting of only 2104 semiconductor devices (about the same number of transistors in the classic Intel 4004 in 1971), and last year PlasticARM, the soft microprocessor developed by the Arm team, consists of approximately 56,340 devices. “In terms of gate count, FlexiCore is an order of magnitude less than the smallest silicon microcontrollers,” said Nathaniel Bleier.

Engineers used PragmatIC’s manufacturing process to fabricate 4-bit microcontrollers on plastic.

FlexiCore also features optimized onboard memory and instruction sets to minimize transistor count and reduce complexity. The researchers also designed the logic components so that they could use a minimum number of transistors. After all, processors are designed to execute one instruction in one clock cycle.

In addition, the team also developed an 8-bit version of FlexiCore, but the effect was not good.

“This is exactly the kind of design innovation needed to support truly ubiquitous electronics,” said Scott White, CEO of PragmatIC Semiconductor.

Using PragmatIC technology, the team produced plastic-coated wafers with 4-bit and 8-bit processors, tested them in multiple programs at different voltages, and bent them mercilessly . The experiment may seem basic, but according to Kumar, it is groundbreaking. Most processors built using non-silicon technologies have such poor yield that results can only be reported from one or at most a few working chips. "As far as we know, this is the first time that data from non-silicon technologies can be reported across multiple chips," Kumar said.

PragmatIC has been working on low-cost chips

Kumar observed that the chip industry aims to balance power and performance metrics, along with some degree of reliability . They are not focusing on cost, consistency and chip thinness. Instead, the focus is on building new computer architectures and targeting new applications.

John A. Rogers, a pioneer in flexible electronics at Northwestern University, called the work impressive and looked forward to future developments in this research.

Of course, this is only what this study has done so far, and there is much more work to be done before the FlexiCore solution or similar solutions are commercially available. However, researchers have tried to optimize their solutions for different processes and target workloads, with some success. There are also questions about how bending affects performance and the durability of plastic chips.

However, with such cheap plastic processors and flexible electronics becoming mainstream, we may soon see the dawn of truly ubiquitous electronics. This kind of chip can be placed on the packaging of almost any product or on a medical patch, and its application fields are no longer limited.

The above is the detailed content of Making chips on plastic for less than 1 cent each. For more information, please follow other related articles on the PHP Chinese website!