Those early supercomputers were built for serious purposes: classified military work, academic research, advanced engineering, and more. In the ‘80s, computers with those capabilities had to run inside vats of cooling liquid to keep their processors from melting down. Today, we can build more powerful computers that fit in your pocket.

Engineers have spent decades shrinking computers, constructing lighter, and more powerful machines. Nearly every remarkable scientific and cultural achievement of the last half century of human history traces back to this progress.

In 1965, just a few years after the semiconductor industry emerged, the engineer and businessman Gordon Moore made a prediction. He said he expected the number of transistors packed onto a single chip to double every two years for the next decade.

It was a bold claim, a vision of the future as bold as the one quantum computing pioneers express today. Moore believed engineers would manage to pull a rabbit out of a hat five times in ten years. But he didn’t go further than 1975 – it seemed clear that at some point the magic would run out and things would slow down.

Moore’s prediction, which came to be known as Moore’s Law, played out for much more than a decade. Every two or three years since the mid-1960s, engineers have managed to double the transistor density of microchips. Transistors that were as wide as wool fibers (20 micrometers) in the late 1960s have shrunk to microscopic scales. In 2021, IBM created a chip with its smallest components just two nanometers wide — narrower than a strand of human DNA.

To keep Moore’s Law alive for the last several decades, engineers have relied on a chemistry trick.

In the 1980s, a team at IBM discovered a new method for printing transistors onto chips. They mixed string-like, branching molecules into a solution, and painted that solution – known as a photoresist – as a thin coating on the surface of a blank chip.

Once the solution dried, the molecules stuck to the surface. Then, the team shone an ultraviolet light on that surface through a patterned screen. The screen made sure that only some of the molecules were exposed to the light, while others were left in shadow. 

The branches on those chemical strands reacted under light. When they reacted, they changed how the photoresist behaved, making it more or less sticky. The researchers washed the UV-exposed photoresist away with water. The molecules that stayed behind formed complex patterns on the silicon wafer.

Today, those patterns act as guides for microchip wiring. Manufacturers print transistors on top of these patterns, using them as guides for the fine computing structures.

“When you look at these patterns under a microscope it’s remarkable how neat those lines are, with crisp edges,” said Jeannette Garcia, Senior Research Manager for Quantum Applications Research and Software at IBM Quantum.

That crispness matters – any imprecision could lead to errors in the manufacturing process and useless microchips.

IBM has worked closely with partners like JSR to refine this process, engineering the photoresists for fine control over the shape of the patterns at nanometer scales.

“With precise chemistry you can get down to these incredibly small feature sizes, no wider than the photoresist polymers,” Garcia said. “That’s how we’ve gotten down to two nanometer-wide components. It drives Moore’s Law.”

Today, the entire microchip manufacturing industry depends on the photoresist process.

 “We at JSR are proud to be one of the leading manufacturers of photoresist solutions in the world,” said Hiroaki Tokuhisa, Chief Technology Officer at JSR. “We supply the chemicals that drive Moore’s law in the 21st century, and work closely with partners like IBM to constantly improve our library of photoresist chemicals.”

Like the microchips they help manufacture, those photoresists have become much more complicated since the early experiments in the 1980s. As the chemistry has evolved to support finer and more delicate patterns, new elements have been added to photoresists to turn them into more precise instruments. For example, researchers have introduced chemical components known as photo-acid generators (PAGs) to the solution.

PAGs act a bit like chemical tugboats, said Garcia, nudging the larger polymers into place. When certain conditions are met, a PAG will spit out a proton that interacts with the polymers in the photoresist, making the molecules soluble so they can be washed away. When manufacturers develop new microchips, they work with JSR to determine the precise photoresist solution needed to get the desired results.

This process can be time-consuming and expensive.

“It’s difficult to predict how a new photoresist will behave until we have created it in the laboratory and rigorously tested it in real-world situations,” said Hiroaki.

The chemistry involved is too complex for even the most powerful supercomputers in the world to simulate effectively.

“We believe this is about to change,” said Hiroaki. “Together with our longtime partners at IBM, we are experimenting with chemical simulations on quantum computers. Already, we’ve shown that quantum computers can simulate small molecules that mimic parts of a photoresist.”

The real world runs on quantum mechanics, and quantum computers could soon be our best tools for simulating it. These computers, now undergoing their own rapid scaling and development process at IBM Quantum, may one day cut through complex problems that stump even classical supercomputers. 

With the aid of computer chemistry simulations, JSR aims to develop new photoresists more quickly and at lower cost – a potential advantage in extending Moore’s Law into the future.

IBM and JSR expect quantum computers to be powerful tools for this kind of chemical simulation once they reach the necessary scale and power. JSR is working with IBM Quantum today to lay the groundwork for that future.

“As quantum computers get more powerful, we want to be ready to use them to support our work,” Hiroaki said.

Recently, a joint JSR-IBM Quantum research team successfully simulated a smaller molecule with similar behaviors to a PAG. This showed that in principle it should be possible to simulate the PAGs themselves as quantum computers scale.

All this work is driving toward a future where quantum-centric supercomputers solve problems that are impossible to solve today, with near-term benefits to chemistry research. For JSR, that’s expected to mean better, faster computer chips produced at lower costs. For other partners, that could mean advances in drug discovery or materials science.

Today, IBM Quantum hosts the world’s most advanced fleet of quantum computing systems and software for executing quantum circuits at scale. Your organization can partner with IBM Quantum to drive research and build quantum skills.

Using technologies cultivated through the development of polymer materials, JSR Corporation (link resides outside ibm.com) develops and supplies many global leading products including lithography materials, CMP materials, process materials, and packaging materials, which are essential to the production of semiconductor chips. JSR’s LCD materials and next-generation display materials are used in the production of LCD and OLED displays.