Skip to Content

IBM’s Chip-Shrinking Secret

New tricks with light and lenses could produce the smallest microprocessors – without revamping the industry.
February 27, 2006

A novel chip-making strategy forged by IBM researchers could break through previous constraints, allowing semiconductor makers to use their same basic tools to continue shrinking chip features. If the strategy pans out, it would enable more advances in computing speed and power – without requiring a long-feared multi-billion-dollar industrial retooling.

The new IBM method boils down to a modification of conventional lithography – part of the process that converts a silicon wafer to a microprocessor. With this modification, the company can create circuit features that are just 29.9 nanometers wide, down from the 65- and 90-nanometer-wide features of today’s mass-produced chips, effectively doubling or tripling chip speed. “The advance is ahead of its time,” says Don Bethune, an IBM optics expert who worked on the project.

While IBM may still be years away from actually manufacturing 30-nanometer chips, the announcement last week smashes industry assumptions. Fashioning circuit features much smaller than 45 nanometers was thought to present a serious challenge because the conventional method for etching features in silicon cannot provide the resolution needed. In fact, many researchers assumed that the 32-nanometer mark would probably require entirely new equipment, materials, and facilities.

With today’s manufacturing technology, ultraviolet light is shone through a stencil-like “mask” that outlines where the circuit patterns are to be made on the silicon wafer. As the feature sizes on the mask become smaller, the type of light used provides a natural limitation. If the wavelength of light is much larger than the lines in the mask, the resulting features on the silicon are blurry. Currently, ultraviolet light with a wavelength of 193 nanometers is used, and a layer of water and a lens effectively shorten the wavelength of the light and focus it tightly enough to etch circuit features down to about 65 nanometers.

While it might seem obvious to use a shorter wavelength of light, this presents huge challenges for manufacturing, says Bethune. Since air molecules tend to absorb the shorter wavelengths of light, the manufacturing process would have to be conducted in a vacuum and under other special conditions, none of which have been fine-tuned for manufacturing. For these reasons, the semiconductor industry has been trying to stave off this transition for years.

IBM’s solution uses the same lithographic process and 193-nanometer light, but with a crucial tweak. The researchers have added a process called “high-index immersion.” It replaces the layer of water with a fluid containing complex organic molecules, says Bethune.

This fluid has a higher index of refraction – a property that can effectively shorten the wavelength of light – than water. In addition, the researchers switched to a lens made from a material that’s also better at refracting light than conventional systems. The team optimized its liquid and lens combination on a customized piece of equipment, called Nemo, which tests next-generation lithography materials.

Although the group’s technology is not yet ready for full-scale manufacturing, its results “demonstrate the potential to be able to do this,” says Fred Zieber, a semiconductor analyst at Pathfinder Research.

Other companies who make chips may also be exploring similar lithography tactics; but Intel, which announced a 45-nanometer “test chip” in January (see “Moore’s Law Lives”), has not given details about its lithography processes for that generation. John Casey, an Intel spokesperson, says they are looking at different options for technology smaller than 45 nanometers, which could include an immersion liquid, as IBM has, or also shorter wavelengths of light.

The lower limit of microprocessor size is still getting closer, though. So it’s an open question how long Moore’s Law – the prediction that the number of transistors on a chip will double every 18 months or so -– will survive. Although the time between generations of new chips may grow longer as the physical limitations become more difficult, Bethune says, conventional lithography could still hold some surprises. “The ultimate limit with 193 nanometer light depends on the materials you use,” he says. “It’s not sharply defined.”

Home page image courtesy of IBM. Caption: Left: a record-small array of lines and spaces that are 29.9 nanometers wide (about 3,000 times narrower than a human hair), created using a variation of optical lithography. Right (same magnification): 90-nanometer-wide features now in mass production in the microchip industry.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

How scientists traced a mysterious covid case back to six toilets

When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.