Nanoimprint lithography hopes to make an impact

After years of development, proponents of a lithographic technology are hoping to make an impression — in more ways than one. Nanoimprint lithography has been around for decades, noted Byung Jin Choi, vice president for technology at Canon Nanotechnologies. He discussed nanoimprinting in a presentation about the technology given at SPIE Advanced Lithography and Patterning. He noted that certain advances position the technology for growth.

Nanoimprint lithography is a method for making make nanometer scale patterns. In this method, systems press a mask into resist and thereby pattern it. Follow-on processes of curing and etch then transfer the pattern onto layers on a wafer. Depositing additional layers and repeating these steps creates a structure. The end result can be the same as that from the projection-based lithography widely used in semiconductor manufacturing to make memory and logic chips.

Nanoimprint lithography could have wide applications, Choi said. “Industries include not only semiconductors but also AR and VR.”

The augmented reality and virtual reality uses come from the technology’s ability to pattern optical materials, producing novel features and properties. These may enable a reduction in the number of optics needed to get light into a user’s eye, making AR and VR headsets lighter and  enhancing performance.

For nanoimprinting to take off, many different elements need to be in place. The list includes the capability to make the mask, with simulations before mask making and verifications after to ensure the mask is correct. The next step is to use the mask to make devices, which in a production environment means worrying about throughput and yield. There may also be an intermediate step in which the mask makes a lower cost replica used for production. If so, that step also has its own requirements.

These various areas and tasks are related and essential, but none are all done by one company. So, there needs to be an ecosystem, an industrial network that supports nanoimprint lithography and advances its capabilities. Canon is working to create that while it also develops its own nanoimprint tool, according to Choi.

In his presentation, he discussed the results of the company’s own systems in producing widely used semiconductor memory and logic chips at different feature sizes. Choi reported that Canon’s tool can meet the line and space requirements for flash memory and perhaps could do so for DRAM memory at commercially important process nodes.

However, during fabrication of a wafer, layers are patterned one at a time. Thus, how well layers align to one another is important, and that is where the nanoimprint lithography machine faces difficulties.

“What is more challenging is the overlay,” Choi said in summing up the situation. He added that Canon is working to get overlay well below two nanometers, a significant improvement.

He reported promising results toward that goal through the application of several techniques. One is the correction of fixed distortion errors in the mask, an approach characterized as very successful overall.

Canon also uses dispensing technology that allows the resist to be substantially thicker in some spots than others. This capability helps because nanoimprinting squeezes the resist out of some places but not others due to peaks and valleys in the mask. Adjusting the beginning resist thickness minimizes this variation and makes the dimensions of critical features on the wafer more uniform.

In talking about the technology’s future, Choi noted that challenges exist that require further research and development. He added, though, that nanoimprint lithography offers more freedom in optical design, a reason for its AR and VR applications.

Nanoimprinting could also offer a significant advantage when it comes to making 3D structures. The mask can directly create a structure in x, y, and z directions instead of going through multiple patterning steps. The resulting simplification would save time and money, according to Choi.

“You can save up to 40 percent, we estimate,” he said of this approach.

Hank Hogan is a science writer based in Reno, Nevada.