Advanced wafer fabs around the world now have robotic help with optimizing critical maintenance tasks on wafer fabrication equipment.

Lam Research’s collaborative robot, Dextro, is the first cobot of its kind, Russell Dover, general manager of the company Customer Support Business Group, told EE Times in an interview. The cobot enables accurate, high-precision maintenance to minimize tool downtime and production variability, he said, which also contributes to significant first time right results that can enhance yield.

Dextro is collaborative in that it was designed to work alongside fab engineers to help complete complex maintenance tasks with precision and repeatability that are beyond human capability. “When we think of industrial robots, we tend to automatically go to car assembly lines where the material is moved to large robot systems and they perform their function and then they move on,” Dover said. The Dextro cobot is run and managed by a person and it must be designed to work in a human environment, he said.

Dextro is a mobile unit with a robotic arm that uses various end effectors as hands to manage critical equipment maintenance tasks that are time-consuming and prone to errors when done manually. For example, it can precisely install and compress consumable components with more than 2× the accuracy of manual application. This precise assembly helps control etch performance at the wafer edge, improving yield.

Dover said maintenance schedules for fab equipment can vary widely, but when it comes to etching tools—which comprises a large part of Lam’s business—they are not only etching materials, but they are etching themselves. “It is said that etch tools eat themselves,” he said. This means they require maintenance more frequently when compared with a metrology tool that is only used on a sampling basis.

This makes maintenance a high value problem, Dover added. “What these tools can achieve is astonishing in terms of their precision in terms of the chips they’re manufacturing, and how you disassemble and reassemble a chamber as part of its maintenance is quite important.” If a mistake is made, he said, it will impact how the equipment performs in production.

Dover pointed out that there is an appetite for maintenance automation while increasing the precision accuracy. “It has a real correlation to the inline performance.”

Another task Dextro is capable of is tightening vacuum-sealing, high-precision bolts to exact specifications, relieving fab engineers of a repetitive task that has up to a 5% error rate when done manually. The cobot can also use automation and cleaning technology to remove side-wall polymer build-up within the chamber, without the burden of disassembling the lower chamber—all while lowering the risk to humans who would need heavy protective breathing equipment to perform the task manually.

One of the reasons Dextro can handle multiple maintenance tasks is because it was designed to be moved to each chamber that needs maintenance, Dover said, rather than just being one unit dedicated to one chamber. “It can be shared across chambers the same way that human personnel would be shared across chambers.”

He added that a key goal while developing Dextro was how to innovate around using multiple end effectors and diverse programming to achieve different tasks. “We wanted a multi-application capable robotic unit to try to achieve as much of the maintenance as possible.”

Dover said developing a cobot that could do more than one task required some key learning for Lam because it is a mobile unit. This means it must physically dock at the chamber, and the robot and its end effectors must be calibrated so it knows where it is and what it is looking at. “We are process engineers and developing that sort of robustness around robotics programming, particularly vision systems, was an area where we had to take initiative and grow.”

Developing Dextro is also fundamentally a change in management exercise that could potentially create resistance as it means going about maintenance differently, Dover added. “That change management barrier has been significantly lowered just because customers are so enthusiastic and excited about the potential that this sort of cobot approach to maintenance can bring.”

FROM EETimes