An aging semiconductor facility and its workforce benefitted from robotic automation.
Learning objectives
- Understand how robotic automation and artificial intelligence-based fleet management improve efficiency and precision in semiconductor manufacturing.
- Recognize the unique challenges of modernizing legacy semiconductor fabs and how mobile cobots can address them.
- Evaluate the strategic and financial impact of automation by shifting the perspective from cost-cutting to long-term gains in productivity, reliability and competitiveness.
Automation insights
- An U.S. semiconductor fab modernized its operations by deploying mobile robots and artificial intelligence-based fleet management software to automate wafer cassette handling, reducing labor strain, increasing precision and eliminating costly production errors.
- The initiative demonstrates how automation, once considered cost-prohibitive, is now essential for improving productivity, overcoming workforce shortages and remaining competitive in chip manufacturing.
There are some 70 existing semiconductor chip plants in America, all of which have opportunities to automate and ramp up production. Automating these facilities eases current workforce issues, boosts reliability and increases overall productivity.
Nowhere else was this evident than at one of the country’s oldest semiconductor fabs that was the first to mass produce the new style chips used in electric vehicles, 5G and 6G smartphones, power grids and data centers.
The facility, which currently employs 1,800 people, was built back in the 1950s and as part of its modernization, the company opted for the combination of collaborative robots, or cobots, on automated mobile robots (AMRs) along with fleet scheduling/management software. It based the decision on the fact that the older facility was designed for human workers and, as such, its production floor layout lacked big, wide aisles and open areas, making it tight and clustered, which limited automation options.
Automation helps with worker shortages
One of the main reasons the semiconductor fab, like many others in the semiconductor industry, looked to automation was in response to the difficulty of finding, hiring and retaining workers for physically demanding roles that require 12-hour shifts on their feet. In addition to wearing complete cleanroom suits, gloves and fully enclosed face masks, workers push carts loaded with semiconductor cassette boxes back and forth between stockers and fab tools/processes — walking several miles per workday doing so — then manually loading and unloading the cassettes that weigh 22 pounds each.
Now, robots equipped with special grippers go to the stocker, retrieve a box, drive to the designated tool, open the box, load the cassettes that hold the wafers and then return the empty box to the stocker. The robots work 24/7 without time lost to bathroom breaks, sick days and other issues associated with human labor.
For semiconductor automation projects, capital expenditure costs are highly dependent on the customer’s goals and the necessary software development needed to integrate the fab’s infrastructure into the robot fleet management software. For example, some customers are content with manually summoning robots, which is a relatively low software burden, while others require the complete integration between the factory’s infrastructure, automated material handling system (AMHS), and KUKA’s fleet manager.
KUKA’s typical goal with these projects is that after the second purchased robot, they are fully amortized within approximately 36 months. The first robot tends to be a bit more expensive mainly because all the nonrecurring engineering work is part of that stage of the project.
Another key reason for automation was to reduce errors and achieve a zero scrap rate. The robots don’t make mistakes and always load the right tool with the right cassette. Automation also reduces mishaps with cassettes and wafers. A cassette holds 25 wafers with each one valued at about $30,000 when completed, so accidents that damage wafers are costly.
About 25 years ago, the facility installed an overhead hoist track for moving wafer boxes around the various work areas, but the cassettes were still loaded and unloaded manually. This is why the facility needed automation that did both transporting and loading/unloading.
Robots enhance automation
To alleviate the various challenges, the KMR iiwa was selected by plant management to provide a nimble, flexible and autonomously navigating mobile robot that can be used to move cassette boxes safely through production processes.
For wafer cassette box handling, the KMR iiwa combines the strengths of KUKA’s sensitive LBR iiwa seven-axis cobot with those of a mobile, autonomous platform that is ISO/FDIS 14644-14: Cleanrooms and associated controlled environments compliant, location independent and highly flexible. The sensitivity of the LBR iiwa allows it to handle delicate wafer boxes safely and without vibration and a low-friction gripper allows for safe handling of 200 and 300 mm wafer boxes, front opening unified pods or standard mechanical interface.
The KMR iiwa is a precise, trackless production assistant that runs on the factory floor, detects obstacles in its path and moves in any direction from a standing start. Additionally, this mobile solution uses inductive charging technology so that when the platform is stationary over a charging station during operation, such as loading or unloading, it is charged without losing production to downtime.
KUKA’s next generation of the KMR iiwa is the KMR iisy. That version safely moves faster in semiconductor facilities using safety scanners and 3D cameras in combination with the LBR iisy cobot. The system detects not only humans around the platform with its light detection and ranging scanner and time-of-flight sensors, but also potential collisions with humans and the cobot. The 3D cameras can spot obstacles up to 6.5 feet above the ground.
Automated mobile robots
Along with the mobile robots, the semiconductor fabincorporated KUKA’s artificial intelligence-based KUKA.AMR fleet management software. The software enables the mapping of numerous industrial scenarios in which complex production environments with dynamic obstacles, persons, door openings and elevators are skillfully mastered. KUKA.AMR controls the fab’s entire AMR fleet autonomously, with obstacles avoided and routes replanned in real time according to traffic to ensure smooth and efficient operation. The semiconductor fabcan carry out the traffic and job management of all its AMRs in a single environment and the simple drag-and-drop configuring of logistics processes provide fast start-ups of the AMRs.
Before modernizing with automation, semiconductor companies need to have scheduler/automatic material handling system software in place to indicate what cassette boxes need to go where.
Additionally, the fabrication facility had to have a way to identify individual boxes other than the use of a piece of paper attached to them, which is extremely difficult or even impossible to use with automation. The alternatives for work in progress range from simple 2D barcodes to radio frequency identifications thatrobots can read.
Once the initial AMRs are deployed, ongoing training empowers the semiconductor fabto become self-sufficient and to equip employees with the technical skills needed for long-term operational independence. Additionally, the KUKA fleet management software is compatible with the facility’s semiconductor SEMI E82: Specification for Interbay/Intrabay AMHS SEM standard fabs used to control the material flow inside the facility, which made the automation transition more seamless.
Robotic automation improves profitability
Regardless of whether a semiconductor company is building from the ground up or modernizing an older existing facility, one constant challenge remains: finding human capital with the willingness and/or training to work while keeping pace with production demands. Automation eliminates the necessity for a human to perform repetitive and monotonous transport tasks and thus the resulting risk of costly mistakes. Robotic automation also works all day without breaks or interruptions to increase speed for improved production.
Additionally, physical plant and space limitations are easily overcome with the right robot automation. What might be a more complex challenge, however, is overcoming the idea that automating is cost prohibitive.
Given the current state of global fab production, it’s not whether automation is a viable option but whether a fab can survive without automating. Automation should be viewed more as an increase in profitability rather than a way to manage or cut costs. Gains in reliability and output generated through automation are direct boosts to a fab’s bottom line.
