Keysight Technologies use simulation-powered digital twin to increase ROI on capital expenditures

How did they do it?

Keysight wanted to identify and remove any barriers or bottlenecks that affected productivity within its fabrication facility to meet its long-term objectives for continued growth.

Its team of engineers had some ideas as to what might be obstructing throughput, including the volume of equipment redundancy, job prioritization, and the number of operators on site, but they were not so eager to commit to any structural changes to the fab layout without first testing them in a digital environment.

Using Simul8, the team highlighted potential throughput increases for specific tools across the factory floor. The model enabled them to continuously review processes within the fab layout and implement changes to increase the rate of throughput.

The Complexities of the Project

Wafer fabrication involves a series of up to 700 different processes to transform sand into a semiconductor slice that can be prepared for microchips and photovoltaics. This intricate work involves a combination of manpower and technical machinery.

Each segmented process on the production line is interdependent, and it is essential that the finished semiconductor is free from damage and imperfections.

However, the scale of the operation means there are risks of disruption throughout production. This could be in the form of machine downtime, operator availability, and work in process (WIP) volume i.e. the number of unfinished commodities in the inventory. Any shift in the status quo has the potential to impact Keysight’s productivity levels.

The Method

A specialist simulation team at Keysight worked with Simul8 to create a fully functioning simulation for Keysight’s wafer fabrication. The simulation was reviewed and improved over three iterations to enhance it with further features, including customized chamber modeling, dual assigned tools to capture redundant systems, and streamlined modeling features to allow results at shorter intervals.

The first iteration looked at the fabrication management policies. This included gauging the correct number of operators on the factory floor and identifying how the production line could handle any increase or decrease in WIP. Next, there was an in-depth look at each individual piece of machinery’s processes before moving onto reviewing assembly and exploring tool utilization and productivity impact.

Using Simul8, the team highlighted potential throughput increases for specific tools across the factory floor. The model enabled them to continuously review processes within the fab layout and implement changes to increase the rate of throughput. One process that was identified as having a bottleneck was photolithography, which involves etching a pattern of an electric circuit on the wafer’s surface, courtesy of a lithography tool.

Unfortunately, Keysight was experiencing issues with the uptime of the lithography tool and the supplier’s response times to maintenance requests, as it would take three days to get a response and seven before an engineer arrived on site to carry out repairs. Times that were unacceptable for Keysight’s needs. Using simulation, they were able to renegotiate maintenance agreements with this supplier to reduce downtime.

The Results

The wafer fabrication simulation has enabled Keysight to better understand the impact of priorities, lot size, and quantity of WIP on the floor. It has helped them determine the best structure to fit production needs, identify critical failure points, and examine the return on investment for key capital expenditures. It is also used to review staffing requirements and identify any gaps in coverage.

The simulation will help the team to review:

• Number of operators required
• Tool redundancy
• Critical KPIs review such as turnaround time and throughput
• WIP impact
• Lot size impact