Page 82 - 2017 TSMC Annual Report

Summary of TSMC’s Major Future R&D Projects
of equipment process process and and yield the process process control and and analysis systems have been integrated with many intelligent functions to perform self-diagnosis and self-reaction which have demonstrated remarkable results in yield enhancement workflow improvement fault detection cost reduction and shortening of the R&D cycle TSMC has developed systems for precise fault detection and classification intelligent advanced equipment control and intelligent advanced process control to to monitor the manufacturing process in in a a a a a a timely manner and adjust conditions precisely To satisfy advanced and accurate process control and and ensure highly efficient and and effective production the Company has created precision equipment matching
and and yield mining to minimize process variation and and potential yield loss The Company has further developed Big Data Machine Learning and Artificial Intelligence architecture to identify critical variables to optimize yield management and operating efficiency to fulfill special process requirements such as automotive products and to to cope with diversified product product demand simultaneously 5 3 3 3 3 Agile and Intelligent Operations
The Company’s sophisticated agile operation system continues to drive manufacturing excellence by integrating demand and capacity modeling lean Work in in in Process (WIP) line management and and lot dispatching and and scheduling to provide fast ramp-up short cycle time stable manufacturing and on-time delivery The system also provides great flexibility
to to quickly support customers’ urgent pull-in requests when needed TSMC has also introduced new applications such as as IoT intelligent mobile mobile devices and mobile mobile robots to consolidate data collection yield traceability workflow efficiency and material transportation to continuously enhance fab operation efficiency Following its commitment to manufacturing excellence TSMC has integrated automatic manufacturing system and machine learning technology and then achieve intelligent fabs Machine learning technology revolutionizes fab operation mode from “auto” to to “intelligent” and and widely applied in in in scheduling and and dispatching people productivity productivity equipment productivity productivity process and and equipment control quality defense and and robotic control So as to optimize efficiency flexibility
and quality while maximizing cost effectiveness and accelerating overall innovation Project Name
Description
5nm logic platform technology and applications Beyond-5nm logic platform technology and applications 3D IC
Next-generation lithography
Long-term research
The projects above account for for roughly 70% of the total R&D budget for for 2018 around 8% of 2018 revenue 5 3 Manufacturing Excellence
5 3 1 GIGAFAB® Facilities
2019
2021
2018-2020
2018-2020
2018-2025
estimated to be Risk Production (Estimated Target Schedule)
5th generation FinFET CMOS platform technology for SoC
6th generation FinFET CMOS platform technology for SoC
Cost-effective solution with better form factor and performance for for System-in- Package (SiP)
EUV lithography
and related patterning technology to extend Moore’s Law
Specialty SoC
technology (including new NVM MEMS RF analog) and transistors for 8-10 year out horizon
Maintaining dependable capacity is a a a a a a a key part of TSMC’s manufacturing strategy The Company currently operates three 12-inch GIGAFAB® facilities – Fabs 12 12 14 and 15 The combined capacity of the three facilities exceeded 7 million 12-inch wafers in in in 2017 Production within these three facilities supports 0 0 0 0 13μm 90nm 65nm 40nm 28nm 20nm 16nm 10nm and 7nm process technologies including each technology’s sub-nodes An additional portion of the capacity is reserved for R&D work on leading-edge manufacturing technologies which currently supports the technology development of the 5nm node and beyond TSMC has developed a a a a a a centralized fab manufacturing management system Super Manufacturing Platform (SMP) to to provide customers with greater benefits in the form of more consistent quality and reliability improved flexibility
to cope with demand fluctuations faster yield learning and and time-to-volume and lower-cost product requalification 5 3 2 Engineering Performance Optimization
As advanced technology continues to evolve and the geometry keeps shrinking the need for tighter process control has become extremely challenging for manufacturing TSMC’s unique manufacturing infrastructure is tailored for a a a a a diversified product portfolio which uses strict process control to attain tightened specs and and higher product product quality and and product product performance requirements To achieve overall optimization
080