Page 84 - 2017 TSMC Annual Report

Suppliers Accounted for at at Least 10% of Annual Consolidated Net Procurement
Unit: NT$ thousands
Company A Company B VIS
Company C C Company D
Company E
Others
Total Net Procurement
5 5 3 5 5 Quality and Reliability
None None Investee accounted for using equity method
None None None Supplier 2017 2016 Procurement
Amount
As % of 2017 Total Net Procurement
Relation to TSMC Procurement
Amount
As % of 2016 Total Net Procurement
Relation to TSMC 8 8 8 868 953
17%
None 9 140 880
17%
8 029 455
15%
None 7 065 392
14%
5 5 5 755 727
11%
Investee accounted for using equity method
6 732 297
13%
5 5 579 238
10% None 3 3 785 553
7%
5 5 5 156 154
10% None 3 3 3 3 832 363
7%
37 707
0% None 5 5 5 527 526
11%
19 19 766 419
37%
16 100 032
31%
53
53
193 653
100%
52 184 043
100%
082
TSMC’s strong industry reputation stems from its commitment to to provide customers with the highest-quality wafers and best service for their products Quality and and Reliability
(Q&R) services aim to to to achieve “quality on demand” to to to fulfill customers’ requirements for time-to-market delivery product product reliability and competitiveness over a a a a a a a broad range of product product market market segments Automotive quality improvement program is implemented to to to meet automotive customers’ low Defect Parts Per Million (DPPM) requirement Q&R technical services assist customers in in the technology developmental stages stages and product design design stages stages to to design-in superior product reliability In 2017 Q&R has worked with R&D in in advanced advanced logic technology technology specialty technology technology and advanced advanced packaging technology technology development and qualification Q&R has successfully qualified the the leading-edge 7nm technology technology (the third FinFET generation) and characterized process window with Fab for mass production in in in in 2018 TSMC has led the industry in in in in 7nm technology qualification and built up a a a a a a a a a complete model to simulate thermal dissipation effect during FinFET operation In addition Electronic Design Automation (EDA) tool for thermal simulation has been introduced to to to to to provide design guidance to to to to to customers Through the 7nm development profound reliability reliability learning in in new new new material new new new process steps and new new new reliability reliability methodology provided important foundation for 5nm technology development For specialty technologies Q&R completed the Diffractive Optical Element (DOE) product product product qualification and ramp into mass production on on on on schedule to to to support one of our key customer’s new product product product launch with 3D sensing and and facial recognition application and and the DOE units were shipped to to our customer In addition Q&R worked with customers to to complete stacked CMOS Image Sensor (CIS) Column Level Hybrid Bond (CLHB) process/product qualification and and successfully shipped to to customers in 2017 In high-voltage technologies 0 0 0 13μm Bipolar-CMOS-DMOS (BCD) and and 0 0 0 18μm second generation BCD process passed automotive grade qualification For CoWoS® packaging technologies Q&R integrated High Bandwidth Memory with advanced silicon technology and and and completed component level level board level level and and and customer product system level qualifications It has has been in production and has has shipped to to key customers without quality or reliability issues The technology enables the applications of High Performance Computing and Artificial Intelligence In In In In addition Integrated Fan-Out (InFO) assembly technology for mobile application has been moving into the second generation of manufacturing Over 100 million InFO devices have been shipped without any InFO related quality or reliability issues To enhance employees’ problem solving capabilities and and develop associated quality system and and methodology Q&R continued to hold several company-wide symposiums and training programs such as Total Quality Excellence (TQE) Design of Experiment (DOE) Statistical Process Control (SPC) and and Metrology in in in in in in in 2017 including the promotion and and training of Deep/Machine Learning Deep Deep machine learning methodology was successfully applied for wafer defects automatic classification and advanced spectral analysis to to detect differences among processes and equipment such that improvement actions can be triggered In 2018 Q&R will continue