Semiconductor & Computer Engineering
Wafer size refers to the diameter of a wafer and is an important parameter as part of the semiconductor manufacturing process. A larger wafer size enables the fabrication of more dies per wafer which translates into cost reduction in high-volume manufacturing.
Motivation and difficulties
Wafer size has increased greatly over the years from 23 mm (0.905 of an inch) in 1960 to 300 mm in nowadays. Foundries make the transition to a larger wafer because of the cost benefits resulting from producing the larger number of dice per wafer, all while maintaining many of the same process steps to produce the them. Unfortunately increasing the wafer size is far from being a trivial process - costing billions in research & development. The transition to 300 mm in the early 2000 cost the industry $10s of billions[1]. Most recently, the transition from 300 mm to 450 mm wafer will have an overall price tag of $25 to $40 billion[2] with other estimates that go as high as $100B. The monumental cost associated with the transition created considerable resistance in the industry fearing insufficient return on investment and incredible risk and uncertainty associated with the transition.
Historical Sizes
| Year | Wafer size[3][4] | Wafer thickness |
|---|---|---|
| 1960 | 23 mm | ? |
| 1962 | 25 mm | ? |
| 1963 | 28 mm | ? |
| 1969 | 51 mm | 275 µm |
| 1972 | 75 mm | 375 µm |
| 1976 | 100 mm | 525 µm |
| 1981 | 125 mm | 625 µm |
| 1983 | 150 mm ("6 inch") | 675 µm |
| 1992 | 200 mm ("8 inch") | 725 µm |
| 2002 | 300 mm ("12 inch") | 775 µm |
| ~2020+ | 450 mm | ~925 µm |
Note that wafers grown using materials other than silicon have different thicknesses than a silicon wafer due to the different stress and strain properties. Wafers must be tick enough to maintain structural integrity during processing and handling.