Isolation method of semiconductor device

摘要

A device isolation method divides a semiconductor substrate into active and inactive regions. A first device isolation layer (38) is formed in a first inactive region using a trench isolation method. Then, local oxidation is used to form a second device isolation layer (46) in a second inactive region which is wider than the first. The dishing phenomenon commonly generated during CMP processing is eliminated, and proper device isolation is realized without exposure of the active region. <IMAGE>