An Al film is formed on a cap wafer and the Al film is patterned into a ring-shaped film. Dry etching is performed by using the ring-shaped film as a mask to form a drum portion enclosing a recess portion to provide a vacuum dome. After forming a depth of cut into the substrate portion of the cap wafer, the cap wafer is placed on a main body wafer having an infrared area sensor formed thereon. Then, the ring-shaped film of the cap wafer and the ring-shaped film of the main body wafer are joined to each other by pressure bonding to form a ring-shaped joining portion.