A first semiconductor chip includes a fixed electrode formed on a first semiconductor substrate and a plurality of first metal spacers formed on a first interlayer dielectric. A second semiconductor chip includes a vibrating electrode formed on a second semiconductor substrate and a plurality of second metal spacers formed on a second interlayer dielectric. The first and second semiconductor chips are metallically bonded to each other using the first and second metal spacers. An air gap is formed in a region of the condenser microphone located between the first semiconductor chip and the second semiconductor chip except bonded regions of the first and second metal spacers.