| 摘要 |
PURPOSE:To make it possible to measure relative displacement with high accuracy, by setting the cycle of one of two diffraction lattices to a value which is non-integer multiple (larger than 1) of the cycle of the other diffraction lattice. CONSTITUTION:Diffraction lattices A, B respectively having cycles dA, dB are arranged on objects 1, 2 parallelly-opposed to each other and the cycle dA is set to a value which is non-integer multiple (larger than 1) of the cycle dB. When coherent parallel luminous fluxes Ii is projected to the diffraction lattice A on the object 1, said luminous flux Ii is diffracted at angles + or -thetam to the surface normal of the diffraction lattice, determined by diffraction order (m), the cycle dA and a wavelength lambda to become I(+ or -m). These luminous fluxes I(+ or -m) are further diffracted by n-order by the diffraction lattice B having the cycle dB on the object 2 to be emitted at angles being + or -thetas to become signal lights I(m, -n), I(-m, n). When the cycle of one diffraction lattice is a value which is non-integer multiple (larger than 1) of the cycle of the other diffraction lattice, these multiple diffracted lights are emitted at angles respectively different from the emitting angles +thetas of signal lights and, therefore, the superposition of diffracted lights different in a route length to the signal lights can be avoided theoretically.
|
