Characterization of cylindrical micro-lenses in transmitted light and with grazing incidence interferometry in reflected light

High numerical aperture cylindrical micro-lenses are needed in collimating the laser light from laser bars in the near infrared. Diffraction limited performance of such collimation lenses can only be obtained if the surface shape deviates strongly from circular symmetry. Therefore, null tests make the use of diffractive optical elements (DOE) necessary. For performance test at the design wavelength the reference DOE produces an ideal cylinder wave which enables the compensation of the wave front coming from the micro-lens to a plane wave. The use of a DOE-master enables beside the null test geometry also a removal of the anamorphic distortion due to the cylinder geometry. The DOEs having a numerical aperture of 0.8 are produced on an e-beam machine. The measurement of the wave aberrations is done with the help of the phase shifting technique. Alignment aberrations are eliminated by a least square fit of suitable misalignment functionals derived from an analytic approximation. The shape of the micro-lenses is tested in reflected light showing surface defects directly. The cylinder symmetry allows for a grazing incidence test using two DOE with nearly constant spatial frequency. There are no limitations concerning the test of high numerical aperture surfaces since the structure of the diffractive elements are parallel curves to the profile curve. The mean spatial frequency of the DOE defines the effective wavelength. Since we use a diffractive interferometer the effective wavelength is identical to the pitch of the DOE. Usually pitches between 4-10 mu m are used resulting in a fringe sensitivity of 2-5 mu m. The test delivers the deviation of the surface from the ideal form. In the case of non-circular symmetry the DOE deviates from the axicon type DOE. In addition to the shape deviations also the radius of curvature at the vertex can be measured. The stage for the cylindrical micro-lens is equipped with a length measuring device using grating references from Renishaw providing a length increment of 0.1 mu m. With the axicon DOE there are two positions where nearly fluffed out fringes can be observed. Starting from the basic test position the lens can be moved until the vertex of the lens coincides with the focal line of the wave generated by the central part of the DOE. The distance between these two positions gives the radius of curvature for the vertex. Modem manufacturing of micro-lenses comprises also hot embossing in plastics or even into glass. Because of the small dimensions of the lenses and the required accuracy of the surface shape also tests of the impressing mould have to be carried out. The DOE approach enables also the test of the embossing form with the same sensitivity as the final test of the lens. Measuring results for the two test methods will be given