Image quality improvement in inspection systems using double integrator illumination

The most annoying problem accompanying production of high-fidelity pattern images in mask defect inspection systems is the generation of virtual images in the imaging process. The focused image pattern on the image acquisition sensor has two images, one true and one virtual. The virtual images are generated under Kohler's illumination using an integrator. The theoretical cause of this virtual image is the periodicity of the integrator. The improvement of image quality gives the mask defect inspection system higher defect detection sensitivity. To reduce virtual images, the double integrator method is applied to the illumination optics. By adopting the double integrator illumination method, virtual images disappear in the imaging field. Further, since this also lowers the power density at bright spots, the interference of lenses in working environments at the aperture stop position between objective imaging lenses is greatly reduced. This paper reports a method by which the ill effects of image quality improvement in the mask defect inspection system can be dramatically reduced. The simulation results when this method is applied to an advanced mask defect inspection system are shown.