In-situ measurement of lens aberrations

Linewidth control across an exposure field is becoming increasingly challenging as design rules shrink. Contributions to linewidth variation can arise from the reticle, the exposure tool and the resist process. For the exposure system, errors may originate from the illuminator set-up, the projection lens aberrations or the stage focus and leveling performance. This paper will describe the characterization of lens aberrations using a new reticle and measurement technique. The technique uses a special reticle, which converts wavefront phase errors to displacements on the wafer. These offsets can be measured using conventional overlay tools with greater speed and accuracy than SEM measurements of small linewidths. Reconstruction of the wavefront using this data provides a more reliable in-situ characterization of aberrations. The principle of the measurement technique and the design of the reticle are described. The technique can be applied using standard photolithography processing and without modification of the exposure equipment. Typical processing conditions are outlined. Preliminary results from a high numerical aperture 248nm lens, using both static and dynamic exposures, showed a wavefront root mean square (rms) error, at a single point, of about 6.7nm and 8.8nm respectively. The repeatability of the method was about 0.1nm rms, due to measurement, and about 0.6nm rms, due to wafer processing. Preliminary results on determining the accuracy of the technique, by comparing measured focus to set focus offsets in the exposure system, showed less than 3% error. Applications of the method to study aberrations over a full scanned field and chromatic aberration were also carried Out In this lens, high order aberrations were found to vary between 7.5nm and 13.8nm but showed little change with wavelength.