Microelectronic Engineering 6 (1987) 597-603 597 North-Holland COMPARISON OF FOCUSED ION BEAM AND LASER TECHNIQUES FOR OPTICAL MASK REPAIR P J Heard*, P D Prewett+ and R A Lawes+ * lon Beam Systems Ltd, Drayton Road, Abingdon, Oxen, England + Rutherford Appleton Laboratory, Chilton, Didcot, Oxen, England ABSTRACT As the critical dimensions of masks and reticles approach i ~m, the limitations of the laser repair technique are becoming increasingly apparent. The use of focused ion beams for mask repair offers a process with significantly improved resolution over that achievable using lasers. A comparison of the two techniques for opaque defect repair reveals a significant degree of mask damage associated with laser repair. In contrast, the FIB process shows much better edge acuity with little damage to the mask. Some staining of the glass substrate does occur but this is found not to be a problem during subsequent printing using standard processes. 1 INTRODUCTION Opaque defects in chrome-on-glass optical mask plates are usually repaired using laser pulses of sufficient energy to cause vaporisation of the excess chromium. The process is fast and simple, but the resolution of the method is limited by the wavelength of the light used, eg 694 nm for a ruby laser. As the critical dimensions of masks and reticles are reduced, a repair process with superior resolution is required. This has led to the development of focused ion beam (FIB) repair, in which a beam of ions is used to mill away excess chromium for the repair of opaque defects, or to produce material deposition to repair a clear defect. A comparison of the techniques of laser and FIB repair of opaque defects has been made by direct observation of the effects of the repair on the mask and of the resist image after contact printing. Both techniques cause damage to the glass substrate. The FIB technique produces a "staining" effect in the glass, in which light transmission through the repaired region is reduced. Laser repair melts the glass substrate and/or redeposits evaporated chromium in the vicinity of the repair. A comparison between the laser and FIB techniques was made by using a specially designed mask with repeated, identical "defects". The mask consisted of an array of die similar to that shown in Figure I, each containing programmed defects such as extensions, ratbites, chrome spots etc. Four chrome extensions and three pinspots were removed and three holes were either laser zapped or FIB milled into the test mask. The objective was to compare the effects of damage, staining and the edge quality produced by the two techniques. 0167-9317/87/$3.50 © 1987, ElsevierSciencePublishers B.V. (North-Holland)