Optical fabrication process steps have remained largely unchanged for decades. Raw glass blanks have been rough-machined, generated to near net shape, loose abrasive or fine bound diamond ground and then polished. This set of processes is sequential and each subsequent operation removes the damage and micro cracking induced by the prior
operational step. One of the long-lead aspects of this process has been the glass polishing. Primarily, this has been driven by the need to remove relatively large volumes of glass material compared to the polishing removal rate to ensure complete damage removal. The secondary time driver has been poor convergence to final figure and the corresponding polish-metrology cycles. The overall cycle time and resultant cost due to labor, equipment utilization and shop efficiency is increased, often significantly, when the optical prescription is aspheric. In addition to the long polishing cycle times, the duration of the polishing time is often very difficult to predict given that current polishing processes are not deterministic processes. This paper will describe a novel approach to large optics finishing, relying on several innovative technologies to be presented and illustrated through a variety of examples. The cycle time reductions enabled by this approach promises to result in significant cost and lead-time reductions for large size optics. In addition, corresponding increases in throughput will provide for less capital expenditure per square meter of optic produced. This process, comparative cycles time estimates and preliminary results will be discussed.
Magneto-rheological finishing (MRF) is a deterministic figuring process capable of quickly achieving extreme surface accuracies. The commercially available Q22 has been instrumental in the manufacture of DUV lithography optics to better than 30 nm P-V figure and 1.0 nm rms microroughness. The requirements for EUV optics, photomask substrates, and silicon-on-insulator (SOI) wafers, however, have taken "extreme accuracy" to new levels. Surface quality is specified over a broad range of spatial frequencies, and allowable error magnitudes shrink ever smaller. These specifications expose some limitations of sub-aperture tool technologies. MRF capabilities, recent developments, and future system improvements that address these concerns are described. We present polishing results on photomasks that pass flatness requirements until year 2010. We further demonstrate extreme precision figure correction capability on SOI wafers, achieving thickness uniformity of better than 2 nm PV and 0.3 nm rms.