Ultrawave Labs (United States)

It is important to thoroughly clean all instruments of all debris, including blood and dental material, prior to sterilisation. Chris Czyrko discusses the role ultrasonic cleaners can play in cross–infection control

The performance requirements for ultra-thick photoresists are rapidly increasing with the dramatic growth in lithographic applications that require electroplating processes. Two of the main applications for ultra-thick photoresists are advanced packaging and nanotechnology (MEMS). Flipchip packaging has become widely adopted to address electrical device performance and chip form factor considerations. The growth in the nanotechnology market is driven by a wide range of products, which include accelerometers, ink jet print heads, biomedical sensors and optical switches. The requirements of thick photoresists for solder electroplating are significantly different from typical thin photoresists used in front end of line applications. As the photoresist becomes thicker, processing times increase for many process steps. Photospeed gets slower due to the requirements for more chemical reactions per area of coating. Coating uniformity and edge bead control also become more difficult as photoresist films get thicker and time delay issues between process steps can arise. This result has led to the requirement for special photoresist formulations for thick photoresist films. These are traditionally positive tone DNQ-Novolak materials such as AZ 50XT. Such materials can be designed to work for a particular range of thicknesses, but as the desired thicknesses increases the processing times can become very long for high volume manufacturing. Many new bumping schemes require photoresists in a 60 to 70 μm thickness range. While DNQ-Novolak chemistry can work, there is a desire for faster alternatives to improve total cost of ownership (COO) of the lithography cell. In order to have fast photospeeds and reasonable processing times a chemistry that is very photo efficient is needed. Negative tone cross linking chemistries, which can give tens of thousands of chemical events for one photochemical event, provide excellent photospeed and process times. Positive tone chemically amplified photoresist provide hundreds or thousands of chemical events per photochemical event. They are somewhat slower in photospeed than free radical materials, but still provide reasonable photospeeds. This paper compares the lithography and processing performance of these two newer types of thick film chemistries with the performance of a state of the art DNQ-Novolak thick film photoresist. The lithographic performance of these three ultra-thick positive photoresists were optimized to control critical dimensions (CD), sidewall angles and aspect ratios. The experimental results includes process latitude studies, electroplating performance and stripping performance. The general result is that negative free radical chemistry has the edge in photo-speed and processing times, but positive photoresist is better for stripping and perhaps for process integration.

The sensitivity of an image position grid to local heat applied to the TIR surface of a prism in a Wynne Dyson design was examined. The mechanisms resulting in the shape and magnitude of the shape change are explored qualitatively and with FEA analysis. The FEA package was used to generate temperature and deformation results, which were then ported to a ray trace program and the impact on the image grid was determined. A device was made to check the validity of the models. The heat load was applied by inserting thermocouple instrumented Watlow cartridge heaters in an aluminum block, placing the aluminum block 127 microns from the TIR surface, and heating an 8mm square surface area of the block to 1.3C, 2.5C and 5.0C. These heat loads resulted in peak image errors of 68nm/C with vector error maps in an annular pattern roughly centered on the geometrical center of the heat load. The models developed were useful for qualitative predictions of the performance, but need tuning to increase the accuracy. Although linear superposition cannot be used to calculate the effect of combined local heat sources, due to the non-linear behavior of the thermal effects, the results can be used to estimate the maximum stray local heat that can be tolerated on this optical surface.