edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 1 24 PROCESSES FOR THINNING AND POLISHING HIGHLY WARPED DIE TO A NEARLY CONSISTENT THICKNESS: PART III Kirk A. Martin RKD Systems, Aptos, California kirk@rkdsystems.org EDFAAO (2024) 1:24-31 1537-0755/$19.00 ©ASM International® INTRODUCTION In Part I of this article series,[1] it was suggested that instead of attempting to thin an entire die to a thickness of less than 10 microns, the entire die be thinned to 50 microns for initial evaluation. The initial evaluation would result in defining specific, small areas of interest. The area of interest would be thinned to the desired, less than 10 microns, remaining silicon thickness (RST). Part I deals with the problems and considerations for thinning a large, warped, die to a uniform thickness of 50 ± 2 microns and suggests processes and procedures for obtaining this result. In Part II,[2] the process and problems of area of interest thinning was discussed. The reasons for thinning only areas of interest were given and basic guidelines and processing limitations described. In this article, the processes and considerations for both global and area of interest are discussed and reference process recipes are given. THE SAMPLE Every failure analysis sample is of extreme value as it is usually one-of-a-kind and the customer who returned it wants to know what happened. No one wants to tell a customer that a returned field failure was lost in processing. Even if it was not a field return, but instead a device that failed the final test, everyone wants to know why, and process loss is not an option. When “failure is not an option,” how do you prevent it? The basic rules for this are very simple. 1. Do the same thing to every sample in exactly the same way. 2. Record what was done in all of the processes and how it was done. 3. Process practice samples first to develop and define the process. 4. When a process is developed on the practice samples, do exactly the same thing for the live sample. This is only simple if the equipment and the processes used are consistent from sample to sample, and the process parameters themselves are not vague variables. Tool rotational speeds and linear travel rates must be defined and optimized. Cutting tool bits need to be sharp and replaced often. Lapping and polishing tools need to be checked for damage and reconditioned, or dressed, or replaced. The abrasives need to be very reproducible in use with the tools. A process step that requires the use of a lubricant must have the lubricant dispensed in the same way on every sample as the lubricant cleanliness directly affects material removal rates and uniformity. During the process, the slurry needs to be refreshed in the same way and time in the process. Additionally, environmental parameters need to be controlled; not necessarily controlled for absolute values, but for changes. All labs have HVAC vents. When the heater is running, the vents will blow warm air or cold air when the air conditioner is active. If the processing is done in proximity to a vent, there will be significant thermal changes and the creation of thermal gradients. This will affect the processing results. An example to demonstrate this effect is the change in length of the tools used in the process. Tools typically have shafts made of stainless steel having a Tc of 17 ppm/K. with 20 mm of shaft exposed to a 5-degree wind from a heater vent, the tool will grow by
RkJQdWJsaXNoZXIy MTYyMzk3NQ==