Gettering property of Cu and Ni was investigated after MCP thinned process wafer. The test samples were prepared to various dopant concentration types of CZ or Epitaxial wafer, and various processed thinning conditions corresponding with a total remaining thickness of 100 µm, 50 µm, and 30 µm by backside grinding. It was found that the 30 µm thinned wafer maintains its Cu gettering ability for heavily boron doped samples but Ni contamination decrease the gettering ability for all samples.

A novel concept of MCP substrates that shows the enhanced internal gettering ability even in thinned state was proposed. To achieve the concept, RTA treatment in high temperature was applied to the CZ-Si substrate with a high initial oxygen concentration. It was confirmed that precipitates were formed as high density even in the near surface shallower than 100 μm, where major portion of internal gettering should take place in the MCP process. Moreover, comparable gettering ability to the EPI wafer was confirmed in our concept after thinning, especially in Ni gettering. In conclusion, it was expected that this concept is one of the possible substitutions of EPI wafers in practical MCP process.

This paper talks about a new class of yield impacting defect types known as polishing induced defects (PID) on polished silicon substrates. These defects were found to cause significant yield drop associated with a pump bias failure in the sub-60nm flash memory process. Traditional incoming wafer quality analysis methods proved inadequate in detecting these defects. Nor had the problem been detected during outgoing wafer quality analysis at the wafer manufacturing plant. At the fab, SEM and AFM were able to verify the defect issue as PID; however, neither of these techniques is production-worthy for routine inspection of incoming wafers. This called for a new non-destructive methodology that not only enables capturing these defects but also provides an ability to reliably distinguish between these device-killing defects and innocuous particles which can be cleanable. This study utilized KLA-Tencor's Surfscan SP2XP unpatterned wafer inspection system, a tool with unique technology to combine multiple illumination angles and multiple detectors with a Rule Based Binning (RBB) algorithm for effective defect detection and classification. Three different rules were developed to improve the classification of PID vs. particles. All three were based on comparing the ratio of scattering signals between detectors or combinations of detectors

not Available.