Abstract

As semiconductor technology keeps scaling down, failure analysis and device characterization become more and more challenging. For foundry FA, transistor device performance is very important with regards to process monitoring and failure analysis. A nanoprobing methodology is widely applied in advanced failure analysis, especially during device level electrical characterization work. There are two popular nanoprobing systems in our lab, one is SEM based and the other is AFM based. Both systems have their advantages in the electrical characterization and fault isolation field. A conventional analysis was performed on a die that failed during functional testing. The die was pulled from a zero yielding wafer where all failing die manifested a VADC failure mode. Preliminary results from the conventional electrical analysis did not reveal any TIVA spots. A circuit analysis done on the failure mode strongly suggested that a transistor defect at a specific location was the most likely cause of failure. However, physical analysis of the suspect transistor did not show any differences at that device between good and failing die. Since the issue caused 100% yield loss on the wafers of interest and there was no TIVA signature from those parts, further analysis of the suspect transistor device was warranted. The nature of the failure suggested that there was a process problem and that the defect was uniform or implant related. In our circuit of interest the multifingered transistor under investigation was a native transistor. Device level characterization of the suspect transistor was required. Since nanoprobing is a direct method used to access the suspected location at the device or contact level, it can be used to measure the actual performance of any suspect transistor. Because an implantation related issued was highly suspected, nanoprobe analysis was the next logical step. Due to the fact that implantation issues are normally invisible and can’t usually be photographed during physical analysis, electrical verification was necessary to prove the bad transistor hypothesis. Additional nanoprobe experiments were designed to check the transistor body effects. The results showed a different body effect between good and bad units which electrically indicated that the implantation dosage was different between passing and failing parts. In order to check each finger of the multifingered transistor, the sample was delayered to contact and nanoprobing was employed again. Surprisingly, different electrical performance was observed between transistor edge fingers and those positioned at the center. Based on this result, a Monte-Carlo computer simulation was studied on ‘Well Ion Implant Trajectories’. It showed significant ion beam scattering by photo resist at the transistor edge which strongly correlated with our nanoprobing analysis results [1] [2].

This content is only available as a PDF.
You do not currently have access to this content.