edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 27 NO. 1 36 resources than training a model from scratch and can be done by FA labs internally on confidential data. However, the deployment of these models on the FA lab infrastructure and their incorporation into the workflows of FA engineers relies on the availability of training data. Therefore, the third part of the meeting continued the trend set in the first UG meeting, where the participants concluded that the establishment of general data exchange standards would ensure FAIR data principles on FA-lab infrastructures. Properly organized business processes equipped with corresponding tools based on common standards can establish the continuous creation of data ready for training AI models, advanced analytics applications, and integration with software systems of other departments, like manufacturing, design, and quality management. ISTFA 2024 FOCUSED ION BEAM (FIB) USER GROUP Chair/Co-Chairs: Valerie Brogden, Steve Herschbein, Michael Wong, and Edward Principe vbrogden@uoregon.edu, steven.herschbein@gmail.com, mike.wong@thermofisher.com, eprincipe@synchres.com FIB User Group presenters and co-chairs. The FIB User Group meeting aimed to facilitate knowledge exchange on emerging technologies and techniques in FIB applications. The agenda included a series of brief presentations by industry experts, followed by a group discussion on the potential for AI in FIB circuit edits. Pawel Nowakowski from Fischione Instruments discussed the applications of scanning transmission electron microscopy (STEM) within SEM/FIB systems, highlighting its benefits for crystallographic orientation mapping via transmission Kikuchi diffraction (TKD). He emphasized the advantages of using STEM in SEM, including cost savings and low voltage analysis, although resolution limitations persist compared to TEM. Pawel proposed using 500 eV argon milling to mitigate gallium-induced damage, demon- strating clearer EBSD maps and enhanced grain visibility. Peter Gnauck, Raith, provided an overview of liquid metal alloy ion sources (LMAIS), emphasizing its capacity for various ions beyond gallium, including bismuth and lithium. He discussed how a Wien filter facilitates selective ion targeting for applications ranging from imaging to milling. Gnauck highlighted the ability of lighter ions, like lithium, to enhance imaging resolution and the potential of reactive ions like lithium for high-resolution SIMS. Lukas Hladik, Tescan, introduced Tescan’s Mistral Plasma FIB Column, which enhances beam current with improved spot profiles for TEM prep. He demonstrated how the Mistral provides finer cuts with reduced amorphous damage, which may eliminate the need for protective deposition. Additionally, he presented Tescan’s integration of AI for automated TEM prep. Steve Herschbein led a discussion on leveraging AI in FIB circuit edits, particularly in planning precise cuts that minimize impact on circuit performance. He proposed an AI-driven tool to simulate and optimize edits, reducing time spent on manual planning. This concept aligns with the “digital twin” approach in AI, where multiple virtual scenarios can be tested before physical alterations. Challenges identified in this discussion suggest that integrating AI with existing customer circuit parameters would require access to sensitive data, which may only be feasible within design houses. Rick Livengood noted the importance of communicating changes in electrical parameters due to FIB edits back to end users. This meeting underscored the value of interdisciplinary discussions and the potential of AI to transform traditional FIB processes, promising both immediate and long-term benefits for the FIB community.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==