Advanced method for 3D FIB tomography AI supported on SRAM I/O circuit block with missing data bits problems

Tuesday, November 18, 2025: 1:30 PM
1 (Pasadena Convention Center)
Dr. Domenico Mello , EM Microelectronic, a Company of the Swatch Group, Marin-Epagnier, Switzerland
Dr. Guillaume Fiannaca , EM Microelectronic, a Company of the Swatch Group, Marin-Epagnier, Switzerland
Dr. Filippo Marinelli , EM Microelectronic, a Company of the Swatch Group, Marin-Epagnier, Switzerland
Dr. Heiko Stegmann , Carl Zeiss Microscopy GmbH, Oberkochen, Germany
Dr. Flavio Cognigni , Carl Zeiss SpA, Milan, ., Italy
Dr. Giulio Lamedica , Carl Zeiss SpA, Milan, Italy
Ms. Angelica Accorinti , University of Rome "la Sapienza", Rome, Italy
Prof. Marco Rossi , Sapienza University of Rome, Rome, Italy

Summary:

In this paper we show how the -tomography FIB, thanks to the support of the new algorithms based on the use of Artificial Intelligence (AI), is changing the way of dealing with Physical Failure Analysis (PFA). In particular, we show how convenient it is to use these methodologies even in problems that happens during device development. Here we address the reconstruction of a I/O circuit block of an SRAM. Tomography unequivocally demonstrated that the problem was not attributable to process or line defects: This results stimulate the designers to revise the estimation of parameter tuning that fit the silicon measurements in terms of minimum size matching. It was only through the study of the 3D model that the solution to this difficult case study could be reached.