<p>Non-destructive failure analysis (FA) of microelectronic devices is more challenging today due to shrinking technology nodes according to Moore’s law (Moore in Electronics 38:114–117, 1965) and increased prevalence of stacked devices composed of assemblies of multiple die as described in “More than Moore” (Lee et al. in IEEE International Electron Devices Meeting (IEDM), 2002). 2.5D and 3D heterogeneous integrated (HI) devices are inherently difficult to perform root cause FA, largely because techniques used for 2D root cause analysis are often not applicable to 2.5D/3D assemblies or require significant sample preparation. The challenges are due to physical and electrical features, such as chip stacking obscuring underlying die or inability to access electrical test points. HI failure analysis (HIFA) roadmaps have identified technology gaps in the FA process flow for 2.5D/3D devices, specifically calling for the development of non-destructive FA techniques (Walraven et al. Heterogeneous Integration Failure Analysis (HIFA) Technology Roadmap for 2.5D/3D Devices SAND2023-11278, 2023). Different root cause FA capabilities such as 3D X-ray tomography/imaging (Microelectronics Failure Analysis Desk Reference, 2019), lock-in thermography (LiT) (Brand and Altmann, 3D hot-spot localization by lock-in thermography, in microelectronics failure analysis: desk reference, 2019), magnetic field imaging (MFI) (Microelectronics Failure Analysis Desk Reference, 2019; Kehayias et al. Proc of the 48th Int. Symp. For Testing and Failure Analysis, 2022), and optical beam-based techniques (Cole et al. Reliability Physics Symposium Proceedings, 1998) leverage physics of signals emitted from and/or transmitted through the device, each technique with its own strengths and challenges. We present a pair of novel techniques for defect localization using radio frequency (RF) electromagnetic (EM) emanations. Either these emanations are emitted by the device under test (DUT) due to an oscillating electrical input (square wave), or they can be directly injected into the DUT via a scanning antenna. In the first case, the DUT emits an electromagnetic spectrum collected by a scanning antenna and spectrum analyzer. In the second case, the DUT acts as the detector, with the input to the spectrum analyzer connected directly to the DUT and the frequency connected to the scanning antenna, injecting the signal into the DUT. These techniques can use either magnetic (H) or electric (E) field antennas to detect or inject EM signals from an HI device and build a 2D or 3D map of the EM response. This EM map corresponds to the current within the device, allowing the user to determine additional current paths (shorts), closed current paths (opens) and activity density. Power spectrum analysis (PSA) (Tangyunyong et al. SAND2016-12238C, Various Applications of Power Spectrum Analysis (PSA), 2016), a technique used to identify electrical differences between groups of parts, is used to compare these spectra at each scan point. The techniques, referred to as EM antenna PSA (EMAPSA) or EM injection PSA (EMIPSA), are described in detail. Results from these techniques are compared to other root cause FA capabilities.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Board-level and Packaged Device Failure Analysis Using Electromagnetic Antenna/Injection Power Spectrum Analysis (EMAPSA/EMIPSA)

  • J. T. Christiansen,
  • J. A. Walraven

摘要

Non-destructive failure analysis (FA) of microelectronic devices is more challenging today due to shrinking technology nodes according to Moore’s law (Moore in Electronics 38:114–117, 1965) and increased prevalence of stacked devices composed of assemblies of multiple die as described in “More than Moore” (Lee et al. in IEEE International Electron Devices Meeting (IEDM), 2002). 2.5D and 3D heterogeneous integrated (HI) devices are inherently difficult to perform root cause FA, largely because techniques used for 2D root cause analysis are often not applicable to 2.5D/3D assemblies or require significant sample preparation. The challenges are due to physical and electrical features, such as chip stacking obscuring underlying die or inability to access electrical test points. HI failure analysis (HIFA) roadmaps have identified technology gaps in the FA process flow for 2.5D/3D devices, specifically calling for the development of non-destructive FA techniques (Walraven et al. Heterogeneous Integration Failure Analysis (HIFA) Technology Roadmap for 2.5D/3D Devices SAND2023-11278, 2023). Different root cause FA capabilities such as 3D X-ray tomography/imaging (Microelectronics Failure Analysis Desk Reference, 2019), lock-in thermography (LiT) (Brand and Altmann, 3D hot-spot localization by lock-in thermography, in microelectronics failure analysis: desk reference, 2019), magnetic field imaging (MFI) (Microelectronics Failure Analysis Desk Reference, 2019; Kehayias et al. Proc of the 48th Int. Symp. For Testing and Failure Analysis, 2022), and optical beam-based techniques (Cole et al. Reliability Physics Symposium Proceedings, 1998) leverage physics of signals emitted from and/or transmitted through the device, each technique with its own strengths and challenges. We present a pair of novel techniques for defect localization using radio frequency (RF) electromagnetic (EM) emanations. Either these emanations are emitted by the device under test (DUT) due to an oscillating electrical input (square wave), or they can be directly injected into the DUT via a scanning antenna. In the first case, the DUT emits an electromagnetic spectrum collected by a scanning antenna and spectrum analyzer. In the second case, the DUT acts as the detector, with the input to the spectrum analyzer connected directly to the DUT and the frequency connected to the scanning antenna, injecting the signal into the DUT. These techniques can use either magnetic (H) or electric (E) field antennas to detect or inject EM signals from an HI device and build a 2D or 3D map of the EM response. This EM map corresponds to the current within the device, allowing the user to determine additional current paths (shorts), closed current paths (opens) and activity density. Power spectrum analysis (PSA) (Tangyunyong et al. SAND2016-12238C, Various Applications of Power Spectrum Analysis (PSA), 2016), a technique used to identify electrical differences between groups of parts, is used to compare these spectra at each scan point. The techniques, referred to as EM antenna PSA (EMAPSA) or EM injection PSA (EMIPSA), are described in detail. Results from these techniques are compared to other root cause FA capabilities.