This article presents a case study involving flash memory bit failures characterized by threshold voltage changes due to positive gate disturb stress. An inconsistency in failing bit behavior, which was found to be dependent on the test mode, was explored to provide an electrical explanation for the failure. The underlying defect was isolated and subsequently identified by physical analysis.

This article describes how physical attacks can be launched on different types of nonvolatile memory (NVM) cells using failure analysis tools. It explains how the bit information stored inside these devices is susceptible to read-out and fault injection attacks and defines vulnerability parameters to help quantify risks associated with different modalities of attack. It also presents an in-depth security analysis of emerging NVM technologies and discusses potential countermeasures.

Analog computing is an important step in taking neural net processing to the next level. However, as this column explains, reliability is intimately linked to performance and efficiency in analog systems, more so than in any modern digital system, and further work is required to understand the relationship.

This column reflects on the effect mobile phones have had on process and packaging technology and failure analysis.

Chip-scale packages (CSPs) make efficient use of space on PCBs, but their small size, multilevel stacking arrangements, and complex interconnects present serious challenges when it comes to testing and failure analysis. This article describes some of the problems encountered when dealing with various types of CSPs and provides practical solutions based on the tools and techniques available in most FA labs. It discusses the causes and effects of package and die related failures and walks readers through the steps involved in decapsulating plastic FBGA packages using conventional etching, polishing, and milling techniques.

Scanning nonlinear dielectric microscopy (SNDM) is a scanning probe technique that measures changes in oscillation frequency between the probe tip and a voltage-biased sample. As the probe moves across the surface of a semiconductor device, the oscillation frequency changes in response to variations in dielectric properties, charge and carrier density, dopant concentration, interface states, or any number of other variables that affect local capacitance. Over the past few years, researchers at Tohoku University have made several improvements in dielectric microscopy, the latest of which is a digital version called time-resolved SNDM (tr-SNDM). Here they describe their new technique and present an application in which it is used to acquire CV, d C /d V-V , and DLTS data from SiO 2 /SiC interface samples.

Chip-scale packages (CSPs) make efficient use of space on PCBs, but their small size, multilevel stacking arrangements, and complex interconnects present serious challenges when it comes to testing and failure analysis. This article describes some of the problems encountered when dealing with various types of CSPs and provides practical solutions based on the tools and techniques available in most FA labs. It discusses the causes and effects of package and die related failures and walks readers through the steps involved in decapsulating plastic FBGA packages using conventional etching, polishing, and milling techniques. It also includes a case study involving a failure caused by improper laser marking.

The second phase of the IARPA Circuit Analysis Tools (CAT) program, which ended in June 2015, focused on the development of prototype tools to demonstrate scalability to the 10 nm node. Our guest columnist, IARPA Program Manager, Carl E. McCants, provides a summary of what the participating teams accomplished.

Various NVM technologies are being explored for neuromorphic system realization, including resistive RAM, ferroelectric RAM, phase change RAM, spin transfer torque RAM, and NAND flash. This article discusses the potential of RRAM for such applications and evaluates key performance and reliability metrics in the context of neural network image classification. The authors conclude that the accuracy-power tradeoff may be further improved using alternative material stacks and multi-layer dielectrics so as to achieve better control of the oxygen vacancy or metallic filamentation process that governs RRAM switching characteristics.

The inverted orientation of a flip-chip packaged die makes it vulnerable to optical attacks from the backside. This article discusses the nature of that vulnerability, assesses the threats posed by optical inspection tools and techniques, and provides insights on effective countermeasures.

A new and improved sample preparation technique was developed by Wang. This technique uses an FIB instrument for the 90° rotation of a small portion of the specimen on the original grid by taking advantage of static force. All sample preparation steps, including thin-section creation and sample tilting, can be accomplished in a single process. The procedure is monitored in a high-resolution FIB instrument to assure a 100% success rate. Figure 1 shows a scanning electron microscope image of a 3D TEM sample with two rotated sections. The original TEM sample is a lift-out sample laid on carbon film.

Several failure analysis case studies have been conducted over the past few years, illustrating the importance of preserving root-cause evidence by means of artifact-free decapsulation. The findings from three of those studies are presented in this article. In one case, the root cause of failure is chlorine contamination. In another, it is a combination of corrosion and metal migration. The third case involves an EOS failure, the evidence of which was hidden under a layer of carbonized mold compound. In addition to case studies, the article also includes images that compare the results of different decapsulation methods.

Circuit boards are vulnerable to a wide range of ill-intentioned modifications done to gain access to information or malevolently influence control. This article describes the various ways attacks on circuit board can occur and presents examples showing how such attacks might look. It also provides general guidelines for protecting circuit-board design integrity.

This case study compares the FA success rate and turn-around time of traditional logic-only and true layout-aware scan diagnosis. It discusses the basic process flow, identifies key success factors, and evaluates physical FA and diagnostic test results obtained from six dies randomly selected from a 9.8 M-gate, seven-metal-layer ASIC manufactured in 90 nm technology. As shown, layout-aware diagnosis reduces the defect search area on the die, in some cases, by an order of magnitude, providing the means to diagnosis-driven yield improvements.

This article discusses the basic principles of dark current spectroscopy (DCS), a measurement technique that can detect and identify low levels of metal contaminants in CMOS image sensors. An example is given in which DCS is used to determine the concentration of tungsten and gold contaminants in an image sensor and estimate the dark current generated by a single atom of each metal.

This article provides a brief introduction to through-silicon via technology, a system-level architecture in which multiple layers of planar devices are stacked with interconnects running in the vertical as well as lateral direction. Some of the different fabrication processes in use are discussed along with related challenges.

Parametric analysis of SRAM cells is widely used to locate faults and analyze device failures, but the pico-probing and bit-line multiplexing required for data acquisition is becoming increasingly difficult. This article explains how the addition of an on-die low-yield analysis circuit eliminates the problem. The simplicity of the measurement circuit and the potential to use a known library of curves, makes low-yield analysis one of the most versatile DFT techniques for cache fault isolation.

The best spatial resolution that can be achieved with far-field optical fault localization techniques is around 20 times larger than the critical defect size at the 45 nm technology node. There is also a limit on the laser power that can be safely used on 45 nm devices, which further compromises fault localization precision. In this article, the authors explain how they overcome these limitations using pulsed laser-induced imaging techniques and a refractive solid immersion lens. Two case studies show how the combination of pulsed-laser scanning optical microscopy and a solid immersion lens improves localization precision and detection sensitivity.

Several technology-focused User's Groups met at ISTFA 2014 to discuss current issues and advances in their areas of interest. This article summarizes key discussion points from the Contactless Fault Isolation User's Group, the Nanoprobing User's Group, the Sample Prep/3-D Package User's Group, and the FIB User's Group.

Faster time-to-production of a new product is the common goal of design houses and foundries. This article demonstrates how foundries can contribute through post silicon validation, which allows design houses to focus on more complicated issues.