Failure physics of integrated circuits — A review

An approach to failure analysis of microelectronic devices is presented. This approach emphasizes the role of the curve tracer as an important tool used to determine what route an analysis should follow. The curve tracer provides information that indicates the nature of the failure mechanism and not just the failure mode, as does functional testing. Therefore, the curve tracer can aid in the decision making process in selecting the proper steps that should be followed to successfully capture the failure mechanism. A general format for failure analysis is outlined in which optional routes exist at decision points, and examples are given as to how the curve tracer can be used to aid in the selection of the appropriate route. It is also shown how several groups of analysis steps can be used to complement each other and thus provide more information about the cause of failure.

Fan-Out (FO) technologies are used for consumer electronic products generally due to the small and very thin features. Some risks would be found when the design transfer from of 2.5D Integrated Circuit (2.5D IC) or system in packages (SiPs) to the FO packages. This paper will share a case study of High Density Fan-Out (HDFO) with thermal cycle test (TCT) and the test vehicle is open/short fail. The test vehicle is examined by curve tracer (I-V), 3D X-ray, Focus Ion Beam (FIB) and Scanning Electron Microscope (SEM) to find out the failure mode. The failure mode is delamination between the u-pad layer and the Redistribution Layer (RDL) interface. The Cu oxide layer is found on the RDL sidewall. The chemical residue leads to the oxidation of Cu on the RDL surface and the poor interface would be delamination gradually during TCT. Root cause of the chemical residue was investigated in this study. The delamination issue was resolved by adding plasma clean.

This paper examines one of the common modes of structural failure in multichip BGAs, determines its locations within the package structure, relates it to the stresses generated in the reliability tests under which it occurs, and by Finite Element simulations, determines an explanation for the failure, and finally proposes a method to avoid this failure mechanism. Several designs of multichip BGA substrates were manufactured and production silicon assembled into them. These were all 14/spl times/22 mm 119 ball PBGA. These were subjected to a set of package reliability tests, until some units failed electrical test. The failed units were analysed and the physical location and shape of the failure was determined in many cases. From this information, the mechanical mode of failure for each unit was determined. In addition there was sufficient information in some of the analyses to provide definite suggestions as to the mechanism of failure. Meanwhile, Finite Element Analysis was performed using simplified representations of the multichip BGAs, in order to find the locations of highest stress, and the expected modes of failure. This data was matched to the failure modes found in the physical analysis. Some novel failure analysis techniques were used to expose the damage in the failed units. A particular failure mode occurred frequently in temperature cycle, and the sites of failure were located by failure analysis. The failure was due to open circuit in the copper tracks in the top layer of the substrate caused by cracking in the solder resist directly underneath the edge of the die attach fillet. Finite element analysis was carried out and the location of the actual failures was found to be a local zone of high tensile stress in the solder resist.

Identifying the failure modes of cloud computing systems is a difficult and time-consuming task, due to the growing complexity of such systems, and the large volume and noisiness of failure data. This paper presents a novel approach for analyzing failure data from cloud systems, in order to relieve human analysts from manually fine-tuning the data for feature engineering. The approach leverages Deep Embedded Clustering (DEC), a family of unsupervised clustering algorithms based on deep learning, which uses an autoencoder to optimize data dimensionality and inter-cluster variance. We applied the approach in the context of the OpenStack cloud computing platform, both on the raw failure data and in combination with an anomaly detection pre-processing algorithm. The results show that the performance of the proposed approach, in terms of purity of clusters, is comparable to, or in some cases even better than manually fine-tuned clustering, thus avoiding the need for deep domain knowledge and reducing the effort to perform the analysis. In all cases, the proposed approach provides better performance than unsupervised clustering when no feature engineering is applied to the data. Moreover, the distribution of failure modes from the proposed approach is closer to the actual frequency of the failure modes.

The aim of this study was to evaluate the effects of a smear layer generated by a high-speed diamond or carbide bur on the durability of microtensile bond strength (μTBS) of a self-etching adhesive to primary dentin. Flat occlusal dentin surfaces of 105 human primary molars were exposed using 600 grit silicon carbide paper before being divided into 2 groups for further grinding with either a highspeed diamond or carbide bur. Ten prepared dentin surfaces treated by each bur were evaluated for the characteristics of the smear layer using a scanning electron microscope (SEM). Seventy-five specimens from each bur-prepared group were applied with a 2-step self-etching adhesive (Clearfil SE Bond(®)) then built up with a resin composite. Each bonded specimen was sectioned into a 1-mm thick slab and trimmed to a dumbbell shape with a cross-sectional area of approximately 1 mm(2). All slabs were divided into 3 groups (n=25) according to 3 storage times of 24 hrs, 3 months, and 6 months, in distilled water at 37°C. After storage, the μTBS was determined using a universal testing machine. All fracture specimens were prepared for observation of failure modes. Ten bonded specimens of each bur group were prepared for observation of the resin-dentin interface using an SEM. Smear-layer thickness, μTBS, and failure mode distributions were statistically analyzed. The high speed carbide bur created a significantly thinner smear layer than the diamond bur (p < 0.05). Dentin surfaces treated with a high-speed carbide bur generally obtained significantly higher μTBS than the diamond bur group (p < 0.05). The μTBS gradually decreased over time such that specimens stored for 6 months had significantly lower bond strength than those stored for 24 hrs (p < 0.05). Self-etching adhesive created a hybrid layer of the same thickness when prepared with either a carbide bur or diamond bur, but the carbide bur group had longer and more resin tags. Highspeed carbide bur groups had a higher μTBS than diamond bur groups for all storage times, and bond strengths decreased over time in both substrate groups. The use of a carbide bur produced a thinner smear layer and therefore is recommended when using this 2-step self-etching adhesive to bond the resin composite to primary dentin.

To evaluate the effects of laser etching on the shear bond strengths (SBS) and failure modes of brackets bonded to fluorosed enamel. This in vitro study included 34 fluorosed and 34 nonfluorosed teeth. Teeth were divided into four subgroups according to the etching procedure: Group (A) normal enamel etched with 37% phosphoric acid for 15 s, Group (B) nonfluorosed enamel etched with erbium:yttrium aluminum garnet (Er: YAG) laser for 15 s, Group (C) fluorosed enamel etched with 37% phosphoric acid for 30 s, Group (D) fluorosed enamel etched with Er: YAG laser for 30 s. After bonding of the premolar metal brackets, specimens were subjected to the thermal cycles. After SBS test, modified adhesive remnant index (ARI) by using stereomicroscope and failure modes with a scanning electron microscope (SEM) was determined. After bonding, one specimen from each group was examined under SEM to identify enamel-resin interfaces. ANOVA and Tukey post-hoc tests were used to compare the SBS values. The Kruskal-Wallis and Chi-square tests were used to analyze the distribution of ARI scores and failure modes of groups. The highest mean SBS value (17.56 ± 1.05 MPa) was found in Group A, while the lowest (12.25 ± 0.96 MPa) in Group D. Significant differences were found in the SBS test and failure modes among all groups. The differences between ARI scores of the groups were not significant. According to our findings, laser etching reduced the SBS of brackets bonded to fluorosed teeth, but provided clinically acceptable SBS values.

Failure analysis and countermeasures of a tunnel constructed in loose granular stratum by shallow tunnelling method

This paper reports the developments of evaluation methods on epitaxially grown superlattices by means of sputterassisted Auger electron spectroscopy (AES), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). AlGaAs/GaAs semiconductor superlattics were grown epitaxially by metal–organic chemical vapour deposition (MOCVD). The layer thickness of the superlattices ranged from a few to ten nanometers. Firstly, developments of Auger depth profiling were tried by using: (1) a differential pumping‐type ion gun instead of a static pressure type to reduce the oxygen adsorbates on the AlGaAs layer; (2) low‐energy Auger signals instead of high‐energy ones to shorten the escape depth; and (3) the lowest ion etching energy of 0.2 keV instead of 1 keV to reduce the surface roughening effects. It is shown that the depth resolution of sputter‐assisted AES is attainable to 1.5 nm. Secondly, high‐resolution SEM can be used as an easy evaluation method by observing the cleaved surface of superlattices, since the layers can be distinguished by signal contrast. Also, TEM can be used as an evaluation method by observing the (110) cross‐section thinned sample. The dark field image has a high contrast between AlGaAs and GaAs using the (002) diffraction. It is confirmed from these AES, SEM and TEM evaluations that the hetero‐interface abruptness of AlGaAs/GaAs superlattices grown by MOCVD is of the order of one monoatomic layer.

ObjectiveThis study evaluated enamel demineralization, shear bond strength, and failure modes associated with orthodontic adhesives containing silver nanoparticles (AgNPs).MethodsForty-eight extracted human premolar teeth were selected and evenly divided into two groups based on the intended assessments, with 28 specimens in each group: Part I focused on enamel demineralization depth, and Part II addressed shear bond strength and failure modes. Specimens were prepared, mounted, and bonded using two adhesive systems: Group A used a conventional adhesive system without AgNPs, while Group B employed a conventional adhesive system mixed with AgNPs. In Part I, specimens underwent artificial demineralization at pH 4.5°C and 37°C for 7 days, followed by buccolingual sectioning. The depth of demineralization was measured using a scanning electron microscope and analyzed with ImageJ software. In Part II, shear bond strength was assessed using a universal testing machine, followed by failure mode evaluation using a stereomicroscope.ResultsResults showed that Group A exhibited a significantly greater demineralization depth compared to Group B (p = 0.000). However, no significant difference was observed between overall location levels (p = 0.093). Additionally, Group B demonstrated significantly higher shear bond strength (p = 0.000) and a more uniform distribution of failure modes compared to Group A.ConclusionsThese findings suggest that incorporating silver nanoparticles into orthodontic adhesives reduces demineralization depth and enhances shear bond strength compared to conventional adhesive systems. In contrast, conventional adhesives were associated with a higher incidence of adhesive mode failure.

Effects of elevated water temperatures on interfacial delaminations, failure modes and shear strength in externally-bonded CFRP-concrete beams using infrared thermography, gray-scale images and direct shear test

Based on the progressive failure analysis method, taking account of 7 kinds of failure modes: stitch thread failure mode, resin extend (or compress) failure mode, fibre extend (or compress) failure mode, extend (or compress) delamination failure mode, a new failure analysis FE model was proposed for predicting the tensile strength of stitch-reinforced composite T-stiffened structures. The 3D solid element and truss element were used to simulate laminate and stitch thread individually in the FE model. The Hashin failure criterion and stiffness degradation technology are used to analyze damage failure of the stitch T-stiffened structure. Comparing the initial and final damage failure strength with those of experiments on stitch composite T-stiffened structures validated the Python secondary development program for the FE model presented in the paper. There was good agreement between the failure strength from the damage failure analysis and that of the experiment. This indicated that the model was correct, and that it could be a reference for stitch-reinforced composite design and analysis.

The community's conventional salt production business is one of the economic drivers of the community. In addition to being able to provide a lot of jobs for people in Indonesia's coastal areas, the people's salt production business is also seen as a way to reduce poverty. Pidie Regency, which is included in the Province of Aceh, is one area that has a wide coastline and can be utilized as a salt farmer business. Pidie is an area that has a salt production business with a total production of 449.05 tons/month. The problem faced in the traditional salt processing business in Pidie District is the unknown level of risk that exists in each salt production process. The purpose of this study is to determine the level of risk in the traditional salt production process using failure mode and effect analysis (FMEA) and probability impact matrix (PIM) in the Saho Adab people's salt business group in Keude Ie Leubeu Village, Kembang Tanjong District, Pidie District. FMEA is a form of qualitative analysis that assesses the severity , occurrence and detection on each activity and then given a value on a predetermined scale. Whereas risk assessment using PIM only assesses the severity and occurrence. Based on the analysis of failure and effect modes, the salt production process has a high level of risk, namely the packaging process with an RPN value (risk priority number) 75, the salt water storage process with an RPN value of 48, the drying process with an RPN value of 45 and the level of the soil process with an RPN 45. While the results of the matrix impact probability obtained by the salt production process that has a high level of risk, namely the process of storing salt with a value of 5, the packaging process with a value of 4, the process of leveling sand with a value of 3,5 and the process of salt water filtration with sand filtration media containing NaCl with a value of 3,5.

Purpose- Quality activities should include all product life processes, starting from the determination of customers' demands and prejudices to the service activities provided to customers. By integrating quality-oriented practices, conducting FMEA, and employing Pareto analysis, the research aims to enhance the system's reliability, efficiency, and overall performance. This, in turn, can lead to improved customer satisfaction, reduced costs, increased productivity, and a more sustainable and successful operation in the packaging department. Methodology- In order to keep up with the rapidly developing technology, to be open to innovations with developing technologies, to be in an advantageous position against competitors and ultimately to satisfy customers and fulfil their satisfaction; it is necessary to meet the demands and needs of customers without error and at low cost. Failure Modes and Effect Analysis is applied to provide the sustainability of the production system without the failures. This study focuses on failure analysis, defining failure modes, and conducting error prevention studies within the context of businesses. The aim is to achieve zero failure and zero error by implementing effective prevention techniques. By identifying the root causes of failures and errors, businesses can develop strategies and procedures to prevent them from occurring in the first place. This proactive approach helps ensure a higher level of quality, reliability, and customer satisfaction. By emphasizing the importance of error and failure prevention, businesses can strive for continuous improvement, minimize disruptions in the production system, reduce costs associated with rework or product recalls, and maintain a strong competitive position. Implementing zero failure-error and zero error-failure prevention techniques is crucial for sustaining a reliable and efficient production system. Findings- This research shows that quality is a science that people meet in reality, not simply a theoretical concept. All organizations in the universe should take their quality management systems one step further with their approach to quality processes. In this context, the concept of zero failure-error in production or service lines has become important. By applying FMEA and considering sustainability aspects, the research aims to enhance the manufacturing process, minimize risks and failures, and contribute to sustainable practices in the food industry. Conclusion- The mentioned sample study on Failure Modes and Effects Analysis (FMEA) in food packaging within the production department of a food sector enterprise in the Marmara region highlights the importance of preventing possible failure modes and their impact on the sustainability of production. One key aspect emphasized in the study is the potential reduction of food and packaging waste by minimizing packaging or substituting oil-based materials with renewable resources. By utilizing environmentally friendly food packaging, various advantages can be achieved, including a decrease in food loss and waste, as well as a reduction in the negative environmental impact associated with managing packaging materials and waste. To accomplish this, it is crucial to address food demands throughout the entire food supply chain, with particular attention to transportation and consumption stages. In order to optimize the sustainability of food and packaging systems as a whole, it is necessary to consider the product and its packaging as an integrated system. By applying FMEA and considering the potential failure modes in food packaging, the study aims to identify areas for improvement and implement measures that enhance sustainability throughout the production process. Keywords: Failure prevention, failure mode and effect analysis, quality, sustainable production. JEL Codes: M10, L15, O14

KOH concentration effect on the cycle life of nickel-hydrogen cells IV. Results of failure analyses

Failure cases play an important role in failure analysis. This paper will focus on how to determine the failure mode of the new failure cases from some existed failure cases. First, the fuzzy relationship matrix of each failure mode and unexpected appearance information can be obtained from the large amount of failure cases have been collected. Secondly, the comprehensive evaluation for the new failure cases can be basically evaluated, the evaluation results corresponding to various failure modes can be calculated. Thirdly, the most probable failure modes will be determined through comparison with the threshold. But if more than one probably failure modes existed, Euclidean algorithm will be used to calculate the relative Euclidean distances of the new failure case to each probably failure modes obtained from above. The failure mode to which the new case belongs can be determined through above several steps.