An Overview on Recent Advances in Statistical Burn-In Modeling for Semiconductor Devices

In semiconductor manufacturing, early life failures have to be screened out before delivery. This is achieved by means of burn-in. With the aim to prove a target reliability level and release burn-in testing of the whole population, a burn-in study is performed, in which a large number of items is investigated for early life failures. However, from a statistical point of view, there is substantial potential for improvement with respect to the modeling of early life failure probabilities by considering further available information in addition to the performed burn-in studies. In this paper, we provide ideas on how advanced statistics can be applied to efficiently reduce the efforts of burn-in studies. These ideas involve scaling the failure probability with respect to the sizes of the different products, as well as taking advantage of synergies between different chip technologies within the estimation of the chips' failure probability level.

Objective To evaluate Class V restorations placed by UK general practitioners comparing those failing or surviving after two years, and to identify factors associated with early failure. Design Prospective longitudinal cohort multi-centre study. Setting UK general dental practices. Materials & methods Ten dentists each placed 100 Class V restorations and recorded selected clinical information at placement and recall visits. Univariate associations were assessed between recorded clinical factors and whether restorations had failed or not at two years. Multi-variable binary logistic regression was also undertaken to identify which combination of factors had a significant effect on the probability of early failure. Results At two years, 156 of 989 restorations had failed (15.8%), with 40 (4%) lost to follow-up. Univariate analysis showed a significant association between restoration failure and increasing patient age, payment method, the treating practitioner, non-carious cavities, cavities involving enamel and dentine, cavity preparation and restoration material. Multi-variable analysis indicated a higher probability of early failure associated with the practitioner, older patients, glass ionomer and flowable composite, bur-preparation and moisture contamination. Conclusions Among these practitioners, both analytic methods identified significant associations between early failure of Class V restorations and the practitioner, cavity preparation method, restoration material and patient's age.

For electromigration in short aluminum interconnects terminated by tungsten vias, the well known “short-line” effect applies. In a similar manner, for longer lines, early failure is determined by a critical value Lcrit for the length of polygranular clusters. Any cluster shorter than Lcrit is “immortal” on the time scale of early failure where the figure of merit is not the standard t50 value (the time to 50% failures), but rather the total probability of early failure, Pcf. Pcf is a complex function of current density, linewidth, line length, and material properties (the median grain size d50 and grain size shape factor σd). It is calculated here using a model based around the theory of runs, which has proved itself to be a useful tool for assessing the probability of extreme events. Our analysis shows that Pcf is strongly dependent on σd, and a change in σd from 0.27 to 0.5 can cause an order of magnitude increase in Pcf under typical test conditions. This has implications for the web-based two-dimensional grain-growth simulator MIT/EmSim, which generates grain patterns with σd=0.27, while typical as-patterned structures are better represented by a σd in the range 0.4 – 0.6. The simulator will consequently overestimate interconnect reliability due to this particular electromigration failure mode.

A layout-level Monte-Carlo simulator capable of predicting the probability of defect-related early failure in the metal-interconnect subsystem of integrated circuits is presented. The interconnect is decomposed into distinct layout-level objects, namely contiguous metal runs and vias, and user-defined cumulative distribution functions are used to determine the probability of failure. The failure distributions for nets, which are sets of interconnected layout-objects, are obtained by combining the distributions of their component objects. System reliability is obtained by sampling from the reliability network comprising all nets. Variance reduction techniques are used wherever possible to improve the statistical confidence in estimates made. A prototype that interfaces directly with commercially used CAD software has been implemented. Results provide a qualitative verification of the methodology and show that failures due to defects are significant during the first few years of system operation.

In such severe core damage accident, as small LOCAs with no ECCS injection or station blackout, in which the primary reactor system remains pressurized during core melt down, certain modes of vessel failure would lead to a high pressure ejection of molten core material. In case of a local failure of the lower head, the molten materials would initially be ejected into the cavity beneath the pressure vessel may subsequently be swept out from the cavity to the containment atmosphere and it might cause the early containment failure by direct contact of containment steel liner with core debris. When the contribution of a high-pressure scenario in a core damage frequency increases, early conditional containment failure probability may become large. In the present study, the verification analysis of PHOENICS code and the combining analysis with MELCOR and PHEONICS codes were performed to examine the debris dispersion behavior during high pressure melt ejection. The PHOENICS code which can treat thermal hydraulic phenomena, was applied to the verification analysis for melt dispersion experiments conducted by the Purdue university in the United States. A low pressure melt dispersion experiment at initial pressure 1.4MPas used metal woods as a molten material was simulated. The analytical results with molten debris dispersion mostly from the model reactor cavity compartment showed an agreement with the experimental result, but the analysis result of a volumetric median diameter of the airborne debris droplets was estimated about 1.5 times of the experimental result. The injection rates of molten debris and steam after reactor vessel failure for a typical PWR plant were analyzed using the MELCOR code. In addition, PHOENICS was applied to a 3D analysis for debris dispersion with low primary pressure at the reactor vessel failure. The analysis result showed that almost all the molten debris were dispersed from the reactor vessel cavity compartment by about 45 seconds after the start of steam release.

Quantum confinement in 3-D device structure together with the newly employed materials like silicon-germanium (SiGe) in advanced technologies (e.g., FinFET, nanowire, nanosheets, etc.) makes transistors seriously suffer from localized self-heating effects in which generated heat within the transistor’s channel is trapped inside. This is mainly due to the much lower channel and surrounding material thermal conductivity and hence lower ability for heat dissipation along with the firm isolation needed for better gate control. Self-heating effects strongly accelerate transistor aging and all the underlying defect generation mechanisms leading to serious reliability problems during the early life of chips. The key challenge in transistor self-heating when it comes to semiconductor testing is the profound difficulty in measuring self-heating directly as generated heat is trapped inside the transistor. Failing in capturing self-heating phenomenon during IC testing would later lead to chips malfunctions at run-time and hence early life failures because of reliability degradations and failure mechanisms will be unexpectedly accelerated akin to excessive internal temperatures. In this paper, we investigate the impact of self-heating effects on n-type and p-type FinFET transistors calibrated with Intel 14 nm measurement data using mature Technology CAD (TCAD) simulations. Then, the industry standard compact model for FinFET technologies (BSIM-CMG) is carefully calibrated to accurately model and reproduce all measurements. This enables circuit’s designers, for the first time, to accurately investigate how emerging self-heating effects in transistors impacts the performance and power of large circuits. This opens new doors for developing novel Design-for-Testing methods that effectively reveal self-heating effects and increase the yield of chips.

Separating HBT wearout from defects during early life operation

A Comparison of Blalock-Taussig, Waterston, and Polytetrafluoroethylene Shunts in Children Less Than Two Weeks of Age

In an attempt to predict early graft occlusion secondary to poor distal run-off, the outflow resistance was measured intra-operatively and compared with the intra-operative angiogram in 50 patients undergoing lower limb bypass procedures. With arterial inflow occluded, 60 mL of plasmalyte was infused via a catheter and the resultant pressure generated in the graft was measured through a second catheter connected to a pressure transducer and recorder. Using the infusion rate and integrated pressure value, outflow resistance was calculated. High resistance measurements were defined as grafts with outflow resistance greater than 1.1 mmHg/mL per min, low measurements were defined as values less than 0.5 mmHg/mL per min and intermediate measurements were those in the range 0.5-1.1 mmHg/mL per min. Angiograms were classified independently of resistance measurements by scoring the numbers of run-off vessels crossing the ankle and stenoses, with subsequent assignment of the patient to one of three groups with either low, intermediate, or high probability of graft failure. Early graft failure was defined as cessation of graft function within 3 months of surgery. Early graft failure occurred in both cases with high outflow resistance and in seven of 44 with low outflow resistance. Of these nine occluded bypasses, six had been placed in the high probability of occlusion category on angiographic appearance and three in the intermediate category. Assessment of angiographic appearance and prediction of patency on a prospective basis had a sensitivity rate of 67% and a specificity rate of 78%. The corresponding sensitivity and specificity rates with resistance measurements were 22% and 100%, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Encapsulant discoloration, a commonly occurring and serious field degradation mode, can be replicated by exposing modules to UV light. Increasing the UV light intensity and the testing temperatures reduces the required testing time. Calculating activation energy using Arrhenius equation for the encapsulant discoloration requires short-circuit current data (or yellowness index data) at multiple temperatures over significant time period. A method of simultaneously obtaining four temperatures in four photovoltaic modules in a single walk-in accelerated UV weathering chamber which drastically reduces the required time and resources is presented. Field-aged modules with visible encapsulant browning were selected for the accelerated testing to determine the wear-out failure mechanisms rather than the early life and midlife failure mechanisms conventionally determined using fresh modules. These modules have been subjected to a very long UV exposure of 450 kWh operating at high temperatures. This work demonstrates that a fast, repeatable, and non-contact UV fluorescence setup can be used in the accelerated testing programs and in powerplants as a standardized technique for an early encapsulant browning detection.

Gold embrittlement in SnPb solder is a well-known failure mechanism in electronic assembly. To avoid this issue, prior studies have indicated a maximum gold content of three weight percent. This study attempts to provide similar guidance for Pb-free (SAC305) solder. Standard surface mount devices were assembled with SnPb and SAC305 solder onto printed boards with various thicknesses of gold plating. The gold plating included electroless nickel immersion gold (ENIG) and electrolytic gold of 15, 25, 35, and 50 microinches over nickel. These gold thicknesses resulted in weight percentages between 0.4 to 7.0 weight percent. Samples were aged for up to 1000 hours and then subjected to a range of environmental stresses, including thermal cycling, and both low and high speed shear testing. Results from thermal cycling indicated an elevated risk of early life failures in SnPb solder joints once gold weight percentage exceeds four percent, which is in line with the results from previous studies. SAC solder showed no indication of early life failures during temperature cycling, yet showed progressive degradation with increasing Au content. The results from low and high speed shear testing seem to suggest that SAC305 solder is more capable of maintaining mechanical properties with increasing gold content. This improvement in gold embrittlement may be partially explained by the additional tin content. Fractography was performed to confirm transitions from ductile to brittle behavior. Metallographic inspections and x-ray mapping were performed to confirm calculated gold content and assess the influence of intermetallic morphology and location on mechanical performance.

In this paper, we illustrate the effectiveness of wafer-level enhanced voltage stress (EVS) along with low voltage sweep (LVS), I/sub DDQ/ and other parametric tests to screen early life failure defects. Our experiment shows temporary undetected defects after repeated exposure to certain applied voltages. We demonstrate a statistical methodology to screen die with suspected early life failure defects.

In this paper presented to the Boston Psychoanalytic Society, Winnicott examines the nature of transference dependence in a patient using clinical material, and he compares this with the stages of infant and child dependency. This vulnerable dependent patient required a phase of deep dependency on his analyst and a regression to the initial difficulties in her early life. All analysts fail and succeed in this work, but in failing and succeeding, they can help towards change. He comments on Dr Zetzel and her views of his work, correcting how he is represented in the area of early environmental failures.

Development of a clinical risk scoring system for artesunate treatment failure prediction in malaria patients

Several probabilistic risk assessments (PRAs) have identified containment loads accompanying reactor vessel failure as a major contributor to the probability of early containment failure during severe accidents. Two significant contributors to these loads are phenomena referred to as steam spike and direct containment heating. To date, direct application of experimental and analytical studies of these phenomena to boiling water reactors (BWRs) are constrained by two limitations: (a) they are based on applications of large, complex containment response analysis computer codes, for which values of many major input parameters are highly uncertain, or (b) they only address pressurized water reactor containment designs. Relatively simple, parametric models are developed which allow a PRA analyst to evaluate the range of conditions under which steam spike or direct containment heating may be important contributors to containment loads for postulated severe accidents in BWRs. The models have been applied to a representative BWR/4 Mark I containment design to illustrate calculated results.