ABSTRACT. Reliability of SiC and GaN power devices from the industrial perspective

Superior performance and proven reliability have been enabling the successful introduction of SiC and GaN power devices into important power electronics applications. This tutorial will review the reliability aspects of wide band-gap semiconductors and provide a deeper insight into the industrial qualification procedures for safe technology and product releases.

GaN (Thomas Detzel)

Compact devices with unique switching performance have been the promise of GaN power technology since several years. And now we are experiencing the exciting time when this is becoming reality. This part of the tutorial will first discuss the main reliability responses of normally-off GaN power HEMTs such as gate module reliability, time-dependent dielectric breakdown, and dynamic Rdson. Then the industrial qualification procedure will be described covering technology, device, and product reliability tests. Focus will be given also to application tests which are of particular importance for the release of this new material system in power electronics.

SiC (Thomas Aichinger)

Reliability aspects and reliability test methods of SiC are discussed and compared with the well-established Si-based power devices. Strong emphasis is put on the gate oxide system of SiC vs. Si MOSFETs. It is shown, that the high reliability and low failure rates of modern IGBTs can be achieved also with SiC MOS devices using electrical screening and an appropriate dimensioning of the gate oxide with respect to the gate use voltage.  Furthermore, the consequences of the higher mechanical stiffness on the packaging technology as well as the issues arising from the handling of 10x higher electrical field in the semiconductor body are discussed and suited test methods are suggested.

ABSTRACT. Dielectrics are the core of most electronic components. Their performance and reliability often determine a technology success and it is for these reasons that they have been deeply investigated since the beginning of microelectronics. In this tutorial we will review the most important dielectric applications, active and passive dielectrics, highlighting the critical issues and best characterization methodology depending on the application. Fast and long lasting electrical tests will be discussed, their correlation and their limits.

ABSTRACT. Today packaging of power devices is a fast developing subject, as new materials and concepts emerge. At the same time reliable operation throughout the lifetime needs to be ensured from the package development. This tutorial will give an overview of the actual trends in power electronic packaging and the related reliability aspect associated within. It will start from a general introduction of the basic methodology. Packaging solution options will be presented ranging from different materials for wire bonding and die attach to different integration concepts (embedding concepts). Associated possible failure mechanisms will be explained as this knowledge is the basis for suitable modelling and test planning. Typical models will be introduced and test examples given. Thus a basis for the Design and Test for Reliability is given.

ABSTRACT. Within outdoor applications like electrical vehicles or a wind turbines the inner electronic system can be exposed to harsh climate conditions from the ambient. Often no dehumidifiers are used to explicitly control the moisture level inside. Therefore condensation of water, corrosion and material swelling can occur which causes disturbances or even a total loss of the system. Here simulation of the moisture transport helps to understand the system and finally protect it from such moisture induced failures. In the past, methodologies were developed which make use of the analogy between moisture transport – considering diffusion processes – and heat conduction. Also known as thermal-moisture analogies, these methods are implemented by using the thermal simulation domain of commercial finite element (FE) solvers and reinterpreting the material properties to describe the moisture diffusion. Because of the fact that moisture and thermal transport are especially at bi-material interfaces not fully analogous, these approaches have specific limitations concerning the allowed transient and spatial boundary conditions. For moisture modeling on system level the computational effort of the FE solver is typically too high, therefore the model complexity has to be simplified. This can be realized by modeling the moisture transport via an equivalent circuit approach in which the moisture resistors and capacitors describe the transport and loading of moisture within the material parts. This approach is suitable even for complex transient mission profiles – given by transient moisture and temperature profiles as input – which cover long periods of time by keeping the computational effort low. This tutorial gives an overview of these common modeling approaches and their specific limitations regarding the allowed boundary conditions and model complexity.

ABSTRACT. Product-Reliability is meeting customers’ expectations throughout expected use time. The Modern product/component Reliability are based on the physics, statistics and the stress/mission profile the component or design will experience.
The method discussed will make it possible to have the important dialog with the management to understand the impact of poor reliability and the necessary priorities that always shall be done in a company. 
The methods discussed here closes the reliability learning loop from setting up the Reliability Goal, following up during the Design Phases, Production Phases and the values can be compared with Field Failures. The method is based on the Physics of Failure parameters, Weibull, Mean Cumulative Function and this make it possible to estimate Warranty cost, which can be transformed to Standard Unit Cost and make it possible to optimize the total product cost. You will have the opportunity to try some of the methods discussed yourself after the tutorial.

ABSTRACT. “Geometrical Scaling” characterized the 70’s, 80’s and 90’s. The NTRS identified major transistors material and structural limitations. To solve these problems the ITRS introduced strained silicon, high-κ/metal gate, FinFET, and other semiconductor materials under “Equivalent Scaling”. Horizontal (2D) features will reach a limit beyond 2020. Flash producers have adopted the vertical dimension. Logic producers will follow. IRDS assessed that “3D Power Scaling” will extend Moore’s Law for at least another 15 years. Furthermore, computing performance will be substantially improved by monolithically integrating several new heterogeneous memory layers on top of logic layers powered by a combination of CMOS and “new switch” transistors.

ABSTRACT. The last decade has witnessed the wind power industry firm its grip on reliability with reported loss of generated energy coming down from double-digits to below 2%, accompanied by a configurable balance between product CAPEX and service & repair cost. Electrics & electronics taking their fair share have undergone a noticeable reliability engineering journey, through all product lifecycles. The presentation will look back at some of these, and forward to next steps.

ABSTRACT. Laser voltage imaging is a commonly used electrical fault isolation technique that is extremely useful in localization of defects in serial chains and circuits generating specific frequencies within modern integrated circuits. The signals are generated by electro-optic mechanisms and the resolution of the tool is therefore optically limited. As such, conventional methods of understanding optical resolution in optical systems such as Rayleigh and Sparrow criteria are currently used to describe the effective resolution. This work attempts to formalize the description of resolution considering the point spread function of the laser and the electrical circuitry the laser interacts with. Finally, the model is compared with real results across multiple laser wavelengths, including those from visible light probing.

ABSTRACT. The work presented in this paper is related to the application of Failure Analysis Laser Voltage techniques (LVx) to a real case of analysis on a scan chain transition fault issue detected during scan chain integrity test. Laser Voltage Imaging (LVI) and Laser Voltage Probing (LVP) techniques were applied with the aim of both identifying and fully characterizing the failing flip-flop of the chain. The relatively simple DFT structure of the device, with a single scan chain embedded in spite of a large area of the digital logic, did not allow the software diagnosis to be accurate, leaving the work of point of failure identification entirely to the following step of fault isolation. In this work we put special focus on the application of the two fault isolation techniques, to identify the failing flip-flop through dichotomy approach, investigating the macro failure mode through a second harmonic analysis and finally characterizing this failure using LVP. The hardware setup required the adaptation of an existing general purpose IDS (Integrated Diagnosis System) format board, to contact and provide few low frequency digital channels to the device, in SOIC8N package and cope with the mechanical constraints of Fault Isolation equipment. The device under test is the digital circuitry controlling a MEMS sensor for automotive market, stacked through bumps with flip-chip approach.

ABSTRACT. Parasitic bipolar effects are a well-known failure mechanism in integrated circuits. They trigger latching phenomena and are of particular interest for higher current applications and under leaky conditions. Photon emission microscopy (PEM) has proven to be a suitable tool to detect the occurrence of parasitic bipolar elements in MOSFETs, such as the parasitic bipolar junction transistor (BJT). Spectral PEM even allows for a more detailed characterization of the respective parasitic bipolar operation modes. In this work, we show that spectral PEM can also be used to characterize bipolar parasitics in modern p-type and n-type FinFets. Additionally, this characterization method allows us to determine whether germanium enriched silicon was used to optimize device performance. We conclude by demonstrating that the FinFET devices under test show a very good suppression of parasitic bipolar effects.

ABSTRACT. The electrolytic capacitor (Al-Cap) is widely considered as one of the reliability-critical components in LED driver applications. A realistic lifetime prediction of Al-Caps would be beneficial to the sizing of capacitors to fulfill a specified design lifetime. This paper investigates the impact of electrolytic capacitor parameter shifts on the LED driver performance and the capacitor reliability. The operation condition, degradation, and tolerance are considered as the main reasons that deviate the parameters from nominal values. A method to model the degradation of Al-Cap parameter is introduced as well. In the case study of a flyback LED driver for street lighting, the contribution of parameter shifts to the electrical performance of LED drivers and Al-Cap core temperature is evaluated based on both theoretical analysis and simulation. A quantitative comparison of the estimated lifetime of Al-Cap in the LED driver is given with and without the consideration of capacitor parameter shifts within the service life. The proposed method enables a more realistic way for electrolytic capacitor lifetime prediction in LED drivers, which would otherwise has the risk of overestimation.

ABSTRACT. As increase the power density of power electronics converters, a capacitor is a major factor in reducing the reliability of high power density power electronics converters. The lifetime and reliability of capacitors are strongly influenced by temperature. To evaluated temperature rise, an accurate loss measurement method is necessary. However, practical capacitor loss measurement system used in power electronics converters has yet to been established because capacitor loss data provided from manufacture is always under sinusoidal excitation and not much to electronics converters excitation. The authors thus propose capacitor loss measurement method for power electronics converters. This measurement method features fast capacitor loss measurement from real circuit with high accuracy and capacitor loss analysis for each switching period of power electronics converters. In order to verify accuracy of loss measurement, measured loss value of filter capacitor used in Pulse Width Modulated (PWM) inverter is compared with calculated value. The experimental results agree with the calculated ones.

ABSTRACT. Condition monitoring plays an important role in estimating health condition of capacitors because ageing of the capacitors is usually accompanied by an increase in equivalent series resistance (ESR) and decrease in capacitance. Either capacitance or ESR cannot be a unique indicator of the lifetime of capacitors in some cases. This paper presents a condition monitoring method of a dc-link capacitor used in a three-phase PWM inverter with a front-end diode rectifier intended for motor drives. The monitoring method extracts both the equivalent-series resistance (ESR) and capacitance of a capacitor under test from the actual ripple current and voltage without disconnecting the capacitor nor introducing additional current injection. The monitoring method, therefore, can be implemented online. Experimental results verify that the monitoring method independently obtains the ESR and capacitance changes of the capacitor under test. This contributes to accurate life time estimation of dc-link capacitors.

ABSTRACT. The reliability and the accompanying fatigue law for solder joints are challenge for the aerospace and defense industries. Due to the service loadings and the lifetime requirements, fatigue laws must take into account plastic and viscous damages as the involved materials exhibit creep and relaxation even at room temperature.

This presentation presents a novel and innovative experimental setup developed for the fatigue analysis of assembled package. The setup controls shear loading of the packages with or without dwell time and controlled temperature. The paper discusses a series of experiments and unveils the correlations between the monitoring of electrical resistance and the mechanical failure criterion. Microscopy failure analyses of the failed cross sections indicate a crack surfaces similar to the ones obtaining during thermal cycling. The monitoring of the electrical resistance is able to detect the failure of a single solder ball. It therefore defines a precise failure criterion, when compared to the direct monitoring of mechanical parameter which leads to an unrealistic number of cycles to failure. The mechanical failure represents more precisely the global failure of the component. Moreover, the analysis shows that the variability of lifetime can be attributed to the variability of the solder joints.

ABSTRACT. In the field, electronic devices are often experiencing switching cycles. With the trend of miniaturization and electrification, automotive electronic packages are facing increasingly critical reliability issues regarding current stressing and the accompanying Joule heating. In this work, the load capability of two surface-mounted shunt components subjected to cyclic current pulses are investigated. Asymmetric self-heating due to Peltier effect is observed with infrared thermography in one of the shunt components, which leads to polarity in the microstructural evolution influencing the failure mode in the solder joint. Finite element method (FEM) is implemented to simulate the coupled electrical, thermal and mechanical fields in the assembly and to assess the response of the solder joints exposed to power cycling. Lifetime data are collected and statistically analysed based on the Weibull distribution.

ABSTRACT. Since lead-free solder was adopted for in-vehicle electronic components, a sufficient number of years have passed to allow examining the condition of lead-free solder joints in components used in the field. For this paper, using an EBSD analysis technique, solder joint degradation analysis was conducted on specimens subjected to accelerated testing and field-aged components. Degradation indicators of the level of grain refining and the amount of strain in the Sn phase in solder were obtained from specimens subjected to accelerated testing. The analysis results of field-aged components revealed increases in the degradation indicators commensurate with the age of components, thereby verifying the validity of the indicators.

ABSTRACT. Wear out in the maximum power point condition of tandem microcrystalline Si / amorphous Si thin film PV modules due to extended light soaking is a well-known effect due to the Staebler –Wronski mechanism [1]. However, several studies [2,3,4,5,6] on a-Si:H single junction and tandem solar cells, have shown that, under suitable conditions of electric field in reverse bias, illumination, especially at well-defined wavelengths (700-800 nm), and temperature conditions, an improvement of the cell main electric parameters is possible. In this work, we report on outdoor tests on stabilized tandem amorphous Silicon PV devices, highlighting the improvement of their electrical performances on going from morning to afternoon illumination. The efficiency improvement observed in the afternoon is stable overnight and observed also in the early stages of the morning outdoor test in the following day. Possible causes of such effect are discussed.

ABSTRACT. The reliability of solar photovoltaic (PV) panels is significantly affected by the formation of hotspots in active operation. In this work, hotspots are analysed in crystalline silicon (c-Si) and amorphous-Si (a-Si) thin-film (TF) PV modules, along with the investigation of the shading effects on their performance and reliability. The c-Si and TF (a-Si) PV panels behave differently in varying shade scenarios with the TF panel technology showing largely superior results in respect to hotspots and the performance. The better performance in TF is attributed to its monolithic panel architecture, which allows better redistribution of the voltages and currents within the module. It was observed that, for a typical module, the overall shadow-induced reverse temperature is lower in TF modules. This work therefore provides useful insights for PV system planning and operation for regions with high average temperatures along with potential issues due to partial shadings.