Reliability is an obvious requirement in particular for the developing use of automotive electronics towards autonomous driving. Micro-electronics must also stay reliable in mounting situations, denoted as “3 rd level reliability”, which has been studied systematically. A methodology combining FEA and high precision optical metrology based measuring techniques was applied to test setups, which achieve "3rd level" component loads either by screwing onto aluminum heatsinks or by superimposed four-point bending loads, as was described in [1]. This paper reports the results of both temperature cycle and temperature shock tests up to 900 cycles -40/125°C, the system level deformation effects at higher cycle loading states as analyzed by the high precision optical measurements, and related FE-analyses, which aimed at adaption of the measuring results from component to system level. From deformation measurements a complicated response of the boards to mounting has been observed, including locally different effective board CTEs as a function of cycle number and an irreversible board stretching. It turned out that standard simulation assumptions like neglecting molding compound initial intrinsic stress, the use of geometrically linear theory, and the assumption of homogeneous temperature fields across the systems do not longer hold for system analyses. Measurements revealed significantly different transient temperature fields at heatsinks and board, dependent on position. The transient temperature gradients across the system were adopted by coupled transient thermo-mechanical simulations. They reflect the transient board deformations different from those simulated by assuming homogeneous temperature changes. Solder joints fatigue loading was considered for these cases and comparison was made to previous results, depending on mounting and w/o or with transient temperature gradients. A significant increase of joint stresses was observed, also with regard to the homogeneous T-distribution assumption. Very early solder fatigue failure was predicted and observed from thermal cycle tests.

Among others, physics of failure related concepts are being developed to address the thermo-mechanical reliability challenges in automotive electronics. Limitations in particular applied for finite element (FE-) analyses are models of limited size, which rarely address the system character of failure. Also in testing a system view on fully mounted electronic control units (ECU) and loaded by environmental and active loading cannot be taken performing end-of-life tests for time limitations. Accelerated testing is done instead, however, mostly on board level. To overcome some of these limitations, a combined measuring-simulation technique is being developed, which is described in the paper. System level view on boards mounted in automotive ECUs is taken by a newly developed high-precision optical deformation measuring system. The multi-sensor measuring method combines a chromatic sensor for topography and warping analysis with an optical sensor for in-plane deformation and strain field analysis. By this combination, a high resolution can be reached for all three components of displacement vectors. Additionally, software tools allow the determination of derived quantities like strains, local curvatures and local warpage radius. The latter can be taken as input for FE-simulations. It is shown that some components, in particular QFNs, are sensitive to thermally induced cyclic warpage even if the corresponding bending deflections are in the micrometers range. Worst case in-plane stretching and cyclic warpage of a board mounted in an ECU have been measured. By corresponding simulations on QFN solder fatigue, mounted on a special test board, the critical fatigue life can be determined dependent on the interaction to the case, which differs by several hundred percent from a free-standing assembly.

During the course of the publicly funded project “HotPowCon” (HPC) a two-phase transient liquid phase soldering technology had been developed [1], with the advantage of processing conditions being close to those for conventional soldering. Further developments of this HPC solder material have been made resulting in a stable process with increased quality of the interconnection layer, i.e. better homogeneity and less voiding. The mechanical properties of the multi-phase materials are different from traditional soft solders and depend strongly on the process, in particular the phase proportions. Materials with different proportions were produced and their properties were studied both experimentally and by FE-simulations. A parametric 3x3 particles model was used for the latter. Measurements on the thermo-mechanical deformation behavior were made by a grey scale correlation method, which is behind the microDAC® system applied. Good agreement was achieved, which allows future HPC solder properties prediction and design.

Developments directed towards autonomous driving require complex smart functionalities at reasonable cost, e.g., combined sensing and high volume data processing. Reliability remains a key issue in that process. However, in various cases dedicated automotive grade components are lacking. Therefore, thermo-mechanical reliability issues are one focus of the European project TRACE, which studies the issues for transfer of consumer electronics (CE) into automotive electronics (AE). Gaps between these use scenarios are figured out and measures to be taken are searched [1].Besides the well-known harsh environmental AE requirements, mounting induced effects on components loadings need to be considered. These mounting conditions superimpose stresses driven by the component-board induced CTE mismatch and are in particular critical for leadless components like QFNs, LGAs; WLPs, characteristic of CE use. For evaluation of this loading scenario, a combined measuring-simulation technique has been developed. It uses an optical multi-sensor metrology system for the thermo-mechanical deformation measurement of electronic components and systems for different size and resolution ranges. An application to critical components in an electronic control unit (ECU) is depicted.The combined experimental-numerical method is applied to test-setups, to figure out effects from board mounting on component reliability and characteristic limits due to mounting. Test-boards with systems in QFNs are analyzed. It is shown, that system effects can have major impact on components stress and solder fatigue life.

The development of automotive electronics (AE) towards autonomous driving applications generates various challenges, in particular also on the reliable functionality. In various cases dedicated automotive grade components are lacking and consumer components have to be used instead, which hardly fulfil automotive standards. Some of the reliability challenges are therefore thermo-mechanical in nature. Some example issues, which are related to consumer electronics (CE) packaging, in particular MEMS-packaging, are given in the paper. Main focus is laid on the development of FE-simulation based evaluation methodologies accompanied by experimental characterization methods, in particular with regard to solder fatigue. It is shown that secondary effects such as, for example, intrinsic warpage of the component and of the circuit board, that are system related effects, can play an important role in AE application. A newly developed optical multi-sensor metrology method is presented for the thermo-mechanical deformation measurement of electronic components and systems for different size and resolution ranges. Thermally induced intrinsic warpage of circuit boards and components were analyzed both by means of the method. It was found that significant intrinsic deformations and warpages can occur and should be considered directly in AE system design and indirectly when evaluation stress risks, e.g. for solder fatigue.

The paper reports on thermo-mechanical performance analyses of power semiconductors. Realistic transient temperature loadings as well as mechanical stresses were simulated by fully coupled electro-thermal-mechanical finite element analyses for power cycling loads. Power cycling tests were run in parallel to the theoretical investigations. The failure modes observed by testing were analyzed and adjusted to FE results. Some of the failures need a sophisticated evaluation strategy, as failure initiates at bi-material free edges, which obey a mechanical stress singularity. Damage mechanical modelling by means of the cohesive zone method (CZM) was adopted along with the coupled finite element analysis (FEA) in those cases. Applications of the methodology are presented for a SiC Mosfet testing sample operating at medium power and a high voltage inverter module with insulated gate bipolar transistors (IGBTs) and diodes, operating at high power. Both modules use silver sintering technology on directly bonded copper (DBC) substrates. Top interconnects are made by wire bonding for the Mosfet test sample but by an electroplating based planar technology for the inverter. Considering electro-thermal results it was calculated that stacks with planar copper interconnects outperform the wire bonded versions by 15-30% dependent on layout and current concerning thermal performance. For the die bonds, networks of cracks in the DCB copper and the silver layer replace the creep-ratchetting mechanism dominant for soft-soldered dies. This failure mode could be attributed to high cyclic in-plane normal stresses leading to subcritical crack growth at high power cycle numbers. The failure mode wire bond lift-off, characteristic for heavy Al wires, was investigated by CZM. The CZM methodology was also adopted to evaluate planar metallization delamination For the latter, a parametric study has been made to optimize the materials choice and the layout of the metallization.

Increasing demands for higher energy efficiency and operating at harsh environments lead to the development of new compact power electronics, which is complemented by new interconnection technologies. Investigations were made on a planar copper interconnection technology. The characteristic difference to other technologies can be seen in the replacement of bonding wires by planar copper interconnects and the high voltage applicability of the resulting modules. A high voltage and temperature resistant polymeric foil provides the insulation. Electrical connection is made by structured electrodeposited copper structures, which allow for additional heat spreading from top of the dies. Investigations on the thermo-mechanical behavior of prototype inverter modules, which use silver sintering and copper wire bonding technology or, alternatively, planar copper interconnection technology are reported. Fully coupled electrical-thermal-mechanical finite element (FE-) simulations were used to get realistic transient temperature loadings as well as mechanical stresses, also including wire heating or heating of the planar metallization, respectively. Improved thermal performance of the planar technology could be shown. A parametric FE-study was made to minimize delamination failure risks of planar structures based on cohesive zone modeling. Studies on processing dependent properties of the key materials sintered silver, electroplated copper, and dielectric foils are reported, which are indispensable for simulation input.

Because of the need for electronics use at temperatures beyond 150°C, new high temperature interconnection technologies emerge, like silver sintering and copper wire bonding. In the project PROPOWER inverter modules with IGBTs and diodes mounted on DBC substrates were investigated as a prototype. Analyses were made to study the effects of the new interconnect technologies on fatigue failure. It is one basic question for the replacement of solders by sinter silver to what extend the failure behaviour depends on the kind of loading, that means passive thermal cycling versus active thermal cycling, and how it relates to solder fatigue failure. These questions were addressed by passive thermal cycling and in particular by different kinds of power cycling. Failure modes different from those observed with traditional interconnection technologies were detected. Finite element (FE-) analyses were made to understand the thermo-mechanical stresses, which cause these failures. To simulate the power cycle loading most realistic, coupled electro-thermal-mechanical analyses were carried out which are in the focus of this paper. The simulation results are compared to failures found by testing.

New demands on the thermo-mechanical design of sintered silver interconnections emerge. Development of this inter-connection technology and both experimental and theoretical studies on their reliability were subjects of the project “PROPOWER”. The focus of this paper is on theoretical analysis of thermo-mechanical reliability risks of a project demonstrator, an insulated-gate bipolar transistor (IGBT) module, subjected to power cycling loadings. Coupled electro-thermal-mechanical analyses have been carried out using the finite element method (FEM). Introduction of a new interconnect material means at the same time introduction of a new constitutive behavior and new failure modes. As the material stiffness increases, the decoupling effect of compliant solder layers reduces and intrinsic mechanical stresses increase in the whole power stack. This leads on one hand to less low cycle fatigue in the interconnect, as plastic dissipation is reduced, but on the other hand to higher failure risks like brittle cracking and sub-critical crack growth. However, if early brittle failure can be avoided by appropriate designs, the new interconnection technology allows an increase in fatigue reliability of several hundred percent. Based on the complex theoretical framework simulation results are validated by testing in order to achieve trustworthy thermo-mechanical reliability predictions. Failures like chip metallization damage and the different damage mechanisms of the die bond if either solder or sinter silver is used are related to the different stress situations in the module.

For high temperature interconnection sintered silver can be used, however, it induces new demands on the thermo-mechanical design. That issue requires knowledge on the thermo-mechanical reliability of silver sintered devices, the subject of this paper. Material characteristics of the sinter layers are needed for simulation, which are addressed in the first part of the paper. Based on material properties of pure silver, for sintered silver with different porosities effective material characteristics have been derived by use of a micromechanical cell model. Shear loadings with in-situ deformation analyses have also been made to investigate sintered silver behavior. A complicated dependence on processing, temperature, and deformation rate is seen. Based on different effective constitutive models for the sintered interconnects, stress loadings are studied for a power module, an IGBT on DCB substrate, for passive and active thermal cycling. For the passive cycle complex interactions of the different layers of the stack are observed, which are not seen in a module with soft solder bonding. This result can be attributed to the missing decoupling by the soft soldering layer. Failure risks are evaluated by both conventional FEA and cohesive zone modeling. A quite different stress situation is depicted for active power cycling. The situation is even more complex and it is obvious from the simulations, that active power cycling can induce failure modes different from passive cycling.

Due to the relatively short use time of lead-free solders, the issue of solder joint service life is not fully resolved yet. In particular, there is a lack of testing data for long term thermal cyclic under benign cyclic conditions as well as a on a reliable predictive model for solder fatigue acceleration. This subject has been addressed by long-term testing. Boards from series production were subjected to field cycles 23°C/93°C, 6 hours. After 3 ½ and 4 ½ years or 4800 and 6500 cycles, respectively, the test boards were analyzed. Both solders under investigation, SAC 305 and Innolot, were additionally tested under -40/150°C, 1hour, test cycling conditions. An acceleration factor of approximately 11 was figured out based on computer tomography and cross sectioning analyses. For SAC acceleration predictions based on different published analytical models (Norris/Landsberg equations) were compared to the testing results. The predicted range of acceleration factors was 4.9-39.4, i.e. this type of prediction can be misleading. A second type of comparison was made based on finite element analysis and a related phenomenological model, which gave predictions of the acceleration at least on the save side. For Innolot less fatigue was generally seen, however, kinds of brittle cracking were observed different for field and test loadings. In a final part of the paper an additional study on the effect of two test cycles, 0/100°C and -40/125°C, 1 hour, on multi row QFN fatigue is discussed which shows that involvement of plastic package with Tg in the cyclic range can cause acceleration totally different from analytical predictions.

In MEMS packaging moisture uptake of packaging materials may lead to severe changes in sensor behavior and package reliability. Knowledge and understanding of the effect of moisture swelling on the overall package deformation and on the sensor system is therefore inevitable for a successful MEMS design and packaging solution. The paper shows a design optimization of a pressure sensor packaging towards minimization of sensor drift due to stresses induced by moisture uptake. The work was done in a combined experimental and simulative approach. Experimental data clearly shows the sensor sensitivity to moisture uptake of polymer based packaging material. Based on the experimental findings the complete sensor package was modeled by means of finite element analysis. In a second step a new packaging geometry was defined to reduce the effect of moisture swelling to sensor signal. Additionally sensitivity of the sensor output signal to viscoelastic properties of three different moulding compounds was simulated. In the result of the work an optimized package design was achieved.

Ensuring reliability of interfaces of dissimilar materials is one of the most critical design aspects of electronic assemblies. A study on the interface delamination phenomenon of different power packages was undertaken. Thermally induced stresses caused by the thermal mismatch between molding compound (MC), Si-die, and Cu- heatsink act as combined shear and normal loadings on the respective interfaces. Interface fracture- or damage mechanics was implemented to avoid the difficulties linked with the singular stress field at bi-material interface edges and to allow for inclusion of different interface adhesion parameter as well as damage progress. The cohesive zone approach was adopted for these purposes. A parametric study with different adhesion properties was performed to determine the critical interface energies causing delamination of different power packages. It turned out that delamination failure is not an issue for good adhesion properties, but depends on the package type. For lower critical fracture energies, delamination sensitive areas were found at the sawing edge of the die and at the MC-heatsink interface. Delamination onset was calculated based on a scalar damage metrics based on the critical interface energies. Delamination progress was followed and the delaminated areas were compared with experimental analyses. It turned out that at the critical interfaces MC-heatsink different types of delamination growth can occur which are stable and unstable delamination growth.

In board manufacturing the shear test is frequently used to evaluate the quality of the solder joints shortly after the soldering process as well as after thermal cycling. The latter evaluation is in particular applied to leadless ceramic components; a certain load drop is linked to solder degradation in a certain region of the joint which is declared “failed” if the load drop of the shear force reaches a predefined limit. However, there are many uncertainties in this evaluation like component misalignment to the shear tool, excessive voiding of the joint or by a fracture mode away from solder bulk failure. As few is known about these effects from a theoretical viewpoint, FE-studies have been performed to close this gap The analyses comprise studies of size effects, misalignment, and voiding on the shear force response. Additionally, the simulated solder creep response was compared to the measured one to check for the validity of creep laws for SAC solder. For a realistic creep rate range reached in low cycle fatigue, very low shear velocities have to be applied to the joints. A new shear tester was designed with shear velocities down to 1 nm/s to reach this goal.

Wafer level packaging (WLP) technologies are cost effective packaging solutions which are used increasingly. Second level reliability, i.e. mainly the thermo-mechanical reliability during thermal cycling, is a major concern of WLP. To avoid excessive solder straining, solder balls have been replaced by resilient interconnects, which can adopt the main part of the thermal mismatch deformation. One solution combining an increased reliability on module level with advantages in processing and the capability of full wafer level test and burn-in is ELASTecreg (ELASTec hArr Elastic-bump on Silicon Technology), particularly developed for memory products. The new failure risks are mainly related to fatigue of the metallic redistribution layer (RDL). Parametric studies using finite element analyses (FEA) were performed to avoid excessive straining of the metal lines. A balance of metal straining and solder straining had to be achieved. Comparisons were made for different soft bump layouts and RDL patterns. Optimal solutions figured out by FEA were also investigated experimentally by thermal cycle tests. However, the thermo-mechanical characteristics like stress-strain behaviour and fatigue resistance of the metallic films are the most important parameters for reliability predictions. In particular, the elastic-plastic properties of thin metallic Cu and Ni films are shown to depend on features like film thickness, grain size and orientation, resulting in a thin film strength exceeding the bulk strength of the same metal by several hundred percent

A wafer level packaging technology ELASTecreg has been developed; which uses a resilient bump contact system. The advantages are twofold; because on the one hand the elastic contact system simplifies wafer probing and on the other hand the elastic interconnects allow an increase in board level reliability. Excessive solder bump straining caused by the mismatch of thermal expansion coefficients (CTE) between silicon and organic board materials can be avoided because of the compliance of the contact system, which can take over the main part of the mismatch deformation. Since the electrical connection is made by an electrodeposited copper/nickel redistribution layer (RDL), placed on top of the bump surface, other failures risks than solder fatigue emerge which were avoided by parametric studies using finite element analyses (FEA). The thermo-mechanical characteristics like stress-strain behavior and fatigue resistance of the RDL metallic films are the most important parameters for reliability predictions by FEA, discussed in some detail. The FEA based prediction that the fatigue performance of a spiral RDL layout is superior is proven experimentally and other reliability test data is provided

Thermal fatigue of solder joints is investigated by means of the finite element method (FEM). During the usual thermal cycle regime with relatively slow temperature ramping rates, solder constitutive response is dominated by secondary creep. In the paper secondary creep laws are given for SnPb and SnAgCu solders, based on our own measurements and literature data. These secondary creep behaviours are compared to other data recently published. Additional primary creep terms are introduced in the creep description. To relate calculation results to fatigue life, criteria of Manson-Coffin type are proposed. These empirical laws are based on either the cyclic inelastic strain or dissipated energy. Effects of the choice of different creep laws on the calculated creep strains and energy densities are studied for two standard components on FR-4 board, a ceramic chip resistor (CC) of size 0805 and a plastic ball grid array with 225 I/O (PBGA 225). It is shown that the choice of the creep law, e.g. the inclusion of primary creep, does significantly affect only the results for the PBGA solder balls.