A fatigue-capable extension of the Advanced Mixed-Mode Bending (AMB) method is presented and demonstrated in pure opening mode [1]. Symmetrical specimens of EMC-PI-EMC containing a notch up to the interface were first quasi-statically preconditioned under displacement control in order to generate a defined initial interfacial crack state. The last stable force level of this precracking step was then used to define a force-controlled harmonic fatigue window. During cyclic loading, force and displacement were recorded at the lower and upper load levels. A progressive increase in displacement under nearly constant cyclic force levels indicated an evolving specimen response during cyclic loading. Since a robust direct crack-length reconstruction from the optical data was not available for all fatigue states, an inverse effective crack-length approach was used [2]. The underlying aeff methodology had been introduced previously for AMB fracture evaluation to improve crack-state identification by correlating experimental and simulated response quantities [2] and [3]. In the present work, this approach was transferred to cyclic loading and used to identify compliance-equivalent crack states from the measured cyclic response. Selected effective crack states were then transferred into finite-element models with premeshed cracks and evaluated by the crack surface displacement extrapolation method in terms of energy release rate and mode mix [4]–[6]. The study demonstrates a compact experimental–numerical workflow for interfacial fatigue assessment with AMB under opening-dominated conditions and provides a basis for future mixed-mode fatigue investigations.

Die Lock-in-Thermographie (LIT) ist ein etabliertes zerstörungsfreies Prüfverfahren zur Detektion von Oberflächen und „subsurface“-naher Fehlstellen anhand von Amplituden- und Phaseninformationen. Die resultierenden Messdaten liegen jedoch typischerweise als zweidimensionale Bildinformationen vor, wodurch eine direkte geometrische Zuordnung der detektierten Anomalien innerhalb dreidimensionaler Bauteilgeometrien nur eingeschränkt möglich ist. Insbesondere im Kontext digitaler Zwillinge und modellbasierter Zustandsbewertung besteht daher ein Bedarf an Verfahren zur räumlichen Integration thermografischer Messdaten in virtuelle Bauteilmodelle. In dieser Arbeit wird ein software-basierter Ansatz vorgestellt, der Lock-in-Thermographiedaten als skalare Oberflächenfelder interpretiert und auf dreidimensionale „Stereolithography“ STL-Bauteilmodelle projiziert. Amplituden- oder Phasenwerte werden normiert, in physikalische Größen überführt und zur Erzeugung eines 3D-Reliefs genutzt, welches die räumliche Verteilung der thermischen Reaktion auf die reale Bauteilgeometrie visualisiert. Die Methode ermöglicht eine intuitive geometrische Lagezuordnung thermografisch detektierter Fehlstellen und stellt eine flexible Schnittstelle zwischen thermografischer Messdatenerfassung und digitaler Bauteilbeschreibung dar.

Lock-in thermography (LIT) is an established non-destructive testing method for detecting surface and near-subsurface defects based on amplitude and phase information. However, the resulting measurement data are typically available as two-dimensional image information, which limits the direct geometric assignment of detected anomalies within three-dimensional component geometries. Particularly in the context of digital twins and model-based condition assessment, there is therefore a need for methods for the spatial integration of thermographic measurement data into virtual component models. In this work, a software-based approach is presented that interprets lock-in thermography data as scalar surface fields and projects them onto three-dimensional stereolithography (STL) component models. Amplitude or phase values are normalized, converted into parametric quantities, and used to generate a 3D relief that visualizes the spatial distribution of the thermal response on the real component geometry. The method enables an intuitive geometric localization of thermographically detected defects and provides a flexible interface between thermographic measurement data acquisition and digital component representation.

Inline inspection of electronic packaging is crucial for ensuring the reliability and safety of modern electronic systems, where both detection accuracy and inspection speed are critical requirements. Infrared thermography has been widely explored for the non-destructive evaluation of packaging structures due to its rapid and accurate inspection capabilities. Among these techniques, pulsed thermography (PT) has demonstrated effectiveness in defect detection; however, conventional thermography relying on external optical heating (e.g., flash lamps or lasers) often shows limited sensitivity to subsurface defects in multilayer electronic packaging structures. For conductive materials, eddy current thermography (ECT) employs eddy currents as an internal heat source to induce localized thermal responses, enabling non-contact and rapid defect inspection. However, for non-magnetic materials with high electrical and thermal conductivity, such as copper (Cu) and sintered silver (SAG), a single short heating pulse is often insufficient to induce a strong thermal response. In this study, we investigate ECT using multi-pulse excitation, which is applied periodically and, when analyzed at the fundamental frequency, is equivalent to lock-in thermography (LIT) with a square-wave signal, for defect detection in electronic packaging, with a focus on subsurface defects within multilayer structures. The influence of lock-in frequency and number of cycles is analyzed for two defect sizes. The goal is to evaluate the potential of ECT for identifying internal defects that may affect the reliability and performance of electronic packaging.

Smart skin technologies constitute a key enabling technology for many topical applications as e.g., for robotics, exo-skeleton, medical or defense applications as human-machine interface. The time-honored way to address this issue is to package sensors on a flex substrate. This paper, in contrast, introduces a novel concept for simple, robust, low cost and conformable tactile smart skin based on low-cost polymer-integrated pressure sensors, an elastomeric coating and readout by a trained neural network, allowing reconstruction of the entire force vector in magnitude, direction and position. We have generated training data by FE-simulation, showing that the concept is indeed robust – not only mechanically against shock – but also computationally and against sensor decalibration and even failure of individual sensors.

Radio-echo sounding (RES) is a widely used tool in the field of glaciology with which critical information about englacial and subglacial conditions can be derived. However, RES observations have historically been limited to zero-offset or small-offset surveys, typically employing one transmitting and one receiving antenna. The poor spatial and azimuthal coverage of the subsurface associated with these sparse geometries limits the ability to robustly constrain key englacial and subglacial properties including ice temperature, bed material composition, water content, ice fabric, and firn density. Furthermore, using radar only in zero- or small-offset configurations limits its potential to provide high resolution imaging of bed geometry. The maximum achievable offset in ground-based radar surveys is typically limited by the relatively high-loss coaxial cable which connects the radar transmitter and receiver. To overcome this limitation, two multi-offset ground-based radar systems, both built around an autonomous phase-sensitive radio-echo sounder (ApRES) as a transmitter, have been developed and deployed by the Radio Glaciology Group at Stanford. The first system forgoes cabled connection between a transmitting ApRES unit and a software-defined radio (SDR) based receiver, instead relying on a post-acquisition processing flow to ensure coherent summation of repeated measurements to achieve sufficient signal-to-noise ratios. The second system replaces the standard high-loss coaxial cable with low-loss fiber optic cable in order to extend the maximum achievable offset between transmitter and receiver. This requires outfitting the ApRES radar system with hardware to convert radio-frequency signals into optical signals that can be transmitted over fiber optic cable (RFoF). Both systems were deployed during the 2023-24 Antarctic field season as part of the Thwaites Interdisciplinary Margin Evolution project in order to collect multi-offset RES data on both floating and grounded ice. These surveys are aimed at detecting englacial temperature anomalies and the estimation of dielectric properties of englacial and subglacial materials through amplitude-versus-offset analysis of radar data. The dense multi-offset coverage in surveys described here was built up by frequent repositioning of only four SDR-based and one ApRES-based receiver; however, future surveys with these systems could have 10s or 100s of radar receivers simultaneously recording, allowing for survey geometries commonly employed in active source seismic imaging to be applied to radar imaging.

Additiv gefertigte Aluminium-Matrix-Composite (AMC)-Materialien bieten durch ihr geringes Gewicht und ihre hohe Festigkeit signifikante Vorteile, führen aber während des Schweißvorgangs zu Porenbildung die die Festigkeit beeinflussen. Im Rahmen der Förderbekanntmachung „WIR! – Wandel durch Innovation in der Region“ des Bundesministeriums für Bildung und Forschung (BMBF) wird im Teilvorhaben „Inline-Prozesskettenentwicklung zum Reparieren von AMC-Werkstoffen sowie Porendetektion mittels aktiver IR-Thermographie zur Prozessüberwachung“ eine Machbarkeitsstudie zur Nutzung der Lock-in-Thermographie (LIT) zur Defekterkennung in AMC-Schweißnähten durchgeführt. Ziel der Machbarkeitsstudie ist es, die Anwendbarkeit der LIT für die Inline-Prozessüberwachung dieses Werkstoffs zu bewerten, da dessen thermische und mechanische Eigenschaften, darunter eine thermische Diffusivität von etwa 8,6 * 1E-5 m²/s, besondere Herausforderungen für die zerstörungsfreie Prüfung (ZfP) darstellen. Die LIT, bei der eine modulierte Wärmequelle verwendet wird, erlaubt die Erkennung von ‚subsurface‘ Defekten durch die Analyse der thermischen Wellen. Es werden verschiedene Algorithmen eingesetzt, um die thermischen Daten zu analysieren und die thermische Empfindlichkeit der verwendeten Kamera zu simulieren. Neben den experimentellen Untersuchungen werden FEM-Simulationen genutzt, um optimale Betriebsparameter wie Modulationsfrequenz und notwendige Laserleistung zu bestimmen. So konnten bereits zuverlässig Defekte ab 2 mm Größe bis zu 1,5 mm Tiefe unter der Oberfläche detektiert werden. Die FE-Simulationsergebnisse zeigen zudem, dass durch Anpassungen der Anregungsfrequenz sowie weiterer Parameter wie Energieeintrag, Kamerasensitivität und Auswertemethoden auch kleinere, tiefer liegende Defekte erkannt werden können. Die ersten Ergebnisse zeigen, dass die LIT auch für additiv gefertigte Metalllegierungen für die präzise Identifizierung verborgener Defekte einsetzbar ist und somit In-line zur Steigerung der Produktionssicherheit sowie der Produktqualität beitragen.

The reliability of thermal interface materials (TIM1) in lidded high-power flip-chip packages is investigated using a thermal test vehicle (TTV) under active power cycling. Two TIM materials with different thermal conductivities were evaluated by combining continuous thermal resistance monitoring with transient thermal measurements at defined cycling intervals. The results show that degradation of TIM1 is highly localized and strongly influenced both by assembly-related factors such as warpage and tolerances, and by the intrinsic material behavior under thermo-mechanical stress. Clear differences in spatial thermal resistance evolution were observed: TIM-A exhibited corner degradation, whereas TIMB showed more pronounced changes in central regions. Complementary accelerated grease pump-out tests confirmed material-dependent tendencies for void formation and displacement. The findings underline the need for location-sensitive assessment to predict TIM1 reliability in large-area dies for high-performance computing applications.

The research paper focuses on physics-based reliability investigation of power module structures at substrate and component level which will be designed for various automotive applications. The direct copper bond (DCB) substrate is attached onto a copper heatsink by sintered silver of different thicknesses. Thermal and electrical conductivity of the sintered silver layers (360 ± 69 W/m•K and 39 ± 6 MS/m respectively) are determined and the microstructure is characterized by scanning electron microscopy (SEM) and electron backscatter diffraction (EBSD). Thermal cycle tests are conducted on the samples by experiment and numerical simulation. The degradation of sintered layer increases with number of thermal cycles. On component level, analysis has been performed on different power modules including die top system (DTS) and copper bond pad metallization (Cu AIT) on SiC chip. Active power cycle tests (APCT) and thermal shock tests (TST) are performed to analyze different reliability issues. Cu AIT has higher lifetime than die top system.

Mismatched thermal expansion coefficients in electronic assemblies induce stresses that can lead to mechanical failure. This problem is intensified in autonomous vehicles due to high power densities, harsh operating environments, and limited cooling capacity. Addressing the heightened reliability requirements of autonomous driving necessitates the development of advanced cooling strategies. To this end, heat pipes for enhanced thermal spreading, as well as different embedding strategies, are investigated in this work. Sintered heat pipes are tested under loads of $1-6 ~\mathrm{W}$ and $20-100 ~\mathrm{W}$ with two different test setups and their performance is reported. Varying thermal conductivities ($791-13181 ~\mathrm{W} / \text{mK}$) are observed depending on the heat flux, heat pipe dimensions and temperatures with an uncertainty range between $5-10 \%$ in most cases. In addition, the importance of heat spreading is shown with a Thermal Test Vehicle (TTV), where a reduction up to 15 % of the maximum junction temperature can be reached under 300 W loading with a copper heat spreader. Finally, two prototypes of heat spreaders with embedded heat pipes are investigated. While it is shown that embedded heat pipes concentrating on the center impede the thermal performance, a reduction up to 10 % in maximum junction temperature is seen in surface embedded heat pipes at coolant temperature range between $50-65^{\circ} \mathrm{C}$ for the given configuration. The results indicate, along with a further potential, a need for improving the thermal contacts and different embedding configurations.

Lock-in thermography (LIT) enables precise, non-destructive defect detection in additively manufactured aluminium matrix composite (AMC) welds. In this work, a specially developed experimental setup is used to optimise the analysis of the thermal response to periodic thermal excitation. By combining experimental investigations with transient thermal simulations and Fourier analyses, defects from 2 mm in size down to a depth of 1.5 mm could be reliably identified. In addition, an algorithm for automated image difference analysis is introduced to facilitate and accelerate detection and depth estimation. The results show that LIT has a high sensitivity to internal material irregularities and has the potential to be used as a method for precise and reliable defect detection for in-line industrial applications. In addition, the measurement time could be reduced down to one minute while using pre-heating of the sample, which increases the efficiency of the method.

Electronic packaging plays a critical role in ensuring the reliability and performance of electronic systems. As various materials are used in these systems, each with distinct properties, failure analysis of each is key to improving overall system reliability. This paper presents a comparative study of two non-destructive testing (NDT) methods—pulsed thermography (PT) and lock-in thermography (LIT) — for failure analysis in electronic packaging materials. The study focuses on defect detection in copper (Cu), aluminum (AI), and polymer, which are widely used in electronic packaging due to their distinct thermal characteristics. Experimental results are evaluated using signal-to-background contrast (SBC). To enhance thermal contrast, three post-processing techniques—pulse phase thermography (PPT), principal component thermography (PCT), and thermographic signal reconstruction (TSR)—were applied to the PT results. These results were then compared with the optimal performance of LIT. The findings demonstrate that both methods have distinct advantages and limitations: PT offers faster detection, while LIT provides higher sensitivity for materials with higher thermal diffusivity. This study enhances the understanding of how these techniques can be effectively applied in electronic packaging failure analysis.

High-Performance Computing (HPC) is already becoming an essential part of many engineering fields, and automotive sector is no exception. With the autonomous driving becoming rapidly the new technological forefront in automotive ecosystem, new peaks in computing power in automobiles will be seen, and relying only on material developments to push this front further might prove to be in vain. It is therefore essential to come up with solutions to address the new challenges rising, such as increasing power densities and thermo-mechanical loads at a system level. To that end, we propose a new modular approach to standard lidded packaging that aims to achieve better thermo-mechanical performance and acts as a reworkable thermal plug. This approach is called CoolStar, where a fractal heat spreader is attached to the liquid cooler with a ultra-low modulus, gel-like gap-filter. Proof-of-concept is presented in this paper, showing that the approach is promising addressing thermo-mechanical stresses at the system level. The idea can also be installed and tested in real-life applications, and further numerical studies along with experimental validation of the concept is essential.

The vision of a deep learning-empowered non-destructive evaluation technique aligns perfectly with the goal of zero-defect manufacturing, enabling manufacturers to detect and repair defects actively. However, the dearth of data in manufacturing is one of the biggest obstacles to realizing an intelligent defect detection system. This work presents a framework for bridging the data gap in manufacturing using the potential of synthetic datasets generated using the finite element method-based digital twin. The non-destructive technique under consideration is pulse infrared thermography. A large number of synthetic thermographic measurements were generated using 2D axisymmetric transient thermal simulations. The representativeness of synthetic data was thoroughly investigated at various steps of the framework, and the image segmentation model was trained separately on experimental and synthetic datasets. The study results reveal that when carefully rendered, synthetic datasets represent the experimental data well. When evaluated on real-world experimental samples, the segmentation model pre-trained on synthetic datasets generalizes well to the experimental samples. Furthermore, another advantage of synthetic datasets is the ease of labelling a large amount of data. Finally, the robustness assessment of the model was done on two new datasets: one where the complete experimental setup was changed, and the other was an open-source infrared thermography dataset

This paper addresses thermal transient IR thermography as full-field and non-destructive failure analytical technique to determine the bond quality of microelectronic joints, exploiting the transparency of a silicon chip for MIR photons. The method is exemplified for C2C bonded electroplated Aluminium joints as they could come in as a novel joining technology for 3D heterogeneously integrated Cryo-SiP architectures for ion-trap based quantum computers. We have evaluated the method experimentally for different thermosonic bond parameters and validated it by metallographic sectioning. First results are shown. We could derive that the phase and amplitude image could be used to detect defects in the joints with good sensitivity and within reasonable testing times. This method could have potential to determine zero-hour quality of such joints during inline testing.

One of the tasks of the research project is to develop large-area sintered joints on copper and aluminum water heat sinks in order to realize corrosion-compatible use in electric vehicles. To this end, strategies are being developed to equalize the strain differences of the different materials in the as-sembly and interconnection technology of the power modules, thus helping to reduce the Coefficients of Thermal Expansion (CTE) mismatch. This will help to extend the lifetime of power modules. Therefore, one goal is to conduct investigations with different materials and metallization between the power electronics and the heat sink, so that an adjustment of the CTE to the packaging and interconnection technology in the active (hot) area can be made. To achieve these goals, the development of suitable large-area sintering pastes and the implementation of necessary modifications to sintering tools and machines are required. Furthermore, the development of suitable failure analysis methods are helpful for the analysis of defects in the samples.

In this work, flow boiling heat transfer to water is investigated experimentally. Flow boiling heat transfer is a topic that is of interest for the cooling of power electronics, high-performance computing in automotive applications, and other electronic packages with particularly high heat flux. To investigate cooling solutions for new electronic packages a closed-loop two-phase flow boiling cooling test stand for rapid experimental characterization and parameter variation is introduced. In order to have a proof of concept for the test stand and to demonstrate comparability with the test setups described in the literature, first, different cooler structures in the mm range are investigated, using copper heater modules driven by heater cartridges. Secondly, the integration of flow boiling cooling together with an industry-oriented electronic power package is demonstrated. A heat transfer coefficient of 24 kW/m(2)K (242 W/cm(2) at 1.5 kW on a chip area of 24.9 x 24.9 mm(2)) was achieved for a plain silicon chip surface at a volume flow rate of 1 l/ min using dielectric water as cooling fluid. The chip temperature and the differential pressure were measured to 137 degrees C and 8 kPa.

Flow boiling heat transfer to water of different cooler surface topologies is experimentally investigated. To have the possibility for investigations of new electronic packages that supports cooling solutions by two-phase flow boiling, a multipurpose closed two-phase flow boiling cooling test stand for rapid experimental characterization and parameter variation is designed. The designed test vehicle holds interchangeable, plug-in heating module that offers the possibility to investigate different cooler structures and surface topologies/chemistry to improve the formation and release of small bubbles. In order to have a proof of concept for the test stand and to show its reproducible to those given in the literature, initially, the investigation is limited to a plain-and dimpled-surface and a pin fin and finned cooler structure in the mm range.

For the attach of dies to a substrate in power electronic assemblies, sintering is one of the common technologies. Sintered die-attach layers always contain pores due to diffusion processes and uncompacted gaps between metal particles in the sinter paste. In copper sinter paste with flake-shaped particles, the sintered structure exhibits regions with varying contents of gap-shaped pores. To meet heat transfer requirements, it is important to investigate whether these pores have a large effect on heat diffusion within the layer. Four different sintering materials were evaluated in generic test assemblies, where half of the samples had the die removed to reveal the uncovered sintered interconnection layer. These layers have been analyzed and lock-in thermography was conducted to identify their heat transferring capabilities. While optically visible structures, that correlate with pore clusters, are found in all samples, the heat transfer was not significantly reduced. Only one sample type with a very strong inhomogeneous structure showed significant influence on the heat dissipation. That infers, that even some stochastic pore cluster structures in the sintered layer of a copper sintered die-attach only have a negligible effect on heat dissipation.

The paper establishes foundational work for lock-in thermography as an in-line quality inspection tool during the production of power electronic devices. It focuses thereby on the fast, reliable, and unequivocal detection of production defects, which reveal themselves through a characteristic transient thermal signature and the analysis of phase and amplitude images. The experimental and finite element analyses were performed at several excitation frequencies and different numbers of cycles for excitation. A finite element model was used to make a comprehensive comparison between simulated outcomes and experimental results done with lock-in thermography (LIT). Different postprocessing techniques are applied to the thermal images to extract the amplitude and phase, such as the Fast Fourier Transform (FFT), the Digital Lock-In Correlation Method (DLCM), and the Four Points Method (4PM). The finite element analysis and experimental results both successfully detected the subsurface defects. The amplitude and phase values were found to be remain constant at specific values without increasing after a number of cycles.