Applied Research/ Programs

Fraunhofer USA Center Midwest CMW

Solid tumors account for the vast majority of the over 600,000 deaths annually from all cancers, excluding non-melanoma skin cancers, in the U.S. They are particularly hard to detect and treat, making them one of the leading challenges facing the U.S. health system. The efficacy of treatment can be improved by minimizing invasive surgery and limiting treatment to the affected tissue site(s). However, this remains a major challenge and is the objective of many innovative cancer therapies. Immunolight, a Michigan-based company, has developed an innovative approach that marks a true paradigm shift in cancer treatment. By repurposing psoralen – a well-established drug traditionally used for skin conditions and cutaneous T-cell lymphoma – they are positioning it as a front-line therapy for a broad range of solid tumors. The treatment is delivered by directly injecting psoralen into the target tumor along with energy converters to activate the drug. This intra-tumoral injection is followed by exposure to a low dose of X-ray energy from a standard radiation therapy system. This process of activating psoralen inside a deeply embedded tumor required multiple breakthroughs in chemistry and physics. Immunolight developed the energy converting materials which efficiently absorb X-ray energy and convert it to UV light inside the tumor. Once activated by UV, psoralen binds to DNA and induces apoptotic cell death, which can lead to a pronounced immune response to cause stabilization of disease or partial or complete remission. Critically, this treatment approach has no substantial side effects, unlike chemotherapy.

Fraunhofer USA CMW supported Immunolight in developing a vacuum thin film coating process to deposit diamond-like carbon (DLC) on the surface of the energy conversion material. The coating must be biocompatible and UV transparent in the 320 to 360 nm range to activate psoralen, which was successfully achieved. In general, DLC coatings have been increasingly adopted for biomedical applications such as surgical instruments and implants due to their hardness in combination with low friction. This amorphous form of carbon is well-suited for the physiological environment of mammalian bodies and can prevent bacterial growth, inflammation, and other allergic rejections. This vertically integrated approach to develop a functional bio-interface was very helpful for Immunolight to finalize the development of their innovative cancer treatment and to speed up the qualification process.

Fraunhofer USA Center Mid-Atlantic CMA

Just as Industry 4.0 has transformed manufacturing by embedding digital technologies into every stage of production, so Quality 4.0 has emerged to use smart technologies to improve product quality and manufacturing operations. It integrates real-time, non-destructive process monitoring, edge computing, and advanced analytics, often powered by machine learning (ML) and artificial intelligence (AI), to create a proactive and predictive approach to ensuring product quality. Instead of traditional end-of-line quality inspections, Quality 4.0 shifts the focus to in-line, process-centric quality assurance (QA), leveraging continuous data collection from smart sensors to capture critical process parameters and IIoT devices to monitor manufacturing processes as they unfold. This approach to QA enables manufacturers to detect deviations in products at an early stage in their manufacture, reveal deviations that may not be detectable by standard screening of the final product, deploy AI/ML-based predictive analytics to forecast final product quality and make process adjustments to mitigate arising deviations. The cumulative effect of this approach is to reduce the likelihood of defects and costly rework, more easily identify root causes of product defects, identify high-risk processes requiring intense quality monitoring, and reduce the need for post-production quality assessments. Thus, implementing Quality 4.0 into manufacturing processes leads to major cost savings.

Computer scientists and engineers at Fraunhofer USA CMA have developed ML and AI tools to introduce Quality 4.0 approaches into two welding use cases important in the automotive industry: spot welding and laser welding. Resistance spot welding is widely used to join metal sheets via localized fusion. Fraunhofer USA CMA utilized AI algorithms to analyze large datasets gathered from strategically placed sensors during production, thus identifying patterns, trends and deviations from established norms and predicting final weld strength and bonding quality. By continuously monitoring process parameters in real-time, AI-powered statistical process control (SPC) systems could alert operators to potential quality issues, enabling timely adjustments to prevent defects and optimize production efficiency. This Quality 4.0 approach has been tested with major automotive manufacturers in the U.S. It reduces the cost of manual weld quality verification by up to 55% and can be integrated into existing spot-welding lines with minimal hardware modifications.

In collaboration with their colleagues at the Fraunhofer Institute for Material and Beam Technology IWS in Dresden, Germany, Fraunhofer

USA CMA has also applied Quality 4.0 to high-precision laser welding, which is used across many industries requiring high strength, minimal distortion and fine detail, including the automotive industry. The team deployed a high-speed thermal camera to capture temperature gradients across the weld zone and an acoustic sensor to record emitted ultrasound waves. A system of AI models were developed utilizing different ML architectures that correlated sensor outputs with key weld parameters. This allowed for parameters such as laser power or scanning speed to be tuned on the fly to address detected anomalies, with the number of weld defects reduced by up to 30%. The approaches deployed here by Fraunhofer USA CMA to introduce Quality 4.0 to welding use cases will also be applicable across a very wide variety of other processes and many manufacturing industries.

Fraunhofer USA Center for Manufacturing Innovation CMI

Biological and chemical assays are critical for sample analysis across the biological sciences, including for diagnostic purposes, and to assess the effects of alternative experimental conditions. In many cases, such as when assessing the effect of a therapeutic drug candidate at the whole cell level, the requisite assays require isolated, viable primary cells extracted from target animal tissues. Such whole cell assays are critical in basic research, and also in the development of drugs and biologics through the pre-clinical and clinical pipelines. They are therefore important in academia and in the biotechnology and pharmaceutical industries.

To provide input for these assays, extraction of the viable cells needs to be a finely tuned process, to ensure tissue dissociation without full homogenization, resulting in mass cell rupture. It also needs to be a relatively rapid process in order not to become a major bottleneck in the process flow. With currently available methods, significant time is expended dissociating tissues to extract viable cells, often with the use of enzymes to digest the extracellular matrix. Indeed, common commercially available platforms take about forty-five minutes to process only eight tissue samples. Furthermore, for widespread practical use, the extraction process should be versatile to accommodate multiple tissue types, and fit into industrial process flows with other standard laboratory equipment.

Fraunhofer USA CMI had previously built a tissue dissociator

capable of simultaneously processing twenty-four samples within

two minutes. Under a project conducted for the Cambridge, Massachusetts-based pharmaceutical company, Moderna, the center recently advanced the development of this instrument to expand the range of processed tissues that it could handle and make the instrument compatible with currently available automation modules. The instrument is comprised of an automated version of a mortar and pestle style design, common to standard grinding processes used in laboratories. The automated version allows for rapid grinding down of tissues without the need for time-consuming enzymatic digestion. The mortar function is provided by a standard disposable 24-well plate, allowing seamless integration with existing workflows and liquid handling systems. The pestle function is provided by a custom manufactured array made from autoclavable materials for easy between sample cleaning and resetting of the system. As a further development, the lengths of pestles were

extended to allow for use with both standard wells and deeper wells for higher volume samples.

Highlighting the flexibility of the platform, system testing generated high numbers of viable cells from a wide variety of tissues, including spleens, lymph nodes, cancerous tumors, and even more challenging tissues such as femurs and nasal cavities. Thus, tissues with tough connective material, small tissues, and calcified bones could be thoroughly dissociated using this equipment. Comparisons to commercially available controls demonstrated higher performance ratings for the Fraunhofer USA CMI developed dissociator. Moderna said of the instrument’s performance, “We finally finished analysis from the most recent studies with the Fraunhofer machine. I am pleased to say that the results show they are working great…thank you for all your help. Please also extend many thanks to the whole team! We are very excited to have these machines in house and see how much it will change our lives.”

Fraunhofer USA Center Mid-Atlantic CMA

In the automotive industry, vehicle body production requires from about 3,500 to 14,000 individual resistance welds, known as spot welds, per vehicle to join sheet metal components. These welds must be verified for quality and structural integrity. Current inspection processes rely on static inspection methods, where all welds are manually checked over several shifts using ultrasound. Additionally, weld integrity is periodically verified through destructive testing, providing critical data on weld quality but further adding to the delay between the production of a weld and confirmation of its quality. The overall process flow for the quality checking of spot welds is highly time consuming and labor intensive, making it a focus point for improvements in efficiency in the automotive industry. Advances in machine learning  offer an approach to greatly reduce this inspection effort and the duration of the feedback process to approve welds and thus optimize the entire welding and associated quality assurance endeavor. Engineers at Fraunhofer USA CMA have worked with colleagues at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA in Stuttgart, Germany and Clemson University to address this opportunity with a major automative manufacturer by collecting extensive data directly from the welding equipment and manual verification systems during production. The team trained artificial intelligence (AI) models on historical data to be able to predict weld quality with high precision in real-time, allowing for targeted manual inspections specifically focused on high-risk welds. This approach also allowed for continuous improvement by refining weld parameter settings to enhance their quality. The outcome of this project was to significantly reduce the time and resources required for manual quality checks while maintaining overall production quality. This effort targets at least a 15% reduction in labor hours dedicated to manual inspections and an estimated return on investment for the manufacturer in under a year. The approach undertaken here should also be applicable to optimizing other production processes requiring inspection and validation, in particular other joining technologies used in the automotive and other manufacturing industries

Fraunhofer USA Center for Manufacturing Innovation CMI

Fiber optic gyroscopes are highly precise and accurate rotation sensors. They are used extensively in navigation and guidance systems installed in aircraft, ships, spacecraft and other complex vehicles. Lower performance range applications include turret stabilization systems, in helmet display stabilization and even the first down marker used in National Football League broadcasts. Such gyroscopes can also be used for research in combined quantum and classical navigation systems such as those envisioned for underwater autonomous vehicles. One of the advantages of fiber-optic gyroscopes over other inertial sensors is low noise emittance, which is important for military applications. The performance of these fiber-optic gyroscopes is closely related to the length of the fiber-optic coil and the accuracy of the winding pattern for that coil. Higher accuracy applications, require coils that can be multiple kilometers in length and successful winding of them requires motion control at the micron level, image processing and high-resolution tension control. Current practices require up to two weeks of highly-skilled manual labor to wind such a coil. Fraunhofer USA CMI engineers are developing a computercontrolled machine for coil winding that employs 17 coordinated servo axes that control its moving parts, allowing for micron level placement of the optical fiber in the required pattern configuration and precise tension control. The coils are wound from the inside out, meaning that the supplied length of fiber is divided into two connected spools before the winding process begins. Based on the required winding pattern the system then winds individual layers from each supply spool while fixing the inactive spool on the winding axis to prevent unwinding. Such an automated system with in-process image processing allows for highly accurate coil winding without physical intervention by operators. This enables strategic grade coils and gyroscopes with the key performance features of low cross-talk between signals in adjacent channels and very low thermal drift. Such high accuracy coil winding results in a low coil rejection rate. Furthermore, the automated system reduces labor costs and increases throughput. Fraunhofer USA is deploying this system to wind coils for various customers in multiple industries, including for an Asian aerospace company

Fraunhofer USA Center Midwest CMW

Transition to a hydrogen economy is viewed as a key strategy to reduce emissions of greenhouse gases and to reduce global warming. The use of hydrogen as a fuel requires fuel cells, typically proton-exchange membrane fuel cells, that are electrochemical cells capable of converting chemical energy into electricity through redox reactions. They utilize hydrogen and oxygen to sustain a chemical reaction to produce electricity continuously. Metallic bipolar plates are integral components of fuel cells, where they evenly distribute fuel and oxidant and collect the electric current that is generated. In a fuel stack, they connect individual fuel cells in series, conducting electricity from the anode of one cell to the cathode of the next. Energy efficient fuel cells require high performance bipolar plates that have several key attributes, including adequate fuel flow, fluid impermeability, hydrophobicity, mechanical stability, thermal transmission, and electrical conductance. These attributes define the design of the metallic bipolar plates, including the overall dimensions of the plates, the layout of the flow field, channel width and depth, and the degree of curvature of channels. Thus, precise channel geometries and tight manufacturing tolerances are essential for fuel cell efficiency. Maximum surface area at minimum weight requires complex plate designs and use of thin metals, typically less than 0.1 mm thick, which must be formed without tearing. High-volume production thereof is challenging. Engineers at Fraunhofer USA CMW have worked with colleagues at Clemson University and the Fraunhofer Institute for Production Technology IPT in Aachen, Germany to optimize approaches for the design and manufacturing of metallic bipolar plates for protonexchange membrane fuel cells. Firstly, flow field and channel geometries were optimized regarding fuel distribution, tooling capabilities and press capacities. Following the forming, processes and fixturing tools were developed for precision laser cutting of external and internal plate features and uniform laser joining of anode and cathode half plates. Assembled plates were coated by physical vapor deposition techniques. The coatings met key U.S. Department of Energy targets for corrosion resistance and interfacial contact resistance. Finally, gaskets were applied by screen printing around the outer periphery and media ports to complete bipolar plate demonstrators. These metallic bipolar plates were developed under a project supported by the South Carolina Department of Commerce and a global manufacturer of automotive components. The plates have applications for a wide range of vehicles and industrial equipment

Fraunhofer USA Center Midwest CMW

There is a significant and urgent need for new materials for a wide range of industrial and medical applications. These are driven by demands for improved sensitivity and specificity among other criteria. To address these, recent research focuses on the development of novel nanoparticles which can overcome limitations with current materials and enhance the precision of surface interactions. Diamond nanoparticles (DNPs) have been demonstrated as an effective material for a wide variety of applications, including bioimaging, drug delivery, and cancer therapies. To realize the commercialization of DNPs for these applications it is necessary to develop synthesis routes that result in particles with well controlled properties and characteristics. Many of the material properties and characteristics of DNPs are dependent on their size, shape, and surface morphology. Currently, the most common methods to produce DNPs are through detonation and ball-milling. These top down approaches,

while successful at producing large quantities of particles, often

feature either broad size or shape distributions, limitations on selection of particle size, and lack of control over composition. Radio frequency (RF) plasma synthesis offers an attractive alternative to these top down approaches since continuous flow-through reactors can be used to controllably produce intrinsic and doped nanoparticles with functionalized surfaces and narrow size distributions.

Under a project funded by the US Department of Energy and in collaboration with colleagues at Michigan State University, Fraunhofer USA CMW is developing a path to synthesize DNPs directly out of the gas phase utilizing a RF plasma process. To date, there has been difficulty in utilizing RF plasma synthesis to generate DNPs. This is in part due to a lack of fundamental knowledge related to the growth mechanisms of nanoparticles in low temperature plasmas. While many mechanisms have been proposed, there is a dearth of systematic studies that integrate experimental data into predictive modeling or simulation. Studies are required that will collect information about the plasma characteristics, gas species, and synthesized material to build predictive models for nanoparticle synthesis. Direct evidence for DNP nucleation and growth directly from the gas phase without the influence of a substrate is fairly limited. Furthermore, information related to the nucleation and growth dynamics of particles in plasma could have important implications for conventional diamond film growth.

The objectives of Fraunhofer USA CMW’s project are to measure various parameters, including the electron and ion energy distribution functions, plasma density, gas temperature, and gas species during DNP synthesis by the RF plasma process, and to map these conditions onto the characteristics of the synthesized nanoparticles, thus identifying process conditions suitable for DNP synthesis. The reactor being used for this work is composed of a quartz tube with external ring electrodes that capacitively couple RF power from a power supply to the plasma. Magnetic coils located externally to the quartz tube are placed in a Helmholtz configuration to provide a uniform magnetic field in the active region of the reactor. Three independent sets of coils, each aligned along a principal axis, generate a magnetic field with controllable strength along each principal axis. A mixture of hydrogen, methane, and argon gases are used for the growth of the DNPs and pressure is monitored and regulated. Particles are collected on substrates by acceleration of their flow through a nozzle where they are deposited via a process of inertial impaction. Initial results have shown that the plasma characteristics can be shifted with the strength of the magnetic field produced by the coils, offering promise for the development of a DNP synthesis pathway. .

Fraunhofer USA Center for Manufacturing Innovation CMI

Millions of existing homes in the northeast of the United States have little or no wall insulation and current technologies to improve insulation either have severe limitations or are cost-prohibitive at large scale. Drill-and-fill wall insulation only increases whole wall thermal insulation to an R-value of about 11 and cannot be used in masonry structures lacking a wall cavity or with no or challenging access to such a cavity. Interior insulation systems are disruptive and invasive, consume valuable indoor floor space and create hygrothermal risk during warmer months. Adding continuous insulation to outside walls is preferable to the above solutions but is currently very costly, since insulating boards need to be custom cut on site by skilled labor and followed up with the application of a stucco finish coat.

Under a project funded by the New York State Energy Research and Development Authority, Fraunhofer USA CMI developed an insulated architectural masonry-like panel block, that when used to clad existing smaller masonry and wood-frame residential buildings, adds continuous exterior insulation to reach a thermal insulation R-value of at least 20 in an appealing and cost-effective manner. The blocks are pre-fabricated using digital processes to analyze terrestrial scans of buildings, allowing for accurate building dimensions to be obtained, an appropriate set of blocks to be designed and fabricated, and an augmented reality experience to be created to guide the installation of the blocks onto the building. This process eliminates on-site cutting of insulation, cladding and trim, and facilitates on-the-job guidance of labor, thus significantly reducing installation costs.

Through this project, Fraunhofer USA CMI identified compression co-molding of the concrete-based manufactured stone shell and foam backing as the most promising process to manufacture the insulated panel blocks. The center then developed a conceptual production molding process and production line for high-throughput manufacturing at minimal possible cost. Cost modeling indicates that the masonry-like panel blocks should be cost-competitive for energy efficient building

retrofits. In the course of completing this project, Fraunhofer USA CMI also demonstrated that their digital processes to analyze scans of wooden buildings could be extended to masonry buildings. The project also allowed for market research, whereby the center identified key customers for the technology among homeowners and contractors and identified additional preferred design attributes, comprising high aesthetic quality, low-visibility seams, and surface durability.

Fraunhofer USA Center Mid-Atlantic CMA 

Team communication and coordination is of critical importance for intelligence gathering by the military and government security agencies, with weaknesses in gathering and processing information often associated with shift handovers, resulting in team cognition challenges. These challenges include inaccuracy blindness, group sharing and storing of knowledge, known as transactive memory systems, and shared mental models. Artificial intelligence (AI) has often been proposed as a possible solution to these problems, since, for example, AI can support teaming by augmenting individuals’ production capabilities, summarize machine read documents and convert them to summary output text, and organize intelligence analysis around entities, such as people and places, rather than freeform text. However, it is not clear how to best align rapidly developing AI technologies with intelligence analysis work. Engineers at Fraunhofer USA CMA have worked on a project with colleagues at the University of Maryland and Duquesne University for the United States Army Research Office to assess the application of AI and machine language analysis to mitigate team communication and coordination problems such as information overload, ignoring potentially relevant data and erosion of trust between team members. The goal of the project was to provide much needed insight into how human teams can work together with AI, especially AI that provides sensemaking support, to improve outcomes in intelligence analysis and avoid exacerbating team interactions. Based on insights gathered from interviews with intelligence analysts, the team developed a software platform and an experimental infrastructure testbed to experimentally study the role of different types of AI during intelligence analyst shift handovers. They also conducted controlled immersive behavioral experiments to test the effect of AI manipulations on sensemaking, problem solving, workload, and transactive memory systems. The testbed consisted of task-relevant input materials, such as mission descriptions and source documents, simulated team members, activity recording tools, such as search tools and scratchpads, experimental monitoring capabilities, such as recording and survey systems, and AI support tools for human analysts, such as AI that can summarize large quantities of information by, for example, constructing topic models. The experiments simulated the 5Vs challenges associated with big data: a high volume of material, a wide variety of material sources, a rapid velocity of information accrual, questionable veracity of some sources, and extractable value being dependent on linking information from multiple sources. The testbed was most recently applied to analyze interactive shifthandovers, comparing relatively simple AI tools with an entity-based AI drawing on developments with ChatGPT and theories on information science and intelligence analysis. The approach shows great promise for assessing AI tools being applied with the goal of improving the efficacy of intelligence analysis.

Fraunhofer USA Center Midwest CMW

There is an increasing need for higher power devices that are more efficient, reduce strain on the electrical infrastructure, and are required for specialized industrial systems. To address this need, new high-power electronic devices are under development for application in a wide range of systems, including telecommunications, motor drives, power grids, electric vehicles, and industrial and locomotive traction control. These new devices require ultra-wide bandgap semiconductors with performance characteristics that are beyond the capabilities of the silicon carbide and gallium nitride materials used in electronic devices today. Various materials, including aluminum gallium nitride alloys, boron nitride, and diamond, are being tested for use in these new semiconductors, as are composites, combining multiple materials, where each material offers unique and necessary performance characteristics to the semiconductor. Engineers at Fraunhofer USA CMW have worked with colleagues at the Fraunhofer Institute for Microstructure of Materials and Systems IMWS in Halle, Germany to develop a technique to create single crystalline diamond nanomembranes that can be integrated with other semiconductor materials to create diodes with improved performance (see schematic figures A and B). The diamond nanomembranes are exfoliated from (i) bulk material via (ii) an ion implantation and (iii) separation process. These nanomembranes can then be transferred and further processed via a commercialized micro-transfer printing process (iv). Early prototypes showed an over ten-fold higher increased forward current density compared to traditional all-diamond diodes. Integration of diamond into these heterogeneous diodes also allowed for improved thermal management. Furthermore, the team demonstrated that chemical-mechanical polishing lowered surface roughness on the diamond nanomembranes, a prerequisite for including them in stacked devices such as thyristors, which are robust and operate like switchable diodes that are commonly used in high-voltage applications to control electric power flow. Under a program funded by the Defense Advanced Research Project Agency (DARPA), the Fraunhofer USA CMW engineers are now working with an industry partner to incorporate these diamond nanomembranes with other ultra-wide bandgap semiconductors that will allow devices to operate at higher voltages, frequencies and temperatures than what can currently be achieved with traditional materials, such as silicon carbide and gallium nitride. Access to such devices will be critical for a transition from a centralized to a distributed power grid. Interfacing renewables, such as solar and wind, as well as battery storage, with the grid at large requires the use of these devices

Fraunhofer USA Center for Manufacturing Innovation CMI

Accurate and rapid detection of specific biomolecules is vital in biomedical research and for the development of effective diagnostics. Current methods for rapid biomarker detection include testing for protein antigens and antibodies using lateral flow assays and testing for DNA and RNA using nucleic acid amplification techniques. The former can be rapidly performed at the point-of-sampling but have relatively high error rates, low sensitivity and poor selectivity. By contrast, the latter have relatively high accuracy, selectivity and sensitivity but need to be conducted by trained technicians in a laboratory and are relatively expensive and time consuming. Thus, there is a market need for diagnostics for a range of biomarkers that are both rapid to perform and inexpensive but that also have a high degree of accuracy and sensitivity and that are highly selective. Scientists at Fraunhofer USA CMW have worked with colleagues at the Fraunhofer Institute for Reliability and Microintegration IZM in Berlin, Germany and the Fraunhofer Institute for Cell Therapy and Immunology IZI in Leipzig, Germany to develop an alternative biosensor technology based on electrochemical detection of proteins binding to a functionalized electrode surface. For initial development of this tool, the team focused on detecting antibodies to the major spike protein antigen of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which is indicative of viral infection, and to the human c-MYC protein, which is indicative of cancer. The team developed a sensor in which a boron-doped diamond surface acting as an electrode was functionalized by click chemistry with known peptide targets of the relevant antibody. Since this mode of functionalizing the sensor surface is modular, it can be tailored to detect any antibody and thus has applications across a wide range of diseases. Differential pulse voltammetry was deployed to measure antibodies in a human serum sample binding to the peptides on the diamond electrode surface. This resulted in a highly selective and quantitative signal. The diamond-based sensor was packaged with a standard microfluidic system to allow for sample flow over its surface, with electrical connections to the electrode made through the back side of the resulting device. The device plugs into a standard cartridge to allow fluid flow through a syringe pump and provides connection to a potentiostat to record the output signal. This packaging scheme allows for a portable and disposable sensor that could be mass produced and read using standard recording equipment at the point-of-sampling. This detection platform is now being further developed with multiple microfluidic channels to allow for internal controls and the analysis of multiple variables within samples. It is also being further validated with various human samples

Fraunhofer USA Center Midwest CMW, Fraunhofer Institute for Manufacturing Technology and Advanced Materials IFAM (Dresden)

The trend toward electronics miniaturization and higher-efficiency thermal heat exchangers presents major challenges in electronic packaging, cooling, and temperature control. As a result, heat flux directly impacts both consumer electronics and the manufacturing tools used to fabricate these devices, driving industrial-scale challenges. On average, global heat flux demands have risen by 7.8% per year over the last decade. This increase is driven by advanced microelectronics applications such as artificial intelligence, and demands are projected to continue growing, with the global thermal management market expected to reach $12-14 billion by 2030.

One leading solution to manage increased heat flux and mitigate damage to electronic components is the integration of advanced heat exchanger materials and custom-shaped structures that cannot be produced through traditional subtractive manufacturing. By leveraging additive manufacturing and combining advanced materials, such as copper–diamond composites or layered copper–ceramic heat spreaders, custom thermal transfer devices can provide improved heat flux management. This reduces thermal mismatches, minimizes assembly-related heat loss, and ultimately extends device lifetimes.

To address these needs and advance multi-material integration of

copper, ceramics, and copper–diamond composites, scientists at

Fraunhofer USA CMW, in collaboration with colleagues at the Fraunhofer Institute for Manufacturing Technology and Advanced Materials IFAM, have developed novel additive manufacturing

approaches. Two distinct processing competencies have been

established to create complex copper and ceramic structures:

Gel Casting: Metal or ceramic powders are cast using a gelatin binder into 3D-printed molds. After post-processing and sintering, this approach yields dense, complex geometries and multi-material stacks composed of copper, alumina, and copper–diamond. Based on the mold structure a variety of complex structures can be fabricated.

Extrusion-Based 3D Printing: A polymer-based feedstock has been developed to suspend metal or ceramic powders within a printable plastic matrix. Using traditional extrusion, complex multilayered structures were fabricated simultaneously. Following debinding and sintering, this method also produced dense copper, alumina, and copper–diamond components of various geometries that could be used for thermal transfer interfaces.

While presenting these results at a conference, our team initiated a collaboration with New Aim3D GmbH, an additive manufacturing company based in Rostock, Germany. This partnership enabled Fraunhofer USA CMW to become the first U.S. customer for their ExAM 255 industrial pellet-based 3D printer and positioned our facility in East Lansing, Michigan, as a demonstration site for interested manufacturers. The ExAM 255 offers production-ready scalability, allowing our research to transition from technologically challenging and expensive filament-based printing to industrially viable pellet-based additive manufacturing of metals and ceramics.

Together, our team has demonstrated novel heat spreaders with broad application potential, enabling our partners to mitigate heat, reduce costs, and unlock new design freedom through additive manufacturing.

Fraunhofer USA Center for Manufacturing Innovation CMI , Fraunhofer Institute for Production Technology IPT

Diseases caused by infectious bacteria and fungi are a highly significant and growing threat. Despite the development of broad spectrum antibiotics over the last century, reduced efforts and successes in developing new antibiotics and increasing ineffectiveness of licensed antibiotics to stem the disease burden are creating a crisis in global health. Indeed, the World Health Organization estimates that antimicrobial resistance and the related phenomena of tolerance and resilience are directly responsible for well over one million deaths and contribute towards an additional five million deaths annually worldwide. Although important in stemming disease, licensed antibiotics are increasingly failing as a major treatment option across the world, with treatment failure primarily caused by two major phenomena, resistance and tolerance, of which tolerance is largely ignored. Tolerance refers to a situation under which bacteria functionally slow down or become dormant in the presence of an antibiotic used for treatment until that antibiotic dissipates to levels that no longer affect the bacteria, after which they revert to their active state and reinfect the patient. Unlike antibiotic resistance, antibiotic tolerance is not screened

for in clinical laboratories and current assays used in research

laboratories are labor intensive, have low throughput and can give highly-variable results between test sites.

To address this issue and develop a technology that can be used to detect antibiotic tolerance in high throughput clinical settings, scientists at Fraunhofer USA CMI have been working with their colleagues at the Fraunhofer Institute for Production Technology IPT in Aachen, Germany to develop an automated high-speed light microscope-based instrument to measure antibiotic tolerance in a platform that is agnostic to the actual antibiotic used. Bacteria from clinical isolates under investigation are grown on media in microwells that allow for phase contrast imaging. Then, image analysis algorithms, trained on large datasets of bacteria, are used to detect tolerance. The bacteria used for training purposes are from clinical samples obtained from a medical school and hospitals in the Boston area. The training datasets comprise images of these bacteria grown in the presence or absence of antibiotics and focus on lag periods and growth rates. The project is currently midway through its anticipated timeframe and is showing promising results. It is anticipated that an automated high throughput system to address antibiotic tolerance will be highly attractive to the clinical microbiology market.

Fraunhofer USA Center Mid-Atlantic CMA, Fraunhofer Institute for Telecommunications, Heinrich-Hertz-Institut HHI

Modern medical optical instruments such as microscopes are becoming increasingly digital, but the data that is generated is not yet fully utilized by the emerging technologies of artificial intelligence (AI), machine learning (ML) and augmented reality (AR). For instance, AI/ML models can rapidly detect patterns in sensor data, such as malignant skin moles observed by microscopy, and digital sensors can exploit wavelengths beyond the visible spectrum to extract information that is invisible to the human eye. This AI-inferred and hyperspectral information can be fully utilized and presented to human users in real-time with AR. For this purpose, appropriate visualization is crucial to make different tissue types distinguishable, such as to differentiate between pathological and at-risk tissue, without overlapping other essential image information. Thus, there is large untapped potential in leveraging novel AI and AR tools for real-time analysis of digital microscopy data, which can assist physicians by providing supplementary information for diagnosis and surgery.

Although the latest AI/ML methods are very powerful and can make highly accurate predictions, they are often opaque. This lack of transparency and interpretability reduces their value to physicians. Furthermore, the AI/ML models can be misled during their

34

training by confounding factors in the training data that are irrelevant to the actual diagnostic or supportive task to be performed by the physician. Additionally, these models are often insufficiently robust, since their performance can decline when given inputs that differ slightly from the training data, such as differences in the configuration during image acquisition. These flaws can lead to incorrect diagnoses and ultimately to a lack of confidence in diagnoses made using AI. Because of the opacity of the AI/ML models that are difficult to interpret, it has been challenging to identify and rule out such flaws.

Computer scientists at Fraunhofer USA CMA have been working with their colleagues at the Fraunhofer Institute for Telecommunications, Heinrich-Hertz-Institut, HHI in Berlin to develop a modular toolbox to integrate state-of-the-art AI models for the classification of microscopic and close-up images with robustness evaluation and improvement methodology along with concept-based explainable AI techniques, thus providing transparent, interpretable and robust AI/AR tools that enhance the efficacy of diagnostics and surgical procedures. Specifically, innovative methods for AI-based image analysis and synthetic image generation, real-time AR assistance, and explainable AI tailored for medical applications, have been developed in-house. Key outcomes for the project include advanced AI models for tissue classification based on multispectral data, data augmentation techniques to enhance AI/ML model training and real-time AR visualization for surgical assistance. Together, these comprise a suite of capabilities to support trustworthy, safe and robust AI in the medical domain.

Fraunhofer USA Center Mid-Atlantic CMA

There are several challenges with laser welding of metal materials and laser cutting of thick metal materials. These include handling materials of diverse thicknesses and qualities, achieving the required detailing and precision in the product, meeting efficiency and time constraint targets, and limiting material wastage. New technologies are being developed to overcome these challenges, including new laser sources with increased power and tailored attributes, high-frequency power modulation of the laser beam, high-frequency oscillation of the laser beam and the focal plane of the laser, and plasma keyhole welding, which allows for the welding of high-alloy and unalloyed materials in a single pass of the laser. However, these new technologies are of high complexity, involving interacting non-linear effects, resulting in unstable processes when they are combined. Therefore, to incorporate these new technologies into a reliable operation requires comprehensive monitoring and control, best achieved through real-time monitoring and computer control guided by artificial intelligence (AI) to recognize deviations from the desired quality attributes and issue commands to promptly rectify deviations and maintain process stability. Engineers at Fraunhofer USA CMA have worked with colleagues at the Fraunhofer Institute for Material and Beam Technology IWS in Dresden, Germany and the Fraunhofer Institute for Applied Optics and Precision Engineering IOF in Jena, Germany to develop controlled optimized laser material processing assisted by AI. The team implemented multimodal process monitoring equipment that provided input data for AI-based process evaluation that then allowed for AI-based closed-loop feedback control. This AI-based solution reduced energy consumption, increased processing speed, improved weld strength, and reduced distortion. Through this project, the team achieved the seamless integration of diverse monitoring technologies, including input from high-speed cameras and microphones, and ultra-fast data processing to provide real-time process control, allowing for dynamic adjustment of laser cutting and welding parameters during operation. The integrated system targets up to 30% higher speeds and up to 40% lower energy use for laser cutting and welding, while it also reduces the risk of cutting and welding failures. It is suitable for a range of materials, including the cutting of thick metal sheets and the welding together of different materials. The built systems are adaptable and scalable to various industrial needs and have broad application potential.

Fraunhofer USA Center for Manufacturing Innovation CMI

Meat, mainly beef, pork, mutton, lamb and poultry, is consumed by humans worldwide and contributes a large proportion of the calorific intake in many countries. This is rapidly increasing as the middle class swells in nations with rapidly rising gross domestic product (GDP). Indeed, the average annual per capita consumption of meat worldwide is more than 40 kg, but this rises to 120 kg in the relatively affluent U.S.. The environmental consequences of such rising meat consumption, both in terms of land usage and greenhouse gas emissions, are considerable and contribute significantly to global concerns over limited availability of arable land and climate change. The generation of antibiotic resistant microbes is a further hazard of the large-scale cultivation of animals for meat. However, cultural drivers make it highly likely that meat consumption will continue to be viewed as aspirational by many human populations for decades to come. One leading potential solution to the land use and environmental consequences of meat consumption is to develop and use animal cell culture techniques to culture meat in bioreactors. However, the economical production of high-quality meat in bioreactors faces several hurdles, including the development of high yielding cell lines, the mass propagation of these cell lines, the fabrication of appropriate scaffolds upon which the cells can adhere, grow and divide to generate meatlike tissue, and appropriate bioreactor design for large-scale culturing, including allowing for flexing of the growing tissue to simulate the exercising of muscle in an animal. Scientists at Fraunhofer USA CMI have worked with colleagues at the Fraunhofer Institute for Molecular Biology and Applied Ecology IME in Aachen, Germany to address the issues of a lack of suitable scaffolds and bioreactors for cultured meat production by developing a wet spinner, a circular braider and a tension bioreactor. This has allowed for the wet spinning of plant based polymers to tensile strengths compatible with cultured meat production. The team also developed porous sponges for compression-based scaffolds and showed them to have compression strengths compatible with the culturing of meat products. In addition, muscle satellite cells were isolated from samples biopsied from cattle. These cells were shown to be able to attach to the fiber scaffolds, where they were induced to form myotubes, a developmental precursor of muscle. The cells could also attach to the sponge-like scaffolds. The team is also optimizing the growth medium for muscle satellite cell cultivation and differentiation. To avoid the use of fetal bovine serum, which is expensive and of animal origin, the necessary growth factors are being produced in a cell-free plant expression system. Together, these developments are highly promising for advancing the cultured meat industry.

Fraunhofer USA Center Midwest CMW

As dramatically demonstrated by the coronavirus disease 2019 (COVID-19) pandemic, current diagnostic assays are ill-equipped to deal with the rapid global spread of emerging and reemerging pathogens. These quantitative assays are incompatible with rapid point-of-infection testing. Rather they require biological samples to be collected and then transported to a testing laboratory for the isolation of nucleic acids and the performance of amplification tests to identify the presence and abundance of a specific pathogen. By contrast, crude antigen lateral flow, or strip, tests can be completed at the point of infection but are not quantitative and have higher error rates. Thus, there is considerable demand for accurate nucleic acid-based assays that can be performed at the point-of-infection. Scientists at Fraunhofer USA CMI have worked with colleagues at Boston University Medical Center, the Fraunhofer Institute for Production Technology IPT in Aachen, Germany and the Fraunhofer Institute for Interfacial Engineering and Biotechnology IGB in Stuttgart, Germany to incorporate printed hydrogels into fluorescent arrays. These arrays combine the accuracy of nucleic acid amplification tests with the simplicity, speed and convenience of lateral flow assays and are thus suitable for rapid, point-of-infection tests to detect viral pathogens. This platform has the added advantages of long-term stable storage of thermolabile reagents within the hydrogel matrix and multiplexing of outputs from multiple pathogens. The sensors detect the presence of specific pathogens in the samples through biological assays based on DNA amplification. Current development with Fraunhofer USA CMW focuses on integration of hydrogels with boron-doped diamond microelectrode arrays. The outcomes of these assays are thus converted into electrochemical signals. The research team initially developed this technology for the detection of four respiratory disease pathogens, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus A, influenza virus B and human rhinovirus. In addition to developing the relevant assays and the hydrogel matrix in which the assay components are encased, the team developed manufacturing processes for the printing of these hydrogel embedded arrays, including defining surface modifications of manufacturing-compatible plastics. Furthermore, the team designed a low-cost, portable reader for the assays and developed algorithms to process the output images to allow for automated analysis. Under a recently awarded grant with the National Institutes of Health (NIH), the research team is now extending this approach to blood borne pathogens, specifically to develop a diagnostic assay to detect human immunodeficiency virus (HIV), the causative agent of acquired immunodeficiency syndrome.

Fraunhofer USA engages with Fraunhofer Institutes primarily through joint projects, through prime and subcontractor project relationships, and through permitted scientific and technology know-how exchange and personnel exchange. These projects help ensure that the applied science and technology being developed that is not subject to restrictions can be translated into products and innovation that benefit society. 

Fraunhofer USA continues to jointly represent the Fraunhofer-Gesellschaft network in the U.S. and has rapidly proven its institutional value as an expert in the U.S. market. Our joint program for internal strategic pre-competitive research continues to prove beneficial for transatlantic knowledge and technology transfer. This program facilitates our readiness to deploy resources across the Atlantic where and when needed. Our continued outreach with Fraunhofer institutes has led to Fraunhofer USA working with about 50 of the 75+ research institutes in Germany at a time, to conduct joint project development activities in the U.S., engage in joint pre-competitive strategic technology development projects, executing on U.S.-industry projects or publicly funded projects.