Projects

Dr.-Ing. Christoph Weigel, Microsystems Technology, Department of Mechanical Engineering
Dr.-Ing. Ulrike Brokmann, Inorganic Non-Metallic Materials, Department of Mechanical Engineering

Glasses represent high-tech materials that are of great current and future importance for microelectronics and innovative computer architectures. However, the required microstructuring of glass by dry etching is a relatively time and energy-demanding process that involves aggressive chemicals.

This project focuses therefore on efficient reactive ion etching (RIE) processes with nanometer-range precision that offer an improved environmental compatibility, which is a key demand for future microelectronic fabrication in diverse fields. It is a project goal to contribute to the gradual substitution of PFAS-containing gases (CF4, C4F8, CHF3 etc.) hand in hand with the reduction of global warming potential (GWP) for plasma structuring of fused silica bulk materials as well as SiO2 thin films by combined alternative chemical and physical attack.  

The project is therefore directly related to other "green electronics" projects that are concerned with the microfabrication of electrical devices and functional structures, which enables a lively interdisciplinary exchange. The project itself covers scientific aspects of materials sciences, glass chemistry, plasma-etching, and microtechnology. 

More information about „Go gRIEn – Green Reactive Ion Etching of Silicon Oxides for Microelectronic Application“

Jun.-Prof. Dr. rer. nat. Robert Geitner, Physical Chemistry / Catalysis, Department of Mathematics and Natural Sciences
Jun.-Prof. Dr. rer. nat. Christian Dreßler, Theoretical Solid State Physics, Department of Mathematics and Natural Sciences

Conventional microelectronic components are based on semiconductors made from metals and silicon. These raw materials are extracted from nature in an energy-intensive and unsustainable manner and therefore cannot be a building block for a low-energy circular economy and green electronics. Electrically conductive polymers made from purely organic elements such as carbon, hydrogen, oxygen and/or nitrogen are available as an alternative.

One current problem with these organic conductors is the loss of electrical conductivity over time. This is caused by the oxidation of the carbon skeleton by atmospheric oxygen, which leads to the formation of hydroxyl and carbonyl groups that interrupt the conjugated system of the conductive polymer. The conjugation is responsible for electron conduction and is therefore decisive for macroscopic conductivity.

As a solution to the problem, it is the goal of the project to develop bio-inspired, self-healing, electrically conductive polymers. These intelligent materials are able to heal molecular damages autonomously, i.e. without external stimulus, so that the service life of the green conductors is sustainably improved.

The project thus contributes to green information technology by substituting critical and energy-intensive materials with intelligent, bio-inspired, green materials and improving their durability so that the material system can be tested for green microelectronics.

More information about "Bio-Inspired, Self-Healing Polymers for Applications in Green Electronics"

Dr.-Ing. Benjamin Spetzler, Micro- and Nanoelectronic Systems, Department of Electrical Engineering and Information Technology
Prof. Dr.-Ing. Patrick Mäder, Data-intensive Systems and Visualization, Department of Computer Science and Automation

This project aims to investigate a novel hybrid-AI method to tailor general-purpose neuromorphic hardware based on two-dimensional MoS₂ memtransistors (next-generation computing). The method builds upon the synergy between materials science and computer science to integrate numerical charge-transport simulations with physics-informed artificial intelligence and measured data. The envisioned modeling framework will be able to quickly perform multi-objective optimization tasks to tailor the system design considering constraints (e.g., of the fabrication processes) and the requirements derived from the learning rules.

If you are eager to pioneer the emerging field of hybrid-AI modeling and set new benchmarks for energy-efficient computing, this project will position you at the forefront of future microelectronics and AI research. It offers an exceptional opportunity to develop robust interdisciplinary expertise, combining modern device technology, computational physics, and cutting-edge machine learning techniques.

More information about "Hybrid-AI Modeling for Biorealistic Neuromorphic Computing"

Prof. Dr.-Ing. Steffen Strehle, Microsystems Engineering, Department of Mechanical Engineering
Prof. Dr.-Ing. Hannes Töpfer, Theoretische Elektrotechnik, Department of Electrical Engineering and Information Technology

Deep Learning is considered one of the areas of innovation, albeit requiring frequently significant amounts of energy for the training of neural networks and the infrastructure operation. Reservoir Computing (RC) offers an interesting alternative for some data domains, but a suitable reservoir is in this case required, enabling a nonlinear transformation of input data.

The project aims to design, model, manufacture, and investigate such RC reservoirs that are based on nonlinear micro-electro-mechanical system (MEMS) oscillators. MEMS oscillators are already established in general and can thus be reconfigured and adapted to fulfill various tasks and requirements. The project encompasses interdisciplinary focuses: energy-efficient computing, bioinspired microelectronics, as well as materials, devices, and technologies, combining modeling and experimental investigation of MEMS-based RC. The required electronic and mechanical simulation tools are available, as well as a comprehensive range of microtechnologies. The project effectively complements also other “green electronics” projects in the RC and memristor fields, as well as those projects dealing with "greener microtechnology." 

A doctoral candidate with skills in both, theoretical and experimental areas in the field of electrical engineering fundamentals/signal processing and microtechnology, e.g., studying electrical engineering, mechatronics, physics (engineering) or microsystems technology, is required. The qualification includes intensive supervision, regular participation in conferences, publication in reputable journals, and participation in training for scientific work and career planning.

Prof. Dr. rer. nat. habil. Kathy Lüdge, Theoretical Physics II, Department of Mathematics and Natural Sciences
Dr. Lina Jaurigue, Theoretical Physics II, Department of Mathematics and Natural Sciences
Prof. Dr. rer. nat. habil. Stefan Sinzinger, Optical Engineering, Department of Mechanical Engineering

The project aims for the investigation of the performance of a reservoir computing system in its dependence on the architecture of the reservoir and the experimental implementation as an optical spatio-temporal recurrent neural network.  An existing experimental setup for a reservoir computing system based on a spatial light modulator in an electronics feedback loop serves as a starting point. The systems will be adapted according the requirements using  optical systems design and three-dimensional beam shaping. The dependence of the performance on the optimised reservoir structure is investigated using benchmark experiments and evaluated in terms of energy efficiency as an example of green electronics.

More information about "Optimized Optical Reservoir Computing"

Dr. rer. nat. Jialan Cao-Riehmer, Physical Chemistry / Microreaction Technology, Department of Mathematics and Natural Sciences
Prof. Dr. rer. nat. habil. Michael Köhler, Physical Chemistry / Microreaction Technology, Department of Mathematics and Natural Sciences
Prof. Dr. rer. nat. habil. Martin Ziegler, Micro- und Nanoelectronic Systems, Department of Electrical Engineering and Information Technology

The transition from living to nonliving matter requires a constant exchange of energy, matter, and information between a system and its surroundings, forming the foundation of life.

The aim of this project is therefore to integrate metabolism into microelectronic networks, enhancing neuromorphic systems with increased conductivity, energy efficiency, adaptability, and responsiveness to the environment. By utilizing electroactive microorganisms, initially through fixed variable resistances and potentially transitioning to memristive components, active bioelectronic networks capable of cathodic and anodic activity shall be created. This allows for electrical stimulation and exploration of intricate emergent dynamics and logical operations.

The project centers on developing an experimental substrate platform, characterizing the electrical and bioelectronic properties of single cells, and analyzing network structures for potential applications in bioelectronic signal processing and communication systems.

It is an exciting opportunity to contribute to cutting-edge research at the intersection of biology and engineering for anyone passionate about advancing knowledge in miniaturized biotechnology and ready to tackle interdisciplinary challenges.

More information about "Metabolism-based Bioelectronic Networks"

Dr. rer. nat. Peggy Reich, Biotechnical Micro- and Nanosystems in Life Sciences, Department of Mathematics and Natural Sciences
Dr. rer. nat. Jörg Pezoldt, Nanotechnology, Department of Electrical Engineering and Information Technology
Dr.-Ing. Heike Bartsch, Electronics Technology, Department of Electrical Engineering and Information Technology

DNA computing is an emerging branch with the potential to solve our problem of high energy consumption in artificial intelligence-based computation. It uses DNA for data storage and information processing, which is biodegradable, renewable, sustainable, and reparable. Furthermore, DNA has memristive properties and a memristive crossbar array features colocation of memory and processing units and is scalable, making it a promising candidate for accelerating machine learning and neuromorphic computing.[i]

In this project, the aim is to develop DNA-based memristive devices through self-organizing processes using contact materials in the form of an electrode array. The addition of short DNA molecules and their self-organized sequence-dependent attachment selectively alters the memristive properties of the DNA. It will be evaluated to which extent these intelligent components are suitable for storing and processing information and for the development of self-learning systems using the example of image recognition.

Jun.-Prof. Dr.-Ing. Patrique Fiedler, Data Analysis in Life Sciences, Department of Computer Science and Automation
Jun-Prof. Hongye Sun, Functional Materials, Department of Electrical Engineering and Information Technology

Multimodal on-skin monitoring of physiological signals during everyday life opens a new era for health care, human–machine interfaces, robotics and prostheses. The foundational elements for on-skin multimodal sensing encompass mechanically stretchable and skin-conformable sensor networks.The project aims to investigate a novel solution based on thermal expandable microspheres in Ag-PDMS to fabricate intelligent stretchable conductors exhibiting tunable electrical conductivities and surface morphologies. Notably, this material and production allows for the selective integration of sensors and conductive tracks at distinct positions into a flexible polymer patch, reducing material waste, ensuring durability and reusability, thus presenting an economical approach to the development of printed on-skin electronics. Moreover, the multimodal approach reduces the overall number of sensors and thus further reduces the technological footprint. Consequently, the proposed novel multimodal sensors will exhibit usability, durability and producibility characteristics superior to existing sensor patches with considerably improved material efficiency during both production and application.

This position offers an exciting opportunity to contribute to innovative and interdisciplinary research involving engineering, computer science and medicine. If you are passionate to advance your knowledge on materials and biomedical engineering within an interdisciplinary and international graduate school, we encourage you to apply.

More information about "Highly Stretchable and Conformable Conductors for On-skin Multimodal Sensing"

Dr. rer. nat. Tzvetan Ivanov, Micro- and Nanoelectronic Systems, Department of Electrical Engineering and Information Technology
Dr.-Ing. Stephan Werner, Electronic Media Technology, Department of Electrical Engineering and Information Technology

How can neuromorphic microsensor technology be used to implement an energy-efficient, robust, and data-saving audio system for detecting, recognizing, and evaluating acoustic anomalies in industrial environments?

To answer this question, investigations on bio-inspired acoustic scene analysis and neuromorphic circuits for audio sensing have to be conducted in a highly interdisciplinary way.

Some of the key aspects of the project will therefore inlude:

(i) Detecting acoustic events and changes in the early stages of the signal processing chain

(ii) Localizing the source of this sound event in a complex and changing acoustic environment

(iii) Forwarding only the associated and relevant data for signal analysis

(iv) Adapting the signal processing and tuning the sensor properties as efficiently as possible to react to changes in the environment or the emitted sounds

More information about "Neuromorphic Acoustic Detection and Recognition"

Prof. Dr. Hannes Töpfer, Advanced Electromagnetics, Department of Electrical Engineering and Information Technology
Dr. Peter Amthor, Distributed and Operating Systems Group, Department of Computer Science and Automation

Modern IT applications are characterized by a dramatic increase in energy consumption, especially due to the rapid penetration of mobile & cloud computing, big data processing, smart cities/smart infrastructure, high-performance computing, or the Internet of Things into society. As transistor scaling is come to some end in the near-term, a further way of reducing the energy for computational operations is seen in abandoning the inefficient, von-Neumann-based hardware architectures based on digitally encoded signals via voltage levels. With a bio-inspired design however, neuromorphic, spike-encoded circuits can be leveraged to overcome this problem by representing analogue values and allowing for their extremely efficient encoding in today's most energy-critical computational problems. Especially, superconductive microelectronics is considered to be a very promising option as the signals are generated as voltage pulses by non-linear oscillators. Signal propagation within the circuits resembles the known processes at biological nerves. As a consequence, they form the basis of a true and generic bio-inspired concept. Furthermore, it is the only known solid-state technology that enables both high speed and energy-efficiency with an energy required of ≈ 10^-19 Joule per bit operation [1].

However, there is no optimum architectural design and no corresponding computational model for using these circuits.

The research goal is to implement neuromorphic principles on circuit-level with superconducting microelectronics, especially using the manipulation of single magnetic-flux quanta as information carriers. Integrated circuits are to be devised and adopted for relevant applications where supremacy in energy-efficiency can be demonstrated.

This project is strongly computer-science-oriented and addresses a highly needed research gap in designing, demonstrating and evaluating spike-encoded dendritic circuits and applying their physical characteristics to a computer architecture and an according computational model for real-world use cases.

[1] D.S. Holmes, A. L. Ripple and M A. Manheimer: "Energy-Efficient Superconducting Computing—Power Budgets and Requirements", IEEE Transactions on Applied Superconductivity, 23 (2013) 3, 1701610 doi: 10.1109/TASC.2013.2244634