TY - JOUR A1 - Martynec, T. A1 - Karapanagiotis, Christos A1 - Klapp, S. H. L. A1 - Kowarik, S. T1 - Machine learning predictions of surface migration barriers in nucleation and non-equilibrium growth N2 - Machine learning is playing an increasing role in the discovery of new materials and may also facilitate the search for optimum growth conditions for crystals and thin films. Here, we perform kinetic Monte-Carlo simulations of sub-monolayer growth. We consider a generic homoepitaxial growth scenario that covers a wide range of conditions with different diffusion barriers (0.4–0.55 eV) and lateral binding energies (0.1–0.4 eV). These simulations are used as a training data set for a convolutional neural network that can predict diffusion barriers and binding energies. Specifically, a single Monte-Carlo image of the morphology is sufficient to determine the energy barriers with an accuracy of approximately 10 meV and the neural network is tolerant to images with noise and lower than atomic-scale resolution. We believe this new machine learning method will be useful for fundamental studies of growth kinetics and growth optimization through better knowledge of microscopic parameters. KW - Machine learning KW - Surface migration barriers KW - Material engineering PY - 2021 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-532523 UR - https://www.nature.com/articles/s43246-021-00188-1 DO - https://doi.org/10.1038/s43246-021-00188-1 SN - 2662-4443 VL - 2 SP - 1 EP - 9 PB - Springer Nature CY - London AN - OPUS4-53252 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Fricke, F. A1 - Mahmood, S. A1 - Hoffmann, J. A1 - Brandalero, M. A1 - Liehr, Sascha A1 - Kern, Simon A1 - Meyer, Klas A1 - Kowarik, Stefan A1 - Westerdicky, S. A1 - Maiwald, Michael A1 - Hübner, M. T1 - Artificial Intelligence for Mass Spectrometry and Nuclear Magnetic Resonance Spectroscopy N2 - Mass Spectrometry (MS) and Nuclear Magnetic Resonance Spectroscopy (NMR) are critical components of every industrial chemical process as they provide information on the concentrations of individual compounds and by-products. These processes are carried out manually and by a specialist, which takes a substantial amount of time and prevents their utilization for real-time closed-loop process control. This paper presents recent advances from two projects that use Artificial Neural Networks (ANNs) to address the challenges of automation and performance-efficient realizations of MS and NMR. In the first part, a complete toolchain has been developed to develop simulated spectra and train ANNs to identify compounds in MS. In the second part, a limited number of experimental NMR spectra have been augmented by simulated spectra to train an ANN with better prediction performance and speed than state-of-theart analysis. These results suggest that, in the context of the digital transformation of the process industry, we are now on the threshold of a possible strongly simplified use of MS and MRS and the accompanying data evaluation by machine-supported procedures, and can utilize both methods much wider for reaction and process monitoring or quality control. T2 - 2021 Design, Automation & Test in Europe Conference & Exhibition (DATE) CY - Online meeting DA - 01.02.2021 KW - Industry 4.0 KW - Cyber-Physical Systems KW - Artificial Neural Networks KW - Mass Spectrometry KW - Nuclear Magnetic Resonance Spectroscopy PY - 2021 UR - www.date-conference.com SN - 978-3-9819263-5-4 SP - 615 EP - 620 PB - Research Publishing AN - OPUS4-52180 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Kowarik, Stefan A1 - Weber, C. A1 - Pithan, L. A1 - Zykov, A. A1 - Bommel, S. A1 - Carla, F. A1 - Felici, R. A1 - Knie, C. A1 - Bléger, D. T1 - Multiple timescales in the photoswitching kinetics of crystalline thin films of azobenzene-trimers N2 - Functional materials that exhibit photoinduced structural phase transitions are highly interesting for applications in optomechanics and mechanochemistry. It is, however, still not fully understood how photochemical reactions, which are often accompanied by molecular motion, proceed in confined and crystalline environments. Here we show that thin films of azobenzene trimers exhibit high structural order and determine the crystallographic unit cell. We demonstrate that thin film can be switched partially reversibly between a crystalline and an amorphous phase. The time constant of the photoinduced amorphisation as measured with real-time x-ray diffraction ($\approx $ 220 s) lies between the two time constants (120 s and 2870 s) of the ensemble photoisomerisation processes that are measured via optical spectroscopy. Our observation of a photoinduced shrinking of the crystalline domains indicates a cascading process, in which photoisomerisation starts at the surface of the thin film and propagates deeper into the crystalline layer by introducing disorder and generating free volume. This finding is important for the rapidly evolving research field of photoresponsive thin films and smart crystalline materials in general. KW - Azobenzene PY - 2017 DO - https://doi.org/10.1088/1361-648X/aa8654 SN - 0953-8984 SN - 1361-648X VL - 29 IS - 43 SP - Article 434001, 1 EP - 8 AN - OPUS4-42501 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Fricke, F. A1 - Brandalero, M. A1 - Liehr, Sascha A1 - Kern, Simon A1 - Meyer, Klas A1 - Kowarik, Stefan A1 - Hierzegger, R. A1 - Westerdick, S. A1 - Maiwald, Michael A1 - Hübner, M. T1 - Artificial Intelligence for Mass Spectrometry and Nuclear Magnetic Resonance Spectroscopy Using a Novel Data Augmentation Method N2 - Mass Spectrometry (MS) and Nuclear Magnetic Resonance Spectroscopy (NMR) are valuable analytical and quality control methods for most industrial chemical processes as they provide information on the concentrations of individual compounds and by-products. These processes are traditionally carried out manually and by a specialist, which takes a substantial amount of time and prevents their utilization for real-time closed-loop process control. This paper presents recent advances from two projects that use Artificial Neural Networks (ANNs) to address the challenges of automation and performance-efficient realizations of MS and NMR. In the first part, a complete toolchain has been realized to develop simulated spectra and train ANNs to identify compounds in MS. In the second part, a limited number of experimental NMR spectra have been augmented by simulated spectra, to train an ANN with better prediction performance and speed than state-of-the-art analysis. These results suggest that, in the context of the digital transformation of the process industry, we are now on the threshold of a strongly simplified use of MS and MRS and the accompanying data evaluation by machine-supported procedures, and can utilize both methods much wider for reaction and process monitoring or quality control. KW - Industry 4.0 KW - Cyber-Physical Systems KW - Artificial Neural Networks KW - Mass Spectrometry KW - Nuclear Magnetic Resonance Spectroscopy KW - Modular Production PY - 2021 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-539412 UR - https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9638378 DO - https://doi.org/10.1109/TETC.2021.3131371 SN - 2168-6750 VL - 10 IS - 1 SP - 87 EP - 98 PB - IEEE AN - OPUS4-53941 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Kern, Simon A1 - Liehr, Sascha A1 - Wander, Lukas A1 - Bornemann-Pfeiffer, Martin A1 - Müller, S. A1 - Maiwald, Michael A1 - Kowarik, Stefan T1 - Artificial neural networks for quantitative online NMR spectroscopy N2 - Industry 4.0 is all about interconnectivity, sensor-enhanced process control, and data-driven systems. Process analytical technology (PAT) such as online nuclear magnetic resonance (NMR) spectroscopy is gaining in importance, as it increasingly contributes to automation and digitalization in production. In many cases up to now, however, a classical evaluation of process data and their transformation into knowledge is not possible or not economical due to the insufficiently large datasets available. When developing an automated method applicable in process control, sometimes only the basic data of a limited number of batch tests from typical product and process development campaigns are available. However, these datasets are not large enough for training machine-supported procedures. In this work, to overcome this limitation, a new procedure was developed, which allows physically motivated multiplication of the available reference data in order to obtain a sufficiently large dataset for training machine learning algorithms. The underlying example chemical synthesis was measured and analyzed with both application-relevant low-field NMR and high-field NMR spectroscopy as reference method. Artificial neural networks (ANNs) have the potential to infer valuable process information already from relatively limited input data. However, in order to predict the concentration at complex conditions (many reactants and wide concentration ranges), larger ANNs and, therefore, a larger Training dataset are required. We demonstrate that a moderately complex problem with four reactants can be addressed using ANNs in combination with the presented PAT method (low-field NMR) and with the proposed approach to generate meaningful training data. KW - Online NMR spectroscopy KW - Real-time process monitoring KW - Artificial neural networks KW - Automation KW - Process industry PY - 2020 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-507508 DO - https://doi.org/10.1007/s00216-020-02687-5 SN - 1618-2642 VL - 412 IS - 18 SP - 4447 EP - 4459 PB - Springer CY - Berlin AN - OPUS4-50750 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - GEN A1 - Kern, Simon A1 - Liehr, Sascha A1 - Wander, Lukas A1 - Bornemann-Pfeiffer, Martin A1 - Müller, S. A1 - Maiwald, Michael A1 - Kowarik, Stefan T1 - Training data of quantitative online NMR spectroscopy for artificial neural networks N2 - Data set of low-field NMR spectra of continuous synthesis of nitro-4’-methyldiphenylamine (MNDPA). 1H spectra (43 MHz) were recorded as single scans. Two different approaches for the generation of artificial neural networks training data for the prediction of reactant concentrations were used: (i) Training data based on combinations of measured pure component spectra and (ii) Training data based on a spectral model. Synthetic low-field NMR spectra First 4 columns in MAT-files represent component areas of each reactant within the synthetic mixture spectrum. Xi (“pure component spectra dataset”) Xii (“spectral model dataset”) Experimental low-field NMR spectra from MNDPA-Synthesis This data set represents low-field NMR-spectra recorded during continuous synthesis of nitro-4’-methyldiphenylamine (MNDPA). Reference values from high-field NMR results are included. KW - NMR spectroscopy KW - Real-time process monitoring KW - Artificial neural networks KW - Online NMR spectroscopy KW - Automation KW - Process industry PY - 2020 DO - https://doi.org/10.5281/zenodo.3677139 PB - Zenodo CY - Geneva AN - OPUS4-50456 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -