Prof. Dr. Helge Stein
Academic Research Focus
- High-throughput systems for synthesis, characterization, assembly, and recycling
- Lab automation systems
- Bayesian optimization methods, and explainable machine learning models
- Lab automation systems
- Bayesian optimization methods, and explainable machine learning models
Fields of Application
Automation, Batteries, Data Science / Machine Learning, Semiconductor Materials
Publications
5.
A Sanin, J K Flowers, T H Piotrowiak, F Felsen, L Merker, A Ludwig, D Bresser, H S Stein
In: Advanced Energy Materials, vol. 15, no. 11, pp. 2404961, 2025, ISSN: 1614-6832.
@article{nokey,
title = {Integrating Automated Electrochemistry and High-Throughput Characterization with Machine Learning to Explore Si─Ge─Sn Thin-Film Lithium Battery Anodes},
author = {A Sanin and J K Flowers and T H Piotrowiak and F Felsen and L Merker and A Ludwig and D Bresser and H S Stein},
url = {https://doi.org/10.1002/aenm.202404961},
doi = {https://doi.org/10.1002/aenm.202404961},
issn = {1614-6832},
year = {2025},
date = {2025-03-01},
journal = {Advanced Energy Materials},
volume = {15},
number = {11},
pages = {2404961},
abstract = {Abstract High-performance batteries need accelerated discovery and optimization of new anode materials. Herein, we explore the Si─Ge─Sn ternary alloy system as a candidate fast-charging anode materials system by utilizing a scanning droplet cell (SDC) as an autonomous electrochemical characterization tool with the goal of subsequent upscaling. As the SDC is performing experiments sequentially, an exploration of the entire ternary space is unfeasible due to time constraints. Thus, closed-loop optimization, guided by real-time data analysis and sequential learning algorithms, is utilized to direct experiments. The lead material identified is scaled up to a coin cell to validate the findings from the autonomous millimeter-scale thin-film electrochemical experimentation. Explainable machine learning (ML) models incorporating data from high-throughput Raman spectroscopy and X-ray diffraction (XRD) are used to elucidate the effect of short and long-range ordering on material performance.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract High-performance batteries need accelerated discovery and optimization of new anode materials. Herein, we explore the Si─Ge─Sn ternary alloy system as a candidate fast-charging anode materials system by utilizing a scanning droplet cell (SDC) as an autonomous electrochemical characterization tool with the goal of subsequent upscaling. As the SDC is performing experiments sequentially, an exploration of the entire ternary space is unfeasible due to time constraints. Thus, closed-loop optimization, guided by real-time data analysis and sequential learning algorithms, is utilized to direct experiments. The lead material identified is scaled up to a coin cell to validate the findings from the autonomous millimeter-scale thin-film electrochemical experimentation. Explainable machine learning (ML) models incorporating data from high-throughput Raman spectroscopy and X-ray diffraction (XRD) are used to elucidate the effect of short and long-range ordering on material performance.
4.
M Vogler, S Steensen, F Ramirez, L Merker, J Busk, J M Carlsson, L Rieger, B Zhang, F Liot, G Pizzi, F Hanke, E Flores, H Hajiyani, S Fuchs, A Sanin, M Gaberscek, I Castelli, S Clark, T Vegge, A Bhowmik, H S Stein
Autonomous battery optimisation by deploying distributed experiments and simulations Journal Article
In: ChemRxiv, 2024.
@article{nokey,
title = {Autonomous battery optimisation by deploying distributed experiments and simulations},
author = {M Vogler and S Steensen and F Ramirez and L Merker and J Busk and J M Carlsson and L Rieger and B Zhang and F Liot and G Pizzi and F Hanke and E Flores and H Hajiyani and S Fuchs and A Sanin and M Gaberscek and I Castelli and S Clark and T Vegge and A Bhowmik and H S Stein},
url = {https://chemrxiv.org/engage/chemrxiv/article-details/66448bfb21291e5d1d55f2a7},
doi = {10.26434/chemrxiv-2024-vfq1n},
year = {2024},
date = {2024-05-16},
journal = {ChemRxiv},
abstract = {Non-trivial relationships link individual materials properties to device-level performance. Device optimisation therefore calls for new automation approaches beyond the laboratory bench with tight integration of different research methods. We demonstrate a Materials Acceleration Platform (MAP) in the field of battery research based on our problem-agnostic Fast Intentional Agnostic Learning Server (FINALES) framework, which integrates simulations and physical experiments without centrally controlling them. The connected capabilities entail the formulation and characterisation of electrolytes, cell assembly and testing, early lifetime prediction, and ontology-mapped data storage provided by institutions distributed across Europe. The infrastructure is used to optimise the ionic conductivity of electrolytes and the End Of Life (EOL) of lithium-ion coin cells by varying the electrolyte formulation. We rediscover trends in ionic conductivity and investigate the effect of the electrolyte formulation on the EOL. We further demonstrate the capability of our MAP to bridge diverse research modalities, scales, and institutions enabling system-level investigations under asynchronous conditions while handling concurrent workflows on the material- and system-level, demonstrating true intention-agnosticism.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Non-trivial relationships link individual materials properties to device-level performance. Device optimisation therefore calls for new automation approaches beyond the laboratory bench with tight integration of different research methods. We demonstrate a Materials Acceleration Platform (MAP) in the field of battery research based on our problem-agnostic Fast Intentional Agnostic Learning Server (FINALES) framework, which integrates simulations and physical experiments without centrally controlling them. The connected capabilities entail the formulation and characterisation of electrolytes, cell assembly and testing, early lifetime prediction, and ontology-mapped data storage provided by institutions distributed across Europe. The infrastructure is used to optimise the ionic conductivity of electrolytes and the End Of Life (EOL) of lithium-ion coin cells by varying the electrolyte formulation. We rediscover trends in ionic conductivity and investigate the effect of the electrolyte formulation on the EOL. We further demonstrate the capability of our MAP to bridge diverse research modalities, scales, and institutions enabling system-level investigations under asynchronous conditions while handling concurrent workflows on the material- and system-level, demonstrating true intention-agnosticism.
3.
F Rahmanian, R M Lee, D Linzner, K Michel, L Merker, B B Berkes, L Nuss, H S Stein
Attention towards chemistry agnostic and explainable battery lifetime prediction Journal Article
In: npj Computational Materials, vol. 10, no. 1, pp. 100, 2024, ISSN: 2057-3960.
@article{nokey,
title = {Attention towards chemistry agnostic and explainable battery lifetime prediction},
author = {F Rahmanian and R M Lee and D Linzner and K Michel and L Merker and B B Berkes and L Nuss and H S Stein},
url = {https://doi.org/10.1038/s41524-024-01286-7},
doi = {10.1038/s41524-024-01286-7},
issn = {2057-3960},
year = {2024},
date = {2024-05-10},
journal = {npj Computational Materials},
volume = {10},
number = {1},
pages = {100},
abstract = {Predicting and monitoring battery life early and across chemistries is a significant challenge due to the plethora of degradation paths, form factors, and electrochemical testing protocols. Existing models typically translate poorly across different electrode, electrolyte, and additive materials, mostly require a fixed number of cycles, and are limited to a single discharge protocol. Here, an attention-based recurrent algorithm for neural analysis (ARCANA) architecture is developed and trained on an ultra-large, proprietary dataset from BASF and a large Li-ion dataset gathered from literature across the globe. ARCANA generalizes well across this diverse set of chemistries, electrolyte formulations, battery designs, and cycling protocols and thus allows for an extraction of data-driven knowledge of the degradation mechanisms. The model’s adaptability is further demonstrated through fine-tuning on Na-ion batteries. ARCANA advances the frontier of large-scale time series models in analytical chemistry beyond textual data and holds the potential to significantly accelerate discovery-oriented battery research endeavors.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Predicting and monitoring battery life early and across chemistries is a significant challenge due to the plethora of degradation paths, form factors, and electrochemical testing protocols. Existing models typically translate poorly across different electrode, electrolyte, and additive materials, mostly require a fixed number of cycles, and are limited to a single discharge protocol. Here, an attention-based recurrent algorithm for neural analysis (ARCANA) architecture is developed and trained on an ultra-large, proprietary dataset from BASF and a large Li-ion dataset gathered from literature across the globe. ARCANA generalizes well across this diverse set of chemistries, electrolyte formulations, battery designs, and cycling protocols and thus allows for an extraction of data-driven knowledge of the degradation mechanisms. The model’s adaptability is further demonstrated through fine-tuning on Na-ion batteries. ARCANA advances the frontier of large-scale time series models in analytical chemistry beyond textual data and holds the potential to significantly accelerate discovery-oriented battery research endeavors.
2.
H S Stein
Nonlinear potentiodynamic battery charging protocols for fun, education, and application Journal Article
In: 2023.
@article{nokey,
title = {Nonlinear potentiodynamic battery charging protocols for fun, education, and application},
author = {H S Stein},
url = {https://chemrxiv.org/engage/chemrxiv/article-details/650ef57ab927619fe7acea3e},
doi = {10.26434/chemrxiv-2023-vj5n0},
year = {2023},
date = {2023-09-25},
abstract = {Most secondary batteries in academia are (dis)charged by applying a constant current (CC) followed by a constant voltage (CV) i.e. a CCCV procedure. The usual concept is then to condense data for interpretation into representations such as differential capacity, or dQ/dV, graphs. This is done to extract information related to phenomena such as the growth of the solid electrolyte interphase (SEI) or, more broadly, degradation. Typically, these measurements take several months because measurements for differential capacity analysis need to be performed at relatively low C-rates. An alternate charging schedule to CCCV is pulsed charging, where CC sections are interrupted by an open-circuit measurement on the second time scale. These and similar partially constant current strategies primarily target diffusive effects during charging and broadly fall into a linear charging category, where the time derivative for the actuated property is mostly zero. Herein, I explore nonlinear charging i.e., the process of actively applying a potential with a nontrivial time derivate and a resulting non-trivial current time derivative to engineer (dis)charge cycles with enhanced information density. This method of nonlinear charging is then used to charge a cell such that some potential ranges in the differential capacity diagram are omitted. This study is purely a simulative endeavor and not backed by experimentation, owing mainly to the lack of facile implementation of arbitrary function inputs for battery cyclers and might point to limitations of the underlying theory. If found to be confirmed through an experiment, this technique would, however motivate a new roadmap to better understand secondary battery degradation inspired by electrocatalyst degradation.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Most secondary batteries in academia are (dis)charged by applying a constant current (CC) followed by a constant voltage (CV) i.e. a CCCV procedure. The usual concept is then to condense data for interpretation into representations such as differential capacity, or dQ/dV, graphs. This is done to extract information related to phenomena such as the growth of the solid electrolyte interphase (SEI) or, more broadly, degradation. Typically, these measurements take several months because measurements for differential capacity analysis need to be performed at relatively low C-rates. An alternate charging schedule to CCCV is pulsed charging, where CC sections are interrupted by an open-circuit measurement on the second time scale. These and similar partially constant current strategies primarily target diffusive effects during charging and broadly fall into a linear charging category, where the time derivative for the actuated property is mostly zero. Herein, I explore nonlinear charging i.e., the process of actively applying a potential with a nontrivial time derivate and a resulting non-trivial current time derivative to engineer (dis)charge cycles with enhanced information density. This method of nonlinear charging is then used to charge a cell such that some potential ranges in the differential capacity diagram are omitted. This study is purely a simulative endeavor and not backed by experimentation, owing mainly to the lack of facile implementation of arbitrary function inputs for battery cyclers and might point to limitations of the underlying theory. If found to be confirmed through an experiment, this technique would, however motivate a new roadmap to better understand secondary battery degradation inspired by electrocatalyst degradation.
1.
I E Castelli, D J Arismendi-Arrieta, A Bhowmik, I Cekic-Laskovic, S Clark, R Dominko, E Flores, J Flowers, K Ulvskov Frederiksen, J Friis, A Grimaud, K V Hansen, L J Hardwick, K Hermansson, L Königer, H Lauritzen, F Le Cras, H Li, S Lyonnard, H Lorrmann, N Marzari, L Niedzicki, G Pizzi, F Rahmanian, H S Stein, M Uhrin, W Wenzel, M Winter, C Wölke, T Vegge
Data Management Plans: the Importance of Data Management in the BIG-MAP Project** Journal Article
In: Batteries & Supercaps, vol. 4, no. 12, pp. 1803-1812, 2022.
@article{nokey,
title = {Data Management Plans: the Importance of Data Management in the BIG-MAP Project**},
author = {I E Castelli and D J Arismendi-Arrieta and A Bhowmik and I Cekic-Laskovic and S Clark and R Dominko and E Flores and J Flowers and K Ulvskov Frederiksen and J Friis and A Grimaud and K V Hansen and L J Hardwick and K Hermansson and L K\"{o}niger and H Lauritzen and F Le Cras and H Li and S Lyonnard and H Lorrmann and N Marzari and L Niedzicki and G Pizzi and F Rahmanian and H S Stein and M Uhrin and W Wenzel and M Winter and C W\"{o}lke and T Vegge},
url = {https://chemistry-europe.onlinelibrary.wiley.com/doi/abs/10.1002/batt.202100117},
doi = {https://doi.org/10.1002/batt.202100117},
year = {2022},
date = {2022-08-28},
urldate = {2022-08-28},
journal = {Batteries \& Supercaps},
volume = {4},
number = {12},
pages = {1803-1812},
abstract = {Abstract Open access to research data is increasingly important for accelerating research. Grant authorities therefore request detailed plans for how data is managed in the projects they finance. We have recently developed such a plan for the EU−H2020 BIG-MAP project\textemdasha cross-disciplinary project targeting disruptive battery-material discoveries. Essential for reaching the goal is extensive sharing of research data across scales, disciplines and stakeholders, not limited to BIG-MAP and the European BATTERY 2030+ initiative but within the entire battery community. The key challenges faced in developing the data management plan for such a large and complex project were to generate an overview of the enormous amount of data that will be produced, to build an understanding of the data flow within the project and to agree on a roadmap for making all data FAIR (findable, accessible, interoperable, reusable). This paper describes the process we followed and how we structured the plan.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract Open access to research data is increasingly important for accelerating research. Grant authorities therefore request detailed plans for how data is managed in the projects they finance. We have recently developed such a plan for the EU−H2020 BIG-MAP project—a cross-disciplinary project targeting disruptive battery-material discoveries. Essential for reaching the goal is extensive sharing of research data across scales, disciplines and stakeholders, not limited to BIG-MAP and the European BATTERY 2030+ initiative but within the entire battery community. The key challenges faced in developing the data management plan for such a large and complex project were to generate an overview of the enormous amount of data that will be produced, to build an understanding of the data flow within the project and to agree on a roadmap for making all data FAIR (findable, accessible, interoperable, reusable). This paper describes the process we followed and how we structured the plan.