Showing 1–14 of 14 results for author: Kuenneth, C

Polymer Informatics Beyond Homopolymers

Authors: Shivank S. Shukla, Christopher Kuenneth, Rampi Ramprasad

Abstract: Polymers are diverse and versatile materials that have met a wide range of material application demands. They come in several flavors and architectures, e.g., homopolymers, copolymers, polymer blends, and polymers with additives. Searching this enormous space for suitable materials with a specific set of property/performance targets is thus non-trivial, painstaking, and expensive. Such a search pr… ▽ More Polymers are diverse and versatile materials that have met a wide range of material application demands. They come in several flavors and architectures, e.g., homopolymers, copolymers, polymer blends, and polymers with additives. Searching this enormous space for suitable materials with a specific set of property/performance targets is thus non-trivial, painstaking, and expensive. Such a search process can be made effective by the creation of rapid and accurate property predictors. In this work, we present a machine-learning framework to predict the thermal properties of homopolymers, copolymers, and polymer blends. A universal fingerprinting scheme capable of handling this entire polymer chemical class has been developed and a multi-task deep learning algorithm is trained simultaneously on a large dataset of glass transition, melting, and degradation temperatures. The developed models are accurate, fast, flexible, and scalable to other properties when suitable data become available. △ Less

Submitted 22 March, 2023; originally announced March 2023.

polyBERT: A chemical language model to enable fully machine-driven ultrafast polymer informatics

Authors: Christopher Kuenneth, Rampi Ramprasad

Abstract: Polymers are a vital part of everyday life. Their chemical universe is so large that it presents unprecedented opportunities as well as significant challenges to identify suitable application-specific candidates. We present a complete end-to-end machine-driven polymer informatics pipeline that can search this space for suitable candidates at unprecedented speed and accuracy. This pipeline includes… ▽ More Polymers are a vital part of everyday life. Their chemical universe is so large that it presents unprecedented opportunities as well as significant challenges to identify suitable application-specific candidates. We present a complete end-to-end machine-driven polymer informatics pipeline that can search this space for suitable candidates at unprecedented speed and accuracy. This pipeline includes a polymer chemical fingerprinting capability called polyBERT (inspired by Natural Language Processing concepts), and a multitask learning approach that maps the polyBERT fingerprints to a host of properties. polyBERT is a chemical linguist that treats the chemical structure of polymers as a chemical language. The present approach outstrips the best presently available concepts for polymer property prediction based on handcrafted fingerprint schemes in speed by two orders of magnitude while preserving accuracy, thus making it a strong candidate for deployment in scalable architectures including cloud infrastructures. △ Less

Submitted 29 September, 2022; originally announced September 2022.

Polymer informatics at-scale with multitask graph neural networks

Authors: Rishi Gurnani, Christopher Kuenneth, Aubrey Toland, Rampi Ramprasad

Abstract: Artificial intelligence-based methods are becoming increasingly effective at screening libraries of polymers down to a selection that is manageable for experimental inquiry. The vast majority of presently adopted approaches for polymer screening rely on handcrafted chemostructural features extracted from polymer repeat units -- a burdensome task as polymer libraries, which approximate the polymer… ▽ More Artificial intelligence-based methods are becoming increasingly effective at screening libraries of polymers down to a selection that is manageable for experimental inquiry. The vast majority of presently adopted approaches for polymer screening rely on handcrafted chemostructural features extracted from polymer repeat units -- a burdensome task as polymer libraries, which approximate the polymer chemical search space, progressively grow over time. Here, we demonstrate that directly "machine-learning" important features from a polymer repeat unit is a cheap and viable alternative to extracting expensive features by hand. Our approach -- based on graph neural networks, multitask learning, and other advanced deep learning techniques -- speeds up feature extraction by one to two orders of magnitude relative to presently adopted handcrafted methods without compromising model accuracy for a variety of polymer property prediction tasks. We anticipate that our approach, which unlocks the screening of truly massive polymer libraries at scale, will enable more sophisticated and large scale screening technologies in the field of polymer informatics. △ Less

Submitted 17 January, 2023; v1 submitted 27 September, 2022; originally announced September 2022.

A general-purpose material property data extraction pipeline from large polymer corpora using Natural Language Processing

Authors: Pranav Shetty, Arunkumar Chitteth Rajan, Christopher Kuenneth, Sonkakshi Gupta, Lakshmi Prerana Panchumarti, Lauren Holm, Chao Zhang, Rampi Ramprasad

Abstract: The ever-increasing number of materials science articles makes it hard to infer chemistry-structure-property relations from published literature. We used natural language processing (NLP) methods to automatically extract material property data from the abstracts of polymer literature. As a component of our pipeline, we trained MaterialsBERT, a language model, using 2.4 million materials science ab… ▽ More The ever-increasing number of materials science articles makes it hard to infer chemistry-structure-property relations from published literature. We used natural language processing (NLP) methods to automatically extract material property data from the abstracts of polymer literature. As a component of our pipeline, we trained MaterialsBERT, a language model, using 2.4 million materials science abstracts, which outperforms other baseline models in three out of five named entity recognition datasets when used as the encoder for text. Using this pipeline, we obtained ~300,000 material property records from ~130,000 abstracts in 60 hours. The extracted data was analyzed for a diverse range of applications such as fuel cells, supercapacitors, and polymer solar cells to recover non-trivial insights. The data extracted through our pipeline is made available through a web platform at https://polymerscholar.org which can be used to locate material property data recorded in abstracts conveniently. This work demonstrates the feasibility of an automatic pipeline that starts from published literature and ends with a complete set of extracted material property information. △ Less

Submitted 26 September, 2022; originally announced September 2022.

Bioplastic Design using Multitask Deep Neural Networks

Authors: Christopher Kuenneth, Jessica Lalonde, Babetta L. Marrone, Carl N. Iverson, Rampi Ramprasad, Ghanshyam Pilania

Abstract: Non-degradable plastic waste stays for decades on land and in water, jeopardizing our environment; yet our modern lifestyle and current technologies are impossible to sustain without plastics. Bio-synthesized and biodegradable alternatives such as the polymer family of polyhydroxyalkanoates (PHAs) have the potential to replace large portions of the world's plastic supply with cradle-to-cradle mate… ▽ More Non-degradable plastic waste stays for decades on land and in water, jeopardizing our environment; yet our modern lifestyle and current technologies are impossible to sustain without plastics. Bio-synthesized and biodegradable alternatives such as the polymer family of polyhydroxyalkanoates (PHAs) have the potential to replace large portions of the world's plastic supply with cradle-to-cradle materials, but their chemical complexity and diversity limit traditional resource-intensive experimentation. In this work, we develop multitask deep neural network property predictors using available experimental data for a diverse set of nearly 23000 homo- and copolymer chemistries. Using the predictors, we identify 14 PHA-based bioplastics from a search space of almost 1.4 million candidates which could serve as potential replacements for seven petroleum-based commodity plastics that account for 75% of the world's yearly plastic production. We discuss possible synthesis routes for these identified promising materials. The developed multitask polymer property predictors are made available as a part of the Polymer Genome project at https://PolymerGenome.org. △ Less

Submitted 22 March, 2022; originally announced March 2022.

Journal ref: Commun Mater 3, 96 (2022)

Copolymer Informatics with Multi-Task Deep Neural Networks

Authors: Christopher Künneth, William Schertzer, Rampi Ramprasad

Abstract: Polymer informatics tools have been recently gaining ground to efficiently and effectively develop, design, and discover new polymers that meet specific application needs. So far, however, these data-driven efforts have largely focused on homopolymers. Here, we address the property prediction challenge for copolymers, extending the polymer informatics framework beyond homopolymers. Advanced polyme… ▽ More Polymer informatics tools have been recently gaining ground to efficiently and effectively develop, design, and discover new polymers that meet specific application needs. So far, however, these data-driven efforts have largely focused on homopolymers. Here, we address the property prediction challenge for copolymers, extending the polymer informatics framework beyond homopolymers. Advanced polymer fingerprinting and deep-learning schemes that incorporate multi-task learning and meta-learning are proposed. A large data set containing over 18,000 data points of glass transition, melting, and degradation temperature of homopolymers and copolymers of up to two monomers is used to demonstrate the copolymer prediction efficacy. The developed models are accurate, fast, flexible, and scalable to more copolymer properties when suitable data become available. △ Less

Submitted 25 March, 2021; originally announced March 2021.

Polymer Informatics: Current Status and Critical Next Steps

Authors: Lihua Chen, Ghanshyam Pilania, Rohit Batra, Tran Doan Huan, Chiho Kim, Christopher Kuenneth, Rampi Ramprasad

Abstract: Artificial intelligence (AI) based approaches are beginning to impact several domains of human life, science and technology. Polymer informatics is one such domain where AI and machine learning (ML) tools are being used in the efficient development, design and discovery of polymers. Surrogate models are trained on available polymer data for instant property prediction, allowing screening of promis… ▽ More Artificial intelligence (AI) based approaches are beginning to impact several domains of human life, science and technology. Polymer informatics is one such domain where AI and machine learning (ML) tools are being used in the efficient development, design and discovery of polymers. Surrogate models are trained on available polymer data for instant property prediction, allowing screening of promising polymer candidates with specific target property requirements. Questions regarding synthesizability, and potential (retro)synthesis steps to create a target polymer, are being explored using statistical means. Data-driven strategies to tackle unique challenges resulting from the extraordinary chemical and physical diversity of polymers at small and large scales are being explored. Other major hurdles for polymer informatics are the lack of widespread availability of curated and organized data, and approaches to create machine-readable representations that capture not just the structure of complex polymeric situations but also synthesis and processing conditions. Methods to solve inverse problems, wherein polymer recommendations are made using advanced AI algorithms that meet application targets, are being investigated. As various parts of the burgeoning polymer informatics ecosystem mature and become integrated, efficiency improvements, accelerated discoveries and increased productivity can result. Here, we review emergent components of this polymer informatics ecosystem and discuss imminent challenges and opportunities. △ Less

Submitted 1 November, 2020; originally announced November 2020.

Polymer Informatics with Multi-Task Learning

Authors: Christopher Künneth, Arunkumar Chitteth Rajan, Huan Tran, Lihua Chen, Chiho Kim, Rampi Ramprasad

Abstract: Modern data-driven tools are transforming application-specific polymer development cycles. Surrogate models that can be trained to predict the properties of new polymers are becoming commonplace. Nevertheless, these models do not utilize the full breadth of the knowledge available in datasets, which are oftentimes sparse; inherent correlations between different property datasets are disregarded. H… ▽ More Modern data-driven tools are transforming application-specific polymer development cycles. Surrogate models that can be trained to predict the properties of new polymers are becoming commonplace. Nevertheless, these models do not utilize the full breadth of the knowledge available in datasets, which are oftentimes sparse; inherent correlations between different property datasets are disregarded. Here, we demonstrate the potency of multi-task learning approaches that exploit such inherent correlations effectively, particularly when some property dataset sizes are small. Data pertaining to 36 different properties of over $13, 000$ polymers (corresponding to over $23,000$ data points) are coalesced and supplied to deep-learning multi-task architectures. Compared to conventional single-task learning models (that are trained on individual property datasets independently), the multi-task approach is accurate, efficient, scalable, and amenable to transfer learning as more data on the same or different properties become available. Moreover, these models are interpretable. Chemical rules, that explain how certain features control trends in specific property values, emerge from the present work, paving the way for the rational design of application specific polymers meeting desired property or performance objectives. △ Less

Submitted 28 October, 2020; originally announced October 2020.

Piezospectroscopy and first-principles calculations of the nitrogen-vacancy center in gallium arsenide

Authors: Nicola Kovač, Christopher Künneth, Hans Christian Alt

Abstract: The nitrogen-vacancy (NV) center occurs in GaAs bulk crystals doped or implanted with nitrogen. The local vibration of nitrogen gives rise to a sharp infrared absorption band at 638 cm$^{-1}$, exhibiting a fine structure due to the different masses of neighboring $^{69}$Ga and $^{71}$Ga host isotopes. Piezospectroscopic investigations in the crystallographic <100> direction prove that the center h… ▽ More The nitrogen-vacancy (NV) center occurs in GaAs bulk crystals doped or implanted with nitrogen. The local vibration of nitrogen gives rise to a sharp infrared absorption band at 638 cm$^{-1}$, exhibiting a fine structure due to the different masses of neighboring $^{69}$Ga and $^{71}$Ga host isotopes. Piezospectroscopic investigations in the crystallographic <100> direction prove that the center has C$_\text{3v}$ point symmetry, which is weakly perturbed by the isotope effect. The stress-induced shifts of some band components show an unusual non-linear behavior that can be explained by coupling between the isotope and the stress splitting. First-principles density-functional theory calculations are in full accordance with the experiments and confirm the C$_\text{3v}$ symmetry, caused by relaxation of the nitrogen atom from the anion lattice site towards the nearest-neighbor Ga plane. The NV center in GaAs is structurally analogous to the same center in diamond. The $-3$ charge state is most stable for nearly all Fermi-level positions. △ Less

Submitted 6 February, 2018; v1 submitted 29 October, 2017; originally announced October 2017.

Impact of Four-Valent Do** on the Crystallographic Phase Formation for Ferroelectric HfO$_2$ from First-Principles: Implications for Ferroelectric Memory and Energy-Related Applications

Authors: Christopher Künneth, Robin Materlik, Max Falkowski, Alfred Kersch

Abstract: The ferroelectric properties of nanoscale silicon doped HfO$_2$ promise a multitude of applications ranging from ferroelectric memory to energy-related applications. The reason for the unexpected behavior has not been clearly proven and presumably include contributions from size effects and do** effects. Silicon incorporation in HfO$_2$ is investigated computationally by first-principles using d… ▽ More The ferroelectric properties of nanoscale silicon doped HfO$_2$ promise a multitude of applications ranging from ferroelectric memory to energy-related applications. The reason for the unexpected behavior has not been clearly proven and presumably include contributions from size effects and do** effects. Silicon incorporation in HfO$_2$ is investigated computationally by first-principles using different density functional theory (DFT) methods. Formation energies of interstitial and substitutional silicon in HfO$_2$ paired with and without an oxygen vacancy prove the substitutional defect as the most likely. Within the investigated concentration window up to 12.5 formula unit %, silicon do** alone is not sufficient to stabilize the polar and orthorhombic crystal phase (p-o-phase), which has been identified as the source of the ferroelectricity in HfO$_2$. On the other hand, silicon incorporation is one of the strongest promoters of the p-o-phase and the tetragonal phase (t-phase) within the group of investigated dopants, confirming the experimental ferroelectric window. Besides silicon, the favoring effects on the energy of other four-valent dopants, C, Ge, Ti, Sn, Zr and Ce, are examined, revealing Ce as a very promising candidate. The evolution of the volume changes with increasing do** concentration of these four-valent dopants shows an inverse trend for Ce in comparison to silicon. To complement this study, the geometrical incorporation of the dopants in the host HfO$_2$ lattice was analyzed. △ Less

Submitted 2 January, 2018; v1 submitted 27 October, 2017; originally announced October 2017.

Symmetry and structure of carbon-nitrogen complexes in gallium arsenide from infrared spectroscopy and first-principles calculations

Authors: Christopher Künneth, Simon Kölbl, Hans Edwin Wagner, Volker Häublein, Alfred Kersch, Hans Christian Alt

Abstract: Molecular-like carbon-nitrogen complexes in GaAs are investigated both experimentally and theoretically. Two characteristic high-frequency stretching modes at \num{1973} and \SI{2060}{cm^{-1}}, detected by Fourier transform infrared absorption (FTIR) spectroscopy, appear in carbon- and nitrogen-implanted and annealed layers. From isotopic substitution it is deduced that the chemical composition of… ▽ More Molecular-like carbon-nitrogen complexes in GaAs are investigated both experimentally and theoretically. Two characteristic high-frequency stretching modes at \num{1973} and \SI{2060}{cm^{-1}}, detected by Fourier transform infrared absorption (FTIR) spectroscopy, appear in carbon- and nitrogen-implanted and annealed layers. From isotopic substitution it is deduced that the chemical composition of the underlying complexes is CN$_2$ and C$_2$N, respectively. Piezospectroscopic FTIR measurements reveal that both centers have tetragonal symmetry. For density functional theory (DFT) calculations linear entities are substituted for the As anion, with the axis oriented along the \hkl<100> direction, in accordance with the experimentally ascertained symmetry. The DFT calculations support the stability of linear N-C-N and C-C-N complexes in the GaAs host crystal in the charge states ranging from $+3$ to $-3$. The valence bonds of the complexes are analyzed using molecular-like orbitals from DFT. It turns out that internal bonds and bonds to the lattice are essentially independent of the charge state. The calculated vibrational mode frequencies are close to the experimental values and reproduce precisely the isotopic mass splitting from FTIR experiments. Finally, the formation energies show that under thermodynamic equilibrium CN$_2$ is more stable than C$_2$N. △ Less

Submitted 4 January, 2018; v1 submitted 21 September, 2017; originally announced September 2017.

A computational study of hafnia-based ferroelectric memories: from ab initio via physical modeling to circuit models of ferroelectric device

Authors: Milan Pešić, Christopher Künneth, Michael Hoffmann, Halid Mulaosmanovic, Stefan Müller, Evelyn T. Breyer, Uwe Schroeder, Alfred Kersch, Thomas Mikolajick, Stefan Slesazeck

Abstract: The discovery of ferroelectric properties of binary oxides revitalized the interest in ferroelectrics and bridged the scaling gap between the state-of-the-art semiconductor technology and ferroelectric memories. However, before hitting the markets, the origin of ferroelectricity and in-depth studies of device characteristics are needed. Establishing a correlation between the performance of the dev… ▽ More The discovery of ferroelectric properties of binary oxides revitalized the interest in ferroelectrics and bridged the scaling gap between the state-of-the-art semiconductor technology and ferroelectric memories. However, before hitting the markets, the origin of ferroelectricity and in-depth studies of device characteristics are needed. Establishing a correlation between the performance of the device and underlying physical mechanisms is the first step toward understanding the device and engineering guidelines for a novel, superior device. Therefore, in this paper a holistic modeling approaches which lead to a better understanding of ferroelectric memories based on hafnium and zirconium oxide is addressed. Starting from describing the stabilization of the ferroelectric phase within the binary oxides via physical modeling the physical mechanisms of the ferroelectric devices are reviewed. Besides, limitations and modeling of the multilevel operation and switching kinetics of ultimately scaled devices as well as the necessity for Landau-Khalatnikov approach are discussed. Furthermore, a device-level model of ferroelectric memory devices that can be used to study the array implementation and their operational schemes are addressed. Finally, a circuit model of the ferroelectric memory device is presented and potential further applications of ferroelectric devices are outlined. △ Less

Submitted 20 September, 2017; originally announced September 2017.

The impact of charge compensated and uncompensated strontium defects on the stabilization of the ferroelectric phase in HfO$_2$

Authors: Robin Materlik, Christopher Künneth, Thomas Mikolajick, Alfred Kersch

Abstract: Different dopants with their specific dopant concentration can be utilized to produce ferroelectric HfO$_2$ thin films. In this work it is explored for the example of Sr in a comprehensive first-principles study. Density functional calculations reveal structure, formation energy and total energy of the Sr related defects in HfO$_2$. We found the charge compensated defect including an associated ox… ▽ More Different dopants with their specific dopant concentration can be utilized to produce ferroelectric HfO$_2$ thin films. In this work it is explored for the example of Sr in a comprehensive first-principles study. Density functional calculations reveal structure, formation energy and total energy of the Sr related defects in HfO$_2$. We found the charge compensated defect including an associated oxygen vacancy Sr$_\text{Hf}$V$_\text{O}$ to strongly favour the non-ferroelectric, tetragonal P4$_\text{2}$/mnc phase energetically. In contrast, the uncompensated defect without oxygen vacancy Sr$_\text{Hf}$ favours the ferroelectric, orthorhombic Pca2$_\text{1}$ phase. According to the formation energy the uncompensated defect can form easily under oxygen rich conditions in the production process. Low oxygen partial pressure existing over the lifetime promotes the loss of oxygen leading to V$_\text{O}$ and, thus, the destabilization of the ferroelectric, orthorhombic Pca2$_\text{1}$ phase accompanied by an increase of the leakage current. This study attempts to fundamentally explain the stabilization of the ferroelectric, orthorhombic Pca2$_\text{1}$ phase by do**. △ Less

Submitted 25 August, 2017; originally announced August 2017.

Journal ref: Appl. Phys. Lett., vol. 111, no. 8, p. 82902, 2017

The Origin of Ferroelectricity in Hf$_{x}$ Zr$_{1-x}$ O$_2$: A Computational Investigation and a Surface Energy Model

Authors: Robin Materlik, Christopher Künneth, Alfred Kersch

Abstract: The structural, thermal, and dielectric properties of the ferroelectric phase of HfO$_2$, ZrO$_2$ and Hf$_{0.5}$ Zr$_{0.5}$ O$_2$ (HZO) are investigated with carefully validated density functional computations. We find, that the free bulk energy of the ferroelectric orthorhombic Pca2$_{1}$ phase is unfavorable compared to the monoclinic P2$_{1}$/c and the orthorhombic Pbca phase for all investigat… ▽ More The structural, thermal, and dielectric properties of the ferroelectric phase of HfO$_2$, ZrO$_2$ and Hf$_{0.5}$ Zr$_{0.5}$ O$_2$ (HZO) are investigated with carefully validated density functional computations. We find, that the free bulk energy of the ferroelectric orthorhombic Pca2$_{1}$ phase is unfavorable compared to the monoclinic P2$_{1}$/c and the orthorhombic Pbca phase for all investigated stoichiometries in the Hf$_χ$Zr$_{1-χ}$O$_2$ system. To explain the existence of the ferroelectric phase in nanoscale thin films we explore the Gibbs / Helmholtz free energies as a function of stress and film strain and find them unlikely to become minimal in HZO films for technological relevant conditions. To assess the contribution of surface energy to the phase stability we parameterize a model, interpolating between existing data, and find the Helmholtz free energy of ferroelectric grains minimal for a range of size and stoichiometry. From the model we predict undoped HfO$_2$ to be ferroelectric for a grain size of about 4 nm and epitaxial HZO below 5 nm. Furthermore we calculate the strength of an applied electric field necessary to cause the antiferroelectric phase transformation in ZrO$_2$ from the P4$_2$/nmc phase as 1 MV/cm in agreement with experimental data, explaining the mechanism of field induced phase transformation. △ Less

Submitted 2 July, 2015; originally announced July 2015.

Journal ref: J. Appl. Phys. 117, 134109 (2015)