-
Deciphering diffuse scattering with machine learning and the equivariant foundation model: The case of molten FeO
Authors:
Ganesh Sivaraman,
Chris J. Benmore
Abstract:
Bridging the gap between diffuse x-ray or neutron scattering measurements and predicted structures derived from atom-atom pair potentials in disordered materials, has been a longstanding challenge in condensed matter physics. This perspective gives a brief overview of the traditional approaches employed over the past several decades. Namely, the use of approximate interatomic pair potentials that…
▽ More
Bridging the gap between diffuse x-ray or neutron scattering measurements and predicted structures derived from atom-atom pair potentials in disordered materials, has been a longstanding challenge in condensed matter physics. This perspective gives a brief overview of the traditional approaches employed over the past several decades. Namely, the use of approximate interatomic pair potentials that relate 3-dimensional structural models to the measured structure factor and its associated pair distribution function. The use of machine learned interatomic potentials has grown in the past few years, and has been particularly successful in the cases of ionic and oxide systems. Recent advances in large scale sampling, along with a direct integration of scattering measurements into the model development, has provided improved agreement between experiments and large-scale models calculated with quantum mechanical accuracy. However, details of local polyhedral bonding and connectivity in meta-stable disordered systems still require improvement. Here we leverage MACE-MP-0; a newly introduced equivariant foundation model and validate the results against high-quality experimental scattering data for the case of molten iron(II) oxide (FeO). These preliminary results suggest that the emerging foundation model has the potential to surpass the traditional limitations of classical interatomic potentials.
△ Less
Submitted 29 February, 2024;
originally announced March 2024.
-
The First Direct Detection of Kirkwood Transitions in Concentrated Aqueous Electrolytes using Small Angle X-ray Scattering
Authors:
Mohammadhasan Dinpajooh,
Elisa Biasin,
Christopher J. Mundy,
Gregory K. Schenter,
Emily T. Nienhuis,
Sebastian T. Mergelsberg,
Chris J. Benmore,
John L. Fulton,
Shawn M. Kathmann
Abstract:
Ion-ion correlations, screening, and equilibrium bulk structure in various concentrated electrolytes are investigated using synchrotron small angle X-ray scattering (SAXS), theory, and molecular simulation. Utilizing SAXS measurements we provide estimates of the Kirkwood Transition (KT) for a variety of aqueous electrolytes (NaCl, CaCl$_2$, SrCl$_2$, and ErCl$_3$). The KT may be defined as the con…
▽ More
Ion-ion correlations, screening, and equilibrium bulk structure in various concentrated electrolytes are investigated using synchrotron small angle X-ray scattering (SAXS), theory, and molecular simulation. Utilizing SAXS measurements we provide estimates of the Kirkwood Transition (KT) for a variety of aqueous electrolytes (NaCl, CaCl$_2$, SrCl$_2$, and ErCl$_3$). The KT may be defined as the concentration above which the ion-ion correlations cease to decay exponentially with a single length scale given by the Debye length $λ_{\rm D}$ and develop an additional length scale, $d=2π/Q_0$ that reflects the formation of local domains of charge. Theoretical models of the KT have been known for decades for highly idealized models of electrolytes, but experimental verification of KT in real electrolytes has yet to be confirmed. Herein, we provide consistent theoretical and experimental estimates of both the inverse screening lengths $a_0$ and inverse domain size, $Q_0$ for the aforementioned electrolyte systems. Taken together, $a_0$ and $Q_0$ are known descriptors of the KT and provide a view into the complexity of ion-ion interaction beyond the well-accepted Debye-Hückel limit. Our findings suggest a picture of interaction for real electrolytes that is more general than that found in idealized models that is manifest in the precise form of the non-local response function that we estimate through the interpretation of the experimental SAXS signal. Importantly, the additional complexity of describing ion-ion interaction of real electrolytes will implicate the short-range ion-ion interactions that can only be computed via molecular simulation and provide a quantitative approach to describe electrolyte phenomena beyond Debye-Hückel theory.
△ Less
Submitted 14 August, 2023; v1 submitted 27 July, 2023;
originally announced July 2023.
-
An Experimentally Driven Automated Machine Learned lnter-Atomic Potential for a Refractory Oxide
Authors:
Ganesh Sivaraman,
Leighanne Gallington,
Anand Narayanan Krishnamoorthy,
Marius Stan,
Gabor Csanyi,
Alvaro Vazquez-Mayagoitia,
Chris J. Benmore
Abstract:
Understanding the structure and properties of refractory oxides are critical for high temperature applications. In this work, a combined experimental and simulation approach uses an automated closed loop via an active-learner, which is initialized by X-ray and neutron diffraction measurements, and sequentially improves a machine-learning model until the experimentally predetermined phase space is…
▽ More
Understanding the structure and properties of refractory oxides are critical for high temperature applications. In this work, a combined experimental and simulation approach uses an automated closed loop via an active-learner, which is initialized by X-ray and neutron diffraction measurements, and sequentially improves a machine-learning model until the experimentally predetermined phase space is covered. A multi-phase potential is generated for a canonical example of the archetypal refractory oxide, HfO2, by drawing a minimum number of training configurations from room temperature to the liquid state at ~2900oC. The method significantly reduces model development time and human effort.
△ Less
Submitted 8 September, 2020;
originally announced September 2020.
-
Ice grains grow by dissolution, ripening and grain boundary migration
Authors:
Henry Chan,
Mathew J Cherukara,
Badri Narayanan,
Chris J Benmore,
Stephen Gray,
Subramanian KRS Sankaranarayanan
Abstract:
Despite the exponential growth in computing resources and the availability of a myriad of different theoretical water models, an accurate, yet computationally efficient molecular level description of mesoscopic grain growth remains a grand challenge. The underlying phase transitions and dynamical processes in deeply supercooled systems are often rendered inaccessible due to limitations imposed by…
▽ More
Despite the exponential growth in computing resources and the availability of a myriad of different theoretical water models, an accurate, yet computationally efficient molecular level description of mesoscopic grain growth remains a grand challenge. The underlying phase transitions and dynamical processes in deeply supercooled systems are often rendered inaccessible due to limitations imposed by system sizes and timescales, which is further compounded by their sluggish kinetics. Here, we introduce a machine-learned, bond-order-based potential model (BOP) that more accurately describes the anomalous behavior, as well as structural and thermodynamical properties of both liquid water and ice, with at least two orders of magnitude cheaper computational cost than existing atomistic water models. In a significant departure from conventional force-field fitting, we use a multilevel evolutionary strategy that trains the BOP model against not just energetics but temperature dependent properties inferred from on-the-fly molecular dynamics simulations as well. We use the BOP model to probe the homogeneous ice nucleation and growth process by performing molecular dynamics on multi-million molecule systems for up to microsecond time scales. These massively parallel, long-time simulations naturally capture the competition between cubic, hexagonal, and stacking disordered ice phases during nucleation and early stages of growth leading to the formation of nanometer sized grains. Subsequently, we elucidate the hitherto elusive mechanism of grain coarsening of the crystallized mixed ice phases, which occur through grain dissolution and Ostwald ripening followed by grain boundary migration.
△ Less
Submitted 1 December, 2016;
originally announced December 2016.
-
Insights into the Hierarchical Structure of Spider Dragline Silk Fibers: Evidence for Fractal Clustering of $β$-Sheet Nano-Crystallites
Authors:
Qiushi Mou,
Chris J. Benmore,
Warner S. Weber,
Jeffery L. Yarger
Abstract:
Spider dragline silk is one of the toughest materials known and understanding the hierarchical structure is a critical component in the efforts to connect structure to function. In this paper, we take the first step in elucidating the hierarchical fractal structure of $β$-sheet nano-crystallites, which form a robust self-similar network exhibiting an non-linear mechanical property. A combined smal…
▽ More
Spider dragline silk is one of the toughest materials known and understanding the hierarchical structure is a critical component in the efforts to connect structure to function. In this paper, we take the first step in elucidating the hierarchical fractal structure of $β$-sheet nano-crystallites, which form a robust self-similar network exhibiting an non-linear mechanical property. A combined small angle X-ray scattering (SAXS) and wide-angle X-ray scattering (WAXS) study of the nano-crystalline component in dragline silk fibers from several species of spiders including, Latrodectus hesperus, Nephila clavipes, Argiope aurantia and Araneus gemmoides is presented. SAXS structure factors exhibit a `lamellar peak' in the q-range from 0.60 to 0.82 nm$^{-1}$ for various spider dragline silk fibers, indicating the presence of strong nano-crystal ordering on the $>$10 nm length-scale. The stochastically reconstructed electron density maps indicate that the $β$-sheet crystals are hierarchically structured as mass fractals and that nano-crystals tend to form 10 to 50 nm sized clusters with long-range crystalline ordering. This nano-crystal ordering along the fiber axis also helps to explain the difference between axial and radial sound velocities recently measured by Brillouin spectroscopy.
△ Less
Submitted 15 July, 2015;
originally announced July 2015.
-
Modelling the atomic structure of very high-density amorphous ice
Authors:
J. K. Christie,
M. Guthrie,
C. A. Tulk,
C. J. Benmore,
D. D. Klug,
S. N. Taraskin,
S. R. Elliott
Abstract:
The structure of very high-density amorphous (VHDA) ice has been modelled by positionally disordering three crystalline phases, namely ice IV, VI and XII. These phases were chosen because only they are stable or metastable in the region of the ice phase diagram where VHDA ice is formed, and their densities are comparable to that of VHDA ice. An excellent fit to the medium range of the experiment…
▽ More
The structure of very high-density amorphous (VHDA) ice has been modelled by positionally disordering three crystalline phases, namely ice IV, VI and XII. These phases were chosen because only they are stable or metastable in the region of the ice phase diagram where VHDA ice is formed, and their densities are comparable to that of VHDA ice. An excellent fit to the medium range of the experimentally observed pair-correlation function g(r) of VHDA ice was obtained by introducing disorder into the positions of the H2O molecules, as well as small amounts of molecular rotational disorder, disorder in the O--H bond lengths and disorder in the H--O--H bond angles. The low-k behaviour of the experimental structure factor, S(k), is also very well reproduced by this disordered-crystal model. The fraction of each phase present in the best-fit disordered model is very close to that observed in the probable crystallization products of VHDA ice. In particular, only negligible amounts of ice IV are predicted, in accordance with experimental observation.
△ Less
Submitted 17 March, 2005; v1 submitted 21 December, 2004;
originally announced December 2004.
