Session P43: Computational design and discovery of novel materials II

Inverse Design of Soft Materials: Crystals, Quasi Crystals, Liquid Crystals

In 1960, Feynman challenged us to think “from the bottom up” and to create new functional materials by directing and manipulating the arrangements of individual atoms ourselves. With recent advances in the synthesis of colloidal nanoparticles and the bottom-up fabrication of nanostructured materials using colloidal self-assembly, we are tantalizingly close to realizing this dream. In this talk, I will show using theory and simulations how one can structure matter over multiple length scales using hierarchical self-assembly. The prediction and design of these structures remains an important challenge for nanomaterials science. I will present a method to predict which structures are stable assuming the shape and interactions between the constituent particles are known, and I will show that particle shape alone can already give rise to a wide variety of structures such as (plastic) crystals [1,2], quasi crystals [3] , and liquid crystals [4,5], which can be classified using machine learning techniques. Subsequently, I will show how one can reverse-engineer the particle shape to stabilize highly exotic liquid crystal phases.


[1] M. Dijkstra, Advances in Chemical Physics 156, 35-71 (2015)
[2] B. de Nijs, S. Dussi, F. Smallenburg, J.D. Meeldijk, D.J. Groenendijk, L. Filion, A. Imhof, A. van Blaaderen, and M. Dijkstra, Nature Materials 14, 56-60 (2015).
[3] D. Wang, T. Dasgupta, E.B. van der Wee, D. Zanaga, T. Altantzis, Y. Wu, G.M. Coli, C.B. Murray, S. Bals, M. Dijkstra, and A. van Blaaderen, arXiv:1906.10088v1 (2019).
[4] M. Chiappini, T. Drwenski, R. van Roij, and M. Dijkstra, Physical Review Letters 123 (6), 068001 (2019).
[5] S. Dussi, M. Dijkstra, Nature Communications 7, 11175 (2016).
  

Hard Particle Colloidal Clathrates with Rotating Guest

Clathrate crystal consisting of polyhedral cages is usually found in gas hydrate or intermetallic systems where the size of particles forming the structures is mostly atomic or molecular length scale. In this work, we report a design strategy of hard colloidal particles that self-assemble into various colloidal clathrate crystals with single or multiple rotating guests. Monte Carlo simulations show that a single component system of hard truncated triangular bipyramids (TBPs) self-assembles into five different clathrates with rotating guests depending on the truncation of the TBP. Truncation of the TBP creates a cavity at the center of the clathrate cage-like motifs, and the cavity is occupied by guest particles when the size of the cavity reaches a comparable size of the guest. Thermodynamic stability of the clathrates is confirmed by Frenkel-Ladd free energy calculations. The dynamics of the guests can be categorized as free rotation, rotation around a fixed axis, or quantized rotation and rattling, with the mode determined by guest/cavity size ratio, prolateness of the cavity and shape of the guest. In addition, when the TBPs are mixed with other hard shapes, clathrates encapsulate these different guest particles.   

Symmetry-based discovery of multicomponent, two-dimensional colloidal crystals

We present a systematic method for computing the ground state phase behavior of multicomponent colloidal materials. In two-dimensions there are exactly 17 “wallpaper groups” which represent distinct combinations of isometries of the Euclidean plane. Using properties of these groups, we develop an algorithm to cover the plane with a fixed number of arbitrary components in all ways that satisfy a desired stoichiometric ratio. These combined symmetry-stoichiometry rules dramatically reduce the number of possible configurations, which generally suffer from a so-called “combinatorial explosion” otherwise making extensive, random structure searching computationally infeasible. With subsequent continuum relaxation, this enumeration approach can predict crystal structures in silico for multicomponent colloidal mixtures. We use this approach to investigate the ground state phase behavior of multicomponent systems inspired by DNA-coated colloidal mixtures, with a focus on stable, low-density “open” crystals. We demonstrate the approach for binary and ternary mixtures at zero ambient pressure to explore how complexity can be achieved through the combination of several components with simple interactions rather than a single component with a more complicated interaction potential.   

Why integer valence point-charge model works so well

Integer valence counting works extremely well in determining stability of chemical compounds. It comes as a surprise that the same integer valence counting also works very well in comparing energies of different lattice structures of the same composition, even though an incomplete (non-integer) charge transfer is expected in realistic systems. Then, why is the point charge model reasonably accurate (even more than employment of real partial charge)? Here, by partitioning the charge density via atom-centered Wannier functions, we justify the integer valence counting approach through a systemic analysis of multi-pole expansion of the Coulomb energy. Our results demonstrate the dominance of the point charge contribution over the finite moment ones, and reveal the important role of local point group symmetry in suppressing first few low-order moments. While the inaccuracy grows expectedly as the system’s covalency increases from LiF, ZnO, GaAs to Si, surprisingly a reasonable accuracy is still achieved by an effective ionic assignment even for Si (as Si⁴⁺Si^4- ). Our formulation also illuminates the removal of the infamous self-interaction in such a simple approach, further explaining its unexpected success in numerous applications.   

Integrated Particle and Field-Theoretic Simulations: A Multiscale Approach to Complex Soft Matter Formulations

Our work investigates the phase-behavior of complex polymeric solutions leveraging the strengths of both particle and polymer field-theoretic simulations. Mesostructured polymeric solutions are difficult to simulate using traditional particle-explicit approaches due to the disparate time and length scales, while the predictive capability of field-theoretic simulations is hampered by the need to specify emergent parameters (e.g., chi parameters) with nonobvious connections to molecular architecture. To overcome the weaknesses of both, we discuss an original way to use small-scale, atomistic simulations to parameterize statistical field theory models. Subsequently, field-theoretic simulations can probe behavior at larger length scales in polymeric solutions efficiently while maintaining a connection to the underlying polymer chemistry. This synergistic approach to polymer simulations opens the door to explore–de-novo–a wide variety of polymeric solution phase behavior. We demonstrate the predictive capability of this approach by reproducing the high-temperature, aqueous, PEO phase diagram, without any experimental data.   

First principles investigations of orthorhombic-cubic phase transition and its effect on thermoelectric properties in cobalt-based ternary alloys.

We screened six cobalt-based 18-VEC systems CoVSi, CoNbSi, CoTaSi (Si-group) and CoVGe, CoNbGe, CoTaGe (Ge-group) by the first-principles approach, with the motivation of stabilizing these orthorhombic phases into the cubic symmetry -- favorable for thermoelectrics. We account the cubic ground state of the Si-group to the interplay of internal pressure and covalent interactions. The principle of reducing covalent interactions will provide insight and could be vital in speeding the search of missing cubic half-Heusler alloys. Meanwhile, the calculated transport properties of all the systems on p-type doping, except CoVSi, are more promising than the well-known CoTiSb. We also provide conservative estimates of the figure of merit, exceeding the CoTiSb. Based on our findings, we suggest possible new phases of ternary compounds for thermoelectric applications.   

Unexpected Photonic Band Gaps in 3D Crystal Structures

Photonic crystals are materials composed of mixed dielectric media that result in the reflection of all electromagnetic waves within a range of wavelengths commensurate with the length scale of the crystal. Such complete photonic band gaps allow for light to be controlled through materials design. Since first theorized in 1987, much effort has been made to define and synthesize photonic crystal structures. In the decades since, many photonic structures have been discovered, often by using naturally occurring crystal structures as templates for design. However, these studies have yet to answer the question: what features of a 3D structure will produce a complete photonic band gap? Here, we present data on over 150,000 potential photonic crystals, and show that complete photonic band gaps are possible for many unexpected structures that have yet to be explored. Our simulations suggest that when designing novel photonic materials, the toolbox of structural templates may be larger and richer than previously thought, widening the field of target crystal structures.   

A mesoscale lattice model and atomistic simulations of controlled drug release

Excipients such as polymers are used for the delivery of poorly soluble drugs to enhance drug bioavailability by limiting drug aggregation/crystallization. I will discuss two distinct approaches, which target different time and length scales of the excipient design problem. First, a two-dimensional lattice model is developed of “ants” (drug) performing a random walk in a lattice containing “walls” (excipient). Ants can be “blind” or “friendly”, corresponding to hydrophilic or hydrophobic drug molecules, respectively. Second, atomistic molecular dynamics simulations are performed on an example system containing acrylic acid oligomers as excipient and doxorubicin as the drug. We mimic drug release by matrix swelling in atomistic simulations using a sequential water removal approach that is a crude but simpler alternative of grand canonical ensemble simulations. Drug release by polymer erosion and drug release in a non-swellable, non-eroding matrix is mimicked in our simulations by variations in polymer/drug concentrations. Together, these two simulation approaches provides design rules for choosing polymer-drug combinations for controlled release.   

Van der Waals Metamaterials

Van der Waals (vdW) heterostructures are a fertile frontier for discovering emergent condensed matter phenomena. They are constructed by stacking elements of a large library of two-dimensional materials, which couple together through vdW interactions. However, fully exploring the vast number of possible combinations within this library is a daunting task. Here we introduce vdW metamaterials to rapidly prototype and screen their quantum counterparts. These layered metamaterials reshape the flow of ultrasound to mimic electron motion. We first present a method to recreate the vdW interaction between layered phononic metamaterials using interlayer coupling membranes, which we can tune to create acoustic analogs of well-known vdW heterostructures, including all configurations of bilayer and trilayer graphene. We then twist coupled metamaterial layers to induce interesting phononic behavior mimicking twisted bilayer graphene. We anticipate vdW metamaterials will inform future electronic devices. Equally, they allow the transfer of useful electronic behavior to acoustic systems, such as flat bands in magic-angle twisted bilayer graphene, which may advance super-resolution ultrasound imaging.   

Development of an implicit solvent model for the interfacial configuration of colloidal nanoparticles and application to the interfacial self-assembly of truncated cubes

This study outlines the development of an implicit solvent model that reproduces the behavior of colloidal nanoparticles at a fluid-fluid interface. The center-point of this formulation is the generalized Quaternion-based Orientational Constraint (QOCO) method. The model captures 3 major energetic characteristics that define the nanoparticle configuration – position (orthogonal to the interfacial plane), orientation, and inter-nanoparticle interaction. The framework encodes physically relevant parameters that provide an intuitive means to simulate a broad spectrum of interfacial conditions. For a wide range of shapes, we are able to replicate the behavior of an isolated nanoparticle at an explicit fluid-fluid interface, both qualitatively and quantitatively. Using the family of truncated cubes as test-bed, we analyze the effect of change in the degree of truncation on the potential landscape. Furthermore, we model the self-assembly of an array of cuboctahedra to provide corroboration to the experimentally observed honeycomb and square lattices. Finally, by exploring a broader range of interfacial conditions, we identify and suggest the assembly mechanism for a set of novel superlattice configurations.   

First-principles study on the stable hydrogen configuration in SrVO2H

Mixed-anion compounds, which contain multiple anionic species in a single phase, such as oxyfluorides and oxyhydrides, have been gathering growing attention as a new platform of materials science [1]. While it is an attracting idea to employ the anionic degrees of freedom for tuning the cation environment, much less is known about mixed-anion compounds than oxides. In particular, understanding and controlling the stable configuration of multiple anions in crystal is one of the most challenging issues. For example, hydrogen atoms are aligned along one direction in some oxyhydrides such as SrVO₂H [2-3], but are randomly distributed in some oxyhydrides such as SrCrO₂H [4]. In this study, we investigate the stable hydrogen configuration in SrVO₂H by performing first-principles calculation of the total energy for several hundreds of possible crystal structures with different hydrogen configurations. In this talk, we will discuss what determines the stability of the hydrogen configuration in this material. References: [1] H. Kageyama et al., Nat. Commun. 9, 772 (2018). [2] F. D. Romero et al., Angew. Chem. Int. Ed. 53, 7556 (2014). [3] J. Bang et al., JACS 136, 7221 (2014). [4] C. Tassel et al., Angew. Chem. Int. Ed. 53, 10377 (2014).   

Understanding Li-ion Diffusion Through Artificial SEI Coating Layers

Li-ion batteries show promise for storing large amounts of energy, yet, they are limited by the capacity of the graphite anode. High capacity alternatives, such as Li or Si, pose additional issues due to volume expansion and dendrite growth. In both Si and Li, Al₂O₃ coatings can be used as a protective layer on the anode [1,2]. We apply a combined theoretical and experimental study to both predict and confirm the structure of this interface, a combination which has been applied numerous times in this field [3]. Using Al K-edge X-Ray absorption spectroscopy calculated with DFT, we have shown how individual layers of Al₂O₃ grow via atomic layer deposition, and how the structure of the layers is dependent on the number of layers. This result was confirmed experimentally, and suggests that the performance of this coating layer is dependent on the number of layers grown. Finally, we calculate the Li⁺ diffusivity across the interface, and show how the interface changes during ion diffusion. These results provide both a computational and experimental explanation for the enhanced performance of Al₂O₃ coated high-capacity anodes.
1. He Y, et al. Adv Mater, 23, 43 (2011)
2. Chen, L, et al. J Mater Chem A, 5 (24), pp.12297-12309 (2017)
3. Harper, AF et al. JM Tech Rev (Under review Oct 2019)   

Finger prints based biasing for finding complex reaction pathways.

Determining the pathway of a reaction/transformation is of great importance in chem-
istry, physics and materials sciences. However, due to the indistinguishability of atoms,
finding complex reaction and transformation pathways, containing a large number of in-
termediate states, is difficult within the existing methods at the density functional theory
level.
we have resolved this issue by introducing a bias that is invariant under atomic index
permutations and that can target a single well defined configuration as the final configu-
ration of a chemical reaction or physical transformation. In this way we can overcome the
index mapping problem. The forces arising from the bias, by construction, do not depend
on the indexing of the atoms. We have thus reduced the combinatorial atomic indexing
problem, that has an exponential scaling, to a global minimization problem on a biased PES
involving an indexing invariant penalty function. The later problem can be solved easily in
practice. The penalty function we propose is universal and can be applied to any reaction
or transformation. We expect that this method will give atomistic insight into complex
reaction pathways i.e. in catalysis as well as complex phase and shape transformations.