Symposium TC1 : In Silico Materials Chemistry

We discuss the results of first principles simulations of interfaces present in organic photovoltaic blends, between nanoparticles and ligands in inorganic, nanostructured solar cells and between photo-absorber, catalysts and water in photo-electrochemical cells. We focus on the identification of descriptors to be possibly used to optimize photo-conversion properties.

CdTe is a widely-used photovoltaic material, due to its high efficiency and low manufacturing cost. The high efficiencies in CdTe thin films are achieved by the CdCl₂ treatment, after which it is believed that Cl atoms substitute the Te atoms at the grain boundaries and enable charge channeling. However, the practical efficiencies of polycrystalline CdTe photovoltaic cells are still well below the theoretical limit, indicating possible room for improvement. A fundamental understanding of the role of dopants at dislocations and grain boundaries on the electronic structure of CdTe using atomistic-level characterization, including microscopy and first principles modeling, is essential to improve the photovoltaic efficiency. In the present work, we constructed atomistic grain boundary and dislocation core models directly from the STEM images using image analysis methods and crystallographic information at the grain boundary interfaces [C. Sun, et al, Sci. Rep. 6 (2016) 27009.]. First principles density functional theory (DFT) calculations are used to compute the electronic structures of large-scale grain models. We incorporate various anion and cation substitutional dopants at the grain boundaries. To understand charge carrier interactions upon doping, we report the changes in the electronic density of states (DOS) and electrostatic potential profiles at CdTe grain boundaries. We predict effective elements to improve CdTe photovoltaic efficiency based on the defect thermodynamics and pertaining changes in electronic structure. The implications of these electronic structure changes at grain boundaries on photovoltaic performance, and corresponding strategies to improve performance, are discussed. ACKNOWLEDGEMENT: We acknowledge funding from the DoE Sunshot program under contract # DOE DEEE005956. Use of the Center for Nanoscale Materials was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. The submitted abstract has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government.

Colored nitrogen-vacancy (NV) centers in nanosize diamonds (~5 nm) are promising probe materials, because their optical transitions are sensitive to mechanical, vibrational and spin changes in the surroundings. In this work, a linear response time-dependent density functional theory approach is used to describe the optical transitions in several NV-doped diamond quantum dots (QDs) in order to investigate size effects on the absorption spectra. By computing the full optical spectrum up to band-to-band transitions, we analyze both the localized “pinned” mid-gap and the charge-transfer excitations for an isolated reduced NV center. Sub-band charge-transfer excitations are shown to be size dependent, involving the excitation of the dopant sp3 electrons to the diamond conduction band. Additionally, the NV-doped systems exhibit characteristic NV-centered excitations whose experimental energies are reproduced well and do not depend on QD size. The local symmetry breaking at the NV center gives rise the observed energy splitting of the vertical excitation energies of the mid-gap transitions. The NV migration dynamics are investigated with the ab initio direct molecular dynamics method. The results show several NV migration pathways, including concerted and asynchronized mechanisms.

In organic photovoltaic (OPV) cells, the photocurrent generation efficiency depends critically on charge–separation (CS) of the electron-hole (exciton) at the interface between electron donor and acceptor phases. This CS process depends on charge transfer (CT) effects taking place at the donor–acceptor interface. In order to theoretically devise new OPV cells with high efficiency, it is thus important to describe these CT effects adequately. In this framework, the Green’s function based many-body perturbation theory (MBPT) constitutes a choice tool for computing the electronic and optical properties of materials [1]. Since the mid-80s, this approach has proven very accurate for extended semiconductors. In particular, the so-called GW/BSE formalism correctly reproduces experimental electronic and optical band gap in many studied systems. Recently, an efficient “auxiliary Gaussian basis” implementation of the GW/BSE formalism (the Fiesta code [2]) has been developed allowing to treat molecular systems consisting of up to a few hundred atoms. The present project aims at studying the electronic and optical properties of the interface between various theoretically designed donors and an acceptor (fullerene C60) using the Fiesta code. Tens of thousands of molecules have been already tested in a multiple-step screening within the Harvard Clean Energy Project (CEP) [3], leading to a ranking of potential candidates for efficient organic photovoltaics. However, the calculations have been performed relying on density-functional theory which is known to underestimate the band gap and not to describe electron-hole interaction and CT effects correctly. Here, we focused on the top 21 donor molecules from the CEP ranking. In a first step, we computed the electronic and optical properties for the donor alone. In a second step, we considered the complete fullerene-donor complexes and computed highly accurate band gaps, exciton binding and charge-transfer energies. The Harvard group computed the equilibrium geometries of the complexes and the UCL team carried out the corresponding GW/BSE calculations. From these results, we were able to evaluate possible efficiency loss pathways.

[1]Electronic excitations: density-functional versus many-body Green’s-function approaches, G. Onida, L. Reining, A. Rubio, Rev. Mod. Phys., 74, 601.
[2] Charge-transfer excitations in molecular donor-acceptor complexes within the many-body Bethe-Salpeter approachX. Blase and C. Attaccalite, Applied Physics Letters 99, 171909 (2011).
[3] Lead candidates for high-performance organic photovoltaics from high-throughput quantum chemistry – the Harvard Clean Energy Project, Johannes Hachmann et al. ,Energy Environ. Sci., 2014, 7, 698-704.

The structure of hybrid perovskites allows to create materials with different stoichiometry changing the ratio of halides, organic cations or metals[1,2]. The opto-electronic properties and the stability of the materials can then be controlled to improve the photovoltaic devices[3]. In this study we employed DFT ab-initio calculations with PBEsol exchange-correlation functional, to understand the thermodynamic properties with the generalized quasi-chemical approximation (GQCA). We first considered as initial test the solid solution MAPb(I_1-xBr_x)₃ , then we extended the procedure to include the other binary perovskite (Cl-I and Cl-Br).
For the three materials we calculated the enthalpy of mixing of the alloys and built the first reported phase diagram for hybrid perovskites. For the initial compound, we can conclude, from the analysis of the phase diagram, that the alloy is more stable in the Br-rich region and phase segregation occurs below the critical temperature of 343K. These observations hold also for the other binary compounds, but we observed an increase in the miscibility as the difference between the halide size decreases.
This behaviour is mainly due to the presence of ordered structures which are stable against the mixing of the pure halide perovskites. Similar results has been obtained previously by other simulations on the completely inorganic perovskite CsPb(I_1-xBr_x)₃[4]. The same trend has been observed experimentally in analogous materials where intermediate halide compositions lead to optical bleaching[5].
These results shed light on the mechanisms behind the observed anion redistribution upon heating and/or illumination responsible for degradation and phase separation that have been reported in devices.


References
[1] H.S. Kim, C., et al., Scientific reports (2012), 2.
[2] H.J. Snaith, The Journal of Physical Chemistry Letters (2013), 4, 3623-3630.
[3] A. Sadhanala, et al., The Journal of Physical Chemistry Letters (2014), 5(15), 2501-250
[4] W.J. Yin, et al., Journal of Materials Chemistry A (2015), 3(17), 8926-8942.
[5] McMeekin, David P., et al., Science 351 (2016), no. 6269, 151-15

Different industrial products require materials with very low thermal conductivity, e.g. thermal-barrier coatings in turbines or thermoelectric materials, or high thermal conductivity, e.g. semiconductor technology and heterogeneous catalysis. However, an ab initio theory that can describe materials belonging to the whole range of such thermal conductivities was lacking so far.
Recently we developed an ab initio Green-Kubo approach (based on the local representation of the stress tensor) together with an asymptotically exact, robust extrapolation approach that enables us to accurately describe and predict thermal conductivities at all realistic and relevant temperatures, and for different classes of materials. The calculations also explain the atomistic and electronic origin of anharmonicities and predict how they can be tailored by doping.
(*) Work done in collaboration with Christian Carbogno (FHI-MPG, Berlin) and Rampi Ramprasad (University of Connecticut).

Conjugated polymers are attracting an increasing commercial interest because of the huge number of applications such as solar cells, field-effect transistors and light-emitting diods.
A deep knowledge of the photoinduced chemical processes characterizing such compounds is necessary in order to improve the design of more performing devices, and this is main reason of a growing scientific interest in such class of molecules on both an experimental and theoretical point of view.
Nowadays, through modern and time-resolved spectroscopic techniques such as femtosecond stimulated Raman spectroscopy¹, it is possible to observe the formation and dynamics of photoinduced species such as excitons, polarons or other transient charged species formed in these systems, although the molecular mechanisms driving their formation, dynamics and decay are still unclear.
We propose an in-silico study performed through the non-adiabatic Ehrenfest dynamics² technique coupled for the first time with a non-stationary signal analysis (the Wavelet Analysis³) to simulate the polaron pair formation and to follow its dynamics and recombination in a model oligothiophene system.
In this study we show that the polaron pair dynamics is coupled with structural changes mainly involving inter-rings coplanarity. The time-resolved vibrational analysis allows to follow the polaron pair induced frequency shifts and to make comparisons with experimental results.
Our work clearly suggests that the coupling of the electron-nuclear dynamics is of key importance to understand the main processes ruling the polaron pair dynamics and recombination. The interesting insights provided by this study in photoinduced processes of conjugated polymers can be employed to improve the design of new devices.


1 – T. J. Magnanelli and A. E. Brgg, J. Phys. Chem. Lett., 2015, 6, 438-445.
2 – X. Li, J. C. Tully, H. B. Schlegel and M. J. Frisch, J. Chem. Phys., 2005, 123, 084106-084112.
3 – A. Petrone, G. Donati, P. Caruso and N. Rega, JACS, 2014, 136, 14866-14874.

The upper efficiency limit of solar water splitting is thought to be 30% for a single semiconductor material with a band gap of 1.6 eV. However, in addition to exhibiting an optimal band gap for solar absorption, semiconductor photoelectrodes must exhibit excellent oxidative/reductive stability in contact with aqueous electrolyte solutions. For thermodynamic stability, a semiconductor’s reductive and oxidative decomposition potentials must be more negative than the semiconductor’s conduction band-edge for water reduction or more positive than the semiconductor valence band-edge potential for water oxidation, respectively. We have predicted the oxidative/reductive stability in contact with aqueous electrolyte solutions for more than 200 semiconductors compounds reported in literature for solar water splitting. We found that all compounds with their valence band maximum (VBM) above 2.48 eV relative to NHE are unstable in aqueous electrolyte solutions, but compounds with their valence band maximum below 2.48 eV relative to NHE may, not necessary, stable. Due to the VBM and Gibbs free energy is correlated for a compound, we shift the VBM upwards in energy by engineering the band structure of a compound accompanying the decrease of its stability. For instance, if we artificially shift the VBM of TiO2 upwards leading to it unstable once its VBM passes around 2.48 eV. Furthermore, oxynitride materials have higher VBM than their oxide counterparts, but the former becomes unstable whenever the latter is stable. Because the conduction band minimum of compounds must also above the water reduction potential (0 eV relative to NHE), the ideal smallest bandgap for solar water splitting electrode compounds must larger than 2.48 eV. The corresponding conversion efficiency is about 7%, which becomes a new upper efficiency limit to solar water splitting.

For optical and excitonic devices understanding the competition between radiative emission rates and various non-radiative decay pathways is key for rational design of device efficiencies. In particular, organic molecules in nonpolar media of the sort that are often used in luminescent solar concentrators or organic light-emitting diodes undergo internal conversion (S1 to S0) and intersystem crossing (S1 to T1).
Estimating the rates of these processes has proven to be very difficult, as heuristics such as the energy gap law can be unreliable. We will discuss these shortcomings, reviewing experimental data in view of new time-dependent density functional theory (TDDFT) electronic structure calculations. Additionally, we will inspect evidence suggesting that higher-lying excited states (e.g., S2, Tn) are needed to explain non-radiative decay rates, a feature that is difficult to observe without using TDDFT.

Sulfur (S) serves as a promising cathode material in Li-ion batteries owing to its abundance on earth, low cost and high theoretical specific capacity ~ 1670 mAhg^-1, which is 3-5 times higher than that of current commercial Li-ion batteries. Nowadays, the most popular strategies of using S cathode are based on producing nanostructured carbon matrices (i.e. hollow carbon nanospheres and nanofibers) to sustain S cathode loading. However, the possible stress evolution and mechanical degradation of the confined S cathode in those carbon matrices have never been explored before. In addition, the associated structural and conductivity changes of the confined S cathode during the lithiation/delithiation process plays a significant role in the battery performance. With the above in mind, here we conduct reactive molecular dynamics simulations to investigate the microstructural and stress evolution of the confined S cathode during lithiation/delithiation process. Simulation results indicate an unusual stress relaxation state in Li_xS compounds at lower Li concentrations (x ≤ 0.7). The strength of corresponding Li-S compounds also increases with respect to the Li concentration.

Understanding structure-property relationships in such applications as energy storage and catalysis represents a significant challenge for theory and experiment. This is further complicated by the fact that the pertinent structural motifs may well be different under the environment in which they operate or evolve dynamically. For experiments, this is driving on-going development of a suite of in situ and operando tools directed to such probes as X-ray photoelectron spectroscopy, X-ray absorption spectroscopy and transmission electron microscopy (including local electron energy loss spectroscopy). Data from each of these tools will provide a slice of information. Integration of this information, in particular the identification of the pertinent structure motifs that drive the target chemical or electronic functionality, is a shared challenge in which theory and simulation can play a significant role. Computations can be used both in the process of structure discovery (identification of plausible structure motifs) as well as in making quantitative predictions of measured observables based on specific structures to close the validation loop. While feasible for some cases, each component of these capabilities requires significant, on-going research to develop.

I will illustrate some of the potential as well as some of the challenges involved in the partnership between computational (‘in silico’) materials simulation and in-situ materials characterization. Specifically, I will discuss recent research directed to discover the relevant local structures in a prototype energy storage material, Li_xAg_yMn₈O₁₆. In this collaboration [1], the central experimental approach is in situ lithiation under transmission electron microscope observation. This data is supplemented by ex situ experiments for reference samples, including electrochemical, X-ray diffraction and extended X-ray absorption fine structure measurements. Using density functional theory based calculations, we analyze the structure motifs as a function of Li and Ag concentration, as well as the role of oxygen vacancies. We develop a pseudoternary phase diagram for Li_xAg_yMn₈O₁₆. Based on computed properties of the stable phases, we are able to interpret the available data in a consistent manner. In particular, we are able to propose a basic interpretation for the in situ observed lithiation dynamics.

[1] Theory: M. Kaltak, M. Fernandez-Serra and Q. Meng; Experiment: F. Xu, L. Wu, J. Huang, A. B. Brady, D. C. Bock, J. L. Durham, A. C. Marschilok, E. S. Takeuchi, K. J. Takeuchi and Y. Zhu.


This work was performed as part of the Center for Mesoscale Transport Properties, an Energy Frontier Research Center supported by the U. S. Department of Energy, Office of Science, Basic Energy Sciences, under award #DE-SC0012673. Work performed in part at the Center for Functional Nanomaterials, which is a U.S. DOE Office of Science Facility, at Brookhaven National Laboratory under Contract No. DE-SC0012704.

Control over crystal structure is one of the primary targets in crystal growth, alongside compositional and morphological selectivity. However, the present lack of predictive handles over crystal structure, as well as a lack of mechanistic understanding of crystal phase selection, precludes the predictive synthesis of novel polymorphic materials. We consider the highly polymorphic class of MnO₂-framework structures and derive the thermodynamic conditions favoring the formation of its most common polymorphs (beta, gamma, ramsdellite, alpha, delta, and lambda) in aqueous conditions, as well as an energetic ranking of the various framework structures across common precursor solution conditions. Our data enables a mechanistic analysis of MnO₂ synthesis routes and phase transformations, previously documented in a number of in-situ studies, by separating thermodynamics from kinetic effects. As a result of our comprehensive energetic analysis, we are able to reproduce specific synthesis conditions which are known to yield each MnO₂ polymorph, offering both a quantitative synthesis map for this important class of functional oxides, and a roadmap for the various phase transformations that may occur in this system

The ability to predict thermodynamic and kinetic properties of electrode materials from first principles is providing opportunities to explore and design new battery concepts. Electrode materials for Li, Na and Mg ion batteries undergo a series of phase transformations as a result of large changes in concentration during each charge and discharge cycle. The mechanisms of these phase transformations remain poorly understood. Non-intercalation reaction mechanisms are often accompanied by substantial hysteresis and sluggish rate capabilities. Experimentally elucidating the rate limiting steps and the kinetic factors responsible for hysteresis is challenging. First-principles computational tools are proving invaluable as a complement to experiment in isolating crystallographic, thermodynamic and kinetic properties of new electrode materials that cause hysteresis and slow charge and discharge rates. In this talk we will illustrate how the application of first-principles statistical mechanics can generate crucial insights about dynamic processes within electrodes and enable the design of new electrochemical energy devices. The approaches rely on coarse graining schemes to connect properties at the atomic and electronic scale to phenomenological descriptions of kinetic processes.

For the practical use of silicon as anodes for Li-ion batteries, understanding their lithiation and delithiation mechanisms at the atomic level is of critical importance. Also, understanding the nature and formation of the solid-electrolyte interface (SEI) formed in Li-ion batteries is very significant for improving their functionality. To accurately predict the lithiation/delithiation behaviors of Si anodes and SEI formations between the anode and electrolytes, a computer simulation method to predict chemical reactions in large-scale systems is necessary. In this aspect, a molecular dynamics simulation with first-principles based reactive force fields (ReaxFFs) should be the best choice. In this talk, I will present recent ReaxFF works regarding lithiation/delithiation of pristine, carbon-coated, and oxidized Si nanowires, along with the SEI formation on Si electrodes. And then, I will introduce a multi-scale simulation platform called iBat (battery.vfab.org) for Li-ion battery that has been developed in our research center.

Knowledge of work functions of different crystallographic orientations and terminations is valuable for engineering surface and interfacial properties for applications including charge injection layers, electrocatalysts, and thermionic electron or field emission-based high power devices. Perovskite oxides containing transition metals are promising materials in a wide range of electronic and electrochemical applications. [1-4] However, neither their work function values nor an understanding of their work function physics have been established. Here, we predict the work function trends of a series of perovskite (ABO₃ formula) materials using Density Functional Theory, and show that the work functions of (001)-terminated AO- and BO₂-oriented surfaces can be described using concepts of electronic band filling, bond hybridization, and surface dipoles.[5] The calculated range of AO (BO₂) work functions are 1.60-3.57 eV (2.99-6.87 eV). We find an approximately linear correlation (R² between 0.77-0.86, depending on surface termination) between perovskite work functions and the relative position of the bulk oxygen 2p band center and the Fermi level, where an oxygen 2p band center closer to the Fermi level results in a higher work function. This correlation with oxygen 2p band center enables both understanding and rapid prediction of trends in work function. Furthermore, we identify SrVO₃ as a stable, low work function, highly conductive material. Undoped SrVO₃ has an intrinsically low AO-terminated work function of 1.86 eV, and Ba-doped SrVO₃ is predicted to have a very low work function of just 1.07 eV. These properties make SrVO₃ a suitable candidate material for a new electron emission cathode for application in high power beam devices, and as a potential electron emissive material for thermionic energy conversion technologies.[5]

References:
[1] T. Susaki, A. Makishima, H. Hosono, Phys. Rev. B 2011, 84, 115456
[2] J. Suntivich, H. A. Gasteiger, N. Yabuuchi, H. Nakanishi, J. B. Goodenough, Y. Shao-Horn, Nat. Chem. 2011, 3, 546
[3] A. M. Kolpak, S. Ismail-Beigi, Phys. Rev. B 2012, 85, 195318
[4] J. Teresa, A. Barthelemy, A. Fert, J. Contour, F. Montaigne, P. Seneor, Science 1999, 286, 507
[5] R. Jacobs, J. Booske, D. Morgan, Adv. Func. Mat. 2016

The prediction of electronic structure, doping, defects, and interface properties based on first principles calculations becomes increasingly instrumental for the design and discovery of oxide semiconductors for a wide range of applications from transparent electronics to solar energy conversion. Relative recent developments in electronic structure calculations have enabled much more quantitative predictions than in the past. A few examples from our recent research activities illustrate these developments:
As an example of an emerging ultra wide gap semiconductor, Ga₂O₃ is recently attracting high interest, e.g., for power electronics. Introducing electron carriers into Ga₂O₃ by doping to a similar level as in, say, ZnO or In2O3 remains a challenge. To aid the optimization in the parameter space (temperature, pO2, choice of dopant, and dopant concentration), we develop a first principles defect phase diagram based on supercell calculations and thermodynamic modeling of (constrained) defect equilibria, taking into account dopant-defect interactions in the relatively complex crystal structure of beta-Ga2O3.
The design of functional semiconductors has long relied on alloying, which was instrumental, e.g., for the lighting revolution by full spectrum light emitting diodes. Traditionally, such alloying is done between isostructural materials (e.g., wurtzite GaN/InN, zinc-blende GaAs/AlAs, etc.). Our recent research explores hetero-structural alloys, where the alloying causes as composition induced phase changes, which can be exploited to enable new functionality. As an example, the rock-salt to wurtzite transition of MnO alloyed with ZnO has enabled initial photo-electrochemical (PEC) solar water splitting applications, due to the superior photon absorption and hole transport properties of the tetrahedral phase [1]. An analysis of the thermodynamic properties reveals that hetero-structural alloys can exhibit a novel phase diagram behavior that enables the synthesis of metastable phases over an unusually wide composition range. These new features can be exploited for the predictive design of novel functional materials away from equilibrium.
[1] H. Peng, P. Ndione, D.S. Ginley, A. Zakutayev, S. Lany, Phys. Rev. X 5, 021016 (2015).

Perovskite oxides (ABO₃), with a transition metal cation on the B site (B = Cr, Mn, Fe, Co, Ni) can exhibit either p- or n-type semiconducting behaviour, depending on their composition. p-type semiconductors present a challenge because in most wide gap binary oxides, the top of the valence band consists primarily of O 2p states. However, in perovskite oxides, the top of the valence band can consist of O 2p and transition metal 3d states. Their p-type conductivity can be improved by doping on the La site with divalent cations (M = Ca²⁺, Sr²⁺, Ba²⁺), where they can find application as cathodes in solid oxide fuel cells. In addition, Sr-doped LaCrO₃ has recently been suggested as a potential p-type transparent conducting oxide (TCO). TCOs have a wide variety of applications in optoelectronics, including photovoltaic cells and flat panel displays. TCO materials have conductivity greater than 1000 S cm^–1, with carrier concentrations in the region of 10²⁰-10²¹ cm^–3 and a band gap of greater than 3 eV, allowing them to transmit visible light. Currently only n-type materials, such as Sn-doped In₂O₃ and Al-doped ZnO, exhibit high electronic conductivity and optical transparency required for use in these devices. TCOs may be used for transparent electronics, but in order to produce pn junctions, a suitable p-type material with properties to match those of the n-type materials is required. For LaCrO_3, Sr doping is reported to effectively dope holes into the top of the valence band, resulting in p-type conductivity.[1] Thin films of Sr-doped LaCrO₃ were shown to exhibit hole concentrations of 10²⁰ to 10²¹ cm^–3 and an increase in conductivity, from 1.2 to 54 S cm^–1 with increase in Sr content from 4 to 50%.

Defect analysis of pure LaCrO₃ and LaCrO₃ containing Sr defects has been carried out using PBEsol + U calculations.[2] In order to understand the origin of the improvement in electronic conductivity and optical properties that have been observed experimentally when LaCrO₃ is doped with Sr, all possible intrinsic and extrinsic defects have been calculated. The chemical potential dependence of defect formation, electronic structure, origin of the charge carriers, and defect stabilities have been investigated, by calculating formation energies and transition levels for each of the defects. This allows the potential of Sr-doped LaCrO₃ as a p-type semiconductor to be assessed. The p-type vacancies have low formation energies, and very low transition levels are observed for p-type defects under oxygen-rich conditions, indicating the potential of LaCrO₃ for p-type conductivity.

[1] Zhang et al., Adv. Mater, 27, 5191-5195 (2015)
[2] J. P. Perdew et al., Phys. Rev. Lett., 2008, 100, 13640

Although promising, lithium-sulfur batteries have serious issues due to dendrite
formation during charge of the battery leading to short-circuit of the battery during cycling and
degradation of the Li-metal anode due to decomposition of polysulfides shuttling from the
cathode to the anode. One additive that has shown positive effects to reduce the dendrite
formation and decomposition of polysulfides over the anode is lithium nitrate (LiNO3 ). Since
experimental results show that LiNO3 increases continuously over time the resistivity of the
anode, its decomposition is expected. Thus we study the reactions of this additive on the metal
anode surface, with or without polysulfide species. The Solid Electrolyte Interphase (SEI)
formed due to electrolyte decomposition and the products of the polysulfide and additives
reactions on the anode surface are critically investigated with density functional theory and ab
initio molecular dynamics simulations.

Rare earth dopants such as dysprosium, gadolinium and yttrium are used to reduce the likelihood of electrical breakdown of barium titanate in capacitors. Experiments suggest [1] that they reduce the oxygen diffusion rate through vacancy trapping and so prevent build-up of defect concentrations at electrodes. Although the rare earths are incorporated into the lattice by a self-compensation mechanism, there remains a residual concentration of oxygen vacancies arising either from non-stoichiometry or the presence of impurities. In this work we use static lattice simulations to investigate how dopants can trap oxygen vacancies and so reduce the probability of electrical breakdown in barium titanate. This requires us to consider both the effect of both cation and oxide migration. The distribution of dopants (whether as isolated ions or clusters) is controlled by the migration of dopants in the lattice during material processing. The resulting defect clusters then trap the oxide vacancies.

Since the defect migration energies were expected to be high (over 4eV in the case of cation diffusion [2]), we have used static lattice methods to investigate the migration behaviour. Using the GULP code [3], we have shown that defect clusters based on mid-sized rare earths (such as dysprosium, gadolinium and yttrium) are effective in trapping oxide vacancies, particularly when present as self-compensating pairs on adjacent A and B type sites in the perovskite lattice. Simple estimates of the cluster cross-section show that small concentrations of rare earth (about 3%) should be sufficient to prevent significant oxygen migration. We also present results on the cation diffusion to show which rare earth cations can diffuse through the oxide on the timescales required to ensure formation of the relevant defect clusters during processing.

References
[1] J. Itoh, D.C. Park, N. Ohashi, I. Sakaguchi, I.Yashima, H. Haneda, J. Ceram. Soc. Japan, 110, 495–500, (2002)
[2] S. Beauchesne and J.P. Poirier, Phys. Earth & Planetary Sci. 55 187-199 (1989)
[3] J.D. Gale and A.L. Rohl, Mol. Simul., 29, 291-341 (2003)

There are a large variety of materials being developed for application in solar cells. The majority are based upon naturally occurring minerals (so-called solar mineralogy). The general procedure has been to take a multi-component system and tune the chemical composition to optimise optical absorption for the terrestrial solar spectrum. Other factors also determine whether a material can be practically employed in a photovoltaic or photoelectrochemical system, for example, the absolute band energies (work functions), defect physics, and chemical stability.

We will present recent progress into computing photovoltaic performance descriptors from materials simulations [1-6], including advances in structure-property relationships in the kesterite (e.g. Cu₂ZnSnS₄) and perovskite (e.g. CsSnI₃ and CH₃NH₃PbI₃) families, in addition to the herzenbergite (SnS) system. New directions in the field, including the development of novel photoferroic materials, will also be addressed.

[1] “Kesterite thin-film solar cells: Advances in materials modelling of Cu₂ZnSnS₄” Advanced Energy Materials 2, 400 (2013)
[2] “Band alignment in SnS thin-film solar cells: Origin of the low conversion efficiency” Applied Physics Letters 102, 132111 (2013)
[3] “Atomistic origins of high-performance in hybrid halide perovskite solar cells” Nano Letters 14, 2584 (2014)
[4] “The dynamics of methylammonium ions in hybrid organic–inorganic perovskite solar cells” Nature Communications 6, 7124 (2015)
[5] “Ferroelectric materials for solar energy conversion: photoferroics revisited” Energy & Environmental Science 8, 838 (2015)
[6] "Quasi-particle electronic band structure and alignment of the V-VI-VII semiconductors SbSI, SbSBr, and SbSeI for solar cells" Applied Physics Letters 108, 112103 (2016)

Hydrophobic blocks may be covalently linked to DNA to create amphiphilic molecules to allow self-assembly. In contrast to DNA tile and origami strategies, DNA amphiphile self-assembly does not require design of specific sequences for base pairing, but utilizes hydrophobic and hydrophilic interactions of the DNA diblocks. However, self-assembly of these materials is usually limited to simple shapes such as micelles. In order to increase the structural complexity of self-assembled structures, researchers have incorporated hydrophobic spaces attached to lipid like tails. However, how these factors influence self –assembly is not well-understood. Simulation techniques offer the opportunity to better understand self-assembly of these systems.

All-atom methods are the most detailed when describing DNA motion and the chemical interactions. However, they are limited to systems containing up to a few million atoms. Large-scale DNA materials typically consist of thousands of molecules. Thus, self-assembly and hybridization for dynamic nanotechnology requires timescales which are not accessible to all atom methods. One solution is to use coarse grained models where many atoms are grouped to create particles that behave approximately the same way. The significant decrease in computing time associated with CG allows for the simulation of larger systems for longer periods of time at the cost of atomic resolution. Our group has previously developed a dissipative particle dynamics (DPD) model to describe the self-assembly of polyelectrolyte diblocks with differing solution ionic strength. However, this method lacks specific chemical detail of the hydrophobic segment. Thus, there exists a need for a model which is able to describe the details of polyelectrolyte assembly, particularly for systems containing complex hydrophobic tail structures.

We developed a new multi-resolution method that allows to capture the self-assembly of micellar core on atomsitic level and coarse-grained represnetation of charged corona. Here, we present results using our hybrid resolution model and and dissipative particle dynamics of the self-assembly of DNA diblock materials. We examine the effect of hydrophobic core identity and sequence of the attached DNA block.

Self-assembling polymeric based macromolecules have shown a great deal of promise for a multitude of novel material applications. While there have been some advancements, many obstacles hinder material development including a shortage in the fundamental understanding of how self-assembly occurs. This establishes the need for developing a method of computationally modeling such interactions. Dissipative Particle Dynamics (DPD) is a coarse-grained meso-scale modeling technique which has been used to study a wide range of phenomena, largely focused on solution and dilute polymeric behaviors. The process of coarse-graining involves using representative particles to act as a substitute for whole molecules of which atomistic resolution is deemed unnecessary. While DPD has proven to be a reliable tool for modeling phase behavior, it has previously lacked the ability to capture long-range interactions without extremely high computational costs. Such interactions are vital to properly reproducing the assembly of relatively complex polymer chains. Very recently we developed a method and used it to examine single-chain polyelectrolytes by implementing an implicit solvent ionic strength model in combination with a soft repulsive potential. This method was able to appropriately reproduce the morphological impacts on polyelectrolyte assemblies by imparting solvent effects on the self-repulsion of the individual electrolyte monomers. Our work utilizes a similar approach to model an amphiphilic self-assembling polymeric system of diblock ssDNA (ssDNA bonded to a synthetic hydrophobic polymer). A two coarse-grained particle per monomer approach was used to represent the diblock ssDNA involving the phosphate and sugar ring constituting one particle while the second consists of the nucleobase. The diblock ssDNA system started from a randomly distributed conformation throughout a periodic system then assembled into micelles. Upon micelle assembly, similarly coarse-grained complimentary ssDNA were added, resulting in a bridged network of micelles. The results from this model allow for the observation of assembly efficiency as a function of the relationship between the length of ssDNA constituting the micelle to that of the ssDNA complimenting those chains. This method ultimately provides insight into promoting and controlling the assembly of these networks and build a foundation to which these components can be used to generate impactful materials.

Diblock copolymers are a subject of wide interest due to the the microstructures that they form. The microstructures of these polymers have current and potential applications in fields such as organic electronics, nanolithography, and catalysis. Controlling the results of the self assembly process is critical for all of these applications. A particularly compelling means of control is the use of confinement. The constraints that arise within a confined space have been shown to open up a much wider range of structures. The challenge becomes finding which microstructures are available and the conditions that lead to a desired one. The size of the confined space, the interactions between the polymer and the walls, and the characteristics of the polymers themselves all influence the final microstructure. Computational simulations are an excellent way to find unique microstructures within this space. We use self consistent field theory to efficiently model polymers and their interactions and predict the resulting configuration. Here we present results of an exploration of diblock polymers under a range of different confinement sizes and geometries.

The unique solvation behavior of C₆₀ fullerene, due in part to the molecule’s size between the colloidal regime and that of the surrounding solvent molecules, has received much attention since C₆₀ was found to exhibit exceptionally promising properties for both biological and electronic applications. Many derivatives of C₆₀, in which one or more chemical adduct has been bonded to the fullerene cage, have been synthesized to enhance these properties and/or solubility to facilitate solution processing. In the following work, we present the results of all-atom molecular dynamics (MD) simulations which have afforded an unparalleled level of environmental control and insight into the little-understood structural and dynamic properties of C₆₀-solvent and derivative-solvent interactions. A novel quantitative analysis method was developed to characterize the distribution of solvent molecule orientations within the solvation shell, which revealed a strong positive correlation between the regularity of solvent molecule orientation and experimentally obtained solubility limits for C₆₀ and its most common derivative, phenyl-C₆₁-butyric acid methyl ester (PCBM) . This analytical procedure has provided insight into the effect of solvent geometry on fullerene solubility and solubilities in specific compounds that out-pace previous theoretical predictions. This granted us the ability to predict the solubility of PCBM in solvents that have yet to be observed experimentally. In addition, we were able to explain the solvent effect on formation of metastable solid C₆₀ complexes, clarifying a direct relationship between solvent geometry, temperature, and the formation of crystalline solvates. We extended our analytical protocol to a variety of other C₆₀ derivatives as well as more complex three-component solvation environments to explore the chemical effect of fullerene adducts and processing additives on the solvation shell structure that is shown to be integral to macroscopic behavior. Our method of solvation shell analysis utilizing MD simulations has thus provided much needed clarity on the dynamic nanoscale solvent environment that has critical for fullerene processing and device performance.

Graphene based surfaces including pristine graphene and graphene oxide (GO) have been spotlighted as a platform for the novel applications with bio-molecules, such as bio-sensors or drug-delivery. In order to improve bio-compatibility and stability of bio-molecule functionalized graphene based surfaces, it is imperative to understand role of surface polarity on physisorption process. However, effect of surface polarity that is responsible for structure and dynamics of bio-molecules has shown incomplete agreements in previous studies, thus it needs to be elucidated. In attempt to address following issues: (1) relationship between bio-molecules’ structure and surface polarity and (2) key interactions that contribute to the bio-functionalization process, we performed all-atom molecular dynamics simulations of two promising bio-molecules, ssDNA and silk fibroin protein on various graphene based surfaces with different oxygen coverages. Simulation results showed that both ssDNA and silk protein lost their unfolded structures on pristine graphene surface via strong hydrophobic interactions with the surface, whereas folded structures were maintained during the entire simulation on GO surface with moderate oxygen coverages. On GO surfaces with high oxygen coverages, strong interfacial hydrogen bonds induced a loss of internal interactions of ssDNA and silk protein. Our detailed explorations demonstrated that balanced vdW and electrostatic interfacial interactions and geometrical hindrance by functional groups on the surface play a key role in adsorption process with a good structural stability of ssDNA and silk protein on graphene based surfaces. Overall, our computational results not only revealed underlying process of physisorption between bio-molecules and graphene based surfaces, but also provided insights in designing many applications of bio-functionalized surfaces.

Point defects and their agglomeration play an important role in determining the properties and the lifetime of a material and device. In particular, in a metal–oxide–semiconductor field-effect transistor (MOSFET), defect accumulation at the metal-oxide interface is believed to be an important reason for oxide breakdown. However, an understanding of the dynamics of point defects at such interfaces is still limited. This paper presents a first-principle calculation of the potential energy surface (PES) for two types of point defects (hydrogen impurity and neutral vacancy) at an Al/SiO₂ interface. Saddle point information acquired from the PES is then used in a novel first-principle, nonequilibrium thermodynamic-ensemble model based on the principle of steepest entropy ascent (SEA) to predict the dynamics of the diffusion and agglomeration process for these two types of point defects. The nonequilibrium SEA model developed provides a complete entropy analysis of the process as well as a set of process reaction rates and reaction rate constants. The role of each entropy contribution, including that of the configurational entropy, is discussed. A time evolution of the distribution of two types of point defect cluster sizes under different environmental conditions is reported.

Ceramic nitrides of silicon continue to find application in the electronic technology arena because of their low-density, excellent thermal stability and high resistivity. Highly uniform SiN_xH_y films, often deposited via PECVD, are also being actively investigated as alternative {Si,N}-based buffer and insulating layers. The silicon diimide Si(NH)₂ end-member, typically produced as a low-density powder, is also a common starting material in the production of high purity Si₃N₄. Although crystalline forms of Si-N-H compounds are less common, they are expected to exist in stable form with crystal structures inherited from their oxide counterparts (e.g., O replace by NH or ND). For example, on the basis of spectroscopic evidence crystalline silicon diimide has been reported as a minor byproduct in the context of Si₃N₄ film growth via PECVD,¹ while Si₂N₂(NH) was explicitly been synthesized via pressurized ammonia reactions and shown to adopt the structure of its silicon oxynitride analog Si₂N₂O.² With these observations in mind we present a comparative PBE-GGA density functional theory study of the structural, thermoelastic and electronic properties of a-SiO₂ and its imide analog a-Si(NH)₂. We show that their crystalline structures, and most of their mechanical properties, exhibit striking similarities due to their common tetrahedral bonding framework. On the other hand, compression reveals that the bulk modulus of Si(NH)₂ is ~ 70 GPa, or about twice that of SiO₂ (39 GPa). Thermochemistry of the compounds, and their constituent elemental molecular and solid reference states, was studied using the PBE0 hybrid functional and yielded standard heats of formation of -840 kJ/mol and -370 kJ/mol for a-SiO₂ and a-Si(NH)₂, respectively. The band structure and RPA dielectric function of both materials were calculated using the TB09 meta-GGA functional, yielding the expected indirect (K→Γ) band gap of 8.8 eV in SiO₂ and an significantly smaller indirect (A→Γ) band gap of 5.5 eV in Si(NH)₂. We will also show that the excitonic contributions to the dielectric response obtained by solving the B-S equation at the meta-GGA level lead to an excellent account of the observed reflectivity spectrum of SiO₂ in the 0-15 eV range, and a systematic enhancement of reflectivity in Si(NH)₂. The static dielectric function of Si(NH)₂ is predicted to be ~35% larger than that of SiO₂ in the quartz structure. Finally, on the basis of detailed DFT-based reaction thermochemistry using the PBE0 functional we will discuss several potential synthesis routes for Si(NH)₂ involving ammonia and/or ether-like NH(SiH₃)₂ molecular sources. ¹ D.V. Tsu, G. Lucovsky and M.J. Martini, Phys. Rev. B 33, 7069 (1986). ² D. Peters, H. Jacobs, J. Less-Common Met. 146, 241 (1989).

Solid-oxide fuel-cells (SOFCs) are an attractive technology for efficient and clean energy conversion which are applicable to a wide variety of fuels (natural gas, liquefied petroleum gas, methanol and hydrogen).¹ Current generation SOFCs require high operating temperatures in the region of 1000°C. This is primarily due to the high temperatures required by the electrolyte in order for oxide ion diffusion to occur. This high operating temperature contributes greatly to the cost and speed of degradation of these devices. LaGaO₃ doped with Sr²⁺ on the A-site and Mg²⁺ on the B-site has been suggested as an alternative electrolyte material which can operate in the intermediate temperature range of 600-800°C.²

Here we present the derivation of a doped-LaGaO₃ force field, where the dopants are Sr²⁺ on the A-site and Mg²⁺ on the B-site. This force field is known as a dipole-polarizable ion model (DIPPIM).³ In the DIPPIM polarization effects resulting from induced dipoles on ions in the system are taken into account. This is of great importance due to the polarizable nature of the O^2- ions in doped-LaGaO₃. The force field is fitted to force, stress and dipole data obtained from DFT calculations. When fitting to ab initio data as opposed to experimental data details on the potential energy surface which are non-equilibrium can also be accounted for. This increases the number of observables which can be reproduced.

We also present the results of initial force field testing which has shown very good agreement with structural data from experiment and molecular dynamics simulations which have shown good agreement with diffusion data from experiment.

[1] Boudghene et al, Renew. Sust. Energ. Rev., 6, 433-455 (2002)
[2] Morales et al, J. Eur. Ceram. Soc., 36, 1-6 (2016)
[3] Castiglione et al, J. Phys: Condens. Matter, 11, 9009-9024 (1999)

Gold nanoparticles (AuNPs) have been sought for medical diagnostics, drug and gene delivery, chemical sensors, and electronics, due to their optical and electronic properties and tunable surface chemistry. AuNPs are often functionalized and exposed to various solvents during their preparation and use; however, fundamental interactions of functionalized AuNPs in different solvent environments have not been investigated in a systematic fashion at the molecular scale. Molecular dynamics (MD) is a computational tool that provides a way to model, simulate, and predict the behavior and properties of AuNPs at this length scale. Thus far, only a few solvent environments have been modeled and investigated with functionalized AuNPs using MD, namely, decane, ethane, saline, and water. To contribute to this area of research, we are investigating a spectrum of different solvent environments ranging from non-polar solvents such as chloroform and toluene to highly polar solvents such as methanol and water. A systematic investigation of AuNPs in a range of different solvent environments can aid solvent selection during the preparation and use of functionalized AuNPs that are presently used and the design of novel functionalized AuNPs and novel solvent systems. One type of parameter employed within an MD model/force field is partial atomic charge. Partial charges directly influence the electrostatic interactions of simulated systems and are often indirectly tethered to other parameters in the model, e.g., bond torsion parameters. These partial charges can greatly dictate the predicted behaviors and properties of modeled systems, especially for polar and charged systems. Different methods have been developed to assign partial charges, for example, the Restrained Electrostatic Potential (RESP) and Austin Model 1-Bond Charge Correction (AM1-BCC) methods. RESP involves the fitting of partial charges in order to best reproduce electrostatic potential (ESP) data from ab initio quantum mechanics (QM) calculations, while AM1-BCC involves arithmetic corrections made to initial partial charges obtained from semi-empirical QM calculations. Using eight solvents, we have assessed measures of partial charge quality such as relative root-mean-squared (RRMS) error between the QM-calculated and partial charge-based ESP data, error between MD-predicted and experimental dipole and quadrupole moments, and error between MD-predicted and experimental Hansen Solubility Parameters (HSPs). RESP led to lower RRMS values compared to AM1-BCC, while similar deviations from experimental electrostatic moments and the dispersion HSP were observed. AM1-BCC, however, led to less error for a combined polar and H-bond HSP value. Additional measures of partial charge quality are being evaluated, which will help to improve the overall quality of the derived partial charges for our AuNP-solvent systems and strengthen the insights and predictions that we aim to contribute to this area of research.

Gas hydrates are crystalline solids that consist of small gas molecules trapped within a network of hydrogen-bonded water molecules that form cage-like structures. They are most notably known as an immense potential source of energy, as many natural gas hydrate deposits have been discovered all over the world in the ocean floor as well as in permafrost regions. However, gas hydrates are also significant in many other areas such as flow assurance, carbon dioxide sequestration, and natural gas transportation and storage. Although research into gas hydrates has been ongoing for the past several decades, their material properties are not well characterized and our understanding of these unique structures at a fundamental level is still lacking. Therefore, this work focuses on investigating the material properties of structure II gas hydrates using first-principles theoretical modeling methods. More specifically, this work uses density functional theory to determine the mechanical properties of structure II gas hydrates, and study this structure at the atomistic scale.

Local or coordination environments (e.g, tetrahedra and octahedra) are powerful descriptors for the structure of inorganic solids. An automatic and robust detection of these environment is an important step towards the data mining of the large databases (experimental or theoretical) currently available to materials scientists.
In this work, we present a tool to automatically determine coordination environments in a given structure. The identification is performed based on the sole consideration of the geometry of the structure. Distortions are taken into account and we allow the description of an environment as a mixture of several environments. The code named ChemEnv is integrated in the python-based open source pymatgen package.
After outlining our algorithm, we will illustrate the approach by presenting a statistical analysis of coordination environments for all oxides from the Inorganic crystal structure database (ICSD). We will discuss the implication of our study to the understanding of crystal chemistry in oxides and outline how this tool can be used to accelerate the materials design process.

Identifying relevant descriptors for material properties enables rapid screening of large numbers of materials and facilitates the design of new materials. One of the leading challenges is determining the best ways to utilize large material data sets for the identification of descriptors of a given property. In this presentation, we demonstrate the use of genetic programming to identify relevant descriptors of dielectric breakdown strength based on eight key features of 86 representative crystalline materials. We identified band gap E_g and phonon cut-off frequency w_max as the two most correlated features, and new classes of phenomenological models featuring functions of the product of two features were uncovered. The genetic programming models were found to outperform other models for descriptor identification, and some of the advantages of the genetic programming approach are discussed.

In order to increase efficiency of energy resource usage, technology of utilizing exhaust heat from factories and vehicles are developed. Installation of thermoelectric modules is one of the methods to reuse waste heat. Bi₂Te₃ is a promising thermoelectric material in low temperatures ranging from a room temperature to a few hundred degrees Celsius. In Bi-Te binary phase diagrams, stoichiometric compounds such as BiTe, Bi₄Te₅, or Bi₅Te₆ exist near the composition of Bi₂Te₃. However, crystal structures have only been reported for BiTe and Bi₂Te₃; those of the other compounds have not been elucidated. Determination of the crystal structures of stable compounds in the Bi-Te system and these electric structures are needed for a rational design of thermoelectric materials based on Bi₂Te₃.
Combining first-principles calculations with a cluster expansion technique is an efficient method searching for stable stoichiometries and structures. Although unit cells of Bi and Te, the end members of the Bi-Te system, have different geometry and symmetry, they can be regarded as distorted simple cubic (A7) structures. BiTe and Bi₂Te₃ are also deemed order phases of binary simple cubic systems with distortion. Therefore, a cluster expansion method can be applied to this system. In this work, we searched for ground state structures in the Bi-Te system using first-principles calculations and the cluster expansion method. In this analysis, we considered all compounds as distorted simple cubic lattices. Electronic structures were also investigated.
First-principles calculations were performed by the projected augmented wave method as implemented in the VASP code. The exchange-correlation term was treated using the Perdew-Burke-Ernzerhof functional, and spin-orbit coupling was taken into account in all calculations. Cluster expansion was performed using the CLUPAN code.
Our structure search revealed an unreported layered structure of composition Bi₄Te as the ground state structure. Some metastable structures, those have slightly higher energy from energy convex hull, were also found in composition range between BiTe and Bi₂Te₃. This result suggests that secondary phases are likely to form in this system. We performed powder X-ray diffraction (XRD) pattern simulations for these structures. All patterns were found to be similar, and it is difficult to distinguish precipitation of secondary phases from Bi₂Te₃ matrix using XRD alone.
The band gap of Bi₂Te₃ was calculated to be 0.14 eV, which is in good agreement with the experimental value. Our calculations show that the other stable or metastable structures in composition range between BiTe and Bi₂Te₃ have no band gap, that is, they are metallic or semi-metallic. For realizing thermoelectric materials with high performance, semiconductive property needs to be controlled and formation of such metallic impurity phase should be avoided. Results of our study point out importance of precise control of chemical composition.

A nano-composite is defined as a multiphase solid material where one of the phases has at least one dimension of less than 100 nanometres (nm) and their large scale properties are critically dependent on how the components interact on the molecular level. We have developed an advanced multiscale simulation environment to predict the properties of such nanocomposites based on their molecular structures and composition. These methods have applications in modelling a wide range of materials and it is our aim to create a “virtual laboratory” to compute the properties of these materials based simply on knowledge of their chemical composition, molecular structure and processing conditions [1]. Here we will present our findings from modelling chemically specific combinations of clay, polymers and organic surfactants. We use our multiscale methods and tools to take us from a parameter free quantum description to atomistic and coarse-grained simulations, thereby leading to predictions of the materials properties of these nanocomposites. Our simulations approach realistic sizes of clay platelets (currently of diameter 100 Å) at low clay volume fractions (5%). These systems exhibit substantial property enhancements when compared to the pristine polymer (elastic properties, gas permeation), but homogeneous dispersion of the clay sheets is required.

Our multiscale approach provides predictions of the melt intercalation behaviour and final morphologies of organo-treated montorillonite clay–polyvinyl-alcohol and montorillonite clay–polyethylene-glycol systems. Many hitherto unobserved processes come into view as a result of this study, including the dynamical process of polymer intercalation and exfoliation into pristine and organo-treated clay tactoids and the ensuing aggregation of polymer-entangled tactoids into larger structures. We study the role of surfactants and are able to elucidate how it facilitates polymer intercalation and ultimately clay sheet exfoliation, which is driven by the attraction of polymer to the clay surfaces [2]. We will also present our recent findings on the use of shear forces to separate aggregated tactoids, where we subject our coarse-grained models to a polymer flow field and observed the role of shear stress in dispersing organic-treated clay sheets.
From our multiscale simulations, we can compute various characteristics of these nanocomposites, including clay-layer spacings, out-of-plane clay sheet bending energies, X-ray diffractograms and materials properties, which we relate to the system's final morphology.

[1] J. L. Suter, D. Groen, P. V. Coveney, Adv. Mater. 2015, 27, 966–984
[2] J. L. Suter, D. Groen, P. V. Coveney, Nano Lett., 2015, 15, 8108–8113.

Reactive transport in bulk materials and at interfaces is ubiquitous to the operation of energy conversion and storage devices. Similarly, materials evolution under extreme conditions, e.g. thermal oxidation of metal alloys in nuclear reactors, involves complex coupled microscopic processes that include ionic and electronic transport across oxide/metal interfaces, and in many cases long-range, vacancy-mediated metal atom diffusion from deep within the alloy phase. In order to gain insights into such reactive transport processes at experimentally relevant length scales a mesoscale model based on density functional theory (DFT) is established. The approach combines quantum DFT for the elementary transport and chemical reactions with mesoscopic Poisson-Nernst-Planck/classical DFT theory of collective long-range interactions, while retaining fully atomistic spatial resolution. The method is validated against experimental data on intergranular oxidation in metal alloys. In particular, the method proved to be instrumental in predicting the composition and microstructure of the forming oxide depending on the chemistry and concentration of the alloying element as well as in predicting the mode of intergranular oxidation – continuous oxidation vs. oxide cluster formation. The application of the approach to the understanding structure-property relation in nanocomposite energy storage materials will be also discussed.

In the past few years, we developed several new transition state searching method (CBM，CBD, BP-CBD, DESW), and based on these methods, we designed a new global optimization method for potential energy surface (PES) search, namely, stochastic surface walking (SSW) method. The SSW method utilizes the approximate normal mode (second derivatives) to guide the PES search and adds gaussian bias potentials to overcome barrier height. A key feature of SSW search is the smooth structure perturbation from one minimum to another and thus it can be used not only for unknown structure search but also for pathway sampling. In this talk, I will show the recent progress of the method development, including the rigid-body implementation and the complex reaction network evolution. Some examples of molecular reactions and solid phase transitions will be presented, which are based on classical potentials and first principles calculations.

Related references：
1. Zhang, Xiao-Jie; Liu, Zhi-Pan* “Variable-Cell Double-Ended Surface Walking Method for Fast Transition State Location of Solid Phase Transition”, J. Chem. Theory Comput, 2015, 11, 4885
2. Zhang, Xiao-Jie; Liu, Zhi-Pan* "Reaction Sampling and Reactivity Prediction Using Stochastic Surface Walking Method",Phys. Chem. Chem. Phys.,2015,17, 2757
3.Shang, Cheng, Zhang, Xiao-Jie and Liu, Zhi-Pan*, “Stochastic Surface Walking Method for Crystal Structure and Phase Transition Pathway Prediction”, Phys. Chem. Phys. Chem. 2014, 16, 17845
4. Zhang, Xiao-Jie; Shang, Cheng and Liu, Zhi-Pan*, “Double-Ended Surface Walking Method for Pathway Building and Transition State Location of Complex Reactions” , J. Chem. Theory Comput, 2013, 9, 5745;
5. Zhang, Xiao-Jie; Shang, Cheng and Liu, Zhi-Pan*, “From Atoms to Fullerene: Stochastic Surface Walking Solution for Automated Structure Prediction of Complex Material”, J. Chem. Theory Comput, 2013, 9, 3252
6. Shang, Cheng and Liu, Zhi-Pan*;“Stochastic Surface Walking Method for Structure Prediction and Pathway Searching”, J. Chem. Theory Comput, 2013, 9, 1838；
7. Shang, Cheng and Liu, Zhi-Pan*;“Constrained Broyden Dimer Method with Bias Potential for Exploring Potential Energy Surface of Multistep Reaction Process”, J. Chem. Theory Comput, 2012, 8, 2215
8. Shang, Cheng; Liu, Zhi-Pan*, “Constrained-Broyden-Minimization Combined with the Dimer Method for Locating Transition State of Complex Reactions” , J. Chem. Theory Comput, 2010, 6, 1136
9. Wang, Hui-Fang; Liu, Zhi-Pan* “Comprehensive Mechanism and Structure-Sensitivity of Ethanol Oxidation on Platinum: New Transition-State Searching Method for Resolving Complex Reaction Network”, J. Am. Chem. Soc. 2008, 130,10996

The recent development of bond-order interatomic potentials for hydrocarbons and silica allows the quantitative study of reaction rates and surface chemistry at the atomic scale. Results obtained in atomic scale simulations may be used to build chemistry models for quartz surfaces in order to account for effects such as radical quenching in plasma and combustion applications. In general, such reactions become increasingly important as the size of combustion devices shrinks and depletion of radical species near walls can dominate flame behavior. In this work, the recombination of atomic hydrogen on the surface of amorphous quartz is quantified using molecular dynamics and Monte Carlo simulations. The interaction of gas-phase radicals with the surface is directly simulated for temperatures between 10 and 600 K at a pressure of 10 atm. In this high-pressure regime, the recombination coefficient of hydrogen falls between 0.1 and 1. Multi-timescale methods are used to study hydrogen-surface interactions at lower pressures. Weak and strong binding sites for atomic hydrogen with densities of approximately 10 nm^-2 are found on the surface using Grand Canonical Monte Carlo (GCMC) simulations. Eley-Rideal rate constants are calculated using GCMC results and semi-equilibrium theory. Monte Carlo Variational Transition State Theory (MCVTST) is used to calculate Langmuir-Hinshelwood and thermal desorption rate constants for hydrogen atoms in strong and weak adsorption sites. Calculated reaction rates are used in a Langmuir kinetics model to estimate the recombination coefficient for hydrogen radicals across a wide range of temperatures (10 < T < 2000 K). The recombination coefficient is found to vary between 10^-4 and 0.8 in these conditions. These values agree with past experimental measurements; results using them as part of a surface chemistry model for continuum combustion simulations will be reported.

The cornerstone of challenges in tissue engineering lies in the unreliable prediction of scaffold degradation concurrent with tissue growth. Experiments based data from in vitro studies provide limited predictive capabilities of the scaffold behavior in vivo. Animal studies also are unable to capture the response in human patients. The degradation of polymers themselves (that constitute components of scaffold material systems) is well understood; however nanocomposite systems present a challenging problem towards the ability to accurately predict mechanical behavior and response over the biological and physiochemical mechanisms of scaffold degradation and tissue formation. The complex multiscale behavior of mechanical manifestations of molecular interactions cannot be captured by simplistic continuum based mechanics models. The complex behavior arises from the degradation characteristics of the polymer, nanoparticles, interfaces and the altered regions of the polymer due to nanoparticle addition as well as cellular processes. Here, we represent a novel multiscale mechanics-based approach to model degradation of nanocomposite bone tissue engineering scaffold system consisting of nanoclay polycaprolactone nanocomposites. The nanocomposite system involves a novel biomimetic route to mineralize hydroxyapatite (HAP) using organically modified nanoclay (montmorillonite) (OMMT). Polymeric scaffolds are synthesized using the biomineralized HAP with polycaprolactone (PCL) for bone tissue engineering applications. These scaffolds have been reported to mediate mesenchymal stem cell differentiation. Molecular models of the OMMT with HAP and PCL have been bu8ilt. We have conducted molecular dynamics simulations of modified nanoclay-HAP-PCL composite Steered molecular dynamics simulation of the nanocomposite system yields molecular scale load deformation behavior that is incorporated into microCT derived microstructure of scaffolds over the degradation timescale. Further, finite element models of the scaffold structures during degradation are built. The degradation model based on damage mechanics for cellular degradation (osteoblasts) is derived and incorporated into the finite element model to enable modeling scaffold degradation during tissue growth. The new degradation/healing model is developed to model the evolution of mechanical properties accurately with time. The combined multiscale model and the degradation/healing model allow us the ability to predict the response of implanted scaffolds over time. The multiscale in silico approach provides a unique capability to design nanocomposite biomaterials.

With emergence of nanotechnology, it is possible to control interfaces and flow at nanoscale. This is of particular interest in the Oil and Gas industry, where nanoscience can be applied on processes such as Enhance Oil Recovery (EOR) and oil mitigation. On this direction, one of potential strategies is the so called Nano-EOR based on surface drive flow, where mobilization of hydrocarbons trapped at the pore scale can be favored by controlling by the chemical environment through “Wetability modifiers”, such as functionalized NPs and surfactants^1,2. The challenge consists then to search for optimal functionalized nanoparticles for oil recovery and mitigation at the harsh conditions found in oil reservoirs. Here, we introduce a hierarchical computational protocol based on the role of NPs interfacial and wetting properties within oil/brine/rock interfaces to the fluid displacement in pore network models (PNMs) ^1,3. This integrated multiscale computational protocol ranges from first principles calculations², to determine and benchmark interatomic potentials, which are coupled with molecular dynamics to characterize the descriptors (interfatial properties and viscosity) ². The MD results are then mapped into Lattice Boltzmann method (LBM) simulation parameters to model the oil displacement process in PNMs at the microscale³. The proposed multiscale protocol can be a resourceful tool to explore the potentialities of chemical additives, such as NP, for the oil recovery process and investigate the effects of interfacial tension and wetting properties on the fluid behavior at both nano and microscales.
[1] De Lara, L.S.; Rigo, V.A.; Miranda, C.R, J. Phys. Chem. C, 120, 6787, 2016.
[2] Rigo,V.A.; de Lara, L.S.; Miranda, C.R, Applied Surface Science, 292, 742, 2014.
[3] Pereira, A.O. ; De Lara, L.S; Miranda, C.R., Microfluidics and Nanofluidics,20, 36, 2016.

Friction at liquid/solid (L/S) interfaces is of critical importance for multiple phenomena, including lubrication, flow of liquids in micro/nanofluidics and flow of ions through small channels in biophysics applications. Friction and adhesion can be controlled by interfacial chemistry, but design of functionalized surfaces and interfaces with optimized friction and slip properties has been hindered by a number of challenges in measuring these properties either in experiments or in simulations. One of the challenges is related to the short time scales accessible to standard molecular simulations. To address this challenge, we have developed a Green-Kubo relation that enables accurate calculations of friction at L/S interfaces directly from equilibrium molecular dynamics simulations on experimentally relevant time scales. The theory has been validated for interfaces with a range of chemical properties and because of the high numerical efficiency of our method, it opens up new opportunities for computational design of functionalized surfaces for L/S applications. The theory can also be used to enable new fundamental discoveries from computer simulations. For instance, we have used it to identify the effects of interfaces on the nearby Brownian motion, and to demonstrate that L/S slip is both an intrinsic property of an interface and that it is not static in time, as had been previously assumed. In another example, we have used multi-scale approach that combines atomistic simulations and kinetic Monte Carlo technique to determine the effects of surface chemistry and humidity on the origin of static friction, which is a highly debated topic in the field of tribology. For most solid/solid contacts, static friction increases logarithmically with time, a phenomenon known as aging. We discovered molecular mechanisms that are responsible for aging of silica/silica contacts in the presence of water and that are based purely on interfacial chemistry. By combing simulations with atomic force microscopy experiments we identified the effects of contact pressure on both friction and aging, which phenomena have important implications for Earthquake mechanics.

Metal-semiconductor (M-SC) contacts play a pivotal role in a broad range of technologically relevant devices. Still, their characterization at the atomic-scale remains a delicate issue. One of the reasons is that the present understanding relies either on simplified analytical models often parametrized using experimental data [1], or on electronic structure simulations describing the interface using simple slab calculations [2]. Here we propose a general strategy to model realistic M-SC interfaces by using density functional theory (DFT) in combination with the non-equilibrium Green's function (NEGF) method as implemented in the Atomistix ToolKit (ATK) simulation software [3]. An accurate description of both sides of the interface is achieved by using a meta-GGA functional [4] optimally tuned to reproduce the SC measured band-gap, and a spatially dependent effective scheme to account for the presence of doping in the SC side. Compared to previous computational methods [2], the present approach has the important advantages of (i) treating the system using the appropriate boundary conditions and (ii) allowing for a direct comparison between theory and experiments by simulating the I-V characteristics of the interface. We apply this methodology to an Ag/Si interface relevant for solar cell applications, and test the reliability of traditional strategies [1,2] to describe its properties [5].
References
[1] S. M. Sze and K. N. Kwok, Physics of Semiconductor Devices: 3rd edition (Wiley, 2006)
[2] C. G. van de Walle and R. M. Martin, Phys. Rev. B 35, 8154 (1987)
[3] “Atomistix ToolKit version 2015.1”, QuantumWise A/S (www.quantumwise.com)
[4] F. Tran and P. Blaha, Phys. Rev. Lett. 102, 226401 (2009)
[5] D. Stradi et al. PHYSICAL REVIEW B 93, 155302 (2016)

Classical molecular dynamics can be used to model large and complex interfaces with the degree of accuracy necessary to allow for the atomic local relaxations that govern the surface stress. This information is, however, difficult to use in continuum based methods that can examine stress effects on large scale phenomena and micro scaled devices. At this larger scale we are often forced to discard the complexity and assume relatively homogeneous stress functions.

We have developed a multi-scale method of simulating interface induced stress in magnetoelastic systems that couples stress effects from the atomic to the continuum scale. We use atomic simulations to compute stress maps for different interfacial regions, providing a more detailed stress map than continuum based methods while avoiding the use of arbitrary loading conditions often used in mechanical modelling. From the virial (atomic) stress map a physically meaningful Cauchy stress map can be extracted through the use of atomic localization functions based on a given mesh for use in a range of finite element packages including micro-magnetics packages. This allows for a flexible multi-scale modelling methodology that can be applied to the study of a range of different interfaces; grain boundaries, free surfaces and twinning. The veracity of this method is demonstrated by comparing simulations of the effective anisotropy in thin films both with and without magnetoelastic energetics in relation to the well defined expression for the contributions of bulk and surface anisotropies.

Au nanoclusters are of technological relevance for catalysis, photonics, sensors, and of fundamental scientific interest owing to unusually large stable planar (2D) structures and diverse geometries from planar to cage-like to compact structures. Computational modeling of these nanoclusters is necessary in order to understand the diverse geometries. With the large configurational degrees of freedom, machine learning techniques are helpful in order to sample configuration space as well as incorporate experimental information. In this talk, we discuss structural determination of Au nanoclusters from single and multi-objective global optimization algorithms, using as inputs high throughput density functional theory (DFT) calculations [1] and a combination of energetic and simulated pair distribution function (PDF) data [2], respectively. In addition, DFT data from thousands of structures are fitted using a genetic algorithm to a hybrid bond-order potential (HyBOP) [3] which is able to predict structural and energetic properties of Au from several-atom clusters to bulk.

[1] A. Kinaci, B. Narayanan, F. G. Sen, M. J. Davis, S. K. Gray, S. K. R. S. Sankaranarayanan, M. K. Y. Chan, "Unraveling the Planar-Globular Transition in Gold Nanoclusters
through Evolutionary Search," Scientific Reports 6, 34974 (2016). DOI:10.1038/srep34974

[2] S. Hills, A. Kinaci, F. Sen, M. K. Y. Chan, in preparation.

[3] B. Narayanan, A. Kinaci, M. J. Davis, S. K. Gray, M. K. Y. Chan, and S. K. R. S. Sankaranarayanan, “Describing the diverse geometries of gold from nanoclusters to bulk − a first principles based hybrid bond order potential,” Journal of Physical Chemistry C 120, 13787 (2016). DOI:10.1021/acs.jpcc.6b02934.

ACKNOWLEDGEMENT

Use of the Center for Nanoscale Materials, an Office of Science User Facility, was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

The submitted abstract has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government.

In this talk, I will review progress in my research group towards extracting design principles for molecular systems directly from the information embedded in machine learning models. In other words, we are trying to answer the question "What did the machine learn, if anything?". I will report our results for organic light emitting devices as well as for other applications.

While intuition-driven experiments and serendipity have guided traditional materials discovery, computational strategies have become increasingly prevalent and a powerful complement to experiments in modern day materials research. A novel approach for efficient materials design is a rational co-design approach, where high-throughput computational screening is used synergistically with experimental synthesis and testing. Here, we demonstrate the utility and promise of such an approach with the example of advanced polymer dielectrics for electrostatic energy storage applications^[1,2,3]. We highlight recent co-design efforts that can potentially lead to replacement of present-day “standard” dielectrics (such as biaxially oriented polypropylene) not only by new organic polymer candidates within known generic polymer subclasses (e.g., polyurea, polythiourea, polyimide), but also by organometallic polymers, a hitherto untapped but promising chemical subspace. Further, we discuss the prospects of significantly accelerating the materials design process using machine learning techniques. Using the vast amounts of computational data that has been generated (and compiled in the form of an online database^[7]), we developed accurate ‘instant prediction’ and ‘design’ models for the relevant properties of the polymer chemical space^{[4,5,6, 7]}. While this materials design philosophy is demonstrated here for polymer dielectrics, it is equally applicable to other classes of materials as well.

References
[1] A. Mannodi-Kanakkithodi et al., “Rational Co-Design of Polymer Dielectrics for Energy Storage”, Adv. Mater. doi: 10.1002/adma.201600377 (2016).
[2] T.D. Huan et al., “Advanced polymeric dielectrics for high energy density applications”, Prog. Mater. Sci. 83, 236 (2016).
[3] V. Sharma et al., “Rational design of all organic polymer dielectrics”, Nat. Comm. 5, 4845 (2014).
[4] A. Mannodi-Kanakkithodi et al., “Machine Learning Strategy for Accelerated Design of Polymer Dielectrics”, Sci. Rep., 6, 20952 (2016).
[5] T. D. Huan, A. Mannodi-Kanakkithodi, R. Ramprasad “Accelerated materials property predictions and design using motif-based fingerprints”, Phys. Rev. B 92, 014106 (2015).
[6] G. Pilania et al., “Accelerating materials property predictions using machine learning”, Sci. Rep., 3, 2810 (2013).
[7] http://khazana.uconn.edu

The discovery of new materials has largely proceeded, so far, by extending known molecular targets with desired properties into new compositional spaces. Starting from an analysis of the limitations intrinsic to ab initio molecular dynamics simulations when used for novel materials design purposes, I will present the effort, undergoing at IBM Research - Zurich, to develop novel approaches for materials understanding and design. The novel platform, based on the use of cognitive computing technologies, leverages the human expertise on first principle atomistic simulations with the large scale data analytics.
I will introduce the basics of cognitive computing coupled to molecular modeling activities and how we can exploit these recent innovations to accelerate the discovery process of new solid state electrolytes for battery applications.

High-throughput, theory-driven computational materials design and discovery has become an essential tool in materials science and engineering. The data-driven and cost-effective framework for accelerated discovery introduced by the Materials Genome Initiative (MGI) has transformed the scale and rate of materials development by exploiting the predictive ability of quantum chemical computational methods. In turn, the drive to rapidly predict, screen, and optimize materials using first-principles calculations has led to large datasets and the construction of open-source databases that are populated primarily by low-cost density functional theory (DFT) total energy calculations (inherently performed at 0 K). Although these datasets are an invaluable resource, their predictive ability at finite temperatures is limited and current methods for evaluating the missing temperature dependencies are computationally prohibitive, limiting the success of MGI approaches for existing and emerging high temperature applications. In this report, we show that compound Gibbs energies of formation can be predicted with small errors at finite temperatures of up to at least 1800 K, and at the same computational expense as a DFT total energy calculation. Our broadly applicable method for evaluating compound Gibbs energies of formation employs readily computable material-specific descriptors and utilizes existing datasets of more than 50,000 inorganic compounds to enable rapid computational prototyping at elevated temperatures. We build upon prior work that has demonstrated the use of total energy calculations to approximate the compound enthalpy of formation at 298 K with mean absolute error relative to experiment (MAE) = 0.054 eV/atom with fitted elemental-phase reference energies (FERE). Our prediction requires only the DFT total energy, the FERE correction, and minimal-computational-cost material-specific predictors to compute the temperature-dependent compound Gibbs formation energy with MAE < 0.1 eV/atom up to at least 1800 K. At 1000 K, the MAE between our prediction and experiment is 0.030 eV/atom for oxides, 0.032 eV/atom for nitrides, 0.069 eV/atom for selenides, 0.058 eV/atom for arsenides, 0.026 eV/atom for phosphides, and 0.039 eV/atom for sulfides. These results bridge the temperature gap between existing MGI inspired computational approaches, which have historically been limited to low temperature, and high temperature materials applications, providing the equivalent of tabulated thermochemical data for more than 50,000 inorganic materials.

We assess the predictive accuracy of perturbation theory based estimates of changes in covalent bonding due to linear alchemical interpolations among systems of different chemical composition. We have investigated σ bonding to hydrogen, as well as σ and π bonding between main-group elements, occurring in small sets of iso-valence-electronic molecular species with elements drawn from second to fourth rows in the p-block of the periodic table. Numerical evidence suggests that first order estimates of covalent bonding potentials can achieve chemical accuracy if the alchemical interpolation is vertical (fixed geometry) among chemical elements from third and fourth row of the periodic table [1]. When applied to solids similar observations hold for carefully chosen systems and properties [2-4].

[1] K. Y. S. Chang, S. Fias, R. Ramakrishnan, O. A. von Lilienfeld J. Chem. Phys. 144 174110 (2016)
[2] M. to Baben, J. O. Achenbach, O. A. von Lilienfeld J. Chem. Phys. 144 104103 (2016)
[3] A. Solovyeva, O. A. von Lilienfeld arxiv.org/abs/1605.08080 (2016)
[4] K. Y. S. Chang, O. A. von Lilienfeld, in preparation (2016)

In this talk, I will describe our efforts to compute the electronic transport properties (using the BoltzTraP code¹) of 60,000 materials from the Materials Project² database (www.materialsproject.org) to discover new bulk thermoelectric materials. Our calculated results will be published openly such that they can be re-used by researchers. I will first present an overview of chemical trends in thermoelectric materials using the electronic transport data set. Next, I will focus on a new YCuTe₂ thermoelectric compound³ discovered jointly through computations and experiments. YCuTe₂ possesses a zT reaching as high as 0.75 and a bulk lattice thermal conductivity as low as 0.5 W/m*K, making it an attractive system for further study. I will present our latest results in understanding how crystal structure, volume, and chemistry affect thermoelectric properties (such as the projected density of states) through the application of statistical learning techniques. Finally, I will describe our open-source automation software for performing many types of calculations and how other researchers can leverage this software.

(1) Madsen, G. K. H.; Singh, D. J. BoltzTraP. A code for calculating band-structure dependent quantities, Comput. Phys. Commun., 2006, 175, 67–71, doi:10.1016/j.cpc.2006.03.007.
(2) Jain, A.; Ong, S. P.; Hautier, G.; Chen, W.; Richards, W. D.; Dacek, S.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; Persson, K. A. Commentary: The Materials Project: A materials genome approach to accelerating materials innovation, APL Mater., 2013, 1, 011002, doi:10.1063/1.4812323.
(3) Aydemir, U., Pöhls, J.-H., Zhu, H., Hautier, G., Bajaj, S., Gibbs, Z. M., Chen, W., Li, G., Ohno, S., Broberg, D., Kang, S. D., Asta, M., Ceder, G., White, M. A., Persson, K., Jain, A., et al. YCuTe2 : a member of a new class of thermoelectric materials with CuTe4 -based layered structure. J. Mater. Chem. A (2016). doi:10.1039/C5TA10330D

Surface segregation refers to the phenomenon that chemical composition at the surface of alloy materials differs from the corresponding value in their bulk region. Knowledge on surface segregation is pertinent to various engineering applications such as adsorption, wetting, oxidation, corrosion, electrical contact, friction and wear, crystal growth, and catalysis. In this presentation, we reported our research on accurately predicting the influence of surface segregation on the functional properties of nanostructured alloy materials using a multiscale computation technique. The employed multiscale computational approach consists of three hierarchical components: (1) developing reliable interatomic potentials for alloys with the modified embedded atom method based on the first-principles computation data, (2) applying these atomic interaction potentials to determine the chemical composition of extended and nanoparticle surfaces of alloys using the atomistic Monte Carlo method, and (3) evaluating properties of surface-segregated nanomaterials using the first-principles and/or mesoscale computation methods. We have successfully applied our multiscale computation to design catalytic and magnetic nanoparticles.

Case I: Platinum (Pt) alloys are the most active catalysts for oxygen reduction reaction (ORR) occurring in proton exchange membrane fuel cells. The surface chemical composition in these catalysts determines the electronic structure and further their catalytic performance for ORR. In the present study, we used our multiscale computation method to elaborate the relation between the surface composition, electronic structure, reaction pathway, and reaction rate of ORR on nanosegregated Pt alloy catalysts.

Case II: For both FePt and CoPt nanoparticles, our multiscale model predicted that it was energetically favorable to exchange the surface Fe and/or Co atoms with the interior Pt atoms and thus to induce the Pt surface segregation. Comparing the magnetic properties of bulk-terminated and surface-segregated nanoparticles, we found that the surface segregation processes in the FePt and CoPt nanoparticles could cause a decrease in their total magnetic moments, a change in their (easy and/or hard) magnetization axes, and a reduction in their magnetic anisotropy.
Hence, our multiscale computational approach is a valuable tool for understanding surface chemistry-property relationships of alloy nanomaterials.

Advances in the theoretical understanding of interfacial electrochemistry have, over the past decade, permitted the extension of periodic Density Functional Theory studies, which have traditionally been applied to probe chemistry at gas/solid interfaces, to electrochemical systems where potential-dependent structural and chemical phase transformations occur at liquid/solid interfaces. Indeed, such techniques have been employed to study a surprisingly wide class of electrochemical processes, ranging from electrocatalysis to corrosion. In this talk, we will discuss how these strategies can be applied to predict the detailed, atomic-scale properties of monolayer (oxy)hydroxide base metal films on precious metal and precious metal alloy substrates as a function of applied potential in alkaline solutions. We will describe the analysis of Moire patterns, variable oxidation states, and three-phase boundaries of these films, and we will demonstrate how these results may be used to predict the shape and oxidation states of supported two-dimensional nanoislands as a function of electrochemical conditions. We will close with some perspectives on the possibilities for alkaline electrocatalyst design offered by these systems.

Cobalt phosphide (CoP) is one of the most promising, earth-abundant electrocatalysts discovered to date for hydrogen evolution reaction (HER), yet the mechanism is not well understood. Since hydrogen adsorption is a key factor of HER activity, here we examine the adsorption of atomic hydrogen on the low-Miller-index surfaces of CoP, including (111), (110), (100), and (011), by using periodic density functional theory. From the calculated Gibbs free energy of adsorption, we predict that (111), (110), and (011) surfaces will have good catalytic activities for HER. From ab initio atomistic thermodynamics, we find that the stabilities of the surfaces at 1 atm H2 and 300 K follow the trend of (111) > (100) ~ (110) >> (011). On the most stable (111) surface, both Co bridge sites and P top sites are found to be able to adsorb hydrogen with a close-to-zero free energy change and the synergy of proximal Co and P atoms on the surface results in a better HER activity. Our work provides important insights into CoP’s excellent HER activity and a basis for further mechanistic understanding of HER on CoP and other transition-metal phosphides.

Metal-organic frameworks with Zr₆-based nodes, such as UiO-66/67 and NU-1000, were investigated as supports for mono- and bimetallic complexes. Quantum chemical calculations provided insight on the structure and reactivity of node-supported metals (e.g. Al, Fe, Co, Ni, Zn, Rh) upon ethylene hydrogenation and oligomerization. The computational results show that catalytic activity and selectivity of these species are influenced by the direct coordination sphere of the active catalytic site. The understanding of these systems helps to rationalize which factors can be modified in catalyst design.

Since the successful mechanical exfoliation in 2004, graphene has become the spotlight in physics and chemistry as well as in materials science. It is the lightest and strongest material with excellent heat and electricity conductions, showing unprecedented applications in composite materials, optical electronics, photovoltaic cells, ultrafiltration, and energy storage. Here, we discuss the applications in CO₂ conversion, which is not well explored as compared with other topics. We focus on porous graphene embedded with metal dimers, and graphene nanoribbons with Cu-termination for converting CO₂ to CH₄ and CH₃OH, which exhibit the advantages of the outstanding selectivity of products, the high dispersity of spatial distribution, and the reduced overpotentials.

ACKNOWLEDGEMENTS
This work was partially supported by grants from the National Research Foundation (NRF) of Singapore under its Campus for Research Excellence and Technological Enterprise (CREATE) program.

The Born-Oppenheimer approximation, which assumes that the electrons are always in their ground state, is widely used in density functional theory (DFT) studies of catalytic processes. These calculations have been quite successful in explaining a large number of catalytic phenomena. However, electronic excitations are easily induced in metal surfaces, due to the lack of a band gap. Here, we show that processes on catalytic surfaces may induce electronic excitations, which are not captured by simulations using the Born-Oppenheimer approximation.

We performed non-adiabatic dynamics using real-time, time-dependent DFT, propagating the nuclei using Ehrenfest dynamics. We study several trajectories of an N₂ molecule interacting with Ru₁₃ and Ru₁₄₇ nanoparticles. These simulations show that, during the adsorption and dissociation processes, a significant amount of energy can be dissipated into electronic degrees of freedom, comparable to the amount of energy dissociated into ionic vibrations. The excitations initially take the form of electron-hole pairs, but after tens of fs they relax into hot electrons. This suggests that the electronic structure of catalytic surfaces may, transiently, be quite different from the ground state.

TiO₂ is the most widely used photocatalyst, but its notoriously large band gap requires technology to enhance photoabsorption rate for higher performed photocatalyst. Doping is one of widely used approaches to achieve this goal. However, roles of dopants are not simply understandable. For example, it was observed that effects of N dopants is turned from positive to negative when the temperature was slightly changed. Even degradation of photocatalytic activity was observed for the sample showing higher photoabsorption intensity. People speculate that dopant structure or intrinsic defect play roles, but it is very difficult task to get information about atomic scale from experiments. Using density functional theory calculations, we find the most probable atomic structure of dopants and intrinsic defects at various environmental conditions, and investigate their effects on photoabsorption rate as well as photocatalytic activity. Specifically, we find that Vc_Ti-N_i pair (Ti vacancy in the vicinity of interstitial N) and W_Ti-N_O pair (W at Ti site and N at O site) are the most probable structures for W-N codoped TiO₂ at normal conditions, and former one may well degrade the photocatalyst while the later one improves it. The concentration of these two structures varies with temperature and pressure, which provides hints for fabricating improved TiO₂. We also find that lifetime and mean free path of photoexcited hole are the hidden but critical parameters.

A key challenge in modern computational materials chemistry is the description of van der Waals interactions in density functional theory simulations, where the failure of conventional exchange-correlation functionals is well-known. While, in the recent years, several methods have been proposed for overcoming these difficulties, the applications are becoming increasingly more demanding as well. An important example is that of ligand-protected nanoparticles, which typically feature metallic, covalent as well as dispersive interactions that should all, ideally, be treated on an equal footing. In this work we show that significant progress can be made in this direction: with the aid of the recently-developed vdW-DF-cx functional we study the phase diagram of self-assembled monolayers of alkanethiolates on gold. This system is important for practical applications and as a general representative of self-assembly at a metal surface. In particular, a quantitative description of the dispersion-driven phase transition between a lying-down and a standing-up monolayer is obtained using an ab initio thermodynamics framework. The results are shown to be in good agreement with experimental data and highlight that accurately accounting for dispersive interactions is both feasible and a crucial part of modeling self-assembled systems.

The existence of domain walls, which separate regions of different polarization, can influence the dielectric, piezoelectric, pyroelectric and electronic properties of ferroelectric materials. In particular, domain-wall motion is crucial for polarization switching, which is characterized by the hysteresis loop that is a signature feature of ferroelectric materials. Experimentally, the observed dynamics of polarization switching and domain-wall motion are usually explained as the behaviour of an elastic interface pinned by a random potential that is generated by defects, which appear to be strongly sample-dependent and affected by various elastic, microstructural and other extrinsic effects. Theoretically, connecting the zero-kelvin, first-principles-based, microscopic quantities of a sample with finite-temperature, macroscopic properties such as the coercive field is critical for material design and device performance; and the lack of such a connection has prevented the use of techniques based on ab initio calculations for high-throughput computational materials discovery. We use molecular dynamics simulations of 90° domain walls (separating domains with orthogonal polarization directions) in the ferroelectric material PbTiO₃ to provide microscopic insights that enable the construction of a simple, universal, nucleation-and-growth-based analytical model that quantifies the dynamics of many types of domain walls in various ferroelectrics. We then predict the temperature and frequency dependence of hysteresis loops and coercive fields at finite temperatures from first principles. We find that, even in the absence of defects, the intrinsic temperature and field dependence of the domain-wall velocity can be described with a nonlinear creep-like region and a depinning-like region. Our model enables quantitative estimation of coercive fields, which agree well with experimental results for ceramics and thin films. This agreement between model and experiment suggests that, despite the complexity of ferroelectric materials, typical ferroelectric switching is largely governed by a simple, universal mechanism of intrinsic domain-wall motion, providing an efficient framework for predicting and optimizing the properties of ferroelectric materials.

Electrocatalytic CO₂ capture and H₂ storage properties of carbon-boron-nitrogen based nanotube/graphene nanomaterial structures are explored computationally. Using density functional theory incorporating long range dispersion corrections, we investigated the predicted adsorption behavior of CO₂, H₂ and other gases on model nanotube and 2D graphene-like structures with varying charge states. Pyridinic nitrogen functionalities are found to induce an increasing CO₂ and H₂ adsorption energy upon electron injection, leading to a highly selective adsorption in comparison with N₂. Similar behaviour has now been found for related 2D C-B-N materials, both conducting and semiconducting in nature. This functionality is intrinsically reversible since capture/release can be controlled by switching the charge carrying state of the system on/off. This phenomenon is verified for a number of different models and theoretical methods, with clear ramifications for the possibility of implementation with a broader class of graphene based membranes.

Creating nanostructures within alloyed bulk thermoelectric materials can greatly decrease the lattice thermal conductivity of the material and thereby increase the thermoelectric efficiency of these materials. However, the rational design of thermoelectric alloys with even larger figures of merit will require a quantitative knowledge of the electronic and thermal properties and phase stability of nanostructured semiconductor materials. Here, we show how first-principles based calculations can reveal the intricate but tractable relationships between properties for optimization of thermoelectric performance. The integrated optimization includes a multipronged strategy: 1) significant reduction of the lattice thermal conductivity with multi-scale hierarchical architecturing, 2) large enhancement of Seebeck coefficients with intra-matrix electronic band convergence engineering, and 3) control of the carrier mobility with band alignment between host and second phases. These techniques can simultaneously enhance the power factor and reduce the lattice thermal conductivity, thereby leading to high efficiency thermoelectric materials.

Recent advances in the synthesis of polar organic materials have introduced strategic alternatives to well-known ferroelectric ceramics, such as PbTiO₃ and BaTiO₃ perovskite oxides. In particular, cocrystals formed by combinations of diisopropylammonium (DIPA) molecules with Chlorine and Bromine counterions were shown to possess large room-temperature spontaneous polarization [[1],[2]]. This opens up exciting new avenues for their incorporation in modern all-organic electronic devices, including capacitors, as well as piezoelectric, pyroelectric, and electro-optical sensors.

In this investigation, we present a systematic density-functional theory (DFT) based study of structural, electronic, dielectric and polar properties of the DIPA-X molecular cocrystals, with X representing various halogens, namely, Chlorine, Bromine and Iodine, as well as the polyatomic nitrate anion. For each of the cocrystals, the polar properties for all of the relevant polymorphs were evaluated and decomposed into contributions from individual structural units by computing the charge centers of their maximally localized Wannier functions (MLWFs). Our calculations reveal antipolar dipole moment arrangements present in all of the cocrystals, with non-zero spontaneous polarization emerging as a result of small cantings of the dipole moments. These insights, combined with the elucidated influence of the coformer/counterion chemical identity on the co-crystal properties, can be used for the design of high-performance flexible-ferroelectric materials for novel all-organic technological applications.







[[1]] D.-W. Fu, H.-L. Cai, Y. Liu, Q. Ye, W. Zhang, Y. Zhang, X.-Y. Chen, G. Giovannetti, M. Capone, J. Li, R.-G. Xiong, Science, 2013, 339, 425.

[[2]] D.-W. Fu, W. Zhang, H.-L. Cai, J.-Z. Ge, Y. Zhang and R.- G. Xiong, Adv. Mater., 2011, 23, 5658.

The successful isolation of graphene from graphite in 2004 opened up a new era of two dimensional (2D) nanomaterials in materials science and nanotechnology. Numerous graphene-based materials and graphene-like single-layer materials have been developed. In graphene, carbon hexagon is the building block, and all carbon atoms are in the form of sp² hybridization and arranged in a hexagonal honeycomb pattern. In 2015, we proposed a 2D carbon allotrope, penta-graphene, which is composed of only carbon pentagons and is a hybridized system with sp² and sp³ bonding (Proc. Natl. Acad. Sci. 112, 2372, 2015). Based on extensive analysis and simulations we demonstrated that penta-graphene has exceptional properties including an unusual negative Poisson’s ratio, a large band gap, and an ultrahigh mechanical strength. More importantly, the unique 2D structure of penta-graphene offers an ideal structural model for developing new 2D materials with fascinating properties. In this talk, I will review recent advances in 2D penta-graphene-like materials including penta-CN₂, penta-Si, penta-BN₂, penta-SiC₂, penta-CB₂, and penta-AlN₂ with a focus on their exotic properties and potential applications. I then will discuss the functionalization of penta-graphene-like materials going beyond graphene.

Although linear elasticity and quasiharmonic models of lattice dynamics are powerful tools for assessing material properties, there are important materials for which these simple approximations are inadequate or fail entirely. In particular, these methods predict dynamical instability of many high-temperature crystal phases, indicating that their zero-K potential energy surfaces are not convex. Unfortunately, most theoretical methods that incorporate anharmonicity are difficult to apply or are semi-empirical, severely encumbering predictive simulation for many materials systems.

We have developed a new simulation framework that enables the prediction of finite-temperature properties of strongly anharmonic and mechanically unstable crystal phases from first principles. Our framework borrows concepts from group theory and alloy theory to construct basis functions of lattice deformation and atomic displacement that are a priori invariant to rigid-body translation and rotation, as well as all space-group operations of the ideal reference crystal. These basis functions can be used to specify order parameters, guide exploration of the crystal potential-energy surface, and efficiently parameterize Hamiltonians with nearly first-principles accuracy. Hamiltonians constructed within this framework can be used within molecular dynamics or Monte Carlo simulation to predict free energies, structural phase transitions, nonlinear elastic properties, or thermal transport coefficients. We will discuss the broad relevance of this approach to thermoelectric semiconductors, metal oxides, and metal hydrides.

Many processes in organic devices, such as charge transport, are critically influenced by the packing of the molecules, and thus ultimately influenced by the crystal structure of the semi-crystalline semiconductors. Semi-crystalline materials are likely to exhibit different polymorphs, depending on the structure of the side chains, the molecular weight and the processing conditions. As a consequence, designing new materials for improved efficiency is a truly multi-parameter problem and often yields disappointing results. Being able to predict the likely crystal polymorph of a given molecule as a function of temperature, or in other words, drawing the phase diagram of the material, is therefore an important challenge. In this work, we choose as a model system 3-hexylthiophene (3HT) oligomers of different length. We represent the molecules by a modified version of the force field by Moreno et al.¹ We investigate potential crystal structures by selecting a number of space groups that are consistent with the likely trans-geometry of the 3HT oligomers and then building a crystal. We find the unit cell parameters for each space group (p-p stacking, lamellar stacking distances and the unit cell angles), by scanning the parameter space using molecular mechanics simulations. We rank the minimised crystal structures according to the Helmholtz free energy, rather than potential energy, in order to evaluate the results against thermodynamic data. Finally, we compare our results with a thermodynamic study of 3HT oligomers.2,3 This step allows us to validate our method and to ultimately accept or reject the proposed crystal structures.

References
1 M. Moreno, M. Casalegno, G. Raos, S. V. Meille, R. J. Po J. Phys. Chem. B, 2010, 114, 1591-1602.
2 F. P. V. Koch, P. Smith, M. Heeney J. Am. Chem. Soc. 2013, 135, 13695-13698
3 F. P. V. Koch, M. Heeney, P. Smith J. Am. Chem. Soc. 2013, 135, 13699-13709

A set of computational modeling methods that contains a brute-force algorithmic generation of chemical isomers, molecular dynamics (MD) simulations, and density functional theory (DFT) calculations is reported, applied to search and evaluate near 3,000 probable molecular structures of polydopamine (PDA) and eumelanin. All probable early-polymerized 5,6-dihydroxyindole (DHI) oligomers, ranging from dimers to tetramers, have been systematically analyzed to find the most suitable structural connections as well as to propose a set of molecular models that represents the chemically diverse nature of PDA and eumelanin. Our results indicate that more planar oligomers have a tendency to be more stable, which is in a good agreement with experimental observations, suggesting that PDA and eumelanin are composed of near planar oligomers that appear to be stacked together via π-interactions to form graphite-like layered aggregates. We also show that there is an elite group of tetramers notably more stable than the others, implying that even though there is an inherit chemical diversity in PDA and eumelanin, the molecular structures of the majority species are quite repetitive. Our results also suggest that larger oligomers are less likely to form, which also agrees with experimental measurements, supporting the existence of small oligomers instead of large polymers as main components of PDA and eumelanin. In summary, this work brings an insight into the controversial molecular structures of PDA and eumelanin, explaining some of the most important structural features of these materials, and providing a set of molecular models for more accurate modeling of eumelanin-like materials.

Polyvinylidene fluoride (PVDF, -CH₂-CF₂-) is a well-known multifunctional polymer with outstanding polar, pyroelectric and ferroelectric properties. By contrast, such properties have never been investigated in analogous polymers derived from 2,3,3,3-tetrafluoropropene (-CH₂-CF(CF₃)-). In this investigation, we have modeled and compared the properties of PVDF, poly(tetrafluoroprene) and their random copolymers of various compositions. This was done by utilizing a plane-wave density functional theory (DFT) approach to study the structure-property relations for the all-trans, or β phase, PVDF crystals where some of the VDF units were converted to fluoropropene by replacing one of the fluorine atoms by the trifluoromethyl (-CF₃) group. The contributions of individual monomer dipole moments within the polymer crystal to its total polarization, were identified from the calculated maximally localized Wannier function (MLWF) centers. The influence of copolymer composition on the magnitude of monomer dipole moments and the value of the resulting total polarization were then examined for a variety of different structural polymer-crystal arrangements.