Thomas Young Centre Student Day 2017

Friday 24th February 2017 09.30 to 16.00

G O Jones Lecture Theatre School of Physics & Astronomy Queen Mary University of

Time 09:30 Welcome with Tea & Coffee 10.00 Introduction to the TYC Day 10.10 Structure prediction and computational materials discovery Professor Chris Pickard, University of Cambridge

Session 1 10.40 Tight-binding approach to polaron states in fullerene adducts Beth Rice, 10.55 A novel method for constructing the micelle intrinsic surface: a means to elucidate true interfacial structure Daniel Allen, Kings College London 11.10 Molecular dynamics study of the effect of lipid composition on small molecules and drugs permeation, through model cell membranes Michail Palaiokostas, Queen Mary’s University 11.25 Accelerating the evaluation of Hartree-Fock exchange forces in molecular dynamic calculations Guido Falk von Rudorff, University College London 11.25 – 11.45 Tea & Coffee Break Session 2 11.45 Plates and Needles: A Computational Study of the Growth of Hexagonal Ice Maxwell Fulford, Kings College London 12.00 Porous organic cages and the Xe/Kr selectivity Marcin Miklitz, Imperial College London 12.15 Atomistic Modelling of Ageing in Ferroelectrics Jacob Chapman, University College London 12.30 Corrosion scale dynamics Markus Tautschnig, Imperial College London 12.45 – 14.15 Lunch Session 3 14.15 Intrinsic charge trapping in amorphous Al2O3 and its role in positive charging Oliver Dicks, University College London 14.30 Multiscale Modelling of Delayed Hydride Cracking Mitesh Patel, Imperial College London 14.45 Inverse Temperature Dependence of Nuclear Quantum Effects in DNA Base Pairs Wei Fang, University College London 15.00 Relativistic Origin of Slow Electron-Hole Recombination in Hybrid Halide Perovskite Solar Cells Pooya Azarhoosh, Kings College London 15.15 Impressions of Materials Modelling Dr Gerhard Goldbeck, Goldbeck Consulting 15.45 Short Break 16.00 Prize Announcement and Refreshments

Poster Presentations

Machine Learning of Quantum Forces: building accurate force fields via "covariant" kernels". Aldo Glielmo King’s College London

In recent years, the construction of data-driven force fields via Machine Learning methods proved to be a promising route in order to bridge the gap between accurate (but slow) quantum chemical calculations and fast (but unreliable) classical interatomic potentials [Behler et al. (2007), Bartók et al. (2010), Li et al. (2015)]

We developed a new scheme [Glielmo et al. (under review), https://arxiv.org/abs/1611.03877] that accurately predicts forces as vector quantities, rather than sets of scalar components, by Gaussian Process (GP) Regression. This is based on matrix-valued kernel functions, to which we impose that the predicted force rotates with the target configuration and is independent of any rotations applied to the configuration database entries. We show that such "covariant" GP kernels can be obtained by integration over the elements of the rotation group SO(n). Remarkably, in specific cases the integration can be carried out analytically and yields a conservative force field that can be recast into a pair interaction form. The accuracy of our kernels in predicting quantum-mechanical forces in real materials is investigated by tests on pure and defective Ni and Fe crystalline systems.

Such learning algorithm can be further used to build a measure of complexity of physical systems. Indeed, this can be defined as the number of canonically sampled data needed to achieve low prediction error with high probability.

Thermal transport across nanoparticle-fluid interfaces Anna Sofia Tascini Imperial College London

The potential uses of nanoparticles in several biomedical applications, such as drug delivery, hyperthermia treatment, magnetic resonance imaging and tissue repair, have been recognized in several works. Many of these applications involve transport of heat from the nanoparticles to the surrounding media and exploit the resulting increase in local temperature to, for example, destroy target cancer cells or modify cell membrane properties to allow drug delivery into the cell. A full description of the mechanism of heat flow across a solid-fluid interface at the nanoscale requires an understanding of the complex interplay between various properties of the interface, such as surface free energies, curvature and microscopic structure. Indeed, the heat transport between two different phases is characterised by a temperature drop at the interface, quantified by a thermal boundary resistance, known as the Kapitza resistance. Using non-equilibrium molecular dynamics computations, we have found that the thermal resistance of the interface depends strongly both on the wetting characteristics of the nanoparticle–fluid interface and on the nanoparticle size1. Strong nanoparticle–fluid interactions, leading to full wetting states in the host fluid, result in high thermal conductances and efficient interfacial transport of heat. Moreover, the strength of the fluid-nanoparticle interactions has been found to influence the variation of the thermal conductance with particle size, with strongly hydrophilic particle showing the strongest curvature dependence.

1 Tascini, A. S., Armstrong, J., Chiavazzo, E., Fasano, M., Asinari, P., & Bresme, F. (2017). Thermal transport across nanoparticle-fluid interfaces: the interplay of interfacial curvature and nanoparticle-fluid interactions. Phys. Chem. Chem. Phys. http://doi.org/10.1039/C6CP06403E

Realistic Membrane Simulations for Nanoporous Materials Aydin Ozcan University College London

Membranes have huge importance for various application areas such pharmaceutical, petro-chemical, materials purification and biomedical industries. For example, the market share of dialysis applications of membranes alone exceeds one billion dollar annually. Therefore, designing target-specific membranes have a great importance and computer simulations open a window to that purpose by describing the interaction between molecules and membranes. In our study, we introduced a new non-equilibrium molecular dynamics simulation method to perform “realistic” permeation simulations for molecules across a membrane. The methodology is based on taking the concentration of species at the inlet and outlet of the membrane as “collective variables” and controlling them by self-adaptive bias forces. We demonstrate the new method for methane, ethane, ethylene permeation and ethane/ethylene separation through a flexible ZIF-8 membrane. The results show that the new method successfully maintains a concentration gradient between the inlet and outlet of the membrane facilitating the diffusion of molecules. The main novelty of the methodology introduced in this study is that it allows continuous steady state simulations of mixture permeation through a membrane while maintaining the concentration of the species at the inlet and outlet of membrane.

Deformation modelling of hardfacing alloys Bartosz Barzdajn Imperial College London

Cobalt based hardfacing alloys are known to have excellent tribological behaviour under sliding conditions. They also have good corrosion resistance and high strength. Therefore, they have been traditionally used in valves for nuclear power plants. Unfortunately, cobalt under neutron radiation forms radioactive isotopes, which leads to radiation build-up as wear particles are released to the circuit. This results in an increased exposure to maintenance crews. As a consequence, iron based alloys (e.g. NOREM, EB 5183, EVERIT 50) have been considered as a potential replacement. This project focuses on a Rolls-Royce patent duplex stainless steel - RR2450. The main goal of this theoretical study is to identify the key characteristics which define a material with excellent tribological properties. Such characteristics may include texture, morphology and, in the case of more complex materials, phase fractions. In our analysis we focus on galling – a type of catastrophic wear, with a view to introduce a physically based definition of galling or relevant performance measures.

In order to achieve our goal, we explicitly modelled, using a crystal plasticity finite element method (CPFE), two representative microscale bodies with rough surfaces that have been brought into contact. We have used Abaqus 6.14 with our own in house developed user material definition (UMAT), that implements physically based slip rule and, implemented in Abaqus pressure-overclosure contact relationship.

Current state of knowledge suggest that, plastic deformation and surface profilometry is playing a critical role in susceptibility to galling. Therefore, after thorough investigation of our models under different loading conditions, we introduce a measure that couples accumulated plastic strain and contact surface profile. Moreover, early analysis suggests that this measure, using framework of survival analysis, can be used to provide reasonable predictions of galling probability for austenitic steels.

Design of Electric Field Controlled Molecular Gates Hosted in Metal-Organic Frameworks Benjamin Tam University College London

Molecular Machines and Metal Organic Frameworks (MOFs) have received much attention in recent years. They are promising for a wide range of potential applications, from gas adsorption to catalysis and from drug delivery to manipulation of chemical structure in nanoscales. Here we describe the concept of hosting a molecular turnstile on hybrid organic – inorganic crystalline porous structures controlled by an electric field in order to perform tasks, such as storage and release of methane molecules. We selected Mg-MOF-74 to host a turnstile type molecular machine. The machine contains a negative charged tri-fluoromethyl groups and positive charge methyl groups that response to electric field. The simulations are based on the universal force field (UFF) and charges are taken through planewave Density Functional Theory (DFT) optimized structure. Proposed molecular turnstiles were stable thanks to strong binding between the bicarboxyl groups and open metal sites (OMS). No methane incursion through the molecular turnstile at the closed position and methane were discharged by rotation of the turnstile due to switch on of the electric field.

The Origin of Uniaxial Negative Thermal Expansion in Layered Perovskites Chris Ablitt Imperial College London

In recent years, uniaxial negative thermal expansion (NTE), the effect whereby a material contracts along one axis with 1 increasing temperature, has been identified in the Ruddlesden-Popper oxide (RP) Ca3Mn2O7 , which is a member of a class of layered perovskites with general formula An+1BnO3n+1. NTE of any sort is extremely rare in pure ABO3 perovskite oxides but has been reported in a number of systems with perovskite-like layered structures, typically along the axis of layering. Taking Ca2GeO4, an RP in which n=1, as the most distinct structure from the n=∞ limit of ABO3 perovskite, we performed first-principles lattice dynamical calculations within the quasi-harmonic approximation and were able to quantitatively reproduce experimentally measured uniaxial NTE in Ca2MnO4 with reasonable accuracy. We find that NTE is driven primarily by soft phonons, behaving as rigid unit modes (RUMs)2 of octahedral-tilt character, as we proposed previously3, but that NTE is only predicted over a wide temperature range once effects from the strongly anisotropic elastic constants are considered. It turns out that the local structure at the interface of layered RP materials gives rise to an atomic mechanism to facilitate a very compliant coupling between crystal axes via internal degrees of freedom. Comparing the anisotropic elastic constants of Ca2GeO4 with those of other layered perovskite materials known to display NTE, we propose that a compliance to cooperative strains of different axes, coupled via internal structural degrees of freedom introduced as a result of layering, is a general feature of uniaxial NTE layered perovskites4.

[1] M. S. Senn et al., Phys. Rev. Lett, 2015, 114, 035701. [2] M. T. Dove et al., Min. Mag. 1995, 59, 629. [3] M. S. Senn et al., J. Am. Chem. Soc. 2016, 138, 5479 [4] C. Ablitt et al., submitted awaiting review.

Embedding tight binding within bond order potential simulations Edmund Simpson King’s College London

In most defect simulations the region of interest is small compared to the size of cell required to reduce strain effects, such as those between dislocations in a periodic cell or between a dislocation and a fixed boundary in an isolated cell. A better approach may be to simulate this inner region with a resource intensive model, like tight binding, surrounded by a region simulated with a much simpler model, like the BOP. Green’s functions are an ideal method for this approach as the interaction between the regions may be included as a perturbation to the GF describing the entire system and can help to further diminish the computational requirements. Once this method is complete it may be used to simulate the interaction between dislocations, grain boundaries and other defects, with interstitial hydrogen in order to study hydrogen embrittlement.

The Application of a Molecular-Continuum Coupling Strategy for the Modelling of Liquid Lubricants Eduardo Ramos Fernandez Imperial College London

The field of nanotribology has (to date) remained slightly detached from mainstream macro-scale tribology, focusing primarily on specialized nano-scale applications. At the macro- and mesoscopic levels, continuum models are often able to correctly model fluids. However, at smaller scales, continuum models do not consider the atomic nature of matter and can sometimes fail to capture the essential physics. In such cases, explicit molecular models must be employed, for example to model a liquid-solid interface. The development of a truly multi-scale approach, which spans nano- to macro-scales, is a decisive step forward in understanding engineering tribological interfaces. Hybrid methods[1], where atomistic simulations such as molecular dynamics (MD) and continuum computational fluid dynamics(CFD) inter-operate, offer a solution that combines the strengths of both paradigms. The aim of this project is to model contact-lubrication problems with a multi-scale simulation methodology. To achieve this, an in-house coupling software (CPL_library [2]) has been used to couple MD (LAMMPS) and CFD software (OpenFOAM). CPL_library implements a well-defined interface to facilitate the communication between two arbitrary MD and CFD codes. This employs a domain decomposition approach with an overlapping region where real time data exchange takes place. The method has been applied to study a parallel flow of n-alkanes in contact with an atomistic surface. A coupled simulation approach allows low shear rates which would not be possible with MD alone and to our knowledge have not yet been explored in the literature.

[1] K.Mohamed, A.Mohamad(2010) MANO,8,3,283–302 [2] www.cpl-library.org

Many-Body calculations using Dynamical Mean Field Theory and Iterative Perturbation Theory Evan Sheridan King’s College London

Dynamical Mean Field Theory (DMFT) is a-state-of-the-art technique used to treat the strong electronic interactions in correlated materials. It maps a quantum many-body problem on a lattice on to a many-body problem on a single site, i.e. an impurity problem. We provide a pedagogical introduction to an implementation of DMFT on the Bethe lattice, and we briefly explain the main concepts behind the DMFT self-consistency equations. In our example, Iterated Perturbation Theory (IPT) is used as a quantum impurity solver. We also present an extension of a quantum monte carlo approach combined with the IPT approximation. Preliminary results obtained within the IPT-QMC method indicate the existence of the expected Mott transition.

Controlling structural change in multiferroic materials by ultrafast laser excitation. Fangyuan Gu Imperial College London

Recently, multiferroics have gained a lot of attention due to their functional properties arising from coupling between different order parameters. Our goal is to understand how optical excitation can be used to effect structural changes in multiferroic materials such as ferroelectric domain reversal and photo-induced domain nucleation. This will help develop basic physics of perovskite multiferroics, interpret results of recent pump-probe experiments and open new possibilities in the design and optimisation of optically controlled devices. In this project, we use electronic structure methods and atomistic simulations to understand the influence of intense ultrafast laser pulses on the domain structure and dynamics in multiferroics. We begin by studying a prototypical perovskite multiferroic, BaTiO3, which has coupled ferroelectric and ferroelastic order. A constrained density functional method is used to calculate the effects on lattice dynamics of above- band-gap electronic excitation. Our preliminary calculations indicate a softening of the inter-atomic forces and a reduction in polarization in the excited system, along with a marked reduction of the energy barrier for ferroelectric domain reversal. This suggests a photo-induced lowering of the coercive field and the phase transition temperature. Furthermore, we studied the dependence of phonon frequencies and structural stability on excited carrier excitation by calculating the full phonon dispersion using the PhonoPy code. This gives us information on timescales of a few phonon periods and length scale of several perovskite unit cells (4-20 ). Future work will be focused on the development of atomistic potentials for photo-excited multiferroic materials, which will be used for studies of significantly longer length- and time-scale structural changes like ferroelectric domain reversal.Å

Shear and bulk viscosities in common fluids – comparison between atomistic and coarse grained models Frederike Jaeger Imperial College London

The shear viscosity is a commonly used quantity in fluid dynamics and many experimental measurements exist. The bulk viscosity, however, is typically neglected in modelling. This is in part due to the lack of experimental results and difficulty in determining accurate values. In the typically used Navier-Stokes equation the shear viscosity, independent of the state of the fluid, is the only quantity breaking universality. The bulk viscosity is neglected entirely. In order to have a complete picture of fluid flow for various applications both viscosities should be included in mathematical modelling. We calculate both the shear and bulk viscosities and their dependence on the density for common fluids using equilibrium molecular dynamics. In addition to using full atomistic models the accuracy of coarse-grained models, here SAFT-gamma- Mie, is also evaluated. We find that the shear viscosities are very similar for fully atomistic and coarse-grained models. However, the bulk viscosities show some differences.

Solvation forces in calcium carbonate nano-confinement of water Goran Brekke Svaland Imperial College London

Structural forces or solvation forces, arise from the ordering of liquid molecules confined between any two surfaces, including two solute molecules. The ordering of the confined liquid depends on surface properties such as surface structure and hydrophobicity. In addition, structural properties of the liquid are very important. Even though the origin of structural forces is well understood, the determination of such forces is difficult experimentally, but its implications are important for understanding of crystal growth and dissolution initiated at the atomic scale. In this research, molecular dynamics simulations of confining calcite and aragonite crystals have been employed to determine the solvation forces and the free energy of separation of the confining crystals in water. Calcium carbonate is an abundant material in the earth’s crust, and is the main building block in shells and bones and other cementing materials. This research is part of the NanoHeal project which aims to gain a greater insight into this fundamental material, and could lead to the discovery of new self-healing cements and cementing pastes nano-tailored for a specific purpose. It could also be the key to determine the fate of shell bearing creatures as the temperature and acidity of the ocean continues to rise. Density profiles and orientational order of the confined liquid are computed relative to the confined region and shows strong surface dependence. Since the solvation force is dependent on the ordering of the liquid, this is also reflected in the resulting disjoining pressure, which is the pressure that needs to be applied in order to keep the surfaces at a certain separation. What will happen when we add salts to the solution? And how does epitaxy impact the solvation forces?

Parrinello-Rahman Molecular Dynamics using a modified GPT Potential Guy Skinner King’s College London

Classical molecular dynamics (MD) is dominated by the use of empirical potentials. However, there remain questions over their ability to describe physics for which they were not expressly fit. Before the advent of density functional theory and modern computer processing power, potentials were commonly found directly from perturbation theory using the generalised pseudopotential theory (GPT). A key assumption underlying these potentials was that the electron density was uniform. Such an assumption breaks down in the presence of point defects and screw dislocations. The same is true for arbitrary strains of the crystal as required by Parrinello-Rahman MD. By allowing the GPT potential to depend on a representation of the local atomic density, these issues can, to a certain degree, be overcome. The effectiveness of these modifications was checked by calculating quantities such as expansivity coefficients, lattice constants and elastic constants at finite temperature and zero stress for pure magnesium

Polaron hopping model on organic semiconducting crystals: rubrene, pentacene, C60 Hui Yang University College London

Organic semiconducting materials (OS) combine several advantages such as easy fabrication, mechanical flexibility and low cost, which has contributed to their widespread applications in light emitting diodes, field effect transistors and photovoltaics. Interestingly, while their essential characteristic is the conduction of electrons, the physical mechanism of charge transport in these materials is still unclear. Here we evaluate the applicability of the commonly assumed polaron hopping model for a number of well known organic semiconductors (rubrene, pentacene and C60). This model is based on the assumption that the charge carrier is localized, i.e. forms a polaron that hops from one molecule to the next. We have calculated the relevant parameters that determine whether a polaron forms or not: electronic coupling matrix element and reorganization energy for the above materials using quantum chemical calculations and molecular dynamics simulations. We find that neither for rubrene nor pentancene the hopping model is justified due to the relatively large electronic couplings in at least one crystal direction. For C60 the couplings matrix elements are smaller and a small but finite barrier for charge transport exists (in any transport direction). Despite the theoretical problems surrounding the polaron transport model, we find that mobilities based on this model (obtained from Kinetic Monte Carlo simulation) reproduce very well the experimental mobility and anisotropy in pentacene and rubrene. However, it fails to reproduce the correct temperature dependence of mobility, predicting a power law decay T-n, n=1, whereas in experiment n=1.5-2. This investigation clearly calls for more advanced simulation approaches, such as non-adiabatic molecular dynamics simulation approach, for modeling charge transport in organic semiconductors.

Non-local Model for Diffusion-Mediated Dislocation Climb and Cavity Growth Iacopo Rovelli Imperial College London

To design efficient thermal recovery procedures for structural materials in fusion energy applications it is important to characterise quantitatively the annealing timescales of radiation-induced defect clusters. With this goal in mind, we present an extension of the Green’s function formulation of Gu et al. for the climb of curved dislocations, to include in the same framework the evaporation and growth of cavities and the effects of free surfaces. This paper focuses on the mathematical foundations of the model, which makes use of boundary integral equations to solve the steady-state vacancy diffusion problem. Numerical results are also presented in the simplified case of a dilute configuration of prismatic dislocation loops and spherical cavities in a finite-size medium, which show good agreement with experimental data on defect annealing in ion-irradiated tungsten.

Polymeric CNT composites: atomistic simulations of interfacial properties Jacek Golebiowski Imperial College London

Functionalized carbon nanotube (CNT)/polymer composites have received significant interest as promising structural materials with applications in the most demanding areas of industry such as aerospace and ballistic protection. In order to optimise the properties of this class of materials, it is imperative to understand how load is transferred through the interface, between the functionalized CNT and polymer matrix with the aim to identify the key factors determining the interfacial shear strength and dominant failure mechanisms.

Computational investigation of the interface requires simulations of tens of thousands of atoms in order to accurately describe the movement of polymer chains; however, critical interfacial failure involves changes in local chemistry such as bond-forming and breaking effects, necessitating a quantum-mechanical (QM) treatment. These issues are addressed by employing a quantum/classical hybrid simulation technique known as `Learn on the Fly' [1]. In this approach, molecular dynamics based on a classical forcefield is used to simulate the majority of the system under strain, while regions of particular interest, such as the functional groups where changes in electronic structure are likely to occur, are investigated using QM methods resulting in an accurate description of bond-breaking processes.

[1] Gabor Csányi, T. Albaret, M. C. Payne, and A. De Vita. Phys. Rev. Let. 93(17):1–4, 2004.

Generation of Oxygen Vacancies in a-HfO2 Jack Strand University College London

Modelling the generation of oxygen vacancies in amorphous HfO2 is important for improving its functionality in transistor and resistive random-access memory devices. Degradation in transistor gate stacks is believed to be rooted in the existence of defects such as oxygen vacancies in the gate oxide. The mechanism of such breakdown is believed to involve already-present defects as well as defects created during device operation. Also, conductive filament formation in HfO2 based RRAM has been shown to be connected to oxygen vacancy generation. Hybrid DFT calculations were used to investigate the generation of oxygen vacancies in a-HfO2 via the creation of Frenkel pairs. Defects were simulated under conditions of electron or hole injection and calculations of the stability of Frenkel pairs and the barrier heights to their formation were made.

Fundamentals of dislocations in motion Jonas Verschueren Imperial College London

Our understanding of dislocation mobility - quantifying the relationship between the force on a dislocation and its resulting velocity - is largely based on experiment. However, the validity of mobility laws extracted from this work breaks down for fast travelling dislocations moving with speeds comparable to the speed of sound in the medium. In the last 20 years, large-scale non-equilibrium molecular dynamics simulations have been used to simulate qualitative mobility laws for fast travelling dislocations. However they have contributed little to our fundamental understanding of dislocation mobility in this regime. We will show how lattice-dynamics models may provide an accurate description of the dislocation - phonon interactions which are widely believed to play a crucial part in descriptions of dislocation mobility. Ultimately, a physically motivated theory of dislocation mobility in the pure-glide regime in good quantitative agreement with existing simulation data is the aim of this project. This could shed light on the phenomenology associated with these fast travelling dislocations. Debate on this topic has been ongoing for over half a century and is problematic given that in this regime, the usual approximations by which elasticity theory is linearised are violated and the quasi-static approximation no longer holds.

Structural and viscoelastic properties of a colloidal suspension using MD-SRD Juan David Olarte Plata Imperial College London

Cement has a very complex structure that varies across length scales, and is composed of different mineral phases bound together by calcium silicate hydrates (C-S-H). During the hydration process of cement, C-S-H precipitates as a colloid of particles of few nanometer, which floculate to form larger units up to 100 nm. Predicting the structural and viscoelastic properties of the binding phase of cement is of industrial interest, and understanding the role of surface interactions in macroscopic behaviour is key in achieving control of the material. We propose a model to describe a colloidal suspension of particles which includes hydrodynamic effects, by coupling molecular dynamics (MD) with stochastic rotation dynamics (SRD). Structural properties and aggregation phenomena are explored as a function of the interaction between colloids. We also explore viscoelastic properties of the mixture using non-equilibrium MD simulations.

Calculating Changes in Entropies and Free Energies of Organic Molecules at Insulating Surfaces Julian Gaberle University College London

The challenges and limitations in calculating free energies and entropies of adsorption and interaction of organic molecules on an insulating substrate are discussed. The adhesion of 1,3,5-tri- (4-cyano-4,4 biphenyl)- benzene (TCB) and 1,4-bis (cyanophenyl)-2,5-bis(decyloxy) benzene (CDB) molecules to step edges on the KCl (001) surface and the formation of molecular dimers were studied using classical molecular dynamics. Both molecules contained the same anchoring groups and benzene ring structures, yet differed in their flexibility. Therefore the entropic contributions to their free energy differ which affects surface processes. Using potential of mean force and thermodynamic integration techniques, free energy profiles and entropy changes were calculated for step adhesion and dimer formation of these molecules. However, converging these calculations is nontrivial and comes at large computational cost. We illustrate the difficulties as well as the possibilities of applying these methods towards understanding dynamic processes of organic molecules on insulating substrates.

A fast method for band edge alignment at solid/liquid interfaces Lars Blumenthal Imperial College London

Identifying efficient systems for storing energy constitutes a major challenge in the transition to sustainable and renewable energy sources. A potential solution is to store energy in a chemical form, for example, in hydrogen gas produced by photoelectrochemical water splitting. However, the efficiencies of current photoelectrochemical water splitting devices are too low for industrial applications. A detailed understanding of the electronic structure of photoelectrodes, in particular the alignment of the electrodes' electronic band edge positions with the relevant redox potentials of water, is required to guide experimental progress towards increased efficiencies. To address this problem, we introduce a new approach based on the combination of many-body perturbation theory within the GW method for the electronic structure of the photoelectrode and joint density functional theory for the description of solid-liquid interfaces. We applied this approach to the prototypical photoelectrode material titanium dioxide in its rutile phase and determined the position of the valence and conduction band edges with respect to the redox potentials associated with the hydrogen and oxygen evolution reactions.

Embrittlement of Ni-based superalloy by oxygen. Luca Cimbaro Imperial College London

Ni-based superalloys are the most frequently used of all the classes of superalloys, because they combine high strength with controlled thermal-expansion features. A notable example of their application lies in the aircraft industry where these superalloys, such as IN718 or RR1000, are employed as main components of the aircraft engines, such as rotor disks. However, in recent years these superalloys at high temperatures, such as 700 °C, have been affected by a new failure process, which shows a time-dependent intergranular crack growth. In order to explain this new failure process, two mechanisms have been introduced, namely Dynamical Embrittlement and Stress Assisted Grain Boundary Oxidation (SAGBO). In this poster, a theory of SAGBO capable of describing the physics of the failure process is outlined.

A Tale of Two Brothers: Understanding the Structure-Dynamics Relationship in Supercooled Water Martin Fitzner University College London

The freezing of water is the most common phase transition on earth with broad impact ranging from industry (e.g. aviation and food sectors) to science (e.g. clouds or cryobiology). The initial spark that starts off the freezing process is known to occur almost exclusively in the presence of foreign materials (heterogeneous nucleation). This opens the question: What is it that these compounds change in the proximity of water and how does it help to promote or inhibit ice nucleation? In this work we aim at shining light on an aspect of nucleation that often goes unnoticed: the liquid dynamics. In recent years some attempts have been made to find meaningful correlation between dynamics and structure, however an unambiguous definition of mobile and immobile molecules was missing. We fill this shortfall by utilizing the recently developed iso-configurational analysis to distinguish between mobile and immobile regions in water supercooled to different degrees. This in turn enables us to establish a clear connection between the dynamics and the structure. Additionally to the appearance of glassy dynamics, a strong distinction between the two regions is found regarding the occurrence of primitive rings and the extent of clustering. Furthermore, we study the dynamics in the vicinity of a pre- formed critical ice-cluster to find a daunting disruption of the dynamics in simulation cells that are usually considered large enough to avoid finite-size effects. These results provide a basis for understanding the dynamics-structure relationship in supercooled water and can be extended to heterogeneous interfaces in the future.

The Role of Water in Material-Driven Fibronectin Fibrillogenesis Mateusz Bieniek King’s College London

An extracellular matrix protein fibronectin (FN) undergoes important quaternary conformational changes upon adhesion to a substrate made of self-assembled monolayers (SAMs) functionalised with ethyl acrylate (EA) group. This conformational transition is necessary to start FN fibrillogenesis or self-association which influences several processes including cell differentiation and migration [1, 2]. Surprisingly, SAMs functionalised with methyl acrylate (MA) group, which differs only by a single carbon, does not trigger the transition. Understanding the cause of the conformational transition will aid the design of better materials to drive fibrillogenesis, which has potential applications in tissue engineering, cell culture and growth factor delivery [3, 4]. We used molecular dynamics to understand the difference in interactions between FN and the two SAMs functionalised with EA and MA. In our simulations, on the two SAMs, we investigate the domains FNIII 9-10 which are known for their crucial role in the beginnings of fibrillogenesis in vivo through the use of integrin receptors [4]. We aim to understand how the domains interact with the two SAMs and how the differences may affect fibrillogenesis. Our initial results, which show that adhesion takes place on EA SAMs, but not on MA SAMs, are consistent with the latest experiments from Prof. Salmeron-Sanchez’s group. Water hydration around the head groups of the two different SAMs is suspected to play an important role in this adhesion discrepancy. We are currently investigating how the structure of water around the functional groups affects the adhesion of the fibronectin domains, which is of significant importance in the beginnings of fibrillogenesis.

[1] Salmerón-Sánchez, M., Rico, P., Moratal, D., Lee, T. T., Schwarzbauer, J. E., & García, A. J. (2011). Role of material- driven fibronectin fibrillogenesis in cell differentiation. Biomaterials, 32(8), 2099–105. https://doi.org/10.1016/j.biomaterials.2010.11.057

[2] Gugutkov, D., González-García, C., Rodríguez Hernández, J. C., Altankov, G., & Salmerón-Sánchez, M. (2009). Biological activity of the substrate-induced fibronectin network: insight into the third dimension through electrospun fibers. Langmuir : The ACS Journal of Surfaces and Colloids, 25(18), 10893–900. https://doi.org/10.1021/la9012203

[3] Llopis-Hernandez, V., Cantini, M., Gonzalez-Garcia, C., Cheng, Z. A., Yang, J., Tsimbouri, P. M., … Salmeron-Sanchez, M. (2016). Material-driven fibronectin assembly for high-efficiency presentation of growth factors. Science Advances, 2(8), e1600188–e1600188. https://doi.org/10.1126/sciadv.1600188

[4] Singh, P., Carraher, C., & Schwarzbauer, J. E. (2010). Assembly of fibronectin extracellular matrix. Annual Review of Cell and Developmental Biology, 26, 397–419. https://doi.org/10.1146/annurev-cellbio-100109-104020

Excited state proton transfer and aggregation induced emission in 2'-hydroxychalcone derivatives Michael Dommett Queen Mary

Fluorophores exhibiting excited-state intramolecular proton transfer (ESIPT) are promising candidates for applications in optoelectronics, lasers and imaging.1 Recently, solid-state emitters based on 2'-hydroxychalcone have been synthesized. 2 The compounds are dark in solution but fluoresce when crystalline, a phenomenon known as aggregation induced emission. Here we combine static calculations and non-adiabatic dynamics to investigate competing non-radiative relaxation pathways in 2'-hydroxychalcones based on intramolecular rotation. In gas phase the relaxation pathway is determined by the electron donating power and position of substituents. In solid state, the non-radiative pathways are inaccessible and relaxation is via fluorescence. The results further our aim to understand the relationship between electronic and packing effects of solid state emitters and aid the design of more efficient optoelectronic devices.3 1 V. S. Padalkar and S. Seki, Chem. Soc. Rev., 2015, 45, 169–202. 2 X. Cheng, K. Wang, S. Huang, H. Zhang, H. Zhang and Y. Wang, Angew. Chemie Int. Ed., 2015, 54, 8369–8373. 3 M. Dommett and R. Crespo-Otero, Phys. Chem. Chem. Phys., 2017.

Modelling Photochemical Porcesses in Molecular Crystals Miguel Rivera Queen Mary University of London

The photochemistry of molecular crystals has recently attracted attention due to their technological applications in crystal lasing, solar cells, OLEDs and field effect transistors. In order to accurately model the phenomena central to these systems, excited state calculations need to be carried out. However this becomes prohibitively expensive when in a periodic environment and there remains a lack of methodology when it comes to modelling excitations in molecular crystals.

To address this issue, existing hybrid methods are being built upon. In this protocol, periodic DFT calculations are combined with refined population analysis methods to provide a background charge distribution for single molecule TDDFT calculations. The resulting Coulomb potential favours photochemical mechanisms present only in the condensed phase. This method benefits from only treating one molecule at a high level of theory while describing the electrostatic influence coming from different cell geometries on the molecule.

Calculations are carried out on two 2′-hydroxychalcone derivatives in order to asses the influence of its crystalline environment on its fluorescence and its conical intersection energies. This poster describes the theory and protocol behind the calculations and the results, comparing them to experimental studies. The limitations of the protocol are discussed and possible improvements are suggested, as well as candidate systems for future development.

The impact of combustion-generated moieties on the degradation of ICE related surface materials Panagiotis Simatos Imperial College London

High temperature in-cylinder processes generate chemical species that, along with fuel residuals, can lead to increased material dependent surface degradation. Moreover, interactions between the generated moieties and lubricants can cause increased friction leading to reduced fuel efficiency. The current study aims to determine sets of chemical species distributions close to and on surfaces, as can result from advanced engine operation modes, by means of detailed chemical kinetics. The study will enable the subsequent detailed analysis of material surfaces. The study therefore aims to provide thermochemical boundary conditions for increasingly accurate micro-scale studies of interactions between fluid phase generated species and the piston-liner material. The chemical kinetics of the material surface degradation and the moieties in adjacent fluid are directly linked. The accurate inclusion of chemical source terms in the fluid phase is achieved using transported joint probability density functions (JPDF) methods. The accurate modelling of the surface chemistry is obtained using a systematic reaction class based approach that provides comparatively accurate heterogeneous reaction mechanisms. The current techniques comprise Variational Transition State Theory (VTST) for determination of absorption and desorption pre-exponentials, two-dimensional collision theory for homogeneous surface pre-exponentials and the unity bond index-quadratic exponential potential (UBI-QEP) for the calculation of the chemical barriers. Additionally, more accurate state of the art Density Functional Theory (DFT) based methods are used to generate supplementary data and to provide potential refinements of rate parameters. The transport to the surface is treated using molecular diffusion approximations and Langevin approaches, whereas in- cylinder transport processes are controlled by turbulent diffusion. The transfer of the relevant information across the layers will be treated using systematic modelling of the associated processes. Due to the high cost of the simulations, the computational tools have been parallelised with an efficiency close to the Amdahl's theoretical curve.

Strong coupling in nanoplasmonic structures Peter Fox Imperial College London

Nanoplasmonics takes advantage of carrier-field excitations at metal-dielectric interfaces. The nanoparticle-on-mirror structure (NPoM) supports extremely large field enhancements in the gap due to the plasmonic interaction between NP and its mirror image. Upon placing a molecule or quantum emitter (QE) in the gap strong coupling may take place, which is the coherent exchange of energy between the plasmonic field and the quantum emitter. The strong coupling leads to a splitting of the excited plasmonic mode into two hybrid modes and may be observed as a peak splitting in the far field scattering cross section. We modelled different widths of a QE skin, corresponding to volumes as low as that of a single molecule, and plotted the scattering cross sections and fields in the gap to determine the extent of strong coupling.

Towards a quantitative understanding of ice nucleating ability Philipp Pedevilla University College London

The freezing of water is arguably the most common phase transition on Earth. Almost all of it happens heterogeneously, that is with the help of a variety of substrates ranging from mineral dust to biological material. Heterogeneous ice nucleation plays a crucial role in many technological applications such as aviation, cryotherapy, fossil fuel extraction to the food industry [1,2]. Last but not least ice formation inside clouds is of paramount importance to Earth's climate [3]. Despite the importance of the process, we still lack a fundamental understanding of what makes ice nucleation particles efficient [4]. Most ice nucleating particles, be it of inorganic or organic/biological nature, interact with water through hydroxyl groups on their surface. Here we address the issue of how different patterns of surface hydroxyl groups affect the ice nucleation rate. We show that it is not the particular arrangement of hydroxyl groups itself that dominates whether or not a material is a good ice nucleator. Instead, we propose a simple model that takes into account the energetic and structural information of water adsorbed on a substrate to predict the ice nucleating efficiency. Using this model, are able to classify correctly 88 % of ice nucleating abilities of over 160 different model substrates.

[1] Koh, C.A., 2002. Towards a fundamental understanding of natural gas hydrates. Chemical Society Reviews, 31(3), pp.157-167. [2] Mazur, P., 1970. Cryobiology: the freezing of biological systems. Science, 168(3934), pp.939-949. [3] Murray, B.J., O'Sullivan, D., Atkinson, J.D. and Webb, M.E., 2012. Ice nucleation by particles immersed in supercooled cloud droplets. Chemical Society Reviews, 41(19), pp.6519-6554. [4] Bartels-Rausch, T., 2013. Chemistry: Ten things

Violations of Lowenstein's Rule in Zeolites Rachel Fletcher University College London

The optimization of commercial catalytic processes, and the rational design and development of novel catalysts, relies on the structural elucidation of real catalytic materials. In zeolites, catalytic activity is directly determined by the distribution of charged tetrahedral units of alumina, and associated counter-­‐cations, throughout the aluminosilicate framework. Yet, little is known about the factors governing framework aluminium arrangement in zeolite catalysts. Furthermore, there are no well-­‐defined designed rules that can be employed to infer aluminium position indirectly, only Löwenstein’s rule of “aluminium avoidance”, which forbids the formation of Al-­‐O-­‐Al linkages, and enforcing a minimum Si/Al ratio of 1 to all zeolite frameworks. Throughout the years there has been very little evidence to suggest that Löwenstein’s rule is violated in zeolite materials, so much so that the rule has come to be regarded as a fundamental law of zeolite science, with many computational studies omitting ‘non-­‐Löwensteinian’ ordered frameworks as potentially stable structures to save ‘unnecessary’ computational expense. Here we describe the unprecedented screening of aluminium distribution in catalytically active zeolite SSZ-­‐13 (CHA), at high and low Si/Al ratios in both its protonated and sodium containing forms, H-­‐SSZ-­‐ 13 and Na-­‐SSZ-­‐13, along with a selection of other zeolite frameworks (LTA, RHO and ABW). Our approach involves the use of periodic DFT (PBE) Implemented in the CP2K code. We confirm, that Löwenstein’s Rule is violated in all the protonated frameworks explored, and that there is a preference for aluminium clustering at low Si/Al ratios. However, this trend cannot be extended to the sodium forms of the zeolites, for which a ‘Löwensteinian’ Structure is favoured. Our Results hence demonstrate the influence of cation identity on framework aluminium distribution. And we hope that in realizing the acidic zeolites that we have predicted, we may open new catalytic routes, processes and materials.

Explicit crystal host effects on excited state properties of linear polyacenes: towards a room-temperature maser Robert Charlton Imperial College London

Maser technology has been held back for decades by the impracticality of the operating conditions of traditional masing devices, such as cryogenic freezing and strong magnetic fields. Recently it has been experimentally demonstrated [1] that pentacene in p-terphenyl can act as a viable solid-state room-temperature maser by exploiting the alignment of the low- lying singlet and triplet excited states of pentacene. To understand the operation of this device from first principles, an ab initio study of the excitonic properties of pentacene in p-terphenyl has been carried out using time-dependent density functional theory (TDDFT) implemented the linear-scaling ONETEP software [2], [3]. In particular, we focus on the impact that the wider crystal has on the localised pentacene excitations by performing an explicit DFT treatment of the p- terphenyl environment. We demonstrate the importance of explicit crystal host effects in calculating the excitation energies of pentacene in p-terphenyl, providing important information for the operation of the maser. We then use this same approach to test the viability of other linear polyacenes as maser candidates as a screening step before experimental testing. [1] M. Oxborrow, J. Breeze & N. Alford. Nature, 488: 353−356 (2012). [2] www.onetep.org [3] T. J. Zuehlsdorff et al. J. Chem. Phys., 139, 064104 (2013).

In silico study of modulation of cardiac myosin dynamics by Omecamtiv Mecarbil Shaima Hashem Queen Mary University of London

The myosin superfamily includes motor proteins that contribute to muscle contraction and force generation. One member of this superfamily, cardiac myosin II, is an allosteric protein of great interest as it is involved in the contraction of the heart muscle. Cardiac myosin II mutations are responsible for the emergence of several cardiac diseases like hypertrophic and dilated cardiomyopathies, which can lead to heart failure and sudden cardiac death. Therapies based on small- molecule effectors of myosin have recently started to be explored and they seem particularly promising (Hwang, 2015). In particular, the sarcomeric modulator Omecamtiv Mecarbil (OM) is currently being tested in clinical trials for the treatment of heart failure. The binding site of OM on cardiac myosin has been recently unravelled by x-ray crystallography (Winkelmann, 2015). The drug was shown to bind to a deep pocket close to key regions of the motor domain involved in the propagation of motion from the actin-binding cleft to the lever arm. This suggests that a possible role for OM is to increase the coupling between these regions and hence their efficiency in propagating structural changes. Moreover, the drug was found to induce subtle conformational changes in distant regions close to the ATP binding site of myosin, suggesting the presence of allosteric effects. The goal of our project is to study the effect of OM on myosin dynamics in order to elucidate its mechanism of action using molecular modelling and Molecular Dynamics (MD) simulations. This is essential for the rational design of drugs that can improve the cardiac function in treating heart failure. Indeed, information on the molecular basis of OM action and its interaction with the other functional sites of myosin can be used to improve existing drugs, decrease their side effects and identify new drug targets.

Charge corrections in periodic boundary conditions Thomas Durrant University College London

The use of periodic boundary conditions (PBC) for charged defects ensures an accurate description of the host crystal's bandstructure, but negatively introduces fictitious interactions with image charges in neighbouring cells. These interactions are large when the defect is in a high charge state, or when some dimensions of the simulation cell are small. This last case is especially severe when expensive hybrid functionals are utilised to study multiple interfaces. Several methods to remove this interaction exist in the literature for the bulk case and are widely used to study charged defects. Equivalent methods for surfaces and interfaces with varying dielectric profile are now starting to be developed. We have been developing a new method that can be applied to this problem, by extending the Density Countercharge (DCC) method of Dabo et al. to crystals. Previous work on this method has been limited to charged molecular fragments in vacuum. Using the newly released electrostatics solver DL_MG, we have extended this method to interfaces and defect clusters. We have validated our electrostatic method, and found a strong correspondence with the Lany-Zunger method for bulk crystals. Now, we are starting to look at interfaces.

Towards variable valence charge polarizable atomic model of non-bulk oxides Vadim Nemytov Imperial College London

The aim of our research is to develop accurate interatomic potentials (IP) for atomistic simulations of heterogeneous oxide materials, where surfaces, interfaces, defects, multiple phases, and/or variations in composition, means that transferability of an IP between different local atomic environments is crucial to its accuracy. Pairwise IPs are unsuitable as they do not describe an ion’s electronic response to local changes in its environment. However, they can be supplemented with various response mechanisms; in our model the ions can self-consistently polarize (PA)[1] and/or vary their charges (Q-Eq)[2]. Supplementing a pairwise IP with the PA has a proven track record of greatly improving accuracy[1,3-6], so the current focus is on bringing further improvement by additionally including the Q-Eq response mechanism. Although the Q-Eq model is widely used, there is scant evidence that it provides a substantially-improved description of bonding. Some published work reveals only negligible improvement as compared to fixed-charges model and sometimes even a small disimprovement[7-9]. Thus, a model that includes Q-Eq requires firstly a clear assessment and understanding of its limitations. Using a large set of thermally-disordered configurations containing ~100 atoms, we compare an IP’s forces, stress tensors, and energy differences, to those calculated using density functional theory (DFT). This provides a quantitative measure of the closeness of the IP and DFT potential energy surfaces (PES). We use the “force-matching” method [10] to search each IP’s parameter space to find out how closely it can fit the DFT PES. By including and excluding various terms in the potential, we quantitatively compare each term’s contribution to the overall accuracy of a polarizable Q-Eq model.

We find that the widely-used Q-Eq model can provide an improvement over the PA model in bulk systems, such as TiO2, BaTiO3, and PbTiO3, at the expense of a substantial computational overhead. On the other hand, we find no significant improvement in heterogeneous systems because the model is unable to describe the change-of-valence levels of charge transfer associated with coordination changes. We have explored some of the reasons for the failure of Q-Eq IPs to model heterogeneous systems. It appears likely that this is due to an overly-simplistic mathematical representation of an ion’s “self-energy” – the dependence of its internal energy on its charge and dipole moment. Within the Q-Eq model, the self-energy is simply a Taylor's expansion about a reference charge. We are exploring self-energies that go beyond perturbation theory.

[1] P. Tangney and S. Scandolo. An ab initio parametrized interatomic force field for silica. The Journal of Chemical Physics, 117(19):8898, October 2002 [2] Anthony K. Rappe and William A. Goddard. Charge Equilibration for Molecular Dynamics Simulations, 1991. [3] P. Tangney and S. Scandolo. A many-body interatomic potential for ionic systems: Application to MgO. The Journal of Chemical Physics, 119(18):9673, oct 2003. [4] X. J. Han, L. Bergqvist, P. H. Dederichs, H. Mu ̈ller-Krumbhaar, J. K. Christie, S. Scandolo, and P. Tangney. Polarizable interatomic force field for TiO {2} parametrized using density functional theory. Physical Review B, 81(13):134108, apr 2010. [5] Joanne Sarsam, Michael W. Finnis, and Paul Tangney. Atomistic force field for alumina fit to density functional theory. Journal of Chemical Physics, 139(20), 2013. [6] Joseph John Fallon. Multiscale theory and simulation of barium titanate. PhD thesis, Imperial College London, 2014. [7] V. Swamy, J.D. Gale, and L.S. Dubrovinsky. Atomistic simulation of the crystal structures and bulk moduli of TiO2 polymorphs. Journal of Physics and Chemistry of Solids, 62(5):887–895, apr 2001. [8] Varghese Swamy, Joseph Muscat, Julian D. Gale, and Nicholas M. Harrison. Simula- tion of low index rutile surfaces with a transferable variable-charge TiO interatomic potential and comparison with ab initio results. Surface Science, 504:115–124, apr 2002. [9] B. S. Thomas, N. A. Marks, and B. D. Begg. Empirical variable-charge models for titanium oxides: A study in transferability. Physical Review B, 69(14):144122, apr 2004. [10] F. Ercolessi and J. B. Adams. Interatomic Potentials from First-Principles Calculations: the Force-Matching Method. Europhys. Lett. 26, 583 (1994)

Design strategies for porous organic cage topologies. Valentina Santolini Imperial College London

Porous organic materials are a class of materials that are constructed by intermolecular packing of discrete organic molecules that contain an internal cavity, and are typically solution processable. Many experimental and computational studies have been conducted on porous cages, however design and prediction of new molecular structures still represents a great theoretical challenge.1 In this context, we developed a computational strategy to generate new porous organic molecules and test their experimental feasibility. This is based on an automatic procedure that assembles building blocks with different numbers of reactive ends into cages with underlying polyhedral topologies.2 Once different topologies are generated from an initial pair of precursors, the software characterises the geometry and energy of each molecule to find which structure is experimentally likely to form. All the cages are assembled according to reversible dynamic covalent chemistry reactions3 that allow for thermodynamic control over the product. For this reason, throughout the procedure the cage with the lowest relative energy is considered the most likely experimental outcome. The solvent is only considered when its scaffolding effect plays an important role in keeping the cavity of the cage open.4 The prediction strategy has been successfully tested on a number of experimentally available structures,2 and is currently being used to predict the formation of a collection of novel organic cages in collaboration with the experimental group of A. Cooper (University of Liverpool). A high-throughput screening robot has been used to synthesise 78 new cages, starting from a combinatorial selection of different functional precursors. The software is working in parallel at the prediction of the cages, and results so far are promising. In the future, we are planning to apply the software to the prediction of multi- component systems (where 3 or more different precursors are involved) and to the discovery of promising candidates for synthesis.

1 K. E. Jelfs et. al., J. Am. Chem. Soc., 2013, 135, 9307. 2 V. Santolini, M. Miklitz, E. Berardo, K.E. Jelfs, Submitted. 3 P. T. Corbett et. al., Chem. Rev, 2006, 106, 3652. 4 V. Santolini, G. A. Tribello, K.E. Jelfs, Chem. Commun., 2015, 51, 15542.

Simulation of Pump-Probe Experiment in Small Tetraheme Cytochrome Xiuyun Jiang University College London

Bacteria like Shewanella and Geobacter are species that express multi-heme proteins which enable them to transfer electrons from the inside of the cell to extracellular solid substrates. Studying long-range electron transfer (ET) events in multi-heme proteins is important for the understanding of geochemical cycling of different metals and the design of promising biological nano-materials. The protein we are focusing on, small tetraheme cytochrome (STC), is one of the smallest multi-heme proteins found in bacteria Shewanella Oneidensis, which has only 4 heme cofactors embedded in one protein domain and is a good model for fundamental studies. In order to obtain a molecular-level insight into the ET process, our experimental collaborators docked a [Ru]-ligand (tris(bipyridine)ruthenium) to the STC protein. A laser flash triggers an ET from the label to the protein and transient absorption spectroscopy is used to monitor time-resolved redox states of the hemes. In this poster we report on the simulation of this experiment in the framework of semi-classical Marcus theory: Classical Molecular Dynamics (MD) simulation together with Fragment-Orbital Density Functional Theory (FODFT) is used to calculate the charge recombination rates from the protein to the ruthenium ligand. Together with previously-calculated inter-heme ET rates we solve for the kinetics of charge transport through the protein using a simple master equation. Our results are compared to and discussed in light of the kinetic traces obtained from pump-probe experiments.

Structure and electronic properties of CdSe magic size clusters Ying Liu Queen Mary University of London

Quantum dots (QDs) are semiconductor nanocrystals that are so small (usually under 10 nm) that they are considered dimensionless and on the length scale fit between small molecules and bulk materials. QDs’ small size significantly affects their structure leading to significant surface reconstruction and stabilising metastable configurations. Their electronic properties are also significantly different from their bulk counterparts as the energy band structure is transformed into discrete energy levels. As a consequence, they offer a number of advantages for potential applications ranging from light harvesting, light emitting diodes to bio-imaging. Colloidal synthesis is one of the most studies and most common approaches to preparation of QDs with the control of size distribution being one of the key issues since narrow size distribution narrow bandwidth of absorption and emission. It has recently been found that growth of QDs in a colloidal solution cannot be adequately described by a classical model of nucleation and growth as along with the standard nanocrystals those of very narrow size distribution (while having certain – magic - size) are produced. We believe that understanding to the structure of these magic size nanoparticles will enable to optimize their synthesis procedures in order to have precisely control over QDs physical properties. Magic size QDs are a class of non-periodic structures, therefore to understand their atomic structure, X-ray Absorption Spectroscopy (XAS), XRD/PDF and Reversed Monte-Carlo simulations (using RMCprofile) will be combined in our project. XRD/PDF will be used as a probe of long-range order, while XAS will provide local atomic and electronic structure information. RMCprofile will be used as a platform for structural data refinement.