Section of Chemistry

Schedule for The Löwdin Lectures 2018

Thursday, November 15th

In room 4101 at the Ångström Laboratory.

13:00 - Introduction

Roland Lindh and Morgane Vacher
Department of Chemistry-Ångström, Uppsala University

13:15 - Proton-Coupled Electron Transfer in Catalysis and Energy Conversion

The 2017 Löwdin Lecturer Prof. Sharon Hammes-Schiffer

Department of Chemistry, Yale University, New Haven, Connecticut, United States

Proton-coupled electron transfer (PCET) reactions play a vital role in a wide range of chemical and biological processes. This lecture will focus on recent advances in the theory of PCET and applications to catalysis and energy conversion. The quantum mechanical effects of the active electrons and transferring proton, as well as the motions of the proton donor-acceptor mode and solvent or protein environment, are included in a general theoretical formulation. This formulation enables the calculation of rate constants and kinetic isotope effects for comparison to experiment. Applications to PCET in enzymes, molecular electrocatalysts for hydrogen production and water splitting, and photoreduced zinc-oxide nanocrystals will be discussed. These studies have identified the thermodynamically and kinetically favorable mechanisms, as well as the role of proton relays, and are guiding the theoretical design of more effective catalysts. In addition, recent developments of theoretical approaches for simulating the ultrafast dynamics of photoinduced PCET, along with applications to solvated molecular systems and photoreceptor proteins, will be discussed.

14:00 - Have you ever pulled on a molecule?

The 2018 Löwdin Lecturer Prof. Markus Reiher
Laboratory of Physical Chemistry, ETH Zürich, Zürich, Switzerland

Technical advances in combination with algorithmic developments have always been a driving force for establishing new approaches in theoretical and computational chemistry. Despite the amazing advances in computer hardware that brought us gadgets such as the iPhone, we still carry out molecular simulations with technology of the 1970s (keyboard) and 1980s (mouse). In this lecture, I will demonstrate how one can harness modern technology in combination with new theoretical concepts to carry out quantum chemical calculations in real time. This brings about new challenges for (i) starting and manipulating the input data and for (ii) perceiving the output data as these steps now become time dominating. I will demonstrate how new computer hardware such as force-feedback haptic devices can be used to allow for molecular structure manipulation in three (rather than two) dimensions and to literally feel the quantum mechanical forces acting on the atoms in a molecule as a descriptor for chemical reactivity.

Friday, November 16th

In room 4101 at the Ångström Laboratory.

09:00 - Opening & Introduction

Roland Lindh and Morgane Vacher
Department of Chemistry-Ångström, Uppsala University, Sweden

09:15 - Exhaustive Exploration of Complex Reaction Networks

Markus Reiher
Laboratory of Physical Chemistry, ETH Zürich, Zürich, Switzerland

A prominent focus of molecular science has been the understanding and design of functional molecules and materials. This brings about new challenges for theoretical chemistry. We are faced with the necessity to obtain theoretical results of predictable accuracy for molecules of increasing size and number. Moreover, the molecular composition, which is required as input for a quantum chemical calculation, might not be known, but the target of a design attempt. Then, the relevant chemical processes are not necessarily known, but need to be explored and identified. Whereas parts of these challenges have already been addressed by the development of specific methods (such as linear scaling or high-throughput screening), the fact that an enormous multitude of structures needs to be considered calls for integrated approaches. This holds particularly true for predictions on complex chemical processes that encode function (e.g., through reaction networks). In my talk, I will review our recent work on these challenges.

10:00 - Break

10:30 - Accelerating inorganic discovery with machine learning

Heather Kulik
Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States

11:00 - Quantum machine learning

Anatole von Lilienfeld
Department of Chemistry, University of Basel, Basel, Switzerland

11:30 - Integrating first principles modeling and experimental characterization via machine learning

Maria K. Y. Chan
Argonne National Laboratory, Argonne, Illinois, United States

12:00 - Lunch break

13:30 - Neural networks learning quantum chemistry

Olexandr Isayev
University of North Carolina, Chapel Hill, North Carolina, United States

14:00 - Machine learning with deep feature graphs leads to interpretable models that predict reaction performance

Lucy Colwell
Department of Chemistry, University of Cambridge, Cambridge, United Kingdom

14:30 - Machine learning algorithms for designing metamaterials

Koji Tsuda
Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Japan

15:00 - Break

15:30 - Machine learning for quantum chemistry

Matthias Rupp
Fritz Haber Institute of the Max Planck Society, Berlin, Germany

16:00 - Multicomponent quantum chemistry: Integrating electronic and nuclear quantum effects

Sharon Hammes-Schiffer

Department of Chemistry, Yale University, New Haven, Connecticut, United States

Nuclear quantum effects such as zero point energy, nuclear delocalization, and tunneling play an important role in a wide range of chemical processes. Typically quantum chemistry calculations invoke the Born-Oppenheimer approximation and include nuclear quantum effects as corrections following geometry optimizations. The nuclear-electronic orbital (NEO) approach treats select nuclei, typically protons, quantum mechanically on the same level as the electrons with multicomponent density functional theory (DFT) or wavefunction methods. Recently electron-proton correlation functionals have been developed to address the significant challenge within NEO-DFT of producing accurate proton densities and energies. Moreover, delta self-consistent-field methods and time-dependent DFT methods within the NEO framework have been developed for the calculation of electronic, proton vibrational, and electron-proton vibronic excitations. These combined NEO methods enable the inclusion of nuclear quantum effects and non-Born-Oppenheimer effects in calculations of proton affinities, pKa’s, optimized geometries, minimum energy paths, reaction dynamics, excitation energies, tunneling splittings, and vibronic couplings for a wide range of chemical applications.

16:45 - Summary & Closing remarks

Roland Lindh and Morgane Vacher