Scientific Highlights

Quantum simulations of light-matter interactions are important tools to predict optimal materials for photovoltaics and photelectrochemical cells. However, these simulations are demanding from a computational standpoint, especially when applied to complex, heterogeneous systems. Recently MICCOM and AMEWS researchers developed a novel formalism to accelerate these simulations using machine-learning algorithms, Their work was recently published in Chemical Science and featured in and UChicago PME News.

Water is ubiquitous on earth, and the knowledge how matter interacts with water is paramount for a wide range of applications, from climate science to energy engineering. Recently, MICCoM scientists studied two very different interactions of water with matter: Ding Pang and Giulia Galli determined the properties of dissolved carbon in water under high pressure in the earth's mantle from first-principles simulations. Understanding what concentration of carbon can be found under these extreme conditions has important implications on how carbon moves from the atmosphere to the earth's mantle. Their study, recently featured in UChicago PME news, found that the carbon concentration deep in the mantle was underestimated by previous models.
In another study of water-matter interactions, MICCoM scientists partnered with scientists at the Argonne National Laboratory to validate first-principles calculations of water/metal-oxide interfaces. These play a crucial role in a wide range of applications, such as photochemical water splitting and electrochemical carbon-dioxide reduction. In a combined theoretical and experimental study, the scientists established a groundbreaking protocol to validate the first-principles simulations. This research provides a pathway for more accurate models of water/metal-oxide interfaces in the future.

Quantum computing promises to greatly advance our capability to calculate properties of molecules and solids beyond the limits of current supercomputers. However, quantum computing is still in it's infancy and the necessary hardware and algorithms are still being developed. MICCoM scientists have recently developed a new quantum-mechanical framework to use quantum computing to determine the properties of spin defects. Combining calculations on classical and quantum hardware, they demonstrated that even the currently available imperfect quantum computers yield accurate and effective results for these complex systems. The new framework, developed by He Ma, Nan Sheng, Marco Govoni, and Giulia Galli, was recently published in a series of papers in npj Computational Materials, Journal of Physical Chemistry and Chemical Physics, and Journal of Chemical Theory and Computation, and was featured in a Argonne National Laboratory press release.

MICCoM has released two python packages for analysis of first principles simulations: PyCDFT (code repository), and PyZFS (code repository) . PyCDFT computes diabatic states using constrained density functional theory (CDFT). It provides an object-oriented, customizable implementation of CDFT, and allows for both single-point self-consistent-field calculations and geometry optimizations. PyZFS, on the other hand, is used for the calculation of the zero-field splitting tensor, D, of molecules and solids, based on wavefunctions obtained from density functional theory calculations. PyZFS features a modular design and utilizes abstract classes for extensibility. It has already been used to predit ZFS tensors for spin defects in semiconductors as well as spin-phonon interactions in solids. Both PyCDFT and PyZFS can use wavefunctions generated by various plane-wave DFT codes as input, including Qbox.

As machine learning (ML) techniques become increasingly ascendant in computational materials science, MICCoM is developing expertise and successfully launched several projects in this field. MICCoM scientists from the group of Prof. Juan de Pablo published a Science Advances paper on using supervised machine learning to obtain electronic structure at coarse-grained resolution. The work was featured by Argonne and UChicago News. It uses artificial neural networks and coarse-graining to compute electronic structure across various molecular conformations. The method was later applied to study optoelectronic properties of conjugated polymers, published in Macromolecules. MICCoM researchers have also contributed perspective articles towards the application of machine learning to collective variables in biomolecular simulation, published in Molecular Physics, as well as multiscale soft materials design, published in Current Opinion in Chemical Engineering. In addition, MICCoM PI Andrew Ferguson was also recently featured in a UChicago news story on "How AI could change science," where he described how using physical laws as inputs to artificial intelligence algorithms could accelerate discovery of new materials. Finally, ML techniques are being applied to electronic structure problems as well, for example the calculation of absorption spectra.

The investigation of salts in water at extreme conditions is crucial to understanding the properties of aqueous fluids in the Earth. In this paper, published in Nature Communications, MICCoM scientists report first principles and classical molecular dynamics simulations of NaCl at temperatures and pressures relevant to the Earth’s upper mantle. The work was UChicago PME news. The study employed the codes Qbox (first principles molecular dynamics) and SSAGES (free energy calculation). Furthermore, it showcased the strategic coupling of the two codes to understand the free energies governing the dissociation process. The coupling of Qbox with SSAGES enabled the identification of two critical metastable states of the salt, which are entropocially and enthalpically favored. They found the minimum free energy path between the metastable points to become smoother at high pressure, as the relative stability of the two configurations is affected by water self-dissociation, which can only be described properly by first-principles simulations.

The design and optimization of photoelectrodes for artificial photosynthesis are critical to achieving sustainable solar to fuel conversion technologies. Most computational screening of materials so far has focused on bulk properties, leaving the critical water/photoanode interface poorly understood. In this paper, published in Nature Materials, MICCoM scientists employed first-principles codes developed in the center to investigate a promising photoanode for water oxidation, tungsten trioxide (WO₃). They identified three major factors determining the chemical reactivity of the material interfaced with water: the presence of surface defects, the dynamics of excess charge at the surface, and finite temperature fluctuations of the surface electronic orbitals. Being generalizable properties, these key computational insights will aid the interfacial molecular engineering of oxide photoabsorbers for water oxidation.

This study employed the MICCoM codes Qbox (first principles molecular dynamics) and WEST (many body perturbation theory). Furthermore, it pioneered the coupling of the two codes in order to obtain quantitative comparisons with photoemission and electrochemical data. The methodological advancements allowed for robust comparison and interpretation of experimental results, furthering the center's goals towards validation of codes.

Semiconducting nanomaterials are critical to emerging optoelectronic and photonic technologies. In such systems, understanding surface chemistry of the interface between nano building blocks and their ligands is an outstanding challenge for rational device engineering. In this study, published in Nature Nanotechnology, a joint computational/experimental team of MICCoM scientists presented an approach to characterize such buried interfaces in solids of prototypical InAs nanoparticles capped with Sn₂S₆^4– ligands. They discovered that that inorganic ligands dissociate on InAs to form a surface passivation layer. A nanocomposite with unique electronic and transport properties is formed, exhibiting type II heterojunctions favourable for exciton dissociation. With these insights, they establish crucial design principles towards attaining desirable electronic and transport properties, while also explaining the origin of unusual measured negative photoconductivity of the nanocrystalline solids.

This study employed the MICCoM code Qbox (first principles molecular dynamics). The close collaboration with experiment furthered the center's efforts towards full validation of theoretical methods and codes. Specifically, it introduced a cogent validation strategy framework, which is generalizable to other systems of semiconducting nanoparticles.

Polyelectrolytes are "smart" materials that find a variety of applications including drug delivery, gene therapy, layer-by-layer films, and fabrication of ion filtration membranes. They fall into two distinct categories - strong or weak - based on their complete or partial ionization in solution, forming polyelectrolyte complexes. However, the thermodynamic origins of polyelectrolyte complex formation have been poorly understood. In this study, published in Journal of the American Chemical Society, MICCoM scientists employed coarse-grained molecular dynamics simulations to study the thermodynamics at play in the complexation processes. They used an expanded-ensemble method to explicitly separate entropic and enthalpies contributions to free energy. In so doing, they discovered that strong polyelectrolytes are primarily driven by entropic effects, consistent with experimental studies. The effect was attributed to gain in entropy from the release of counterions upon complex formation. On the other hand, they found that complex formation for weak polyelectrolytes, where counterions are not as strongly bound, is more influenced by enthalpic effects. Their findings will guide the molecular engineering of tunable, weak polyelectrolytes for various applications in molecular capture and release and layer-by-layer assembly of films.

This study employed the MICCoM code SSAGES (free energy calculation). Free energy profiles along the center-of-mass distance between polymer chains were then obtained through the adaptive bias force algorithm.

The burgeoning age of high performance computing and big data is revolutionizing both the methods and pace of scientific research. However, the development of digital data infrastructures has not kept up with this rapid growth, especially in the field of computational research for energy application. The community sorely needs open-source strategies for documenting codes, curating workflows, and publishing computational data. In view of these challenges, MICCoM has created the open source software Qresp: “Curation and Exploration of Reproducible Scientific Papers." In this study, published in Scientific Data, MICCoM scientists introduce Qresp in detail. The tool allows users to curate the codes/data of a completed publication, while storing all data locally. Towards this end, Qresp has an intuitive graphical user interface (GUI) designed to simply prompt user input information about a publication in order to generate metadata in JSON format. On the other hand, once a paper and its codes/data are curated, Qresp's GUI may also be used as a searchable explorer of metadata or a means to dive into the entire workflow and procedures of an individual paper, from code inputs to finished figures.

The Qresp project is one of the cornerstones of the MICCoM data effort. Many publications from the center have already been curated with Qresp and are being used for both validation efforts and pedagogy.