Graph neural networks (GNN) offer a potential method of bypassing the Kohn-Sham equations in density functional theory (DFT) calculations by learning both the Hohenberg-Kohn (HK) mapping of electron density to energy, allowing for calculations of much larger atomic systems and time scales and enabling large-scale MD simulations with DFT-level accuracy. In this work, we investigate the feasibility of GNNs to learn the HK map from the external potential approximated as Gaussians to the electron density 𝑛(𝑟), and the mapping from 𝑛(𝑟) to the energy density 𝑒(𝑟) using Pytorch Geometric. We develop a graph representation for densities on radial grid points and determine that a k-nearest neighbor algorithm for determining node connections is an effective approach compared to a distance cutoff model, having an average graph size of 6.31 MB and 32.0 MB for datasets with 𝑘 = 10 and 𝑘 = 50 respectively. Furthermore, we develop two GNNs in Pytorch Geometric, and demonstrate a decrease in training losses for a 𝑛(𝑟) to 𝑒(𝑟) of 8.52 · 10^14 and 3.10 · 10^14 for 𝑘 = 10 and 𝑘 = 20 datasets respectively, suggesting the model could be further trained and optimized to learn the electron density to energy functional.
Using DFT calculations and GAMESS computational software, porphine and its derivatives were analyzed for unique sites to accept the adsorbates As(III), As(V) and P(V) in order to compare resulting adsorption energies and determine if any of these molecules prefer arsenic oxyanions over phosphate. Pure porphine preferred As(III) over P(V) with a resulting adsorption energy of -0.7974 eV. Of the functionalized porphyrins tested, carboxyl porphyrin preferred As(V) over P(V) with a total adsorption energy of -0.7345 eV. Ethyl, methyl, chlorine and amino porphyrin all preferred As(III), with energies of -0.7934, -0.8239, -0.7602, and -0.8508 eV, respectively. Of the metalated porphyrins tested, copper and vanadium porphyrin preferred As(V) over P(V) with adsorption energies of -0.7645 and -2.0915 eV. Chromium, iron and magnesium porphyrin all preferred As(III) over P(V) with energies of -0.5993, -1.4539, and - 1.0790 eV, respectively.
Chapter 2 describes the study on the Pt and Pt-Me (Me: Co, Ni) alloy nanoparticles supported on the pyrolyzed zeolitic imidazolate framework (ZIF) towards ORR. The Co-ZIF and NiCo-ZIF were synthesized by the solvothermal method and then mixed with Pt precursor. After pyrolysis and acid leaching, the PtCo/NC and PtNiCo/NC were evaluated in proton exchange membrane fuel cells (PEMFC). The peak power density exhibited > 10% and 15% for PtCo/NC and PtNiCo/NC, respectively, compared to that with commercial Pt/C catalyst under identical test conditions.
Chapter 3 is the investigation of the oxygen vacancy (OV) effect in a-MnO2 as a cathode catalyst for alkaline membrane fuel cells (AMFC). The a-MnO2 nanorods were synthesized by hydrothermal method and heated at 300, 400 and 500 ℃ in the air to introduce the OV. The 400 ℃ treated material showed the best ORR performance among all other samples due to more OV in pure a-MnO2 phase. The optimized AMFC electrode showed ~ 45 mW.cm-2, which was slightly lower than that with commercial Pt/C (~60 mW.cm-2).
Chapter 4 is the density functional theory (DFT) study of the protonation effect and active sites towards ORR on a-MnO2 (211) plane. The theoretically optimized oxygen adsorption and hydroxyl ion desorption energies were ~ 1.55-1.95 eV and ~ 0.98-1.45 eV, respectively, by Nørskov et al.’s calculations. All the configurations showed oxygen adsorption and hydroxyl ion desorption energies were ranging from 0.27 to 1.76 eV and 1.59 to 15.0 eV, respectively. The site which was close to two Mn ions showed the best oxygen adsorption and hydroxyl ion desorption energies improvement with the surface protonation.
Based on the results given in Chapters 1-4, the major findings are summarized in Chapter 5.
A proof of principle demonstration of a resonator that can switch from a high-Q “on state” to a low-Q “off state” at reduced temperatures is demonstrated in (Al1-xFex)2O3 and La(Al1-xFex)O3. The Fe3+ ions are in a high spin state (S=5/2) and undergo electron paramagnetic resonance absorption transitions that increase the microwave loss of the system. Transitions occur between mJ states with a corresponding change in the angular momentum, J, by ±ħ (i.e., ΔmJ=±1) at small magnetic fields. The paramagnetic ions also have an influence on the dielectric and magnetic properties, which I explore in these systems along with another low loss complex perovskite material, Ca[(Al1-xFex)1/2Nb1/2]O3. I describe what constitutes an optimal microwave loss switchable material induced from EPR transitions and the mechanisms associated with the key properties.
As a first step to modeling the properties of high-performance microwave host lattices and ultimately their performance at microwave frequencies, a first-principles approach is used to determine the structural phase stability of various complex perovskites with a range of tolerance factors at 0 K and finite temperatures. By understanding the correct structural phases of these complex perovskites, the temperature coefficient of resonant frequency can be better predicted.
A strong understanding of these parameters is expected to open the possibility to produce new types of high-performance switchable filters, time domain MIMO’s, multiplexers, and demultiplexers.