| 研究課題/領域番号 |
19K15397
|
|---|
| 研究機関 | 東京大学 |
|---|
研究代表者 |
|
|---|
| 研究期間 (年度) |
2019-04-01 – 2023-03-31
|
| キーワード | nanostructure / graphene / ortho-para / hydrogen / DFT |
|---|
| 研究実績の概要 |
This year, the orientation-dependent potential energy surface (PES) of H2 on transition metal (TM) functionalized graphene was constructed by means of DFT calculations. Previous results show that Pd functionalized graphene is the most energetically and structurally stable. The PES of H2 with parallel orientation has a deep potential well at around 1.7 Angstrom H2-Pd bond. The H2 is in the molecular chemisorption state. On the other hand, the PES of H2 with perpendicular orientation has a shallower potential well located at around 2.2 Angstrom above the Pd atom, indicating physisorption. The corresponding binding energies relative to the isolated gas phase H2 and TM-graphene are -813 and -296 meV, for parallel and perpendicular orientations, respectively. The obtained rotational hindering potential is around 1.6 eV which is significantly higher than on metal surfaces.
Furthermore, the neural network potential (NNP) for graphitic carbon nitride (gCN) was constructed. The symmetry functions were used to describe the atomic environment and NNPs were trained using structures, energies and forces obtained from first principles calculations. The training results in root mean square error of around 1.1 meV for the energy and 57 meV for the force. Both pristine and hydrogenated gCN were considered. To test the accuracy of the NNP, the vibrational properties of gCN was investigated. The resulting phonon dispersion exhibits good agreement with DFT calculations. Further, the calculated lattice thermal conductivity at room temperature is also in good agreement experimental results.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
The orientation-dependent potential energy surfaces (PES) have already been obtained and the rotational-vibrational properties calculations can proceed smoothly. Similarly, the nuclear spin conversion calculations have been refined and can already be applied to TM-functionalized nanostructures. On the other hand, functionalization of nanostructures with TM remains a challenge which is hoped to be addressed by employing machine learning methods.
|
| 今後の研究の推進方策 |
In the future research, the rotational and vibrational properties as well as ortho-para conversion of H2 adsorbed on TM-functionalized nanostructures will be calculated. Furthermore, a more general neural network potential (NNP) that can be used to predict the adsorption properties of H2 on TM-functionalized nanostructures will be constructed. This NNP can significantly improve the calculation time needed to obtain the orientation-dependent potential energy surface needed for the determination of rotational and vibrational states of adsorbed H2.
|
| 次年度使用額が生じた理由 |
Due to the coronavirus pandemic, attending conferences in person both locally and abroad was not possible, therefore, the amount allocated for travel expenses were not used. This year I plan to use the funds for in-person conferences, open access journal publications and purchases of computers and memory drives.
|
