| 1. |
- Dobryden, Illia, et al.
(författare)
-
Local Wear of Catechol-Containing Diblock Copolymer Layers : Wear Volume, Stick-Slip, and Nanomechanical Changes
- 2021
-
Ingår i: The Journal of Physical Chemistry C. - : American Chemical Society. - 1932-7447 .- 1932-7455. ; 125:38, s. 21277-21292
-
Tidskriftsartikel (refereegranskat)abstract
- Polymers containing catechol groups have gained a large interest, as they mimic an essential feature of mussel adhesive proteins that allow strong binding to a large variety of surfaces under water. This feature has made this class of polymers interesting for surface modification purposes, as layer functionalities can be introduced by a simple adsorption process, where the catechol groups should provide a strong anchoring to the surface. In this work, we utilize an AFM-based method to evaluate the wear resistance of such polymer layers in water and compare it with that offered by electrostatically driven adsorption. We pay particular attention to two block copolymer systems where the anchoring group in one case is an uncharged catechol-containing block and in the other case a positively charged and catechol-containing block. The wear resistance is evaluated in terms of wear volume, and here, we compare with data for similar copolymers with statistical distribution of the catechol groups. Monitoring of nanomechanical properties provides an alternative way of illustrating the effect of wear, and we use modeling to show that the stiffness, as probed by an AFM tip, of the soft layer residing on a hard substrate increases as the thickness of the layer decreases. The stick-slip characteristics are also evaluated. © 2021 The Authors.
|
|
| 2. |
- Feltrin, Ana Carolina, et al.
(författare)
-
Transformation of metastable dual-phase (Ti0.25V0.25Zr0.25Hf0.25)B2 to stable high-entropy single-phase boride by thermal annealing
- 2021
-
Ingår i: Applied Physics Letters. - : AIP Publishing LLC. - 0003-6951 .- 1077-3118. ; 119:16
-
Tidskriftsartikel (refereegranskat)abstract
- Transition metal borides have a unique combination of high melting point and high chemical stability and are suitable for high temperature applications (>2000 °C). A metastable dual-phase boride (Ti0.25V0.25Zr0.25Hf0.25)B2 with distinct two hexagonal phases and with an intermediate entropy formation ability of 87.9 (eV/atom)−1 as calculated via the density functional theory (DFT) was consolidated by pulsed current sintering. Thermal annealing of the sintered dual-phase boride at 1500 °C promoted the diffusion of metallic elements between the two boride phases leading to chemical homogenization and resulted in the stabilization of a single-phase high-entropy boride. Scanning electron microscopy, in situ high temperature x-ray diffraction, and simultaneous thermal analysis of the as-sintered and annealed high-entropy borides showed the homogenization of a dual-phase to a single-phase. The experimentally obtained single-phase structure was verified by DFT calculations using special quasirandom structures, which were further used for theoretical investigations of lattice distortions and mechanical properties. Experimentally measured mechanical properties of the single-phase boride showed improved mechanical properties with a hardness of 33.2 ± 2.1 GPa, an elastic modulus of 466.0 ± 5.9 GPa, and a fracture toughness of 4.1 ± 0.6 MPa m1/2.
|
|
| 3. |
- Hedman, Daniel, 1989-, et al.
(författare)
-
Impact of training and validation data on the performance of neural network potentials : A case study on carbon using the CA-9 dataset
- 2021
-
Ingår i: Carbon Trends. - : Elsevier. - 2667-0569. ; 3
-
Tidskriftsartikel (refereegranskat)abstract
- The use of machine learning to accelerate computer simulations is on the rise. In atomistic simulations, the use of machine learning interatomic potentials (ML-IAPs) can significantly reduce computational costs while maintaining accuracy close to that of ab initio methods. To achieve this, ML-IAPs are trained on large datasets of images, which are atomistic configurations labeled with data from ab initio calculations. Focusing on carbon, we use deep learning to train neural network potentials (NNPs), a form of ML-IAP, based on the state-of-the-art end-to-end NNP architecture SchNet and investigate how the choice of training and validation data affects the performance of the NNPs. Training is performed on the CA-9 dataset, a 9-carbon allotrope dataset constructed using data obtained via ab initio molecular dynamics (AIMD). Our results show that image generation with AIMD causes a high degree of similarity between the generated images, which has a detrimental effect on the performance of the NNPs. But by carefully choosing which images from the dataset are included in the training and validation data, this effect can be mitigated. We conclude by benchmarking our trained NNPs in applications such as relaxation and phonon calculation, where we can reproduce ab initio results with high accuracy.
|
|
