Computational Attempts to Approach Magnetism in Metal Organic Frameworks (MOFs)

The approach to magnetic materials with DFT is a challenging task that is often overlooked.
This slides are about our journey to try to model some properties.

Transcript

1. Computational Attempts to Approach Magnetism in Metal Organic Frameworks (MOFs) Dr. Federico Brivio - Federico.Brivio@natur.cuni.cz 14th International conference on materials chemistry Raffaello Sanzio - La scuola di Atene
2. 1 Intro Nachtigall’s group: Chemistry department - Albertov, 8 Resarch: 2D-materials Mesoporous materials Metal- Coordination chemistry Particular focus catalysis |
3. 2 Personal Background Background: Bachelor/Master Materials Science Photovoltaics, Batteries, conventional crystals Elettronic and Thermodynamic properties |
4. 3 Metal-Organic Frameworks MOF = metal nodes interconnected by organic linkers: - Tunability - Incorporation of Mt properties - Porousity They are suitable for different applications: - Catalysis - Gas storage/chromatography - drug delivery |
5. 4 Kagome-MOF We considered two new synthetized MOF with Kagome lattice with Cu and Co as metal center, linker: tetrazole-5-carboxylate ethyl ester Structure hexagonal lattice Very stable (Solvent, pH, T) Magnetic properties |
6. 5 Frustration Magnetic system can have several magnetic dispositions, in the case of triangular lattice, it is difficult to define Antiferromagnetic disposition. |
7. 6 Magnetic Properties - Co AFM behaviour low T! But hysteresis suggests FM fraction. Kagome lattice generates frustration but long range order. Co2+, 3 d7 Small hysteresis low T = FM fraction Good fit with Curie-weiss law χmol = C/(T − Θ) AFM ordering Θ = −94K f = Θ/TN = 14.1 |
8. 7 Magnetic Properties - Cu Complex behaviour: stronger interaction and frustration Cu2+, 3 d9 Small hysteresis low T = FM fraction Good fit with Curie-weiss law χmol = C/(T − Θ) Stronger interaction Θ = −531K f = Θ/TN = 10.2 |
9. 8 Structure DFT needs some PRECISE input. This first step should be trivial, but it is delicate. Goal is MODEL not reality. Issues overlooked: Light elements not detected Different possibilities for oxidation states, H2 O vs OH– Disorder Thermal effects |
10. 9 Structure Optimization We considered several structures: different oxidation states, n. H2 O Chemically reasonable. Symmetry was constrained to P3 |
11. 10 Electronic structure DFT core business is electronic structure, and it is pretty reasonble: 1. DFT simply works! This is a source of abuse of DFT usage 2. Plenty of cases where DFT is not ideal 3. Our case is about magnetic properties! Issue of Magnetic material: electron are over-delocalized(wrong description of the exchange-kinetic energy component) DFT master recipe to improve results: 1. GGA $ * 2. GGA + U $ *** 3. HYBRIDS $$$ *** |
12. 11 Cococcioni’s Method Hubbard scheme has some limitations: Semi-empirical Constant correction The idea: effective Hubbard parameter system dependent! Performed at a linear response regime UI = ∂EGGA ∂(nIf )2 − ∂EGGA 0 ∂(nIf )2 = χ−1 0 − χ−1 (1) Specific for your system! Physical Review B 71 (3), 035105 |
13. 12 Electronic structure Issue of SCF convergence for B systems. 1. GGA-PBESol metallic state 2. GGA+U Did semi-metallic Co-MOF GGA Co-MOF GGA+U |
14. 13 Electronic structure Cu is affected by the same limitations. They d9 configuration results in a different configuration. Even large U parameters can not converge (PhysRevB.84.115108: Ud = 9.79eV , Up = 8.47eV ). Cu-MOF GGA Cu-MOF GGA+U |
15. 14 Cococcioni’s Method Results 1. Hybrid functional did not converged (–> we did not try to change HF-parameter) 2. GGA+UC also did not We as a last resort used HSE06+U (Cococcioni) Co-MOF Cu-MOF |
16. 15 Conclusion Extreme care in the definition of structures When dealing with magnetic system most approach can fail There is a strong temptation to ’tune’ |
17. 16 Acknowledgement Thank you for your attention! I also want to thanks: - Prof. Nachtigall and the whole research group - CUCAM project - EU - European structural and investing funds and the MSMT This Presentation is powered by L A TEX- Beamer Class |
18. 17 Download Slides will be available here: blogfederico.gitlab.io |
19. 18 extradata |
20. 19 extradata |