journal article·2019-02-28Open access

Overcoming Distrust in Solid State Simulations: Adding Error Bars to Computational Data

Francesca Peccati, Rubén Laplaza, Julia Contreras-García

Abstract

Simulations are progressively becoming an everyday tool for the understanding of materials. DFT is without any doubt the working horse in the vast majority of cases in which quantum mechanical effects need to be included for a correct description of the chemical problem. However, unlike other quantum-mechanical methods, DFT approaches cannot be improved in a systematic manner. This leads to distrust and the feeling that calculation parameters can be tuned with the explicit aim of matching the experimental results, even at the expense of the quality of the simulation. More specifically, it would appear that changing the functional can provide any desired quantity to match the experiment. Hereby, we focus on the nature of systematic simulation errors as compared to reproducibility–experimental ones. We show that DFT errors need to be estimated from bracketing physical descriptions. Highlighting the role of the main error affecting DFT calculations, the delocalization error, we show that HF and LDA provide bracketing limits and we develop a simple and robust procedure to quickly provide error bars that help understanding the inherent error of a given calculation. Moreover, for challenging experimental quantities, the experimental and computational error bars become of the same order of magnitude. Overall, this approach should help overcoming distrust in solid state simulations through a detour to the overwhelming question: “which functional to choose?”.
Source: openalex_inverted_index · Quality: baseline

Research context

OpenAlex topics: Machine Learning in Materials Science · X-ray Diffraction in Crystallography · Electronic and Structural Properties of Oxides

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

Laboratoire de Chimie Théorique, LCT, Sorbonne Université, CNRS, F 75005 Paris, France

Institut des Sciences du Calcul et des Données, ISCD, Sorbonne Université, F-75005 Paris, France

Departamento de Quı́mica Fı́sica and Instituto de Biocomputación y Fı́sica de Sistemas Complejos (BIFI), Universidad de Zaragoza, 50009 Zaragoza, Spain

Funding

Agence Nationale de la Recherche

References (22)

Show source-provided references
1. 10.1039/C6SC03759C10.1039/C6SC03759C ↗
2. 10.1119/1.218633310.1119/1.2186333 ↗
3. 10.1038/s41586-018-0337-210.1038/s41586-018-0337-2 ↗
4. 10.1107/S010876810000269X10.1107/S010876810000269X ↗
6. 10.1098/rspa.1913.004010.1098/rspa.1913.0040 ↗
7. 10.1051/m2an/201703510.1051/m2an/2017035 ↗
8. 10.1126/science.115872210.1126/science.1158722 ↗
9. 10.1021/ar500171t10.1021/ar500171t ↗
10. 10.1063/1.177497610.1063/1.1774976 ↗
11. 10.1021/ct100653210.1021/ct1006532 ↗
12. 10.1063/1.46788210.1063/1.467882 ↗
13. 10.1002/jcc.2519510.1002/jcc.25195 ↗
14. 10.1107/S010827010300968510.1107/S0108270103009685 ↗
15. Hönerlage B. · 2010 · 44
16. 10.1021/acs.jctc.6b0030410.1021/acs.jctc.6b00304 ↗
17. 10.1021/ct200561610.1021/ct2005616 ↗
18. 10.1103/PhysRevB.74.24511510.1103/PhysRevB.74.245115 ↗
19. 10.1016/j.jssc.2011.03.02710.1016/j.jssc.2011.03.027 ↗
20. 10.1107/S205225251402548210.1107/S2052252514025482 ↗
21. 10.1016/j.ssc.2006.05.04310.1016/j.ssc.2006.05.043 ↗
22. 10.1021/acs.jctc.8b0047710.1021/acs.jctc.8b00477 ↗