Benchmarking of density functionals for a soft but accurate prediction and assignment of 1H and 13C NMR chemical shifts in organic and biological molecules
详细信息 查看全文
- 作者：Enrico Benassi
- 刊名：Journal of Computational Chemistry
- 出版年：2017
- 出版时间：January 15, 2017
- 年：2017
- 卷：38
- 期：2
- 页码：87-92
- 全文大小：343K
- ISSN：1096-987X
文摘
A number of programs and tools that simulate 1H and 13C nuclear magnetic resonance (NMR) chemical shifts using empirical approaches are available. These tools are user-friendly, but they provide a very rough (and sometimes misleading) estimation of the NMR properties, especially for complex systems. Rigorous and reliable ways to predict and interpret NMR properties of simple and complex systems are available in many popular computational program packages. Nevertheless, experimentalists keep relying on these “unreliable” tools in their daily work because, to have a sufficiently high accuracy, these rigorous quantum mechanical methods need high levels of theory. An alternative, efficient, semi-empirical approach has been proposed by Bally, Rablen, Tantillo, and coworkers. This idea consists of creating linear calibrations models, on the basis of the application of different combinations of functionals and basis sets. Following this approach, the predictive capability of a wider range of popular functionals was systematically investigated and tested. The NMR chemical shifts were computed in solvated phase at density functional theory level, using 30 different functionals coupled with three different triple–ζ basis sets.