Relativistic Environmental Effects in ²⁹Si NMR Chemical Shifts of Halosilanes: Light Nucleus, Heavy Environment

Abstract

Relativistic calculations of (29)Si NMR shielding constants (chemical shifts) in the series of halosilanes SiX(n)H(4-n) (X = F, Cl, Br and I) are performed within a full four-component relativistic Dirac's scheme using relativistic Dyall's basis sets. Three different theoretical levels are tested in the computation of (29)Si NMR chemical shifts in comparison with experiment: namely, four-component relativistic GIAO-DFT, four-component relativistic GIAO-RPA, and a hybrid scheme of a nonrelativistic GIAO-MP2 with taking into account relativistic corrections using the four-component relativistic GIAO-RPA. The DFT results give larger relativistic effects as compared to the RPA data which might be rationalized in terms of the manifestation of correlation effects taken into account at the DFT level and not accounted for at the uncorrelated RPA level. Taking into account solvent effects slightly improves agreement with experiment, however, being not a matter of principle. Generally, relativistic pure nonempirical wave function methods perform much better as compared to relativistic DFT methods when benchmarked to experiment.

Four‐Component Relativistic DFT Calculations of 13C Chemical Shifts of Halogenated Natural Substances

“…[14] In heavier atoms (e.g., 183 Wo r 195 Pt), [15] such effectsm ay amount to thousands of ppm but are partly compensated for,i nt he chemical shift calculation, because the reference compound is affected by roughly the same amount. [16] In contrast, no such compensation occurs for 13 Ca toms, because the reference compound (tetramethylsilane) is nota ffected. Thus, even if s SO is much smaller for 13 Ca toms (see below), it needs to be calculatedi n order to attain aq uantitative prediction.…”

“…[16] In contrast, no such compensation occurs for 13 Ca toms, because the reference compound (tetramethylsilane) is nota ffected. Thus, even if s SO is much smaller for 13 Ca toms (see below), it needs to be calculatedi n order to attain aq uantitative prediction. Other instances, notably the extremelyl arge effectse ncountered in heavy-atomh ydrides, have been successfully modeled, [12,17] but the involved structures are not present in organic molecules.…”

“…Quantum chemical calculations, especially those based on density functional theory (DFT), have provedt ob eq uite effective at predicting the NMR properties of aw ide variety of complexes and molecules. Thus,t he "DFT-NMR"a pproach has become as taple technique for the elucidation of the structure [1][2][3] and stereochemistry [1,4] of organic molecules,w hich is generally carried out through the analysis of 1 Ha nd 13 CNMR chemicalshifts and couplings. [5] Am ajor strength of quantum mechanicalm ethods, as opposedt oe mpirical ones, is that they dispense with such notions as atom type,h ybridization,o rs train and (at least in principle) deal with all moleculars tructures on an equal footing.…”

“…[2,3] Moreover,acareful discussiono fthe significance of the statistical parameters and the most appropriate ones in order to assignag iven experimental 1 Ho r 13 CNMR spectrum to one or more putativestructures has been presented. [6][7][8] Nevertheless,ap erceived weakness [2,4] concerns the predictions for ar ather common functionality,t hat is, the 13 CNMR shifts of carbona toms bonded to heavy halogen atoms:B r and I, and to al esser extent Cl. Typical DFT approaches do not perform well in these cases and lead to shifts that are too positive (deshielded) with respect to the experimental results.…”

“…Typical DFT approaches do not perform well in these cases and lead to shifts that are too positive (deshielded) with respect to the experimental results. Indeed, trendsi nt he 13 CNMR shifts of halomethanes have long been known to be anomalous. [9][10][11][12] The shifts of fluoroand chloromethanes follow an intuitive trend of the d value increasingw ith the number of Xa toms, whichf its with electronegativity concepts, across a1 00 ppm range;i nc ontrast, bromomethanes go through am aximum and span no more than 30 ppm, and iodomethanes show an opposite trend, with as trong shielding effect upon an increase in the number of I atomsa cross approximately 270 ppm (see Figure S1 in the Supporting Information).…”