Probing Halogen Bonds by Scalar Couplings

As halogen bonding is a weak, transient interaction, its description in solution is challenging. We demonstrate that scalar coupling constants (J) are modulated by halogen bonding. The binding-induced magnitude change of one-bond couplings, even up to five bonds from the interaction site, correlates to the interaction strength. We demonstrate this using the NMR data of 42 halogen-bonded complexes in dichloromethane solution and by quantum chemical calculations. Our observation puts scalar couplings into the toolbox of methods for characterization of halogen bond complexes in solution and paves the way for their applicability for other types of weak σ-hole interactions.


The iodine basicity scale
The iodine basicity (e.g. pKbI2 and (I-I) scale) gives Lewis base-type dependent correlations to observables that describe the non-covalent binding of Lewis bases to diiodine. NMR-spectroscopy detects a timeaveraged signal of the species involved into the binding equilibrium. The pKbI2 scale describes the same equilibrium process. The iodine basicity observed for a halogen bond acceptor is interconvertible via a linear correlation justifying the pKbI2 2 dependency of the correlations described in this work. 1 There are a variety of basicity scales available, with the book "Lewis Basicity and Affinity Scales" by C. Laurence and J-F. Gal 2 providing a helpful overview of the scope of each.

Experimental Setup and Details
The Spectrometer settings relevant for the measurement accuracy of 1 JF,Cs 19 F NMR spectra have been recorded for CD2Cl2 solutions at 25.0 °C, acquiring 128 scans with 1 sec relaxation delay at 376.25 MHz using specifications given in Table S1. Table S2.

S3
The filename format of the NMR rawdata, which is provided at the open access respository Zenodo with DOI: 10.5281/zenodo. 4698893: [

The method used for error estimation in this study
The reported error of the 1 JF,C (see 1.2.1) is the averaged ( ̅ ) value, based on ̅ = (∑ ) "AVERAGE", and its standard deviation  as = This method consistently gives the largest within the dataset/sample size, compared to alternatives tested. Alternative methods for describing variations of a very small datasets do treat differently, e.g. giving the ± range of the averaged value ( ( ) ( ) ) instead of accessing variance.

Curve fitting algorithms
To fit functions to the dataset, a given function has been minimized by reducing its overall R 2 value without applying weighing methods, using the software Origin 2018.
Comparison of the observed errors for the  1 JF,C estimations for 1iodopentafluorobenzene and 1-iodoheptadecafluorooctane The pyridine data sets with 1-iodopentafluorobenzene and 1-iodoheptadecafluorooctane have been compared to estimate which of the two halogen bond donors gives more reliable  1 J values. Despite the closer proximity of the observed  1 JF,C measured for 1-iodoheptadecafluorooctane higher relative errors have been observed. We therefore decided to conduce the further study with 1-iodopentafluorobenzene.

S11
The influence of moisture To estimate the influence of moisture on the detected 1 JF,C, we recorded the 19 F NMR spectrum of 1iodoheptadecafluorooctane in dry CD2Cl2 (Eurisotop® lot: T1071) and in mixtures of non-dried CD2Cl2 (Eurisotop® lot: T1071) and CH2Cl2 (VWR, Lot 20G034019). The influence of water residues was insignificant (Table S1), with the variations between the samples being within the error limits. Averaging the coupling constants read on several individual peaks of a multiplet for a single coupling constant gives a lower error than the digital resolution of the spectra would permit.   Figure S1. Summary of the Lewis bases evaluated for halogen bonding. S13 1-Iodoperfluorobenzene: a typical example Enlarged multiplets of the 19 F NMR spectrum of 1-iodopentafluorobenzene without and in the presence of 4methoxypyridine. The multiplet patterns don't change in the presence of the base. The 19 F NMR spectra corresponds to the data given in Figure S2 and Figure S3.   The  1 JF,C and  values for 1-iodopentafluorobenzene induced by Lewis bases Figure S4: The  1 JF,C and  values for 1-iodopentafluorobenzene induced by the presence of a base, based on Table S11.  Table S11.

S18
Subset analysis of  1 Jortho-F,C values for 1-iodopentafluorobenzene induced by Lewis bases Figure S6: The  1 JF,C and  values for 1-iodopentafluorobenzene induced by Lewis bases, based on Table S11. Top-left:  1 Jortho-F,13C vs pKBI2 grouped according to the type of base employed. The data points have been fitted to ax 2 +b. top-right: The change of 1 JF,C in the ortho-position of iodopentafluorobenzene,  1 Jortho-F,C, as a function of the Lewis basicity, pKBI2 2 , upon binding to various Lewis bases. Errors are given as standard deviations; pKBI2 = 0 refers to a K = 1. The data corresponding to the pyridines is shown in red (R 2 = 0.98), to amines in blue (R 2 = 0.60), to N-oxides in green (R 2 = 0.94), whereas to all data in black (R 2 = 0.71).

1-Iodoheptadecafluorooctane: A typical example
Expansions of the 19 F NMR spectrum of 1-iodoheptadecafluorooctane without and in the presence of 4methoxypyridine. The multiplet patterns don't change upon addition of a Lewis base. The 19 F NMR spectra correspond to the data given in Figure S7 and Figure S8.

1-Iodoperfluorooctane: Base induced changes in the 19 F NMR
Changes in selected chemical shifts and coupling constants of 1-iodoperfluorooctane are given in Table S14.
The changes in coupling constant observed for the terminal CF3 group are within the error of the experiment. Significant chemical shift changes of the terminal CF3 group has been rationalized by Ciancaleoni et. al. as secondary interaction with the aromatic π-system leading to an orientation of the chain above the aromatic cycle and therefore remote Fluorine resonances experience a shielding effect. 3   Table S14. The average changes in the 19 F-13 C-1 J coupling constants ( ) as well as their standard deviations (STDEV) of 1-Iodoperfluorooctane (c = 0.200 mol/L) by the presence of selected halogen bond acceptors (c = 0.500 mol/L). Samples have been prepared at least twice.

S22
The  1 JF,C and  values for 1-Iodoheptadecafluorooctane induced by Lewis bases

Evaluating the impact of Non-Directional Effects
Addition of Lewis bases does not affect the 1 JF,C of the terminal -CF3 group of IC8F17 whereas it influences in 1 JF,C 1 JF,C and 1 JF,C in a Lewis basicity dependent manner ( Figure 4, main text). Significant change, 2.15 Hz, was observed for the 1 JF,C of the -CF3 group upon addition of 2.5 eq. pentane to the solution of the halogen bond donor, whereas the 1 JF,C of the and -CF2 groups were not affected by n-pentane significantly. This observation indicates that the remote -CF3 group serves as an internal reference to detect non-directional solvent effects introduced by changes in the bulk properties of the solution e.g. polarity. Thus, halogen bonding is observable close to the halogen bond donor site at the and -CF2 groups but not at the remote -CF3 group whereas bulk solvent effects are detectable on the -CF3 but not on the and -CF2 groups.
Iodopentafluorobenzene lacks a comparable reference position. However, upon addition of 2.5 eq. n-pentane only minor,  1 JF,C < 0.07 Hz, were observed at the ortho and para-1 JF,C whereas that at the the meta position has been slightly more sensitive ( 1 JF,C = 0.13 Hz). Halogen bonding shows the opposite effect (larger change at the ortho and para-1 JF,C) whereas no significant Lewis basicity dependent change at the meta position.
Overall, no significant influence of solvent polarity has been observed upon n-pentane addition, suggesting that the Lewis bases used in this study did not induce changes in 1 JF,C by changing the solvents bulk properties.

Computational Details
All geometry optimization calculations for all halogen bonded complexes investigated in this work were carried out utilizing the B3LYP 4-5 functional augmented with Grimme's D3 6 dispersion correction in combination with the large correlation consistent Dunning's aug-cc-pVTZ 7-8 basis set. Scalar relativistic effects for heavy atoms (e.g. I) were assessed by utilizing the Stuttgart-Dresden (SDD) 9-10 effective core potential. The B3LYP-D3 functional was chosen as it is known to adequately account for electron correlations for S23 systems exhibiting noncovalent interactions. [11][12] Dichloromethane solvation effects were included using the polarizable continuum model (PCM) of Tomasi and co-workers. 13 Vibrational frequency calculations were followed at the same level of theory to confirm the optimized geometry corresponding to geometry minima.
All calculations were performed using the Gaussian 16 Rev. C.01 package. 14 The geometries were optimized using an ultrafine grid and tight convergence criteria for the forces and displacements. 15 Natural population analysis and second-order perturbation of the Fock matrix analysis of two interacting orbitals were carried out utilizing the NBO7 program. 16 Topological analysis of electron density were carried out using the AIMALL version 19.10.12 program. 17 The nature of halogen bonding interactions were characterized through the energy density (Hc) at the NI and OI bond critical points, where a negative value of the energy (Hc < 0) indicates covalent bond and a positive value of the energy (Hc > 0) points to electrostatic interactions. 18 The associated binding energies (E) were calculated by taking the energy difference between halogen bonded complexes (A···B) and its isolated components (A and B) at their equilibrium geometries, Due to the size of the molecules, the coupling constant calculations of 1-iodoperluorooctane were replaced with 1-iodoperfluoropropane. Table S15. Contributions of the Fermi-contact (F,C), the spin dipolar (SD), the paramagnetic (PSO), and the diamagnetic spin-orbit (DSO) components to the change of 1 JF,C ( 1 JF,C) of the halogen bond donor upon halogen bonding. The slope of the computed contribution of the component as a function of  1 JF,C are given, with the variance of the linear fits being shown in brackets. The first three rows show data computed for 1-iodopentafluorobenzene (IC6F5), whereas the next two data for 1-iodoperfluorooctane (IC8F17). Visualisation of the data used for extraction of these values can be found in section 4.1. S25 Table S16. Bond distances (R), 19 F-13 C-1 J-coupling constants (JF,C), binding energies (E), electron densities () and energy densities (H) at the C-F and NI bond critical point, calculated at the B3LYP-D3/aug-cc-pVTZ-pp  Table S18. Decomposition terms of calculated 1 JF,C coupling constants into Fermi contact (FC), spin-dipolar (SD), paramagnetic spinorbital (PSO), diamagnetic spin-orbit (DSO), and the total 1 JF,C coupling constants at the ortho position as well as the second-order perturbation of the Fock Matrix between  orbital of C-I and * orbital of C-F, calculated at the B3LYP-D3/aug-cc-pVTZ-pp   a The aromatic rings were placed in the XY plane (in-plane), whereas the  orbitals are located in the out-of-plane (Z-axis). The  orbitals are assumed to have sp2-hybridization with equal contribution from 2s, 2px, and 2py.  orbital occupation a The aromatic rings were placed in the XY plane (in-plane), whereas the  orbitals are located in the out-of-plane (Z-axis). The  orbitals are assumed to have sp2-hybridization with equal contribution from 2s, 2px, and 2py.
Figure S11: Trends between calculated 1 JF,C coupling constants and natural occupation of 2s and 2p orbitals of the meta-C-atom in pentafluoroiodobenzene on halogen bond formation with different pyridine bases. a The aromatic rings were placed in the XY plane (in-plane), whereas the  orbitals are located in the out-of-plane (Z-axis). The  orbitals are assumed to have sp2-hybridization with equal contribution from 2s, 2px, and 2py.
Figure S12: Trends between calculated 1 JF,C coupling constants and natural occupation of 2s and 2p orbitals of the para-C-atom in iodopentafluorobenzene on halogen bond formation with different pyridine bases.  Table S27. Decomposition terms of calculated 1 JF,C coupling constants into Fermi contact (FC), spin-dipolar (SD), paramagnetic spinorbital (PSO), diamagnetic spin-orbit (DSO), and the total 1 JF,C coupling constants at the carbon  position as well as the second-order perturbation of the Fock Matrix between  orbital of C-I and * orbital of C-F and lone pair of I with  * of C-F orbital, calculated at the B3LYP-D3/aug-cc-pVTZ-pp

Spin-Dipolar Paramagnetic
Spin-orbit      Table S34. Decomposition terms of calculated 1 JF,C coupling constants into Fermi contact (FC), spin-dipolar (SD), paramagnetic spinorbital (PSO), diamagnetic spin-orbit (DSO), and the total 1 JF,C coupling constants at the ortho position as well as the second-order perturbation of the Fock Matrix between  orbital of C-I and * orbital of C-F, calculated at the B3LYP-D3/aug-cc-pVTZ-pp

C6F5-IBase
Fermi Contact    Table S40. Decomposition terms of calculated 1 JF,C coupling constants into Fermi contact (FC), spin-dipolar (SD), paramagnetic spinorbital (PSO), diamagnetic spin-orbit (DSO), and the total 1 JF,C coupling constants at the ortho position as well as the second-order perturbation of the Fock Matrix between  orbital of C-I and * orbital of C-F, calculated at the B3LYP-D3/aug-cc-pVTZ-pp

C6F5-IBase
Fermi Contact  Decomposition of the computed scalar couplings into FC, SD, PSO and DSO terms     S44 Figure    S46 Figure S24: The calculated 1 JF,C couplings vs calculated N..I for Lewis base-IC8F17 adducts for alpha and beta-F.
The calculated 1 JF,C vs N..I agree with the experimentally observed trends in  1 JF,C vs pKbI2 2 for IC6F5 , thus 1 JF,C ortho > para >meta (Figure 2 and Figure S23). For IC8F17, the  1 JF,C follows the trend alpha>beta ( Figure 4 and Figure S24). We observed good correlations for 1 JF,C for the -F and -F of IC8F17 and for the ortho-F and para-F of IC6F5, while weak correlation for the meta-F of IC6F5 in both the experimental (R 2 =0.13 with pKBI2 ) and calculated (R 2 =0.40 with N..I) datasets.