Synlett 2023; 34(10): 1129-1134
DOI: 10.1055/a-1928-2473
cluster
Dispersion Effects

London Dispersion Stabilizes Chloro-Substituted cis-Double Bonds

a   Institute of Organic Chemistry, Justus Liebig University, Heinrich-Buff-Ring 17, 35392 Giessen, Germany
,
a   Institute of Organic Chemistry, Justus Liebig University, Heinrich-Buff-Ring 17, 35392 Giessen, Germany
,
b   Institute of Inorganic and Analytical Chemistry, Justus Liebig University, Heinrich-Buff-Ring 17, 35392 Giessen, Germany
,
a   Institute of Organic Chemistry, Justus Liebig University, Heinrich-Buff-Ring 17, 35392 Giessen, Germany
› Author Affiliations

This work was supported by the priority program ‘Control of London Dispersion in Molecular Chemistry’ (SPP1807) of the Deutsche Forschungsgemeinschaft.
 


Abstract

We present a combined experimental and computational study on the thermodynamic stability of cis- and trans-alkenes substituted with dispersion energy donor (DED) groups. To investigate the role of noncovalent interactions on equilibrium of cis- and trans-alkenes we utilized hydrochlorination reactions. While the general assumption is that increasing steric bulk favors the trans-alkene, we observe an equilibrium shift towards the more crowded cis-alkene with increasing substituent size. With the aim to quantify noncovalent interactions, we performed a double mutant cycle to experimentally gauge the attractive potential of bulky substituents. Additionally, we utilized local energy decomposition analysis at the DLPNO-CCSD(T)/def2-TZVP level of theory. We found LD interactions and Pauli exchange repulsion to be the most dominant components to influence cis- and trans-alkene equilibria.


#

It is an accepted view in organic chemistry that trans double bonds are more stable than their cis analogues.[1] Based on the hard-sphere model, intramolecular steric hindrance is believed to alter the stability of the cis-isomer more significantly with increasing size of attached alkyl groups than for the ‘unfolded’ trans isomer. The notion of steric hindrance has a firm place in strain analyses, such as allylic,[2] syn-pentane,[3] and 1,3-diaxial[4] strain and is solely based on repulsive interactions due to close alkyl contacts. For instance, the A values are considered to be a direct measure of steric bulk.[4] However, recent studies demonstrated that London dispersion (LD) interactions[5] have to be taken into account as a key counterpart to steric repulsion.[6] That is, bulky alkyl groups can provide considerable stabilization due to LD. LD interactions were successfully utilized to stabilize weak bonds such as in hexaphenylethane derivatives[7] or cause molecules to dimerize by forming very close H···H contacts.[8] As well as carbon double bond, the nitrogen congeners usually favor the unfolded trans conformation.[9] For instance, Wegner et al.[10] demonstrated the importance of LD on the thermal isomerization reaction from trans- to cis-azobenzene by attaching bulky dispersion energy donor (DED) groups[11] in all-meta positions. Herein, we investigate the relative thermodynamic stability of trans- and cis-alkenes by utilizing a hydrohalogenation[12] reaction (Scheme [1]) and take into consideration the role LD plays in stabilizing the diastereomers.

Zoom Image
Scheme 1 Hydrochlorination reaction of all-meta-substituted diphenylalkynes 1-R2 . The reaction proceeds via a syn addition of HCl and forms a mixture of cis- and trans-alkenes (2-R2 ), which can equilibrate.

The surface-mediated syn addition of HCl generated in situ from the reaction of oxalyl chloride with water first gives the kinetically favored cis-alkene. Under acidic conditions the double bond isomerizes via a protonation–deprotonation mechanism. Kropp et al.[12] demonstrated that alkyl groups directly attached to the double bond favor the trans-alkene. However, a systematic study on the role of noncovalent intramolecular interactions has not been reported. Herein, we focus on stilbene derivatives using DEDs in all-meta positions. While parent stilbene favors the unfolded trans form[1] [13] (the enthalpy difference is 5.7 kcal mol–1 derived from heats of hydrogenation,[14] 3.8 kcal mol–1 from heats of combustion,[15] and 2.3 kcal mol–1 from temperature-dependent measurement of the iodine-catalyzed isomerization reaction[15]) over the more crowded ‘folded’ cis-stilbene due to conjugation of the planar trans-stilbene and intramolecular steric hindrance in cis-2-R2 .[16] The inclusion of chlorine disrupts planarity (dihedral angle α = ca. 36°) of the trans isomer and therefore lowers the energy gain due to conjugation by around 2.5 kcal mol–1 (see the Supporting Information (SI) for details). Nevertheless, repulsive steric interactions in cis-2-R2 are not affected much by chlorine incorporation.

To measure the effects of DEDs on the thermodynamic stability of stilbene derivatives, we synthesized all-meta-substituted diphenylalkynes[17] 1-R2 with methyl (Me), ethyl (Et), iso-propyl ( i Pr), and tert-butyl ( t Bu) groups attached. We generated the 1-bromo-3,5-disubstituted benzene derivatives via literature procedures[7a] [18] and utilized a Pd-catalyzed decarboxylative coupling reaction[19] with 2-butynedioic acid to give all-meta-substituted diphenylalkynes 1-R2 (Scheme [2]). Unsymmetrical diphenylalkynes were synthesized in a coupling reaction with phenylpropiolic acid.

Zoom Image
Scheme 2 Synthetic procedure for the preparation of all-meta-substituted diphenylalkynes and molecular structures derived from single-crystal X-ray diffraction of 1-Me2 and 1- t Bu2 . Thermal ellipsoids drawn at 50% probability level.

To test the suitability of the hydrochlorination and to gather information on thermodynamic equilibria of the stilbene derivatives, we chose 1,2-bis(3,5-dimethylphenyl)ethyne (1-Me2 ) as a model system.[20] We adopted the procedure of Kropp et al.[12] utilizing alumina for surface activation. The hydrolysis of oxalyl chloride was used to generate HCl in situ. The reaction was monitored via gas chromatography (GC–MS). After adding oxalyl chloride to a suspension of alkyne and alumina in DCM the starting material was consumed after around 20 min. Alumina adsorbs HCl in DCM solution, thereby enhancing its acidity and reactivity towards alkynes.[12b] The syn addition of HCl with the alkyne gives cis-2-Me2 . With an excess of reagent, the system equilibrates between the two diastereomers resulting in trans-2-Me2 as the main product. While equilibrium was reached after around 2 h, the mixture was allowed to react for 4 h in total. Figure [1] shows the composition of the reaction mixture with cis-alkene (red markings) as the kinetic and trans-alkene (blue markings) as the thermodynamic product. The starting material (grey markings) is consumed quickly. While the equilibrium shifts towards the unfolded trans-2-Me2 , the ratio and therefore energy difference between both conformers can be determined. Accordingly, the hydrochlorination is a means to an end to equilibrate the stilbene isomers.

Zoom Image
Figure 1 Composition of the reaction mixture with respect to reaction progress t. 1,2-Bis(3,5-dimethylphenyl)ethyne (1-Me2 , grey markings) converts into the cis- (red markings) and trans-alkene (blue markings).

The thermochemical results (ΔG R–R – ΔG H–H) of the computational and experimental study are depicted in Figure [2]. By representing the relative Gibbs free energies of the substituted and unsubstituted system, the trend already highlights the role DEDs play in the equilibrium (for absolute values, see the SI). While negative energy values correspond to a shift towards the more crowded cis diastereomer, positive values denote a shift to the trans isomer. Accordingly, the experimental data (Figure [2], blue markings) show a shift to the cis isomer with increasing polarizability of the substituents, although the absolute energies shows a preference of the trans isomer for all derivatives (for absolute values, see the SI). While Me substituents do not affect the equilibrium in comparison to the parent system (ΔG Me–Me – ΔG H–H = –0.1 ± 0.2 kcal mol–1), bulkier substituents all favor the folded isomer (negative energy values). This trend correlates well with polarizability α and therefore can be traced back to the substituents acting as DEDs.[5] Hence, the largest effect can be observed with tert-butyl substitution (ΔGt Bu–tBu – ∆G H–H = –0.7 ± 0.2 kcal mol–1). While bulkier DEDs shift the equilibrium to the more crowded cis-stilbene derivative, the chlorine atom appears to have a minor impact on the equilibrium. With its high polarizability chlorine can compete with alkyl substituents and diminishes the energetic preference for the cis derivative.[21]

Zoom Image
Figure 2 Gibbs free energies for the equilibrium of all-meta-substituted diphenyl cis- and trans-alkenes relative to the unsubstituted alkenes plotted vs. the polarizability α. Blue markings correspond to experimental values. Computations were performed at the B3LYP-D3(BJ)/def2-TZVPP (grey markings) and B3LYP/def2-TZVPP (red markings) level of theory. The PCM solvent model (DCM) was utilized for single-point energy computations with thermal corrections added from DFT frequency computations. Data points in the shaded area correspond to repulsive interactions.

We also performed a computational study on the thermodynamic stability of stilbene derivatives to assess the cis and trans diastereomer equilibrium. We utilized the Conformer–Rotamer Ensemble Sampling Tool[22] (crest) to identify conformers lowest in energy for the cis- and trans-alkenes. The preoptimized diastereomers were further optimized in the gas phase with Gaussian16.[23] We chose the B3LYP[24] functional including and excluding Grimme’s D3(BJ) correction[25] in conjunction with Ahlrich’s def2-TZVPP[26] basis set. To address solvent effects, single-point energy computations were performed utilizing the PCM model[27] with DCM as solvent. We verified these results using ωB97X-D[28] and higher-level computations such as single-point energies at the DLPNO–CCSD(T)/def2–TZVP[29] level of theory (see the SI for details).

The exclusion of LD interactions by utilizing the B3LYP functional without dispersion correction (red markings) predicts the trans-alkene to be lowest in energy for all compounds studied. Additionally, an incorporation of bulky substituents, such as iso-propyl or tert-butyl groups in all-meta position, results in positive energy values (up to +0.3 kcal mol–1). This would imply the intuitively often preferred repulsive nature of the intramolecular interactions in the cis-alkene. On the contrary, the inclusion of Grimme’s D3(BJ) correction (grey markings) leads to a significant stabilization in favor of the cis diastereomer. While the parent trans-stilbene is computationally favored by ΔG eq = +0.4 kcal mol–1, bulky substituents shift the equilibrium towards the cis-alkene. The largest effect is associated with the most polarizable tert-butyl groups with ΔGt Bu–tBu – ∆G H–H = –2.4 kcal mol–1 (for absolute values, see the SI). The close correlation with the polarizability α (Figure [2]) hints to attractive interactions relating to LD. By comparing computed with experimental data (blue markings), it is apparent that only the inclusion of dispersion corrections can help rationalize the observed trends. While computed energies arising from B3LYP/def2-TZVPP are not affected by the attached substituents, the B3LYP-D3(BJ)/def2-TZVPP level of theory is in line with the relative experimental energies. Both linear regressions (blue and grey markings) show the same sign and shift the equilibrium towards to more sterically hindered alkene. Nevertheless, the computed energetic preference due to the all-meta substitution pattern is around 70% more pronounced than our experimental findings suggest.[30] The highest shift towards the cis-alkene was observed with tert-butyl substitution (ΔGt Bu–tBu – ΔG H–H = –0.7 ± 0.2 kcal mol–1). While the absolute energies of the B3LYP-D3(BJ)/def2-TZVPP computations favor the cis isomer, the experimental study shows the trans isomer to be lowest in energy regardless of the substituents attached. Therefore, we conclude that an attenuation of intramolecular noncovalent interactions in DCM influences the equilibria of stilbene-type molecules significantly.[30] [31] This effect is not captured with solvent inclusion in the computations resulting in unsatisfactory computation of the absolute energy differences.

To investigate the origin of stabilization of the cis-alkene with bulky substituents attached, we conducted a qualitative analysis of noncovalent interactions. We generated intramolecular noncovalent interactions (NCI) plots[32] to highlight attractive and repulsive regions between the disubstituted phenyl moieties. The results for the cis diastereomers of 2-H2 (top) and 2- t Bu2 (bottom) are depicted in Figure [3]. While the trans-alkenes show no NCI interactions between the phenyl moieties (this is also confirmed by LED computations, see the SI), the cis isomers feature one large isosurface between adjacent phenyl moieties. With strong repulsive interactions color-coded in red and strong attractive interactions in blue, both alkenes show red isosurfaces in close proximity to the double bond, but blue ones in the exterior of both phenyl moieties. This is in line with the finding for hexaphenylethane.[7] The incorporation of bulky substituents results in an additional green isosurface corresponding to LD interactions. Accordingly, cis-2- t Bu2 is stabilized by close σ–σ contacts of around 2.3 Å in distance; these originate from the methyl groups of the tert-butyl substituents.

Zoom Image
Figure 3 Noncovalent interaction (NCI) plot of cis-2-H2 (top) and cis-2- t Bu2 at B3LYP-D3(BJ)/def2-TZVPP. Isosurfaces (isovalue s(ρ) of 0.5, ranging from ρ = –0.05 a.u. to +0.05 a.u.) are color-coded in red (strongly repulsive), blue (strongly attractive), and green (weak NCI).

Apart from qualitative investigations, we quantitatively assessed the role DEDs play in the equilibrium. We performed a double-mutant cycle[33] (DMC) to dissect the interaction energy ΔΔG R–R which corresponds to the energy gained or lost due to two substituents interacting with each other. ΔΔG R–R is dissected from unsubstituted and singly substituted systems according to the following equation:

ΔΔG R–R = ΔG R–R – ΔG R–H – ΔG H–R + ΔG H–H

The results of the DMC give an experimental estimate of the attractive (negative energy values) or repulsive (positive energy values) interactions exclusively between DEDs, thereby, excluding interactions between substituent and the opposing phenyl moiety (Figure [4]). Due to a large error estimate (summation of four ΔG errors) the results of the analysis have to be treated with caution (see the SI for error estimation). Nevertheless, the analysis qualitatively confirms the DED capacities of alkyl groups. The total interaction energies ΔΔG R–R between the substituents are all attractive (negative) and, consequently, stabilizing for the cis-alkene. Both, Me–Me and Et–Et contacts (ΔΔG R–R = ca. –0.1 kcal mol–1) only faintly favor the cis-alkene. This is not surprising since LD interactions are highly distance-dependent (r –6). With a distance of around d Me–Me = ca. 4.2 Å and d Et–Et = ca, 3.1 Å, the substituents do not fall into the van der Waals minimum (d ideal = ca. 2.5 Å) range.[31b] Additionally, the latter suffers an entropic penalty due to the increased flexibility of the Et substituent. A similar effect was observed in the conformational analysis of all-meta-substituted diphenylthiourea.[34] Azobenzene derivatives[35] as well as hierarchically assembled dinuclear titanium(IV) helicates[36] confirm the effect for longer alkyl chains. The highest stabilizing interaction in favor of the cis-alkene again is observed for the close t Bu– t Bu (dt Bu–tBu = ca. 2.3 Å) contacts by ΔΔGt Bu–tBu = –0.6 kcal mol–1. The observed trends stem from a combination of attractive LD interactions between bulky DEDs and the solvophobic effect in polar solvents. Both effects increase with the size of alkyl substituent attached.

Zoom Image
Figure 4 Double mutant cycle (top) utilized for the determination of ΔΔG R–R and results of the analysis (bottom). Negative energy values correspond to stabilizing interactions between the substituents in the cis isomer.

We performed a computational and experimental study on the role LD interactions play in the thermodynamic equilibrium of cis- and trans-alkenes, utilizing surface-mediated hydrochlorinations to study the effects of bulky substituents. In contrast to the notion that steric bulk favors the unfolded trans-alkene due to steric hindrance in the cis-olefin, we found a counterintuitive shift towards the more crowded cis-alkene with increasing substituent size. We highlight LD interactions as the main source of stabilization. To confirm these findings, we evaluated a double-mutant cycle to quantify the stabilizing interactions between polarizable alkyl groups attached in all-meta positions. The most prominent effect was observed with tert-butyl substitution. By analogy, an LED analysis provides additional evidence that LD interactions only affect the folded cis-alkene via intramolecular σ–σ contacts.


#

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

We thank Steffen Wagner for performing GC-MS measurements and Jan M. Schümann, Bastian Bernhardt, Henrik F. König, and Finn M. Wilming for fruitful discussions.

Supporting Information


Corresponding Author

Peter R. Schreiner
Institute of Organic Chemistry, Justus Liebig University
Heinrich-Buff-Ring 17, 35392 Giessen
Germany   

Publication History

Received: 07 July 2022

Accepted after revision: 19 August 2022

Accepted Manuscript online:
19 August 2022

Article published online:
30 September 2022

© 2022. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


Zoom Image
Scheme 1 Hydrochlorination reaction of all-meta-substituted diphenylalkynes 1-R2 . The reaction proceeds via a syn addition of HCl and forms a mixture of cis- and trans-alkenes (2-R2 ), which can equilibrate.
Zoom Image
Scheme 2 Synthetic procedure for the preparation of all-meta-substituted diphenylalkynes and molecular structures derived from single-crystal X-ray diffraction of 1-Me2 and 1- t Bu2 . Thermal ellipsoids drawn at 50% probability level.
Zoom Image
Figure 1 Composition of the reaction mixture with respect to reaction progress t. 1,2-Bis(3,5-dimethylphenyl)ethyne (1-Me2 , grey markings) converts into the cis- (red markings) and trans-alkene (blue markings).
Zoom Image
Figure 2 Gibbs free energies for the equilibrium of all-meta-substituted diphenyl cis- and trans-alkenes relative to the unsubstituted alkenes plotted vs. the polarizability α. Blue markings correspond to experimental values. Computations were performed at the B3LYP-D3(BJ)/def2-TZVPP (grey markings) and B3LYP/def2-TZVPP (red markings) level of theory. The PCM solvent model (DCM) was utilized for single-point energy computations with thermal corrections added from DFT frequency computations. Data points in the shaded area correspond to repulsive interactions.
Zoom Image
Figure 3 Noncovalent interaction (NCI) plot of cis-2-H2 (top) and cis-2- t Bu2 at B3LYP-D3(BJ)/def2-TZVPP. Isosurfaces (isovalue s(ρ) of 0.5, ranging from ρ = –0.05 a.u. to +0.05 a.u.) are color-coded in red (strongly repulsive), blue (strongly attractive), and green (weak NCI).
Zoom Image
Figure 4 Double mutant cycle (top) utilized for the determination of ΔΔG R–R and results of the analysis (bottom). Negative energy values correspond to stabilizing interactions between the substituents in the cis isomer.