London Dispersion Favors Sterically Hindered Diarylthiourea Conformers in Solution

Abstract

We present an experimental and computational study on the conformers of N,N′‐diphenylthiourea substituted with different dispersion energy donor (DED) groups. While the unfolded anti–anti conformer is the most relevant for thiourea catalysis, intramolecular noncovalent interactions counterintuitively favor the folded syn–syn conformer, as evident from a combination of low‐temperature nuclear magnetic resonance measurements and computations. In order to quantify the noncovalent interactions, we utilized local energy decomposition analysis and symmetry‐adapted perturbation theory at the DLPNO‐CCSD(T)/def2‐TZVPP and sSAPT0/6‐311G(d,p) levels of theory. Additionally, we applied a double‐mutant cycle to experimentally study the effects of bulky substituents on the equilibria. We determined London dispersion as the key interaction that shifts the equilibria towards the syn–syn conformers. This preference is likely a factor why such thiourea derivatives can be poor catalysts.

Steric crowding counterintuitively favors the folded and apparently more sterically hindered syn–syn conformer of N,N′‐diphenylthiourea derivatives. Combined experimental and computational analyses identified London dispersion interactions to be the decisive factor favoring the conformer equipped with bulky dispersion energy donor groups.

Introduction

In the field of enzyme catalysis, Fischer's “key and lock” hypothesis ^[1] or the more sophisticated “induced fit” model by Koshland ^[2] perennially highlight the importance of conformational flexibility and catalytic activity. The structural dynamics of peptides allow enzymes to bind and to recognize substrates effectively and convert them into products. Thus, a specific conformer of the catalyst is needed to exploit transition state stabilization and energetic differentiation among a series of possible transition state geometries. Conformational structure–property relationships can be probed with small molecules as well. The restricted bond rotation within the thioamide functional group offers three differently populated conformers (Scheme 1). ^[3] While the mechanism for anion recognition or catalytic activation of a substrate due to hydrogen‐bonding is most effective via the open anti–anti conformer, an analysis of the conformational landscape of thiourea derivatives is an essential part to understand the origin of their catalytic activity and any limitations thereof. ^[3] Here, we present a study of all‐meta‐disubstituted diphenylthiourea ^[4] derivatives to elucidate the conformational preferences dependent on noncovalent interactions including London dispersion (LD). ^[5] Since the compounds discussed in this work both are less catalytically active than commonly exploited thiourea catalysts ^[4c] and poor anion receptors,[ ^4h , ⁶ ] we hypothesize this is in part due to the population of a conformer that does not allow double N−H bonding to Lewis‐basic atoms or groups in the substrate. ^[3]

Scheme 1.

Lowest energy conformers of diphenylthiourea derivatives with the anti–anti (left), syn–anti (center), and syn–syn (right) conformers. The shown values correspond to the shortest σ–σ distance d _σ–σ contact for each conformer of 1‐R¹R² computed at B3LYP‐D3(BJ)/def2‐TZVPP.

In recent years, a number of studies demonstrated that the catalytically active anti–anti diphenyl(thio)urea conformer is not necessarily the predominant conformer in the gas phase and in solution.[ ³ , ^6a , ⁷ ] Infrared and temperature‐dependent NMR measurements in different solvents demonstrated the presence of multiple conformers for diarylthiourea derivatives. ^[7] An exception to this conformational flexibility is the well‐known N,N′‐bis[3,5‐bis(trifluoromethyl)phenyl]thiourea catalyst with the anti–anti conformer being predominant in, for example, tetrahydrofuran (THF) at elevated temperatures. ^[3] While the experimental evidence points to the fact that the anti–anti conformer of this thiourea catalyst is catalytically most active, other substitution patterns are likely to display a different conformational landscape, which, in turn, is likely to result in reduced catalytic activity.

Most recently, Sandler et al. ^[6a] highlighted the relationship of conformational effects and the anion binding affinity of receptor molecules such as diphenylthiourea.[ ^4h , ^6b ] Whereas urea and squaramide derivatives prefer the anti–anti conformer due to intramolecular CH‐carbonyl hydrogen bonding, diphenylthiourea does not benefit as strongly from this stabilization since its phenyl moieties are twisted out of plane. ^[4c] Consequently, diphenylthiourea populates the syn–anti and syn–syn conformers, thereby lowering its anion binding affinity. ^[6a] To explain the enantioselectivity of an asymmetric Henry reaction, Heshmat proposed cinchona‐thiourea catalyst substrate activation via the syn–anti conformer. ^[8] Experimental data suggest a similar trend. In an extensive study of crystal structures of urea and thiourea derivatives, Luchini et al. ^[9] showed that around 60% of all thiourea motifs crystallize in a syn–syn or syn–anti fashion. On the other hand, 98% of urea derivatives are reported to have an anti–anti conformation in the solid state. ^[9] Solid state and gas phase IR ^[10] and NMR ^[11] studies in solution support this trend for urea derivatives as well. For diarylthiourea derivatives, IR measurements suggest a significant shift to the syn–syn conformer in solution ^[7] but a systematic NMR study determining the role and the apparent intramolecular stabilization of the syn–syn conformer has not been reported.

In order to investigate the equilibria depicted in Scheme 1, we treated the N,N′‐diphenylthiourea derivatives as molecular balances. ^[12] By increasing the size of the all‐meta‐substituted aryl dispersion energy donors (DEDs),[ ⁵ , ¹³ ] we observed a systematic and counterintuitive shift of the equilibrium toward the folded and more crowded syn–syn conformer. The increasing number of close σ–σ contacts is indicative of the prevalence of attractive LD ^[14] interactions rather than Pauli (exchange) repulsion. This effect was recently emphasized in a study of the equilibria of 1,4‐ and 1,6‐di‐tbutyl cyclooctatetraene in a large series of solvents of very different polarities showing that intramolecular LD interactions do not cancel in solution. ^[15]

Results and Discussion

To dissect the influence of each DED, we synthesized a logical series of diphenylthiourea derivatives with methyl (Me), ethyl (Et), iso‐propyl ( ⁱ Pr), and tert‐butyl ( ^t Bu) substituents. In brief, the all‐meta‐substituted N,N′‐diphenylthioureas were synthesized via a two‐step addition of aniline precursors to thiophosgene. ^[3] Prior, the all‐meta‐substituted aniline precursors were generated via bromination and de‐diazotization reaction of 2,6‐disubstituted aniline derivatives (for details, see Supporting Information). ^[16] To gather as much information as possible, we generated all R¹ and R² combinations of these groups and measured ¹H NMR spectra in THF. The choice of solvent was based on its physical properties (i.e., low melting point) and the fact that all diphenylthiourea derivatives remained soluble during the low‐temperature NMR measurements.

The restricted bond rotation of all N,N′‐diphenylthiourea derivatives required low‐temperature NMR measurements (performed at 193 K) in order to freeze the C−N bond rotation. The lowest temperature possible to hold up over a longer period of time in the NMR was 193 K. Intrinsic reaction coordinate (IRC) computations suggest activation barriers of 10.3 kcal mol⁻¹ (corresponding to a rate constant of 4.0×10⁻³ s⁻¹) and 9.0 kcal mol⁻¹ (1.3 s⁻¹) for unsubstituted diphenylthiourea. ^[3] We first tested our approach with N,N′‐bis(3,5‐di‐tert‐butylphenyl)thiourea 1‐ ^t Bu₂ (1‐R¹R² ) and the parent N,N′‐diphenylthiourea 1‐H₂ . For both derivatives the singlet N−H signal splits into three separate signals upon cooling, two of which belong to the same conformer (blue marking, Figure 1). Additionally, the aromatic signals (Figure 2) split into four, and the aliphatic tert‐butyl signals into two separate NMR peaks.

Figure 1.

NMR measurements at 193 K of symmetrically substituted N,N′‐diphenylthiourea derivatives 1‐R¹R² in THF and molecular structure of 1‐ ^t Bu₂ . For simplicity, the NH signals of symmetric 1‐R¹R² are depicted only. Thermal ellipsoid plot of the molecular structure obtained by single‐crystal X‐ray diffraction was drawn at 50% probability level. The blue markings correspond to the NH signals of the syn–anti and the green markings to the syn–syn conformer. Note that the anti–anti conformer is not populated and has therefore been omitted.

Figure 2.

NMR measurements at 193 K of the aromatic signals of 1‐ ^t Bu₂ (grey) and computed spectra for the syn–anti (blue), syn–syn (green), and anti–anti (red) conformers in THF (SMD solvent model) at the B3LYP‐D3(BJ)/def2‐TZVPP level of theory. For the full spectral data see the Supporting Information.

Accordingly, these signals were assigned to the syn–anti conformer since it is the only structure with inequivalent N−H, aromatic, and tert‐butyl protons. While the parent 1‐H₂ (purple NMR, Figure 1) considerably favors the syn–anti conformer by around 2.3±0.1 kcal mol⁻¹ (all energies were determined via K _eq at 193 K), the NMR of 1‐ ^t Bu₂ (black NMR, Figure 1) shows a distinct symmetric conformer. Nevertheless, 1‐ ^t Bu₂ favors the syn–anti conformer by around 0.5±0.0(3) kcal mol⁻¹. The computed NMR signals (Figure 2) suggest that the new signals belong to the syn–syn conformer (green spectrum), which also helped us assign the syn–anti (blue spectrum) and disregard the anti–anti (red spectrum) conformer. Whereas the N−H proton shift is difficult to determine by NMR computations, ^[17] the aromatic and aliphatic C−H signals were assigned to the syn–syn conformer.

Concentration dependent measurements showed no change in signal ratios with the lowest concentration being 15.5 mM (0.01 mmol). This is in line with NMR measurements investigating the complexation of thiourea catalyst with lactones, where it was found that the anti–anti conformer is catalytically most active. ^[3] Consequently, aggregation in solution was deemed to be unimportant. To ensure that equilibrium had been reached, we equilibrated each NMR sample for one hour at 193 K. Since the barrier height for rotation around the thioamide bond is around 10 kcal mol⁻¹, equilibrium was reached after around 5 min (see Supporting Information for details). After transferring the samples to the NMR spectrometer, they were further equilibrated until the temperature stabilized at 193 K. Figure 1 displays the N−H proton splitting for symmetric 1‐R¹R² . While 1‐H₂ shows only low concentrations of a second conformer, bulky substituents such as those with tert‐butyl groups clearly affect the conformer distributions.

Figure 3 displays a summary of the experimentally determined ΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2\text{-}\mathrm{HH}}$ values of the equilibrium between syn–syn and syn–anti 1‐R¹R² relative to parent N,N′‐diphenylthiourea 1‐H₂ . Consequently, 1‐H₂ is depicted as ΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2\text{-}\mathrm{HH}}$ =0.0±0.2 kcal mol⁻¹ in Figure 3 (rightmost data point). While 1‐H₂ favors the syn–anti conformer by around 2.3±0.1 kcal mol⁻¹ (see Supporting Information for absolute energy values), substituents in all‐meta position shift the equilibrium towards the syn–syn conformer (ΔG<0). In contrast to the often encountered view that large groups repel each other, Figure 3 illustrates that bulky groups favor the conformer that displays close alkyl–alkyl contacts (d _σ‐σ=2.61 Å for 1‐ ^t Bu₂ ). The unsymmetric functionalization in 1‐R¹H (blue bars) and 1‐HR² (rightmost block of columns) only has a small effect on the equilibrium position (up to ΔG ${}_{{}^t\mathrm{BuH}\text{-}\mathrm{HH}}$ =−0.6±0.2 kcal mol⁻¹). The shift in energy towards the syn‐syn conformer can be rationalized by attractive σ–π interactions between substituents and opposing phenyl moiety. Thereby, a decrease in distance between substituent and phenyl moiety systematically increases the stabilizing σ–π interactions. A similar effect was already observed and quantified by Shimizu et al. for a para substitution pattern utilizing molecular torsion balance. ^[18] Here, distance dependence of σ–π interactions was documented for a para substitution pattern with the largest and bulkiest alkyl groups forming the strongest stabilizing interactions. These observations are consistent with the recent concept of DEDs in which bulky alkyl groups form stabilizing dispersion interactions.[ ⁵ , ¹³ ]

Figure 3.

Experimentally determined Gibbs free energy values ΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2\text{-}\mathrm{HH}}$ for the equilibrium between syn–syn and syn–anti 1‐R¹R² at 193 K′ gray lines indicate error bars. ΔG<0 corresponds to a shift towards the syn–syn conformer: The lower and more negative the ΔG expressed, the more favored the syn–syn conformer. Note that the supposedly catalytically active anti–anti conformation is not populated at all.

By systematically increasing the substituent size on both phenyl moieties, the equilibrium shifts further to the more crowded syn–syn structure. The introduction of additional CH₃ groups increases the number of close intramolecular alkyl–alkyl contacts in the syn–syn conformer, thereby reducing the distance between substituents (Scheme 1). An increasing number of noncovalent contacts at distances of around 2.5 Å has proven to be effective in stabilizing labile compounds such as hexaphenylethane ^[19] or rationalizing isomerization energies of linear and branched alkanes. ^[20] The largest difference in energy due to incorporation of a methyl substituent can be observed from 1‐ ^t BuH (ΔG ${}_{{}^t\mathrm{BuH}\text{-}\mathrm{HH}}$ =−0.6±0.2 kcal mol⁻¹) to 1‐ ^t BuMe (ΔG ${}_{{}^t\mathrm{BuMe}\text{-}\mathrm{HH}}$ =−1.2±0.2 kcal mol⁻¹) with around −0.6 kcal mol⁻¹ stabilization due to σ–σ contacts. ^[21] Additional methyl groups shift the equilibrium further towards the syn–syn conformer by around −0.1 kcal mol⁻¹. Consequently, the most prominent effects can be observed for 1‐R^1t Bu derivatives (orange bars), which shift the equilibria significantly towards the syn–syn conformer (up to ΔG ${}_{{}^t\mathrm{Bu}{}^t\mathrm{Bu}\text{-}\mathrm{HH}}$ =−1.7±0.1 kcal mol⁻¹). Hence, the experimental data suggest that increasingly larger alkyl substituents act as stabilizing DEDs rather than as repulsive steric bulk.[ ⁵ , ¹⁹ , ²² ] Correlations of our experimental findings with the molecular volume or in the total molecular dipole moment of each conformer are insufficient to rationalize the trends observed (see Supporting Information).

To support these findings, we performed a computational study focusing on the role of intramolecular noncovalent interactions. To be able to switch dispersion corrections on and off, we utilized density functional theory (DFT) to investigate the equilibria depicted in Scheme 1. After an initial conformer analysis using the Conformer–Rotamer Ensemble Sampling Tool ^[23] (crest) program, the lowest conformers were further optimized with Ahlrich's def2‐TZVPP ^[24] basis set. The B3LYP ^[25] functional was utilized with and without (Supporting Information) Grimme's D3 ^[26] correction including Becke–Johnson ^[27] (BJ) damping. All geometry optimizations were performed in the gas phase under standard conditions. The gas phase structures were utilized for single‐point energy computations to account for solvation effects and entropy at 193 K. The polarizable continuum model (PCM) ^[28] was used with THF as solvent and thermal corrections added from DFT (gas phase) frequency computations. Additionally, the B3LYP‐D3(BJ) (gas phase) optimized structures were utilized for single‐point energy computations at the DLPNO‐CCSD(T)/def2‐TZVPP level of theory. ^[29] This analysis follows that of Sandler et al. (Supporting Information) ^[6a] who demonstrated that the B3LYP functional in conjunction with medium‐sized basis sets is an appropriate approach for geometry optimizations of thiourea derivatives and, that DLPNO‐CCSD(T)/large basis set is an excellent approximation to its canonical counterpart. Since LD interactions are in a first approximation temperature independent, the results of the thermochemical analysis of the equilibrium fit qualitatively to gas phase computations (see Supporting Information). ^[14] The thermochemical results (ΔG _eq) for the symmetric and unsymmetric N,N′‐diphenylthiourea molecular balances are depicted in Figure 4. While the anti–anti conformer is highest in energy (red markings) for all systems and cannot be observed by NMR, the syn–syn (green markings) conformers are generally favored. Computations on B3LYP/def2‐TZVPP excluding the LD corrections predict the syn–anti/anti–syn conformers to be favored by around 3–4 kcal mol⁻¹. Including LD, the unsubstituted balance already slightly favors the syn–syn conformer (ΔG _eq≈−0.3 kcal mol⁻¹). Increasing alkyl substitution shifts the global energy minimum further from the syn–anti towards the syn–syn conformer. The largest effect can be observed for 1‐ ^t Bu₂ (ΔG _eq≈−2.8 kcal mol⁻¹). These results fit qualitatively well to our experimental data, albeit the attenuation of the attractive interactions due to solvent effects is higher than predicted by the computations. ^[30]

Figure 4.

Gibbs free energies at 193 K for the equilibrium of the syn–syn (green markings) and anti–anti (red markings) conformers relative to the syn–anti/anti–syn conformers of 1‐R¹R² at the DLPNO‐CCSD(T)/def2‐TZVP//B3LYP‐D3(BJ)/def2‐TZVPP level of theory including a solvent correction (THF) at the B3LYP‐D3(BJ)/def2‐TZVPP utilizing the PCM model. Thermal corrections were added from DFT optimizations at 193 K. 1‐H₂ and 1‐ ^t Bu₂ are highlighted for clarity.

To assess these counterintuitive results, we visualized the intramolecular noncovalent contacts (Figure 5) utilizing noncovalent interaction (NCI) plots ^[31] to highlight the main source of thermodynamic stability of 1‐ ^t Bu₂ by depicting the reduced density gradient in regions of low electron density. While strongly attractive and repulsive interactions are color‐coded in blue and red, respectively, green isosurfaces can be assigned to weak NCIs. The anti–anti conformer of 1‐ ^t Bu₂ features a mixture of red and blue isosurfaces due to the substitution pattern and a CH⋅⋅⋅S contact ^[4c] a green contact area is not visible. On the other hand, the syn–anti conformer already shows small green areas between bulky ^t Bu substituents and the opposing phenyl group. Finally, the syn–syn conformer shows large green isosurfaces implying significant intramolecular NCIs. An incorporation of bulky alkyl groups increases the number of noncovalent contacts via close σ–σ (i.e., ^t Bu− ^t Bu in Figure 5) and σ–π ( ^t Bu–π) contacts of both substituents. This analysis qualitatively supports experimental and computational findings.

Figure 5.

Noncovalent interaction (NCI) plots of the anti–anti (left), syn–anti (center), and syn–syn (right) conformers of 1‐ ^t Bu₂ at B3LYP‐D3(BJ)/def2‐TZVPP. Isosurfaces (isovalue s of 0.2, ranging from sign(λ ₂)ρ=−0.05 a.u. to +0.05 a.u.) are color‐coded red (indicating strong repulsion), blue (strong attractive interactions), and green (corresponding to weak NCI).

To quantify the amount of LD interactions between each substituent, we dissected the energy values ΔΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2}$ from singly substituted molecular balances. ^[32] Hereby, two substituents R¹ and R² are mutated separately to investigate the impact of each substituent on the thiourea molecular backbone. According to the following equation the interaction energy ΔΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2}$ between two substituents can be determined as follows:

$${\displaystyle \Delta \Delta G{}_{\mathrm{R}{}^1\mathrm{R}{}^2}=\Delta G{}_{\mathrm{R}{}^1\mathrm{R}{}^2}-\Delta G{}_{\mathrm{R}{}^1\mathrm{H}}-\Delta G{}_{\mathrm{HR}{}^2}+\Delta G{}_{\mathrm{HH}}}$$ (1)

While this application of Hess's law (also referred to as double mutant cycle) gives an experimental estimate of the role each DED plays, the results have to be treated with caution due to a large error estimate (see Supporting Information for details). Nevertheless, Figure 6 qualitatively supports our findings that sterically hindered diphenylthiourea derivatives favor the syn–syn conformer. In general, all calculated energies are negative implying a stabilizing effect between the alkyl groups. Especially for large moieties, a stabilization of the syn–syn conformer can be observed (ΔΔG ${}_{{}^t\mathrm{Bu}{}^t\mathrm{Bu}}$ =−0.5±0.3 kcal mol⁻¹). Therefore, around 30% of the observed Gibbs free energy values (ΔΔG ${}_{{}^t\mathrm{Bu}{}^t\mathrm{Bu}\text{-}\mathrm{HH}}$ =−1.7±0.1 kcal mol⁻¹) can be assigned to stabilizing alkyl–alkyl contacts. The remaining 70% consists of σ–π interactions between ^t Bu and the opposing phenyl moiety. ^[21] The smallest effect was measured for the 1‐Me₂ molecular balance (ΔΔG _MeMe=−0.1±0.4 kcal mol⁻¹). In comparison to ΔΔG ${}_{\mathrm{R}{}^1\mathrm{Me}}$ (yellow bars), ΔΔG ${}_{\mathrm{R}{}^1\mathrm{Et}}$ (purple bars) does not profit from an additional CH₃ group. This can be rationalized with an entropic penalty ^[33] due to increasing flexibility of the ethyl substituent.

Figure 6.

Double mutant cycle (top) to dissect the interaction energy ΔΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2}$ and results of the analysis (bottom); gray lines indicate error bars. ΔΔG ${}_{\mathrm{R}{}^1\mathrm{R}{}^2}$ describes the relative interaction energies of R¹–R² contacts of the syn–anti and syn–syn equilibrium at 193 K. Negative energies correspond to stabilizing interactions between both groups.

With the aim to dissect the intramolecular interaction energy into its main contributors, we employed symmetry‐adapted‐perturbation theory ^[34] (SAPT) analysis as implemented in PSI4. ^[35] The scaled version was used according to Sherrill et al. ^[36] to improve the performance of the decomposition method. We focused solely on the interaction between the two substituted phenyl moieties. As a starting point, we took the B3LYP‐D3(BJ)/def2‐TZVPP optimized geometries and removed the thiourea moiety. The resulting phenyl radicals were saturated with hydrogen yielding a benzene dimer in geometry of the syn–syn, syn–anti and anti–anti conformer. This approach allows us to transfer the intramolecular into intermolecular interactions between two substituted benzene molecules. While the electronic constitution of benzene varies from the electronic structure within diphenylthiourea, this method was solely used to identify the main source of thermodynamic stability. Figure 7 displays the energy decomposition of the total interaction energy (E _tot) between two di‐substituted benzene molecules based on their geometry in the syn–syn conformer (for other conformers see Supporting Information). While inductive effects (E _ind, blue markings) only play a minor role in the dimerization of substituted benzene, electrostatic (E _estat, grey markings) as well as LD interaction (E _disp, green markings) are essential to understand the interaction energy between two benzene molecules. Both energies, E _estat and E _disp, stabilize the benzene dimer due to alkyl substitution with LD interactions as the major component (up to E _disp=−21.1 kcal mol⁻¹ for 1‐ ^t Bu₂ ). Nevertheless, only a combination of both energies overcompensates the destabilizing contributions of Pauli exchange repulsion (E _exch, red markings). Especially for 1‐H₂ , repulsive interactions (E _exch=+15.3 kcal mol⁻¹) disfavor the aggregation of benzene and override all stabilizing effects (E _tot=+0.7 kcal mol⁻¹). While Herbert et al. ^[37] identified LD as the main attractive component in cofacial π‐stacking (via σ–π contacts) of benzene, this effect alone is not strong enough to stabilize 1‐H₂ . The geometry of close benzene dimers enforced through the thiourea molecular backbone is therefore not ideal to afford the perfect balance between attractive and repulsive contacts. With increasing substituent bulkiness repulsive interactions increase (up to E _exch=+26.3 kcal mol⁻¹ for 1‐ ^t Bu₂ ) but do not overcompensate the attractive interactions.

Figure 7.

sSAPT analysis of two 1,3‐disubstituted benzene molecules in the geometry of the syn–syn thiourea conformers at sSAPT0/6‐311G(d,p) at 298 K. The dashed lines are used to guide the eye.

After establishing that LD interactions are the major factor for the conformational preference of diphenylthiourea derivatives, we set out to quantify the magnitude of LD interactions between the aromatic moieties without changing the electronic structure of N,N′‐diphenylthiourea. While the double mutant cycle (Figure 6) represents the total interaction energy (sum of all attractive and repulsive components) between DED groups attached, the overall energy gain due to LD interactions was dissected using a Local Energy Decomposition (LED) analysis ^[38] as implemented in ORCA. ^[39] Therefore, we fragmented every N,N′‐diphenylthiourea molecular balance into three parts (F1, F2, and F3). During this process all bonds are cleaved homolytically resulting in large electrostatic interactions between all fragments. Consequently, we investigated only the gain in energy due to LD interactions between F1 and F2. Figure 8 shows the results of the analysis for the syn–syn conformers (see Supporting Information for other conformers).

Figure 8.

LD interaction energies derived from LED analysis of two 1‐R¹R² substituted phenyl moieties in syn–syn conformer at DLPNO‐CCSD(T)/def‐TZVP//B3LYP‐D3(BJ)/def2‐TZVPP at 298 K.

The LED analysis fits qualitatively to the results of computational and experimental data very well. In comparison to the SAPT analysis, LED suggests lower LD contributions (around 6 kcal mol⁻¹), but this is due to the different models used. Accordingly, 1‐H₂ and the semi‐substituted 1‐HR² series benefit the least from LD interactions (between −4.0 to −7.3 kcal mol⁻¹). On the other hand, substitution on both phenyl moieties results in higher LD interaction energies up to E _disp=−13.7 kcal mol⁻¹ for 1‐ ^t Bu₂ . This effect is most prominent in the syn–syn conformer. All methods utilized to quantify noncovalent interactions demonstrate the role of LD on the conformational preference of N,N′‐diphenylthiourea derivatives. The experimental and computational data suggest simple additivity of the DED strength due to an increasing preference of the syn–syn conformer with growing steric bulk. The double mutant cycle highlights both, σ–σ and σ–π contacts as the origin of stabilization.

Conclusion

We performed a systematic experimental‐computational study on the folding equilibria of all‐meta substituted diphenylthiourea derivatives investigating the impact of steric bulk on the conformer preferences. In stark contrast to the broadly accepted dominance of Pauli repulsion dictating conformations, we identified LD interactions as the main contributor that counterintuitively stabilizes the syn–syn conformers. Therefore, LD proves to be a powerful interaction to shift equilibria towards apparently more crowded conformers.

A double‐mutant cycle allowed us to quantify and differentiate between attractive σ–σ and σ–π contacts as origin of stabilization. The most prominent shift towards the folded syn–syn conformer was observed when attaching bulky tert‐butyl substituents to diphenylthiourea. An SAPT analysis reveals a combination of electrostatic and LD interactions counteracting Pauli repulsion. The LED analysis helped quantify intramolecular LD interactions and confirmed tert‐butyl substituents to be highly effective DEDs.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

L. Rummel, M. H. J. Domanski, H. Hausmann, J. Becker, P. R. Schreiner, Angew. Chem. Int. Ed. 2022, 61, e202204393; Angew. Chem. 2022, 134, e202204393.

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.