Mechanistically Guided Design of Ligands that Significantly Improve the Efficiency of CuH-Catalyzed Hydroamination Reactions

Abstract

Using a mechanistically guided ligand design approach, a new ligand (SEGFAST) for the CuH-catalyzed hydroamination reaction of unactivated terminal olefins has been developed, providing a 62-fold rate increase over reactions compared to DTBM-SEGPHOS, the previous optimal ligand. Combining the respective strengths of computational chemistry and experimental kinetic measurements, we were able to quickly identify potential modifications that lead to more effective ligands, thus avoiding synthesizing and testing a large library of ligands. By optimizing the combination of attractive, non-covalent ligand-substrate interactions and the stability of the catalyst under the reaction conditions, we were able to identify a finely-tuned hybrid ligand that greatly enables accelerated hydrocupration rates with unactivated alkenes. Moreover, a modular and robust synthetic sequence was devised, which allowed for practical, gram-scale synthesis of these novel hybrid ligand structures.

Graphical Abstract

INTRODUCTION

In 2013, Buchwald,¹ and Miura and Hirano² independently demonstrated that copper hydride complexes (LCuH) can catalyze the chemo- and enantioselective hydroamination reactions between olefins and hydroxylamine esters (Figure 1a). Since then the generality and applicability of this approach has been well demonstrated for a variety of substrates such as styrenes,³ vinylsilanes,⁴ alkynes⁵ and occasionally unactivated olefins,⁶ highlighting its enormous potential. Despite these achievements hydroamination reactions catalyzed by LCuH are not without limitations.⁷ For example, transformation of coupling partners such as cyclic, internal, and some unactivated terminal olefins often require elevated temperatures and increased reaction times compared to those of activated substrates.⁸ In particular, an efficient anti-Markovnikov hydroamination reaction with unactivated terminal olefins is highly desirable, because the products of these reactions are frequently found in bioactive molecules.⁹ These compounds are traditionally prepared by transforming carbonyl compounds, amides or alkyl electrophiles into their corresponding amine products.¹⁰ From a strategic standpoint, hydroamination reactions between olefins and electrophilic amine sources provide one of the most straightforward and general avenues to access these important motifs, especially since the precursors are typically stable, readily available, and easy to handle.¹¹

Figure 1.

a) LCuH-catalyzed anti-Markovnikov hydroamination reaction. b) Proposed catalytic cycle for LCuH-catalyzed anti-Markovnikov hydroamination reaction. c) SEGPHOS L1 and DTBM-SEGPHOS L2 ligands.

Recently, several experimental^12,13 and computational¹⁴ mechanistic investigations have appeared on both CuH catalyzed hydroamination and hydroboration reactions, revealing the same basic catalytic cycle, comprised of four elementary steps: hydrocupration (I), oxidative addition (II), reductive elimination (III) and σ-bond metathesis (IV) (Figure 1b).¹⁵ These studies demonstrated that the rate determining step (RDS) can vary between different olefinic substrates. Specifically, the RDS for activated substrates, such as styrenes, is often the catalyst regeneration by σ-bond metathesis, whereas it changes to the hydrocupration step for unactivated internal or terminal alkenes,^{12, 14a} indicating that the lower reactivity observed for unactivated olefins is the direct result of higher barriers for hydrocupration.

Typically, reaction development in this area has relied on empirical observations pertaining to which catalytic system provides the fastest reaction rates.¹¹ For example, reactions with SEGPHOS L1 supported LCuH catalysts are often found to be slower; whereas, the use of sterically more demanding DTBM-SEGPHOS L2 derivative is often key to achieving higher reactivity, especially in the reactions of unactivated olefins (Figure 1c).^14a

Beginning last year, our laboratories sought to unravel the theoretical foundations that lead to favorable hydrocupration events between LCuH (L = L1 and L2) and various unactivated olefins, by performing the ligand-substrate interaction model analysis on this crucial step (Figure 2a).^14a Using this approach, the contributions of the different types of catalyst-substrate interactions to the overall activation energy (ΔE^‡) were split into three categories: (1) the distortion energy required for the LCuH and the substrate to reach their transition state geometries (ΔE_dist); (2) the through-space interactions between the ligand and the substrate (ΔE_int-space); and (3) the through-bond interactions between the CuH moiety and the substrate (ΔE_int-bond). In the hydroamination reactions when SEGPHOS L1 and DTBM-SEGPHOS L2 ligands were employed, the ΔE_dist and ΔE_int-bond terms were not found to correlate with the overall activation energies (ΔE^‡); however, excellent linear correlations were observed with ΔE_int-space. This suggested that the t-butyl substituents at the 3- and 5-positions on the P-aryl₂ groups in DTBM-SEGPHOS L2 promote stabilizing non-covalent interactions (Figure 2b). Indeed, dissecting the ΔE_int-space term into its individual components revealed that attractive London dispersion forces (ΔE_disp) between the 3,5-di-t-butyl substituents on the P-aryl₂ groups and the substrate were the main contributing factor to achieve high catalyst activity with the DTBM-SEGPHOS L2 ligand. While the London dispersion interactions are relatively weak (0.5~1.5 kcal/mol for interactions with each t-Bu substituent),^14a,16 collectively they significantly reduce activation barriers via transition state stabilization.^17,18 Moreover, these conclusions were experimentally validated through ligand synthesis and subsequent kinetic analysis.^14a

Figure 2.

a) Ligand-substrate interaction model to study the origin of reactivity in hydrocupration. b) London dispersion interactions lowering the hydrocupration barrier for L2CuH.

Building upon this knowledge, we undertook the challenge of designing a new family of ligands based on SEGPHOS L1 to more efficiently facilitate the copper-hydride catalyzed anti-Markovnikov hydroamination reaction with terminal olefins. We theorized a more effective ligand system can be rationally designed by retaining the stabilizing dispersion effects of DTBM-SEGPHOS L2 while incorporating other types of stabilizing through-bond and/or through-space interactions.¹⁹ Specifically, we surmised that other types of weak non-covalent interactions with the olefin substrate,²⁰ may be harnessed by installation of hetero-atom-containing substituents on the P-aryl₂ groups. In addition, the through-bond stabilization between the CuH moiety and the substrate in the hydrocupration transition state can be fine-tuned by altering the electronic character of the ligands. However, when designing catalysts capable of promoting reactivity through an assortment of stabilizing interactions, infinite possibilities are conceivable. With the unique ability to computationally quantify and experimentally verify these interactions, an iterative catalyst design approach was envisioned (Figure 3).^21,22 This approach comprised of four stages: (1) experimentally identify a suitable class of ligand derivatives; (2) using computational analysis to understand what key interactions can stabilize the transition state; (3) using this knowledge to computationally predict a more effective ligand and (4) experimentally test the ligand providing feedback for the next round of ligand optimization.

Figure 3.

Project outline.

3.1. Kinetic and computational analysis of SEGPHOS Ligands.

3.1.1. Preliminary Experimental Investigations with Symmetric SEGPHOS Ligands.

As described above, previous investigations indicated that primarily bulky substituents at the 3 and 5-positions on the P-aryl₂ groups were critical in facilitating the hydrocupration event with terminal olefins.^14a This finding directed our preliminary studies to investigate SEGPHOS derivatives with substituents possessing different steric (TMS) and electronic (CF₃) properties at these positions (Scheme 1).²³ To kinetically quantify and compare the effects of these ligands on the hydrocupration event 4-phenyl-1-butene (1) and O-benzoyl-N,N-dibenzylhydroxylamine (2) were selected as model substrates, because a first order dependence on the olefin had been shown previously for the hydroamination reaction with DTBM-SEGPHOS L2.²⁴ The initial rates were measured for the reaction with each ligand by monitoring the formation of hydroamination product 3 under typical copper hydride hydroamination conditions (1.0 mol% Cu(OAc)₂, 1.1 mol% ligand, 0.36 M in THF, 23 °C) utilizing dimethyoxymethylsilane (DMMS, 3.0 equiv/1) as the stoichiometric reductant. To allow for a straightforward comparison, the rate of hydroamination was measured first with DTBM-SEGPHOS L2 (8.60 ± 0.05) x 10⁻⁶ M/s so that the rates could be normalized (Scheme 1).

Scheme 1.

Initial kinetic analysis for symmetric ligands L2, L3 and L4.

Following our standard protocol, the reaction employing TMS-SEGPHOS L3 was found to be 3.3 times faster than that with DTBM-SEGPHOS L2, suggesting that the larger TMS substituents, with a Taft value of E_s´ = 1.79, have stronger interactions with the olefin substrate than the t-Bu groups (E_s´ = 1.49) in DTBM-SEGPHOS L2 (Scheme 1). ²⁵ Interestingly, the hydroamination with the CF₃-SEGPHOS L4 derivative underwent the hydrocupration event 6.7 times faster than L2 and 2.0 times faster than L3 even though the CF₃ groups (E_s´ = 0.78) are smaller and presumably less polarizable.²⁶ This suggested that the fluorine-containing substituents have additional stabilizing effects that are stronger than simple London dispersion interactions as observed with L2 and L3. The origin for this significant and unexpected rate increase for L4 was revealed by computational investigations as detailed below.

3.1.2. Computational Analysis of the Origin of Reactivity with Symmetric SEGPHOS Derivatives.

The preliminary experimental studies revealed promising results with the CF₃-SEGPHOS L4 derivative. However, it was unclear what further modifications could lead to additional reactivity enhancement.^19,27 Although successful predictions of new transition metal catalysts from computational results alone are still rare,²⁸ several examples have recently been described wherein a combination of computational and experimental evaluations has led to the discovery of catalysts with improved reactivity and selectivity.^19,21 Such synergetic efforts effectively utilize the predictive power of computation, while the experimental verification helps resolve the uncertainty of calculated energies and issues that cannot be readily addressed by computations alone, such as catalyst decomposition.²⁹

3.1.2.1. Computational Methods

Geometry optimizations and single-point energy calculations were carried out using Gaussian 09.³⁰ Geometries of intermediates and transition states were optimized using the B3LYP functional³¹ with a mixed basis set of SDD for Cu and 6–31G(d) for other atoms in the gas phase. Vibrational frequency calculations were performed for all of the stationary points to confirm if each optimized structure is a local minimum or a transition state structure. Truhlar’s quasi-harmonic corrections³² were applied for entropy calculations using 100 cm⁻¹ as the frequency cut-off. Solvation energy corrections were calculated in THF solvent with the CPCM continuum solvation model³³ based on the gas-phase optimized geometries. The ωB97X-D functional³⁴ with a mixed basis set of SDD for Cu and 6–311+G(d,p) for other atoms was used for solvation single-point energy calculations. The computed gas-phase activation energy (ΔE^‡) was dissected using the following ligand-substrate interaction model analysis.³⁵

$$\Delta E^{‡}=\Delta E_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}}+\Delta E_{\text{int-bond}}+\Delta E_{\text{int-space}}$$ eq.1

The distortion energy (ΔE_dist)³⁵ is the sum of the energies required to distort the LCuH catalyst and the substrate into their transition state geometries. ΔE_int-space was calculated from the interaction energy of a supramolecular complex of the phosphine ligand and the olefin substrate at the transition state geometry but in the absence of the CuH moiety (ΔE_int-space = E_lig+sub − E_lig − E_sub). Then, the through bond interaction was calculated from ΔE_int-bond = ΔE^⧧ − ΔE_dist − ΔE_int-space. The ΔE_dist and ΔE_int-space were both calculated using the ωB97X-D functional with the SDD basis set for Cu and 6–311+G(d,p) for other atoms. The ωB97X-D functional was chosen because it has been shown to accurately describe non-covalent interactions,³⁶ which we expected to be important in this system. The computed free energy barriers using this method provided very good agreement with the experimental reaction rate constants (see SI for details and comparison with results from other functionals and solvation models). The through-space interaction energy (ΔE_int-space) between the ligand and the substrate is further dissected according to the following equation:

$$\Delta E_{\text{int-space}}=\Delta E_{\text{Pauli}}+\Delta E_{\text{elstat}}+\Delta E_{\text{pol}}+\Delta E_{\text{ct}}+\Delta E_{\text{disp}}$$ eq. 2

In accordance with our previous study, the dispersion energy component (ΔE_disp) was obtained from the difference of interaction energies calculated using MP2 and HF. The MP2 calculations were performed with Q-Chem 5.0 using the SOS(MI)-MP2 method in combination with the dual-basis set approach utilizing the db-cc-pVTZ basis set.³⁷ The ΔE_Pauli, ΔE_elstat, ΔE_disp, ΔE_pol, and ΔE_ct terms in eq 2 were calculated using the second-generation energy decomposition analysis based on absolutely localized molecular orbitals³⁸ (ALMO-EDA) method implemented in Q-Chem 5.0.³⁹ The second generation ALMO-EDA provides further decomposition of the Pauli and electrostatic interaction (ΔE_rep) term into Pauli repulsion (ΔE_Pauli) and electrostatic (ΔE_elstat) energies, which is important in the analysis of through-space electrostatic interactions with the fluorinated ligands. To avoid double counting of dispersion, HF method with the 6–311G(d,p) basis set was employed in the energy decomposition analysis (EDA) calculations.

3.1.2.2. Computational Analysis of Symmetric-SEGPHOS Ligands.

In order to fully understand the underlying principles and interactions that lead to the enhanced rate, an in-depth computational analysis was performed to study the origin of the different hydroamination reactivities between the DTBM-SEGPHOS L2 and CF₃-SEGPHOS L4-supported CuH catalysts.

The activation energies of the rate-determining hydrocupration transition states were computed using propene (4) as the model substrate with the method outlined above (Table 1). The computed barrier of hydrocupration with the CF₃-SEGPHOS L4CuH complex was in good agreement with the experimentally observed rate increase with L4 compared to DTBM-SEGPHOS L2CuH (ΔΔG^‡_comp = 1.5 kcal/mol vs ΔΔG^‡_exp = 1.1 kcal/mol). In order to quantify the different factors that lead to the improved reactivity, the ligand-substrate interaction model analysis was employed to dissect the overall hydrocupration activation energies (Eqs. 1 and 2, see Computational Methods for details). Energy-decomposition analysis of the hydrocupration transition state with L4CuH revealed that the increase in the reaction rate was due to significantly stronger through-bond interactions (ΔE_int-bond) resulting in an extra 2.3 kcal/mol stabilization of TS-4 compared to the DTBM-SEGPHOS-bound TS-2. This is because of the electron-withdrawing nature of the CF₃-substitutents which consequently results in enhanced Lewis acidity of the CuH catalyst and more favorable binding of the olefin substrate (see SI for details). While the through-space interaction energies (ΔE_int-space) are comparable in TS-2 and TS-4, the origins are different. Using the second-generation ALMO-EDA methods, the ΔE_int-space term was further dissected into its individual energy components (Eq. 2). While TS-2 is stabilized by stronger attractive London dispersion (ΔE_disp = –13.3 kcal/mol for TS-2 compared to –10.7 kcal/mol for TS-4), electrostatic interactions are more favorable in TS-4 (ΔE_elstat = 0.3 kcal/mol for TS-2 compared to –1.5 kcal/mol for TS-4). The optimized geometry of TS-4 revealed multiple C–F···H–C contacts, which are responsible for the through-space electrostatic interactions between L4 and the olefin substrate thereby lowering ΔE^‡ (Figure 4).

Table 1.

Activation free energies of the hydrocupration transition states and energy components derived from the ligand-substrate interaction model. ^a

ligand	DTBM (L2)	CF3 (L4)	i-C3F7 (L5)	DTBM-i-C3F7 (L6)
hydrocupration transition state	TS-2	TS-4	TS-5	TS-6
ΔG‡solv	20.2	18.7	17.2	17.0
ΔE‡	−0.1	−1.0	−3.4	−3.0
distortion (ΔEdist)	28.6	29.5	28.8	27.9
through-bond interaction (ΔEint-bond)	−23.9	−26.2	−26.2	−25.4
through-space interaction (ΔEint-space)	−4.8	−4.3	−6.0	−5.6
Pauli repulsion (ΔEPauli)	9.0	8.4	7.9	7.8
electrostatic (ΔEelstat)	0.3	−1.5	−1.2	−0.3
London dispersion (ΔEdisp)	−13.3	−10.7	−11.9	−13.0
charge transfer (ΔEct)	−0.2	−0.2	−0.1	0.0
polarization (ΔEpol)	−0.6	−0.4	−0.4	−0.4
ΔΔG‡comp	0.0	−1.5	−3.0	−3.2
ΔΔG‡exp	0.0	−1.1	−2.4	−2.4

^aAll energies are reported in kcal/mol. The activation energies (ΔG^‡_solv and ΔE^‡) are with respect to the separated CuH catalyst and propene (4). ΔΔG^‡_comp values were calculated by subtracting ΔG^‡_solv-L2 from ΔG^‡_solv-LX. ΔΔG^‡_exp were derived from the experimental relative rate constants (k_rel).

Figure 4.

Optimized geometries of hydrocupration transition states with the DTBM-SEGPHOS (TS-2) and CF₃-SEGPHOS ligands (TS-4). Distances are in Ångström [Å].

Although the use of CF₃-SEGPHOS L4 leads to a relatively moderate increase of reactivity, the computational analysis suggested types of modifications that might result in a more effective ligand. Considering that the CF₃-SEGPHOS L4 ligated LCuH complex has weakened dispersion interactions when compared to the L2CuH complex, we hypothesized that the installation of a larger perfluorinated substituent would be beneficial. Since the i-C₃F₇ group is sterically more demanding than CF₃, we assumed that it should increase stabilizing London dispersion, while maintaining the favorable through-space electrostatic attractions and through-bond electronic effects.

Indeed, the calculated hydrocupration transition state TS-5 indicated that the use of i-C₃F₇-SEGPHOS L5 as the ligand led to an additional 1.5 kcal/mol lower activation energy compared to the hydrocupration with L4CuH (Table 1).

The ligand-substrate interaction model analysis validated our hypothesis, as the ΔE_dist and ΔE_int-bond terms of TS-5 remained largely unchanged when compared to TS-4. Meanwhile, the through-space interaction of TS-5 was 1.7 kcal/mol more stabilizing. Further dissection of the through-space interactions revealed that the primary reason for the increased reactivity was due to the increased London-dispersion interactions (ΔE_disp) in TS-5. To validate this computational prediction, we needed to experimentally measure the reactivity of i-C₃F₇-SEGPHOS-supported CuH catalyst L5CuH.

3.1.4. Synthesis and Kinetic Analysis of Hydroamination with the Symmetric i-C₃F₇-SEGPHOS Ligand.

Informed by the computational predictions described above, we set out to synthesize ligand L5. Adopting a closely related report by Yu, we were able to prepare L5 from dibromide 5 and bis-(3,5-i-C₃F₇-C₆H₃)₂PBr (6) in a single step (Scheme 2a). ⁴⁰

Scheme 2.

Synthesis of SEGPHOS derivatives L5 and L6 (see supplementary information for details).

When L5 was employed with the standard catalytic conditions, vide supra, the formation of hydroamination product 3 was observed to be 61 times faster than with L2, indicating that increased London dispersion interactions were indeed facilitating the hydrocupration event. However, only a short burst of reactivity was observed under the reaction conditions employing L5. This suggests that the L5CuH complex, although an active catalyst, was not stable under the reaction conditions (Figure 6, red curve).⁴¹ This catalyst decomposition is most likely the consequence of the diminished Lewis basicity of the phosphorus atoms in L5, due to the electron-withdrawing nature of the i-C₃F₇ substituents which results in weaker binding to the copper center. In order to exhibit both high reactivity and stability, the Lewis acidity of the copper center needed to be finely tuned.

Figure 6.

Combined data for the formation of amination product 3 (see, supplementary information for details).

3.1.5. Hybrid-SEGPHOS Ligands

To harness the increased reactivity that we observed using the i-C₃F₇ substituents without sacrificing the stability of the resulting complex, we had two options: either to synthesize and test various new derivatives with different substituents, in order to find a suitable ligand that provides a catalyst system that combines high activity and stability, or exchange one P-aryl₂ substituent for a more electron-donating group in order to stabilize the resulting copper complex. To avoid significant structural changes at the 3- and 5-positions of the aryl groups, we reasoned that the merger of DTBM-L2 and i-C₃F₇-L5, the ligands with higher catalyst stability and reactivity, might result in the perfect balance of their respective beneficial interactions. This hypothesis found further support in examining the transition-state structure TS-5, in which the improved through-space ligand-substrate interactions primarily arise from the C–F···H–C interactions in the 1^st and 4^th quadrants (Figure 5). The i-C₃F₇ groups in the 2^nd and 3^rd quadrants are further away from the substrate, and thus are less significant in promoting the hydrocupration step. Therefore, exchanging the P-aryl₂ groups in the 2^nd and 3^rd quadrants was not expected to significantly impact the enhanced reactivity gained from the i-C₃F₇ moieties.

Figure 5.

Optimized geometries of hydrocupration transition states with the i-C₃F₇-SEGPHOS (TS-5) and the hybrid DTBM-i-C₃F₇-SEGPHOS ligand (TS-6 and TS-6a). Distances are reported in Ångström [Å].

3.1.5.1. Computational Studies of Hydrocupration with Hybrid-SEGPHOS Ligands

The computational investigations showed that the hydrocupration barrier for the hybrid SEGPHOS derivative L6CuH was similar to that of the symmetric derivative L5CuH (see Table 1). In the lowest energy transition state structure with L6 (TS-6, Figure 5), the methyl group on propene (4) prefers to be placed in the i-C₃F₇-occupied 1^st quadrant, rather than the DTBM-occupied 3^rd quadrant (TS-6a, Figure 5), indicating the C–F···H–C non-covalent interactions with the i-C₃F₇ group are more favorable than the C–H···H–C interactions with the t-Bu group. Further energy decomposition analysis showed similar through-space interaction energies (ΔE_int-space) in TS-6 and TS-5 (Table 1). While electrostatic interactions in TS-6 were slightly decreased relative to those in TS-5, London dispersion interactions were increased as a result of the larger t-butyl substituents in the 2^nd and 3^rd quadrants of TS-6. This finding indicated, that a comparable energy barrier of hydrocupration might be obtained from L6CuH.

3.1.5.2. Synthesis of Hybrid-SEGPHOS Ligands

To verify our hypothesis, a practical synthetic sequence had to be developed to prepare this hybrid ligand. After extensive experimental effort, a modular three-step sequence was established (Scheme 2b). The installation of the bis-(3,5-CF₃-C₆H₃)₂P subunit was achieved by trapping mono-magnesiated 5 with freshly prepared bis-(3,5-i-C₃F₇-C₆H₃)₂PBr (6). After, the introduction of the DTBM-P(O) moiety, via a palladium catalyzed cross-coupling reaction with DTBM phosphine oxide 7, and subsequent reduction, L6 was obtained in 26% yield over 3 steps.⁴²

3.1.5.3. Kinetic Analysis of Hybrid-SEGPHOS Ligands

Following our standard kinetic protocol, L6 was employed with our usual catalytic conditions and the formation of hydroamination product 3 was found to be 62 times faster than that when using L2, indicating that the rate enhancement observed with the symmetric L5CuH complex was maintained (Figure 6, black curve). We also noted that no detectable catalyst decomposition was observed with L6 under the reaction conditions, validating our hypothesis that the hybrid system could maintain stability without sacrificing reactivity.³⁸

3.1.6. Demonstration of Hybrid-SEGPHOS Ligand L6 under Preparative Conditions.

In order for this newly developed ligand to be useful in a synthetic context, the observed rate increase would need to be maintained at preparatively relevant scales and on substrates bearing functional groups. After slight optimization of the reaction conditions, the scope of olefins was established using hydroxylamine ester 2 as the amine source (Table 2). The hydroamination of terminal olefins that contained various functional groups were surveyed at room temperature. Epoxide 8, ester 9, silyl ether 10, and ketal 11 all provided the desired tertiary amine product in excellent yield. Moreover, substrates that contained a variety of heterocycles, such as piperazine 12, morpholine 13, and thiophene 14 also underwent smooth hydroamination at room temperature. Stronger Lewis bases found in heterocyclic compounds like indole 15, benzothiazole 16, pyrimidine 17, and in quinolines 18 and 19 slightly inhibited the reaction, and thus their reactions required slightly elevated temperatures (40 ºC) to reach full conversion within 3 hours. To compare our new catalyst to the current state-of-the-art catalyst, L2, epoxide 8 and ester 9 were subjected to these reaction conditions employing DTBM-L2 as the ligand.⁴³ Diminished yields of 17 and 29% were observed compared to 89 and 94% with L6, respectively. This demonstrates that the rate enhancement using this catalyst system is maintained under preparative reaction conditions.

CONCLUSION

This study demonstrates how the combination of mechanistic insights, computational prediction, and experimental verification can successfully benefit ligand development. Using this synergistic approach we were able to discover a new hybrid ligand L6 that is capable of promoting the anti-Markovnikov hydroamination of unactivated, terminal olefins with a 62 fold rate increase compared to DTBM-SEGPHOS L2. By employing energy decomposition analysis methods, we were able to deconvolute each individual energy contributions of the steric, electronic, and dispersion effects that comprise the hydrocupration barrier. During the course of our investigation we identified that in addition to London dispersion, both electrostatic C–F···H–C non-covalent interactions and inductive effects of the i-C₃F₇ substituents are capable of lowering the energy barrier for hydrocupration even further. Ultimately, the merger of both DTBM and i-C₃F₇ substituents was key to success in designing L6 with balanced stability and reactivity. Furthermore, a modular and robust synthetic sequence to access these novel hybrid ligand structures was devised, that allowed for its gram-scale synthesis. In addition, the effectiveness of the catalyst system employing L6 was proven under preparative conditions. We anticipate that this rational ligand design approach can be utilized in other catalytic systems providing accelerated reaction development.

Scheme 3.

Isolated yields are reported as the average of two runs. Standard reaction conditions: terminal olefin (0.50 mmol), Bn₂NOBz (2) (0.60 mmol), Cu(OAc)₂ (2.50 mol%), L6 (2.55 mol%), DMMS (1.50 mmol), THF (1.0 mL), 23 °C, 3h. ^a40 °C. ^b DTBM-SEGPHOS was used in place of L6 and NMR yields are provided.