Evaluation of Pd→B Interactions in Diphosphinoborane Complexes and Impact on Inner‐Sphere Reductive Elimination

Abstract

The dative Pd→B interaction in a series of ^RDPB^R’ Pd⁰ and Pd^II complexes (^RDPB^R’=(o‐PR₂C₆H₄)₂BR’, diphosphinoborane) was analyzed using XRD, ¹¹B NMR spectroscopy and NBO/NLMO calculations. The borane acceptor discriminates between the oxidation state Pd^II and Pd⁰, stabilizing the latter. Reaction of lithium amides with [(^RDPB^R’)Pd^II(4‐NO₂C₆H₄)I] chemoselectively yields the C−N coupling product. DFT modelling indicates no significant impact of Pd^II→B coordination on the inner‐sphere reductive elimination rate.

Ligand effects: The borane acceptor in diphosphinoborane ligands discriminates between the oxidation state Pd^II and Pd⁰, stabilizing the latter. The impact of the Pd→B interaction on inner‐sphere C−N bond reductive elimination of N,N‐dimethyl‐4‐nitroaniline was investigated (see scheme).

Introduction

Z‐type acceptor ligands have attracted considerable attention over the past decade.1 Their coordination to transition metals grants access to complexes with unusual coordination geometries2 and electronic properties by formation of dative M→Z bonds. Group 13 acceptor ligands, with a special focus on boranes, have been particularly well studied. M→Z bonds can stabilize low oxidation states at the coordinated transition metal.3 Thus, facile access to complexes featuring transition metals with formally negative oxidations states is realized (Figure 1 a).4 This stabilization of low oxidation states appears to inhibit oxidative addition reactions.3b, 3e, 5 However, we demonstrated that this obstacle can be overcome for complex 1 by addition of catalytic amounts of acetate, which competes with Pd⁰ for the free coordination site at the borane, thus reversibly breaking the Pd⁰→B interaction (Figure 1 b).3b This concept allowed for the application of 1 in catalytic allylic amination, and most recently of 2 in the catalytic hydro‐/deutero‐dechlorination of aryl chlorides.3e Alternatively, bifunctional substrate activation across the M→Z interaction has been described.3a, 6 The aptitude of hydride,7 halide8 and carbon group9 migration between the Z‐type ligand and the coordinated transition metal has initiated further applications. Catalytic processes have concentrated on transformations in which the catalyst is not required to change its oxidation state quickly, but rather profits from an electronic fine‐tuning by electron‐withdrawing Z‐ligand coordination.10 Successful applications include CO₂ hydrogenation11 and hydrosilylation,3d, 12 enyne cycloisomerization13 and alkyne hydroamination.14 Michaelis used the heterobimetallic Ti^IV/Pd^II complex (Figure 1 c), developed by Nagashima,15 for allylic amination of allyl chlorides with hindered secondary amines.5b, 16

Figure 1.

M→Z interaction: stabilization of low oxidation states and impact on oxidative addition and reductive elimination.

Combined experimental and computational investigations indicated a rate enhancement of 10³⁻–10⁵ of the outer‐sphere reductive C−N bond elimination, due to the electron‐withdrawing Pd^II→Ti^IV interaction.5b, 17 This result agrees with previous investigations performed with Pd η³‐allyl and Ni η³‐allyl complexes, which showed favored reductive outer‐sphere reductive elimination in the presence of less electron‐donating spectator ligands.18

We speculated that the electron‐withdrawing properties of the borane functionality in diphosphinoborane (DPB) ligands enhances the rate of inner‐sphere reductive elimination from Pd complexes due to 1) overall reduced electron density at the Pd^II center and 2) increasing of the Pd→B interaction strength during reductive elimination. We determine how the oxidation state of Pd and co‐ligands affect the strength of the Pd→B interaction in DPB complexes. NBO/NLMO calculations and solid‐state structures are used to assess the strength of Pd→B interactions. The value of the ¹¹B NMR chemical shift as a probe is discussed. The reductive elimination of N,N‐dimethyl‐4‐nitroaniline from [(^PhDPB^Ph)Pd^II(4‐NO₂‐C₆H₄)NMe₂] (5) was studied and modelled with DFT calculations to investigate the assumed influence of the borane acceptor.

Results and Discussion

Syntheses and reactivity of [(DPB)Pd] complexes

A series of [(^PhDPB^Ph)Pd^II] complexes was synthesized to examine a possible correlation between the nature of ligands at Pd and the strength of the Pd^II→B interaction (Scheme 1).

Scheme 1.

Synthesis of [(^PhDPB^Ph)Pd^II] complexes.

Complex [(^PhDPB^Ph)Pd^IICl₂] (7) was produced by reaction of ^PhDPB^Ph ligand with [(cod)PdCl₂] in DCM and was isolated in 74 % yield (Scheme 1). Single crystals were grown from CH₂Cl₂/benzene and analyzed by X‐ray diffraction (Figure 2). A typical square‐pyramidal coordination around the palladium was observed around the Pd^II center. The chloride ligands are located in cis‐configuration at the basal position, and the borane adopts the apical position. The Pd,B distance of 2.762(3) Å is shorter than the sum of the van der Waals radii (3.28 Å),19 but elongated compared to the sum of the covalent radii (2.23 Å).20 A long Pd,C51 distance of 3.405(3) Å seems to rule out a η²‐(B,C) type coordination to the Pd^II center. A slightly increased pyramidalization at the boron atom is observed (ΣB_α=355.4°) compared to complex [(^iPrDPB^Ph)PdCl₂] (ΣB_α=359.9°).21

Figure 2.

Left: thermal ellipsoid plot of the solid‐state structure of 7 at the 50 % probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (°): Pd1−Cl1=2.3355(7), Pd1−Cl2=2.3628(7), Pd1−P1=2.2558(8), Pd1−P2=2.2932(8), Pd1−B1=2.762(3), Pd1−C51=3.405(3), P1‐Pd1‐P2=95.49(3), C51‐B1‐C61=118.3(3), C51‐B1‐C71=118.2(3), C71‐B1‐C61=118.8(3).22 Middle: Ball and stick display of [(^PhDPB^Ph)PdCl]‐dimer (9) generated by symmetry. Right: thermal ellipsoid plot of the asymmetric unit of 9 at the 50 % probability level. Hydrogen atoms and crystal CH₂Cl₂ are omitted for clarity. Selected bond lengths (Å) and angles (°): Pd1−Cl1=2.3781(11), Pd1−Cl1^†=2.3928(13), Pd1−P1=2.2638(13), Pd1−P2=2.3084(11), Pd1−B1=2.721(5), Pd1−C1=3.338(4), P1‐Pd1‐P2=95.38(5), C11‐B1‐C41=117.5(4), C1‐B1‐C11=119.4(4), C1‐B1‐C41=118.9(4).23.

The ligand backbone is twisted (dihedral angle C62‐C61‐C71‐C72: 35.6(3)°) to allow for a P‐Pd‐P angle of 95.49(3)°. This twist renders the two phosphine groups diastereotopic. The ³¹P NMR spectrum of 7 in CD₂Cl₂ displays two broad resonances of equal integral at δ=39.0 and 48.2 ppm. A series of ³¹P VT NMR spectra was recorded (Figure 3), covering a temperature range from −29.8 to 35.1 °C. The two singlet resonances coalesced into a single resonance (δ=48.2 ppm) at elevated temperatures. The rate constants of the dynamic process were determined by line‐shape analysis using Bruker's TopSpin software. An Arrhenius plot analysis gave an activation energy of E_a=9.3±0.5 kcal mol⁻¹ with a pre‐exponential factor of A=(14±7) x 10⁹.

Figure 3.

³¹P VT NMR analysis of 7 in CD₂Cl₂. Left: recorded ³¹P NMR spectra. Middle: simulated ³¹P NMR spectra. Right: Arrhenius plot.

We suggest that the observed dynamic process in the ³¹P NMR spectrum of 7 is caused by an interconversion of 7 with its enantiomer ent ‐7 (Scheme 2).

Scheme 2.

Proposed interconversion between 7 and ent ‐7 by twisting of the DPB ligand.

In order to accommodate for the small P‐Pd‐P angle of 95.49(3)°, the σ‐symmetric ^PhDPB^Ph ligand is twisted. As a result, its B−Ph group points towards one of the two phosphine groups, rendering them chemically inequivalent. This assumption is in line with the observed two ³¹P NMR resonances at low temperatures. Twisting of the C62‐C61‐C71‐C72 dihedral angle converts 7 into its enantiomer ent ‐7, presumably via a σ‐symmetric transition in which the B−Ph group is orientated between the two chloro ligands.

Complex 8 was synthesized in the same fashion as 7 from [(cod)PdBr₂] and was isolated in 67 % yield. The ³¹P NMR spectrum displays two broad resonances of equal intensity at δ=45.2 and 38.1 ppm (CD₂Cl₂), suggesting a similar dynamic process as in 7. Due to the poor solubility of both 7 and 8, no ¹¹B NMR spectra could be obtained.

Cationic complex [(^PhDPB^Ph)Pd^IICl]SbF₆ (9) was produced in 51 % isolated yield by halide abstraction from 7 with AgSbF₆ (Scheme 1). Single crystals were grown from CH₂Cl₂/hexane and analyzed by X‐ray diffraction (Figure 2). In the solid state a chloro‐bridged dimer [(^PhDPB^Ph)Pd^II(μ‐Cl)]₂(SbF₆)₂ is observed with an inversion center between the two Pd^II centers. Within the dimer, the Pd^II center is coordinated in a square‐pyramidal fashion with the borane located in the apical position. The Pd, B distance in complex 9 is 2.721(5) Å, which is slightly shorter than in [(^PhDPB^Ph)Pd^IICl₂] 7 (2.762(3) Å). However, pyramidalization of the borane is almost identical (ΣB_α=355.8°). The absence of a relevant η²(B,C)→Pd^II interaction is suggested by the long Pd1,C1 distance of 3.338(4) Å. The Pd,B distance and lack of significant pyramidalization at the borane suggest a weak Pd^II→B interaction, which is in line with a broad resonance in the ¹¹B NMR spectrum at δ=65 ppm (ω _1/2=1900±500 Hz).

The ligand backbone is twisted similarly to that in 7 (dihedral angle C42‐C41‐C11‐C12 of 33.5(5)° (9) vs. 35.6(3)° in 7), resulting in an almost parallel orientation of the B−Ph with the Pd1−Cl1 bond (dihedral angle C1‐B1‐Pd1‐Cl1 of 10.6(3)°). The ³¹P NMR spectrum of 9 displayed only a singlet resonance at δ=49.9 ppm which suggests a quick interconversion between the two diastereotopic phosphine donors in solution.

Cationic allyl complex [(^PhDPB^Ph)Pd^II(η³‐C₃H₅)]SbF₆ (10) was synthesized by reaction of AgSbF₆ with zwitterionic allyl complex [{(o‐PPh₂C₆H₄)₂B(OAc)Ph}Pd^II(C₃H₅)] (4) (Scheme 1) and was isolated in 38 % yield by crystallization from CH₂Cl₂/hexane. Figure 4 depicts its solid‐state structure. The Pd^II center in complex 10 is located in a trigonal‐pyramidal environment in which the borane occupies the pseudo‐apical position and the C₃H₅‐ligand and the two phosphines are located in the trigonal‐planar positions. A weak Pd^II→B interaction is indicated by a Pd,B distance of 2.676(5) Å, which is in line with a minor pyramidalization at the borane center (ΣB_α=354.7°) and a broad ¹¹B NMR resonance at δ=62 ppm (ω _1/2=1200±100 Hz). A large Pd,C22 distance of 3.066(6) Å eliminates the possibility of a strong η²(B,C)→Pd^II interaction. The η³‐coordinated C₃H₅‐ligand is disordered. Using the borane as a reference point, a 39:61 mixture of the exo‐ and endo‐isomers is observed. A wider P‐Pd‐P angle of 102.86(5)° is realized by a decrease in the twisting of the ligand backbone (dihedral angle C18‐C17‐C28‐C33 of 24.04°). The observed disorder of the C₃H₅‐ligand is in good agreement with the observed NMR spectra. In the ³¹P NMR spectrum (CD₂Cl₂), two singlet resonances are observed in a 40:60 ratio (δ=28.1 and 26.9 ppm) and two sets of C₃H₅‐units are detected in the ¹H NMR spectrum. DFT calculations (BP86/def‐SV(P)) based on the solid‐state structures of 10‐endo and 10‐exo indicate a small Gibbs free energy preference of ΔG=0.74 kcal mol⁻¹ for 10‐endo, predicting a 29:71 ratio at 298 K.

Figure 4.

Thermal ellipsoid plot of the solid‐state structure of 10 at the 50 % probability level. Hydrogen atoms and one molecule of CH₂Cl₂ are omitted for clarity. Selected bond lengths (Å) and angles (°): Pd1−B1=2.676(5), Pd1−C22=3.066(6), Pd1−P1=2.304(1), Pd1−P2=2.340(1), Pd1−C1=2.191(5), Pd1−C2a=2.186(12), Pd1−C2b=2.192(7), Pd1−C3=2.201(4), P1‐Pd1‐P2=102.86(5), P1‐Pd1‐B1=82.1(1), P2‐Pd1‐B1=75.1(1).24.

To explore the potential influence of the Pd^II→B interaction on reductive elimination proceeding via an inner‐sphere mechanism, complex [(^PhDPB^Ph)Pd^II(4‐NO₂‐C₆H₄)I] (5) was reacted with lithium amides. Complex 5 was reacted with LiNMe₂ (1.1 equiv) at room temperature in [D₈]THF (Scheme 3).25

Scheme 3.

Reductive elimination from 5 and independent synthesis of 11.

A conversion of 84 % was observed ³¹P NMR spectroscopically after 1 h. Two complexes were formed with singlet resonances at δ=31.1 (70 %) and 38.3 ppm (14 %). After a total of 4.5 h, all resonances in the ³¹P NMR spectrum disappeared in favor of the singlet at δ=31.1 ppm. ¹¹B NMR spectroscopy suggested formation of a zero‐valent palladium complex by a broad resonance at δ=19 ppm (ω _1/2=400±100 Hz). The concurrent formation of the expected reductive elimination product N,N‐dimethyl‐4‐nitroaniline was confirmed by GC/MS analysis, using an independently prepared sample as a reference. The absence of an intermediate complex cis‐[(^PhDPB^Ph)Pd^II(4‐NO₂‐C₆H₄)NMe₂] suggests that transmetalation is rate‐limiting in this transformation. The intermediate occurrence of the ³¹P NMR resonance at δ=38.3 ppm is possibly due to a reversible reaction of LiNMe₂ with complex 6. In a control experiment complex [(^PhDPB^Ph)Pd⁰(pyridine)] (1) was reacted with LiNCy₂ and LiNMe₂ in [D₈]THF. In both cases ca. 7 % of a new complex at δ=38.5 (s) and 37.7 ppm (s) were observed.

Complex 6 decomposed within hours with simultaneous precipitation of palladium black. Addition of PMe₃ as a stabilizing co‐ligand led to the formation of complex [(^PhDPB^Ph)Pd⁰(PMe₃)] 11. The ³¹P NMR spectrum of 11 showed a doublet at δ=35.3 and a triplet at −40.1 ppm (J=15.1 Hz) in a 2:1 ratio, which is consistent with the expected κ³P‐coordination. The broad resonance in the ¹¹B NMR spectrum at δ=25 ppm (ω _1/2=400±100 Hz) suggested a strong Pd⁰→B interaction. Complex 11 could also be synthesized independently by reaction of PBP pincer 12 with PhLi and PMe₃, or reaction of 1 with PMe₃, thus confirming unambiguously the identity of 11 (Scheme 3).

Complex 5 reacted in a similar fashion with LiNCy₂ (26 % 6 after 3 h) and LiNHtBu (14 % 6 after 5.5 h). However, the reaction proceeded slower with these sterically more demanding substrates. The reaction of complex 5 with LiNHtBu was monitored for 96 h by ³¹P NMR spectroscopy (46 % conversion towards 6) without any side products being observed (cf. Table S1). This is in line with the assumption of a rate‐determining transmetalation followed by a quick reductive elimination.

Analyses of Pd→B interactions

The solid‐state structures of Pd^0/II DPB complexes were analyzed to identify factors which affect the strength of Pd→B interactions. In addition to the new Pd complexes presented in this work (6–10), the structurally characterized DPB complexes cis‐[(^PhDPB^Ph)Pd^II(4‐NO₂‐C₆H₄)I] (5),9d [(^PhDPB^Ph)Pd⁰(pyridine)] (1),3b [(^PhDPB^Me)Pd⁰(PMe₃)] (13)9d and [(^CyDPB^Ph)Pd⁰] (3)3c (Figure 4) were included to cover a broad range of B‐/P‐substituents and co‐ligands at the Pd^0/II center. The shorter Pd,B distances and higher degree of borane pyramidalization (Table 1) confirm a significantly stronger Pd,B interaction in Pd⁰ complexes, than in Pd^II complexes. Surprisingly, within a given oxidation state only a very moderate variation of the Pd→B bond strength is observed, regardless of substituents at the borane and phosphines, or the number and nature of co‐ligands (Pd⁰: ΣB_α=338–346°, d(Pd⁰,B)=2.194(3) −2.243(2) Å vs. Pd^II: ΣB_α=354–356°, d(Pd^II,B)=2.676(5) −2.762(2) Å). Remarkably, even the generation of cationic Pd^II complexes (9 and 10) has no significant impact on the strength of Pd^II→B interactions. The oxidation state at Pd is unambiguously the dominant factor for the strength of the Pd,B bond.

Table 1.

Experimental and computational analysis of the Pd→B interactions.^[a]

	7	8	9[e]	10‐endo	5	1	13	3	6
d(Pd,B) [Å] (XRD/DFT)	2.762(3) 2.740	– −2.654	2.721(5) 2.554	2.676(5) 2.731	2.7402(4) 2.781	2.194(3) 2.193	2.278(3) 2.360	2.243(2) 2.264	– −2.253
(Pd,Cipso) [Å] (XRD/DFT)	3.405(3) 3.256	– −3.292	3.338(4) 3.112	3.066(6) 3.259	3.346(4) 3.440	2.463(3) 2.865	2.815(2) 2.685	3.079(2) 3.054	– −2.768
ΣBα [°] (xrd/dft)	355/355	–/352	356/355	355/355	354/351	346/346	338/341	341/343	–/349
11B NMR (δ, ω 1/2)	–	–	65 ppm 1900 Hz	67 ppm 1400 Hz	63 ppm 3000 Hz	20 ppm 400 Hz	25 ppm 500 Hz	22 ppm 800 Hz	19 ppm 400 Hz
E 2(Pd,B)[b] [kcal/mol]	11.46	10.42	11.41	8.04	8.72	23.46	19.53	46.83	42.12
NLMO %B[c]/Pd[c]	6.6/91.9	6.3/92.2	5.4/92.9	3.7/93.9	4.7/93.4	16.0/78.7	15.0/81.5	15.5/81.7	14.3/83.0
occ. B[d]	0.391	0.387	0.400	0.360	0.353	0.618	0.621	0.498	0.519
occ. Pd[d]	1.859	1.865	1.870	1.887	1.879	1.666	1.702	1.686	1.704
B‐hybrid % (s/p)	7.6/92.4	7.2/2.7	7.2/92.7	6.7/93.3	6.4/93.6	11.6/88.4	13.9/86.1	12.8/87.2	10.7/89.3
WBI (Pd,B)	0.2164	0.2063	0.2119	0.1738	0.1801	0.4207	0.3634	0.5032	0.4604
WBI (Pd,Cipso)	0.0079	0.0079	0.0208	0.0093	0.0062	0.0697	0.0171	0.0103	0.0325

[a] Structure optimization: Turbomole 7.0.1, BP86/def‐SV(P); NBO analysis: Gaussian 09/NBO 6.0, BP86/6‐31G(d), MWB10 (P,Cl), MWB28 (Pd, Br), MWB46 (I). [b] NBO stabilizing energy E₂ associated with the Pd→B interaction. [c] Contribution of the donor/acceptor NBO to the NLMO. [d] Occupancy of the donor/acceptor NBO. [e] Calculated structure parameters of 9 are based on the monomer.

The Pd→B interactions were further analyzed using QM calculations. Complexes 1, 3, 5–11 and 13 were geometrically optimized using Turbomole 7.0.1 (BP86/def‐SV(P)). A good agreement was observed between the optimized structures and their corresponding solid‐state structures (Table 1). Complexes 6 and 8 were constructed based on the solid‐state structure of complexes 1 and 7. The Pd→B interactions were further analyzed using NBO/NLMO calculations. In all cases, an NBO donor/acceptor interaction was found between an occupied d‐orbital at Pd and an unoccupied p‐orbital at B (Figure 5). For all examined complexes no relevant η²(B,C)‐coordination was found in the NBO calculations. The Wiberg bond index for Pd,C_ipso was below 0.02, with the exception of Pd⁰ complexes 1 (0.0697) and 6 (0.0325). Reactivity studies of [(DPB)Pd]‐complexes presented in this paper thus appear to be unaffected from significant η²(B,C)‐coordination.

Figure 5.

Graphical representation of the NLMOs associated with the Pd→B interactions in [(^PhDPB^Ph)Pd(0/II)] complexes.

The NBO stabilizing energy of this Pd→B interaction varied depending on the Pd oxidation state. For Pd^II→B interactions, a narrow range of NBO stabilizing energies between 8.04 and 11.46 kcal mol⁻¹ was observed. Surprisingly, generation of cationic complexes (9, 10‐endo), exchange of chloro‐ligands by bromide (8) or iodide/aryl (5) had very little effect. In the case of Pd⁰→B interactions, significantly higher NBO stabilizing energies of 19.53–46.83 kcal mol⁻¹ were found. Regardless of the oxidation state at Pd an approximately linear correlation between the Pd,B distance and the NBO stabilizing energy (E ₂) associated with the Pd,B interaction was observed (Figure 6) for 16 valence electron (VE) complexes 1, 5, 7, 8, 10 and 13. The Pd,B distance appears to be dictated by the Pd,B bond strength, and not by constraints imposed by the chelating ligand. Substitution of PPh₂‐groups (6) by PCy₂‐groups (3) had only a minor effect. The E ₂ values for the Pd⁰→B interaction in the 14 VE complexes 3 (46.83 kcal mol⁻¹) and 6 (42.12 kcal mol⁻¹) significantly deviate from this correlation and are almost twice as much as for 16 VE complexes 1 (23.46 kcal mol⁻¹) and 13 (19.53 kcal mol⁻¹). Neither the ¹¹B NMR chemical shift, Pd,B distance or pyramidalization at B indicate a change of the Pd⁰→B interaction strength in this magnitude between the 14 VE and the 16 VE complexes (Table 1). This discrepancy might be explained by the difficulty to compare the 2^nd order perturbation interaction energies from NBO analysis from 14 VE with 16 VE complexes.

Figure 6.

Left: correlation between solid state Pd,B distances and δ(¹¹B). Right: correlation between calculated Pd,B distances and NBO stabilizing energies.

The ¹¹B NMR resonances are shifted linearly towards higher field with an increasing Pd,B distance for Pd⁰ complexes, regardless of the valence electron count at the Pd center (Figure 6). Complex [(^PhDPB^Ph)Pd⁰(PPh₃)] (2) reported by Kameo and Bourissou3e also fits perfectly into this correlation (d(Pd,B)=2.294(2) Å, δ(¹¹B) 27 ppm). In contrast, the ¹¹B NMR resonance shifts linearly towards lower field with an increasing Pd,B distance in case of Pd^II complexes. ¹¹B NMR spectroscopy therefore can be used as a tool to assess the strength of Pd→B interactions within a given ligand system, provided that the oxidation state at the Pd center is taken into account. However, given the difficulty to determine the precise δ(¹¹B) of [(DPB)Pd^II] complexes (poor solubility and ω _1/2 >1000 Hz ), a certain error for weak Pd^II→B interactions needs to be factored in.26

Quantum chemical calculations (DFT) were used to model the inner‐sphere reductive elimination of N,N‐dimethyl‐4‐nitroaniline from complex 14‐B (Scheme 4). C−N bond formation is predicted to proceed via an inner sphere reductive elimination with a low activation barrier of ΔG ^≠=+7.90 kcal mol⁻¹ (transition state 15‐B), yielding Pd⁰ complex 6 and N,N‐dimethyl‐4‐nitroaniline (overall ΔG=−58.75 kcal mol⁻¹). In order to understand how the Pd^II→B interaction affects the reductive elimination, the reaction was also modeled for bis[(2‐diphenylphosphino)phenyl]ether (DPEphos) complex 14‐O and diphosphinoamine complex 14‐N. DPEphos is well established as an effective ligand in palladium catalyzed Buchwald–Hartwig‐type coupling reactions,27 and commands very similar structural features to ^PhDPB^Ph (Table 2). However, DPEphos cannot mimic the potential steric effect of the B−Ph group on the coordinated reactive ligands. For this reason, the diphosphinoamine ligand (o‐PPh₂C₆H₄)₂NPh28 has also been included in the theoretical considerations, as its N‐Ph bridgehead gives a good model of the B‐Ph group in 14‐B. Elimination of N,N‐dimethyl‐4‐nitroaniline from complexes 14‐O and 14‐N gave very similar Gibbs free reaction energies of ΔG=−38.52 kcal mol⁻¹ and ΔG=−38.63 kcal mol⁻¹, respectively. No Pd^0/II→E interactions were observed in complexes featuring DPEphos and the diphosphinoamine ligand (Table 2, WBI(Pd,E)=0.005, E=O, N). Given the high structural similarity of complexes 6, 16‐O and 16‐N the increase of ΔG by ca. 20 kcal mol⁻¹ in case of the ^PhDPB^Ph ligand is a good approximation for the increase of the Pd⁰→B interaction strength in 6 compared to the Pd^II→B interaction strength in complex 14‐B. When switching from ^PhDBP^Ph to DPEphos, a small decrease of ΔΔG ^≠=0.41 kcal mol⁻¹ was found for the reductive elimination barrier (Scheme 4). This was surprising, as a more facile reductive elimination was expected from 14‐B than from 14‐O, due to 1) an electronic effect by Pd→B coordination and 2) increased steric bulk of the DPB ligand imposed by the B‐Ph group. In case of diphosphinoamine complex 14‐N the reductive elimination barrier decreased to ΔG ^≠=5.54 kcal mol⁻¹ (ΔΔG ^≠=2.46 kcal mol⁻¹), possibly as a result of the increased steric pressure imposed by the N‐Ph group (Table 2). Reductive elimination from 14‐E (E=B, O, N) proceeds via structurally early transition‐state 15‐E (Figure 7).

Scheme 4.

Reductive elimination of N,N‐dimethyl‐4‐nitroaniline from PEP complexes 14‐B, 14‐O and 14‐N.

Table 2.

Computational analysis of C−N bond formation from complexes 14‐B, 14‐O and 14‐N.^[a]

E=B, O, N	14‐B	15‐B	6	14‐O	15‐O	16‐O	14‐N	15‐N	16‐N
d(Pd,E) [Å]	2.845	2.947	2.253	3.343	3.349	2.955	3.360	3.381	3.023
d(C,N) [Å]	2.904	2.084	–	2.816	2.077	–	2.801	2.068	–
d(Pd,C) [Å]	2.042	2.059	–	2.036	2.051	–	2.033	2.051	–
d(Pd,N) [Å]	2.102	2.108	–	2.091	2.102	–	2.089	2.100	–
∢(P,Pd,P) [°]	101.2	101.0	147.1	100.4	102.0	136.4	97.5	98.8	132.9
q(Pd) [b]	+0.376	+0.330	+0.055	+0.318	+0.275	−0.162	+0.320	+0.276	−0.123
q(E)[b]	+0.722	+0.735	+0.527	−0.498	−0.496	−0.485	−0.448	−0.448	−0.444
WBI(Pd,E)[c]	0.193	0.162	0.460	0.005	0.005	0.005	0.005	0.005	0.005
ΣBα [°]	355.4	354.6	348.8	–	–	–	–	–	–

[a] Structure optimization: Turbomole 7.0.1, BP86/def‐SV(P); NBO analysis: Gaussian 09/NBO 6.0, BP86/6‐31G(d), MWB10 (P), MWB28 (Pd). [b] Natural population analysis (NPA) charge. [c] Wiberg bond index.

Figure 7.

Calculated intermediates of reductive elimination from 14‐B (top), 14‐O (middle) and 14‐N (bottom). For clarity the H atoms are omitted, and only the C_ipso atoms of the Ph‐groups at B and P are shown. Red: NPA charges, blue: bond distances.

Unexpectedly, the Pd→B interaction is slightly weakened in transition‐state 15‐B, compared to starting complex 14‐B, as indicated by a slightly elongated Pd,B distance (2.947 Å) in 15‐B compared to 14‐B (2.906 Å). Similarly, the Wiberg bond index for the Pd→B interaction is reduced to 0.162 in 15‐B (14‐B: 0.176), and the NPA charge at the borane remains unchanged (14‐B: +0.737 vs. 15‐B: +0.735). The increase of the Pd→B interaction strength occurs after the reductive elimination, explaining why the inner‐sphere reductive elimination of the C−N bond does not kinetically profit from the substantial increase of the Pd→B strength in the course of the reaction.

To rule out effects originating from restraints imposed by a chelating ligand frame work, the reductive elimination of N,N‐dimethyl‐4‐nitroaniline was also modeled using cis‐[(PMe₃)₂Pd^II(4‐NO₂C₆H₄)NMe₂] (17, ΔG=37.47 kcal mol⁻¹) and its BH₃ adduct [(PMe₃)₂(BH₃)Pd^II(4‐NO₂C₆H₄)NMe₂] (17‐B, ΔG=49.19 kcal mol⁻¹) as substrates (cf. Scheme S1). Again, a more favorable transition state was found for the acceptor free complex 17 (ΔG ^≠=+7.35 kcal mol⁻¹), than for the borane adduct 17‐B (ΔG ^≠=+8.55 kcal mol⁻¹).

Conclusions

The strength of Pd→B interactions in [(DPB)Pd] complexes depends primarily on the oxidation state of Pd. In contrast, modifications of the DPB ligand or co‐ligands have only a minor effect. ¹¹B NMR spectroscopy has been established as a useful tool to assess the strength of Pd→B interactions in solution. Reaction of lithium amides with [(^PhDPB^Ph)Pd^II(4‐NO₂C₆H₄)I] (5) chemoselectively yields the C‐N coupling product and [(^PhDPB^Ph)Pd⁰] (6). Inner‐sphere reductive C−N bond elimination was modelled with DFT methods for the ^PhDPB^Ph ligand. In contrast to reports on acceptor promoted outer‐sphere reductive C−N bond elimination,5b, 17 no significant effect of the borane acceptor on the inner‐sphere reductive elimination rate was found. This is explained by the fact that the strengthening of the Pd→B bond occurs after the reductive elimination.

Experimental Section

General

All manipulations were performed under an argon atmosphere using standard Schlenk line and glovebox techniques. Glassware was oven dried at 120 °C overnight and dried with a heat gun under vacuum prior to use. Tetrahydrofuran was dried by an MBraun solvent purification system. Benzene and n‐hexane were dried over sodium, distilled under argon prior to use and stored over activated molecular sieves (4 Å).

CD₂Cl₂ and C₆D₆ were degassed employing the freeze‐pump‐thaw technique and stored over activated molecular sieves (4 Å). [D₈]THF was dried over activated molecular sieves (3 Å), distilled under an argon atmosphere and degassed employing the freeze‐pump‐thaw technique. ^PhDPB^Ph, [(^PhDPB^PhOAc)Pd(C₃H₅)] (4), [(^PhDPB^Ph)Pd(4‐NO₂C₆H₄)I] (5) and [{(o‐PPh₂C₆H₄)₂BPh}PdI] (12) were synthesized according to published procedures.3b, 9d

NMR‐experiments were performed in Wilmad^® quick pressure valve NMR tubes. ¹H, ¹¹B{¹H}, ¹³C{¹H}, ¹⁹F{¹H}, and ³¹P{¹H} NMR spectra were recorded on a Bruker Avance II (400.1 MHz, probe: BBO) or a Bruker Avance (400.3 MHz, probe: ATM BBFO) spectrometer. ¹H and ¹³C{¹H} NMR spectra were referenced to residual solvent resonances as implemented in MesReNova 10.0.2. Infrared spectra were recorded on an Avatar 360 FT‐IR E.S.P. device by Nicolet. CHN combustion analysis were carried out on an Elementar EL device by Elementar Analysesysteme GmbH.

Deposition Number(s) 1987620 (7), 1987625 (9) and 1987626 (10) contain(s) the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/structures.

Reactivity studies

A solution of the respective lithium amide (5.7 μmol, 1.1 equiv) in [D₈]THF (0.25 mL) was added dropwise over a period of 4 min to a stirred solution of nitroarene complex 5 (5.0 mg, 5.2 μmol, 1.0 equiv) in [D₈]THF (0.25 mL). The resulting mixture was stirred for another 5 min and then transferred into an NMR tube. Reductive elimination was monitored by ³¹P NMR spectroscopy.

Synthesis of [(^PhDPB^Ph)PdCl₂] (7)

CH₂Cl₂ (8 mL) was added to a mixture of ^PhDPB^Ph (400 mg, 0.665 mmol, 1.0 equiv) and [(cod)PdCl₂] (187 mg, 0.665 mmol, 1.0 equiv). The mixture was stirred for 30 min at room temperature. Yellow crystals (380 mg, 0.482 mmol, 74 %) were formed by overlaying the solution n‐pentane (16 mL). Single crystals suitable for X‐ray diffraction were grown from a solution of [(cod)PdCl₂] (9.7 mg, 34 μmol, 1.0 equiv) and ^PhDPB^Ph (21.2 mg, 34.7 μmol, 1.0 equiv) in CD₂Cl₂ (0.7 mL) overlaid with benzene (0.3 mL). ¹¹B and ¹³C NMR data have not been collected due to poor solubility. ¹H NMR (400.13 MHz, CD₂Cl₂, 25 °C): δ 7.81–7.76 (m, 2 H), 7.55 (tdd, J=7.3, 3.0, 1.1 Hz, 3 H), 7.50–7.46 (m, 3 H), 7.46–7.38 (m, 6 H), 7.35–7.14 (m, 13 H), 6.97–6.78 (m, 5 H), 5.32 (s, 2 H, CH₂Cl₂). ³¹P{¹H} NMR (161.98 MHz, CD₂Cl₂, 26 °C): δ 44.5 (s, w_1/2=570 Hz). IR (KBr): $\tilde{\nu }$ =3643‐3284 (w), 3049 (w), 1587 (w), 1497 (m), 1433 (vs., sh), 1223 (s), 1158 (vw), 1128 (w), 1093 (vs.), 987 (w), 889 (vw), 864 (vw), 754 (s), 744 (s), 733 (m), 688 (vs.), 667 (w), 611 (m), 600 (s), 542 (m), 523 (vs.), 505 (m) cm⁻¹. Elemental analysis calcd (%) for C₄₂H₃₃BCl₂P₂Pd⋅CH₂Cl₂: C 59.18, H 4.04, found: C 59.61, H 4.33.

Synthesis of [(^PhDPB^Ph)PdBr₂] (8)

The ^PhDPB^Ph ligand (200 mg, 0.328 mmol, 1.0 equiv) and [(cod)PdBr₂] (122.7 mg, 0.328 mmol, 1.0 equiv) were solved in DCM (10 mL) and stirred at r.t. for 30 min. The solution was overlaid with n‐hexane (20 mL) yielding title compound 8 as orange crystals (192.0 mg, 0.219 mmol, 67 %). ¹¹B and ¹³C NMR data have not been collected due to poor solubility. ¹H NMR (400.30 MHz, CD₂Cl₂): δ 7.85–7.76 (m, 3 H), 7.59–7.19 (m, 30 H). ³¹P{¹H} NMR (162.04 MHz, CD₂Cl₂): δ 45.2 (bs, 1P, w_1/2=450 Hz), 38.1 (bs, 1P, w_1/2=450 Hz). IR (KBr): $\tilde{\nu }$ =3424 (s), 3048 (m), 1621 (w), 1587 (w), 1478 (m), 1455 (w), 1432 (s), 1311 (w), 1237 (w), 1220 (s), 1205 (m), 1187 (m), 1153 (w), 1126 (m), 1092 (s), 1027 (w), 1000 (m), 887 (w), 863 (w), 753 (s), 741 (s), 713 (m), 699 (s), 690 (s), 667 (m), 610 (s), 600 (s), 539 (s), 522 (s), 505 (s), 465 (m) cm⁻¹. Elemental analysis calcd (%) for C₄₂H₃₃BBr₂P₂Pd⋅0.25CH₂Cl₂: C 56.51; H 3.76, found: C 56.72, H 3.83.

Synthesis of [(^PhDPB^Ph)PdCl]SbF₆ (9)

Complex 7 (200 mg, 254 μmol, 1.0 equiv) and AgSbF₆ (87.2 mg, 254 μmol, 1.0 equiv) were stirred in DCM (15 mL) for 40 minutes. The suspension was filtered through a syringe filter (0.2 μm, PTFE membrane). The clear solution was overlaid with n‐hexane (30 mL) yielding the title compound 9 as long colorless needles (128 mg 130 μmol, 51 %). ¹H NMR (400.30 MHz, CD₂Cl₂): δ 7.97–7.92 (m, 2 H), 7.80 (tdd, J=7.5, 2.8, 0.9 Hz, 2 H), 7.69 (dd, J=7.6, 2.6 Hz, 2 H), 7.65 (t, J=7.5 Hz, 2 H), 7.55 (tt, J=7.4, 1.4 Hz, 1 H), 7.47–7.34 (m, 6 H), 7.27–7.16 (m, 10 H), 7.00 (dt, J=7.6, 2.4 Hz, 4 H), 6.83 (dd, J=12.4, 7.9 Hz, 4 H). ¹¹B{¹H} NMR (128.43 MHz, CD₂Cl₂): δ=65 (bs, w_1/2=1900±300 Hz). ¹³C{¹H} NMR (100.67 MHz, CD₂Cl₂): δ=δ 141.79, 135.43 (d, J=8.5 Hz), 134.88 (d, J=11.1 ‐Hz), 134.25, 133.69 (d, J=19.5 Hz), 133.22 (d, J=17.4 Hz), 132.49 (d, J=3.7 Hz), 129.67 (d, J=8.9 Hz), 129.33–128.82 (m), 128.10, 127.13, 126.74, 126.16. ³¹P{¹H} NMR (162.04 MHz, CD₂Cl₂): δ 49.9 (s, w_1/2=30 Hz). IR (KBr): $\tilde{\nu }$ =3441 (s), 3058 (w), 1588 (w), 1482 (w), 1435 (s), 1230 (m), 1200 (w), 1125 (w), 1034 (m), 1001 (w), 867 (vw), 752 (s), 702 (s), 692 (s), 659 (vs.), 614 (m), 538 (s), 517 (s), 697 (w) cm⁻¹. Elemental analysis calcd (%) for C₄₂H₃₃BClF₆P₂PdSb⋅0.25 C₆H₁₄: C 51.75, H 3.64, found: C 51.77, H 3.785.

Synthesis of [(^PhDPB^Ph)Pd(C₃H₅)]SbF₆ (10)

Allyl complex 4 (120 mg, 143 μmol, 1.0 equiv) and AgSbF₆ (49.0 mg, 143 μmol, 1.0 equiv) were solved in CH₂Cl₂ (7 mL) and stirred at r.t. for 20 min. The suspension was filtered through a syringe filter (0.2 μm, PTFE membrane). The clear solution was overlaid with n‐hexane (10 mL). The obtained crystals showed insufficient purity and were crystallized again under the same conditions yielding 10 as slightly yellow crystals (50.2 mg, 53.8 μmol, 38 %). ¹H NMR (400.30 MHz, CD₂Cl₂): δ 7.72–7.59 (m, 4 H), 7.58–7.53 (m, 2 H), 7.53–7.44 (m, 13 H), 7.43–7.29 (m, 6 H), 7.23–7.15 (m, 2 H), 7.05–6.87 (m, 5.5 H), 6.78–6.67 (bs, 2 H), 5.88–5.70 (bs, 0.7 H), 3.77–3.61 (bs, 1.3 H), 3.59–3.33 (bs, 1.3 H), 3.03–2.85 (bs, 0.9 H), 2.49–2.29 (bs, 1.2 H) (fractional integrals are a result from signal splitting caused by a dynamic process). ¹¹B{¹H} NMR (128.38 MHz, CD₂Cl₂): δ 64 (bs, w_1/2=1550±50 Hz). ¹³C{¹H} NMR (100.67 MHz, CD₂Cl₂): δ 141.1, 140.2, 136.1, 135.5, 135.3, 135.0, 134.4, 134.3, 134.0, 133.2 (t, J=5.8 Hz), 132.3, 132.2, 132.1, 131.6, 131.5, 131.2, 131.0, 129.6 (t, J=5.3 Hz), 129.3, 128.9, 123.1, 80.4, 80.2. ³¹P{¹H} NMR (162.04 MHz, CD₂Cl₂): δ 28.1 (s, 0.6P), 26.9 (s, 0.4P). IR (KBr): $\tilde{\nu }$ =3430 (s), 3000 (m), 1588 (m), 1480 (m), 1458 (w), 1434 (s), 1268 (m), 1227 (s), 1127 (m), 1095 (m), 1031 (w), 999 (w), 950 (vw), 875 (w), 772 (w), 754 (m), 742 (m), 733 (m), 695 (s), 659 (vs.), 609 (s), 537 (m), 521 (s), 478 (w), 430 (w) cm⁻¹. Elemental analysis calcd (%) for C₄₆H₄₀BCl₂F₆P₂PdSb: C 51.22, H 3.74, found: C 51.04, H, 3.86.

Synthesis of [(^PhDPB^Ph)Pd] (6)

A solution of LiNMe₂⋅THF (0.7 mg, 6 μmol, 1.1 equiv) in [D₈]THF (0.25 mL) was added over a period of 3 min to a solution of complex 5 (5.0 mg, 5 μmol, 1 equiv) in [D₈]THF (0.25 mL). The combined solutions were transferred to an NMR tube and NMR spectra were recorded after 1.5 and 4.5 h. ¹¹B{¹H} NMR (128.38 MHz, [D₈]THF): δ 19 (bs, w_1/2=550 Hz±50 Hz). ³¹P{¹H} NMR (162.04 MHz, [D₈]THF): δ 30.93 (s).

Synthesis of [(^PhDPB^Ph)Pd(PMe₃)] (11)

A solution of PhLi (3.2 mg, 38 μmol, 1.2 equiv) in THF (0.5 mL) was slowly added to a solution of complex 12 (25 mg, 33 μmol, 1.0 equiv) in THF (0.5 mL). After stirring for 10 min at r.t. a solution of PMe₃ in toluene (1.0 m, 50 μL, 50 μmol, 1.5 equiv) was added. The precipitate was removed by filtration and the solution was concentrated in vacuo. The resulting solid was washed with pentane and dried in vacuo (20.7 mg, 26.1 μmol, 79 %). ¹H NMR (400.13 MHz, C₆D₆): δ 8.34 (d, 2 H, J=7.8 Hz), 7.69–7.58 (m, 4 H), 7.44–7.37 (m, 2 H), 7.36–7.28 (m, 4 H, Ar‐H), 7.12 (t, 2 H, J=6.7 Hz), 7.09–7.05 (m, 13 H), 6.85 (m, 2 H), 6.68 (pt, 4 H, J=7.8 Hz), 0.64 (d, ² J _P‐H=5.0 Hz, 9 H, PMe₃). ¹¹B{¹H} NMR (128.38 MHz, C₆D₆): δ 25 (bs, w_1/2=740 Hz ±50 Hz). ¹³C{¹H} NMR (100.62 MHz, C₆D₆): δ 168.7 (bs), 143.2 (d, J=16.3 Hz), 143.0 (d, J=16.3 Hz), 141.5 (td, J=15.2, 2.0 Hz), 138.9 (t, J=13.5 Hz), 135.8 (t, J=6.4 Hz), 135.7 (t, J=2.7 Hz), 133.5 (t, J=7.7 Hz), 133.0 (dt, J=16.7, 5.0 Hz), 132.3 (s), 132.3 (s), 132.4 (t, J=6.7 Hz), 129.5 (s), 129.0 (s), 128.6 (s), 127.2 (s), 126.1 (t, J=2.8 Hz), 125.2 (s), 18.1 (dt, J=11.8, 2.2 Hz, PMe₃). ³¹P{¹H} NMR (162.04 MHz, C₆D₆): δ 35.44 (d, ² J _P‐P=14.1 Hz, 2P, ArPPh₂), −40.13 (t, ² J _P‐P=14.2 Hz, 1P, PMe₃).

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.

Supplementary

F. Ritter, L. John, T. Schindler, J. P. Schroers, S. Teeuwen, M. E. Tauchert, Chem. Eur. J. 2020, 26, 13436.