Twisted Amide Xantphos Ligands: Wide Bite Angle, Twist-Controlled Diphosphines

Abstract

Xantphos ligands are among the most significant bis(phosphines) developed in the field of homogeneous catalysis. Twisted amides have been established to have a pronounced impact on the geometry of substituents attached to the amide linkage. Herein, we describe the synthesis, structural characterization and activity of novel, easily accessible twisted amide Xantphos ligands. These ligands feature controlled P–P distance and increased P–M–P bite angles guided by the amide bond twist and geometry. Notably, N-t-BuC(O)-Xantphos, which exhibits significant amide nitrogen pyramidalization (χ_N = 32°), showed enhanced catalytic efficiency toward C–C, C–P, and C–B bond-forming reactions. Steric properties and coordination chemistry with Au(I), Cu(I), and Pd(II) are presented. Considering the critical role of phosphine ligands and twisted amides in chemistry, the concept of controlling bis(phosphines) geometry by amide bond twist is anticipated to gain rapid and widespread adoption.

Developed in 1995 by van Leeuwen and co-workers, Xantphos ligands have been indispensable in the field of organic, organometallic and inorganic chemistry owing to their capacity to enforce large P–M–P angles and tunable electronic properties (Fig. 1A).^1–5 Among bis(phosphines), the structural rigidity of Xantphos ligands renders them highly effective for coordinating with an array of metals, enabling numerous catalytic applications.^1,6–8 Moreover, the tunable electronics of Xantphos by heterocyclic backbone substitution allow for additional control of ligand properties. Geometrically, the heterocyclic scaffold induces large P−M−P bite angles, which have proven critical in stabilization of various metal centers and catalytic processes.^1,4,5,9–11 Furthermore, Xantphos ligands are optimized to achieve large bite angles, favouring formation of trigonal or tetrahedral metal–L complexes and enabling the formation of catalytic intermediates relevant in chemical processes, which are difficult to accommodate with other ligand classes.^1,12–14

Fig. 1.

A. Xantphos ligands; B. Twisted amides; C. Merging twisted amides with ligand design: twisted amide Xantphos ligands.

On the other hand, twisted amides have attracted major attention as an effective means to control the geometry and reactivity of typically planar amide bonds (Fig. 1B). Geometrically, archetypal amides are planar around the amide linkage due to N–C(O) stabilization via n_N→π*_C=O resonance.^15–21 However, distortion of the planar amide bond through geometric effects gives rise to unique twisted amides of prime relevance in organic synthesis, biochemistry, drug discovery, proteomics, molecular folding and catalysis.²²⁻²⁵⁴ With reduced amide conjugation, twisted amides exhibit enhanced N-nucleophilicity and C=O electrophilicity akin to amino-ketones rather than conventional amides.^26,27 However, despite their intriguing properties, the concept of controlling molecular bond geometry by twisted amides remains unexplored in ligand design.

Drawing inspiration from the elegant studies on diphosphine Xantphos ligands^1–5 (Fig. 1A) and building on our expertise in ligand design^28–34 and twisted amides^15–27 (Fig. 1B), we aimed to advance the concept of electronically modified, easily accessible twisted amide Xantphos ligands. We hypothesized that incorporating a twisted amide geometry with a pyramidalized nitrogen would fold the phenyl rings on the phenoxazine scaffold, thereby increasing the distance between the phosphine atoms. Further, the increased P−P distance would lead to an increased P−M−P bite angle (Fig. 1C), resulting in a modular approach to controlling metal coordination by remote twisted amide geometry.

Twisted amide Xantphos ligands are readily prepared via cost-effective, scalable and operationally-simple acylation of N-Xantphos (Scheme 1). Analogues with different steric and electronic properties based on amide NC(O)–R substitution (R = Me, t-Bu, Ph) have been selected for the study, resulting in tunable diphosphine topology. As shown, the synthesis of twisted amide Xantphos is achieved by deprotonating N-Xantphos using LiHMDS at −78° C followed by acylation with acid chlorides. After optimization, we were pleased to find that N-MeC(O)-Xantphos (2a), N-tBuC(O)-Xantphos (2b) and N-PhC(O)-Xantphos (2c) could be prepared in excellent 91%, 88% and 90% yields. Ligands 2a, 2b and 2c were fully characterized by X-ray crystallography (Fig. 2). Table 1 summarizes Winkler−Dunitz distortion parameters (τ, χ_N, χ_C) and relevant bond lengths (N−C(O), C=O, C−C). The table also includes geometric parameters of representative planar amides for comparison.

Scheme 1.

Synthesis of Xantphos ligands. Conditions: 1 (1.0 equiv), LiHMDS (1.7 equiv), acyl chloride (1 equiv), THF, −78° C.

Fig. 2.

X-ray crystal structures of 2a–2c. Diphenyl angle (angle between the benzene rings) and P–P distances are shown. 2a: CCDC 2525811; 2b: CCDC 2525812; 2c: CCDC 2525814.

Table 1.

Summary of Structural Parameters of 2a−2c and Representative Planar Amides^a

Amide	N–C(O) (Å)	C=O (Å)	C–C (Å)	τ [°]	χN [°]
2a	1.390(4)	1.209(6)	1.489(7)	1.11	1.38
2b	1.401(2)	1.215(2)	1.547(3)	9.92	31.9
2c	1.392(2)	1.217(2)	1.496(3)	3.91	12.8
1-methylpyrrolidin-2-onea	1.307(8)	1.241(8)	1.498(9)	1.6	1.9
formamideb	1.318(3)	1.228(3)	-	0	0

^aref. 20;

^bref. 21.

With the presence of twisted amide, the diphenyl angle (angle between the benzene rings in phenoxazine) on the phenoxazine scaffold are altered from 35.71° (2a) to 37.29° (2b) to 40.35° (2c). Consequently, the P–P distances are changed from 4.296 Å (2a) to 4.268 Å (2b) to 4.161 Å (2c) due to diaryl rotation of the phenoxazine scaffold enforced by the twisted amide geometry.

The P–P distances in 2a–2c are larger than the P–P distance of 4.046 Å in Xantphos,³⁵ while the P–P distances in 2a,b are also larger than in N-Xantphos of 4.255 Å,³⁶ consistent with the expectation that delocalization of the lone pair at nitrogen into the amide bond would increase the P–P distances. There is a negative correlation between the diphenyl angle and the P–P distance (R² = 0.98) with the larger angle corresponding to a shorter P–P distance and a positive correlation between the diphenyl angle and the C=O bond length (R² = 0.80). Further, there is an excellent correlation between the N–C(O) and C–C bond lengths (R² = 0.99) and between the N–C(O) bond length and the additive Winkler-Dunitz parameter (τ+χ_N) (R² = 0.96), indicative of the well-defined twisted amide geometry (see SI). The additive Winkler-Dunitz (τ+χ_N) value of 41.8° for 2b corresponds to a significant distortion of the amide bond, despite the rigidity of phenoxazine ring.

The effect of amide bond geometry on distortion depends on the class of amides investigated.^22,23 In the present case, there is a positive correlation between the N–C bond length and the C–C distance, and specifically the additive Winkler-Dunitz distortion parameter. There is a correlation between the changes in amide distortion and steric properties of the diphosphine. Detailed examination of distortion reveals that the main contributors are ω¹ (C₁-C-N-C₂) and ω³ angles (O-C-N-C₂) (Table 2). The steric effect arises from the interaction between the C-substituent and the flanking N-aryl rings. Further insight is gained by evaluation of the dihedral C_(CO)-N-C_(Ar)-C_(Ar) angles. 2a: 60.0°/57.3° (avg. 58.7°), 2b: 50.4°/57.8° (avg. 54.1°), 2c: 50.9°/51.8° (avg. 51.4°), which negatively correlate with the diphosphine dihedral angle (2a: 35.7°, 2b: 37.3°, 2c: 40.4°).

Table 2.

Summary of Structural Parameters and Dihedral Angles of 2a–2c.

Amide	ω1 [°]	ω2 [°]	ω3 [°]	ω4 [°]	θ1 [°]	θ2 [°]
2a	169.1(3)	−167.2(3)	−9.8(5)	11.7(5)	60.0(5)	57.4(5)
2b	151.8(2)	171.7(2)	−23.6(3)	−12.9(3)	50.3(3)	57.8(3)
2c	−169.8(2)	177.6(2)	10.4(3)	−2.6(3)	50.9(3)	51.8(3)

The structure of 2b (N-tBu-C(O)-Xantphos) shows that this compound is most twisted in the series (τ+χ_N) = 41.8°, with χ_N of 31.9° as the major contributor of distortion. This is followed by 2c (N-Ph-C(O)-Xantphos) (τ+χ_N) = 16.7°, while 2a (N-Me-C(O)-Xantphos) is nearly planar (τ+χ_N) = 2.5°. Accordingly, amide 2b shows the longest N–C(O) bond length of 1.401 Å and the longest C–C bond length of 1.546 Å, while the length of the C=O bond is 1.215 Å. Analysis of the pyramidalization at nitrogen angle reveals significant pyramidalization (χ_N = 31.9°). The availability of fully characterized 2a−2c allows us to compare the effect of distortion on pyramidalization at nitrogen. As shown in Fig. 3 and SI, a plot of pyramidalization at nitrogen (χ_N) vs twist (τ) gives an excellent linear correlation (R² = 0.96), indicating that the N-pyramidalization is followed by N–C(O) bond rotation in this series.

Fig. 3.

Graphical Illustration of C_(CO)-N-C_(Ar)-C_(Ar) Dihedral Angles.

With access to N-tBuC(O)-Xantphos, we next characterized its Au(I), Cu(I) and Pd(II) complexes (Scheme 2). The gold complex [Au(N-tBuC(O)-xantphos)Cl] (3) was prepared using AuCl(SMe₂) and 2a in a 1:1 ratio in THF. Palladium complex [Pd₂(N-tBuC(O)-Xantphos)Cl₄] (4) and copper complex [Cu(N-tBuC(O)-Xantphos)I] (5) were prepared using Pd(NCMe)₂Cl₂ and CuI in CH₂Cl₂ and acetonitrile. Importantly, all complexes were found to be stable to air and moisture. Complexes 3−5 were fully characterized by X-ray crystallography (Fig. 4 and SI). [Au(N-tBuC(O)-Xantphos)Cl] features a the Au−Cl bond length of 2.501 Å and the Au−P bond lengths of 2.314 Å and 2.329 Å. The P−Au−P angle of N-tBuC(O)-Xantphos is 127.32°, which is larger than the P−Au−P bite angle of 116.63° of Xantphos,³⁷ indicating that the increased P−P distance results in a larger P−M−P bite angle. [Cu(N-tBuC(O)-Xantphos)I] crystallized as an unsymmetrical iodo-bridged dimer. The complex is characterized by Cu−P bond lengths of 2.259 Å/2.261 Å (Cu1) and 2.261 Å/2.263 Å (Cu2). The Cu1–I1 bond associated with the bridging iodide is shorter (2.530 Å), whereas the Cu2–I1 distance corresponding to the collapsed interaction is longer (2.773 Å) (see SI).

Scheme 2.

Metal Complexes of N-tBuC(O)-Xantphos. a. AuCl(SMe₂) (1 equiv), 2a (1 equiv), THF, rt, 8 h; b. Pd(NCMe)₂Cl₂ (2 equiv), 2a (1 equiv), CH₂Cl₂, rt, 20 min. c. CuI (1 equiv), 2a (1 equiv), MeCN, 50 °C, 15 h. 3: CCDC 2525815; 4: CCDC 2528040; 5: CCDC 2525813.

Fig. 4.

X-ray crystal structures of complexes 3-5. 50% ellipsoids. Hydrogen atoms have been omitted. 3: CCDC 2525815; 4: CCDC 2528040; 5: CCDC 2525813.

The P−Cu−P angles are 121.54° (P₂Cu−I) and 117.46° (P₂Cu−I₂), which can be compared with the P−Cu−P angle of 118.41° of Xanthphos.³⁸ [Pd₂(N-tBuC(O)-Xantphos)Cl₄] features two bridging and two terminal chlorides with two palladium atoms bridging the phosphines. This complex is characterized by a Pd-Pd bond length of 2.861 Å and Pd−P bond lengths of 2.219 Å and 2.228 Å. The dihedral P−Pd−Pd−P angle is 2.8°, indicating coplanarity of P−Pd atoms, while the average P−Pd−Pd angle is 94.35°.

Further insight is gained from the %buried volume analysis (Fig. 5).³⁹ For [Au(N-tBuC(O)-Xantphos)Cl], the X-ray analysis revealed the (%V_bur) of 56.2% with SW 57.0%, NW 54.3%, NE 55.2% and SE 58.3% for each quadrant (Fig. 5A); [Cu(N-tBuC(O)-Xantphos)I] is characterized by the (%V_bur) of 51.4% with SW 55.1%, NW 42.3%, NE 52.9% and SE 55.2% (Fig. 5B); for [Pd₂(N-tBuC(O)-Xantphos)Cl₄] the (%V_bur) is 31.6% with SW 29.3%, NW 30.7%, NE 36.6% and SE 30.0% for each quadrant (Fig. 5C), indicating significant hindrance and a very considerable steric flexibility of metal coordination (57.3% to 31.6%).

Fig. 5.

Topographical steric maps of [Au(N-tBuC(O)-Xantphos)Cl] (3), [Cu(N-tBuC(O)-Xantphos)I] (4) and [Pd₂(N-tBuC(O)-Xantphos)Cl₄] (5) showing % V_bur per quadrant.

We have obtained additional spectroscopic characterization to evaluate the amide C=O properties (Table 3). Since it is known that IR spectroscopy is not always a reliable method to evaluate C=O properties in acyclic twisted amides,^22,23 we have compared the ¹³C NMR carbonyl shifts of the ligands 2a-2c and their corresponding metal complexes. These values clearly indicate a decrease of amide conjugation with an increase of steric rotatio, while it is established that R = aromatic cannot be compared directly with R = aliphatic in acyclic twisted amides²³ and an increase of conjugation upon metal coordination.

Table 3.

Summary of Spectroscopic Properties of 2a−2c and 3–5.

Amide	13C NMR (C=O, ppm)	IR (νC=O, cm−1)
2a	169.37	1682.0
2b	179.30	1668.0
2c	168.58	1668.3
3: [Au(I)(2b)Cl]	178.66	1667.1
4: [Pd2(II)(2b)Cl4]	178.28	-
5: [Cu(I)(2b)Cl]	178.67	-

The activity was briefly evaluated in Pd- and Cu-catalyzed functionalization of alkynes (Scheme 3). As such, hydrocarboxylation with formic acid catalyzed by palladium (entry 1),⁴⁰ hydrophosphinylation catalyzed by palladium (entry 2),⁴¹ and hydroboration of internal (entry 3) and terminal (entry 4)^42,43 alkynes, all proceeded in good to excellent yields using N-tBuC(O)-Xantphos. We have conducted comprehensive studies across all Xantphos-amide ligands 2a-2c as well as reference N-Xantphos and Xantphos. In all cases, the original twisted amide Xantphos 2b (R = t-Bu) outperformed other twisted amide Xantphos ligands as well as control ligands. These results demonstrate that the N-tBuC(O)-Xantphos is effective in C–C, C–P and C–B bond forming reactions that are routinely employed in industrial and academic settings and might offer a significant potential for catalytic reaction development.

Scheme 3.

Catalytic Activity of Xantphos-Amide 2a-2c, and Reference Ligands.

In conclusion, we have developed a novel class of easily accessible twisted amide Xantphos ligands. These ligands merge the concept of controlled amide bond geometry enabled by bond twisting to ligand design of rigid bidentate diphosphines, leading to diaryl folding, larger P–P distances and enhanced P–M–P bite angles. X-ray crystallographic analysis revealed that the N-tBuC(O)-Xantphos shows a significant distortion of the amide bond with nitrogen pyramidalization as the major contributor. The synthesis and characterization of Au(I), Cu(I) and Pd(II) complexes has been reported. Considering the critical role of phosphine ligands and twisted amides in chemistry, we anticipate that the approach of controlling ligand geometry by amide bond twist will find rapid and widespread adoption.

Data availability

The data supporting this article have been included as part of the Supplementary Information. Supplementary information: CIF files for the crystal structures of 2a-2c and 3–5, experimental details, and analytical data.