Well-Defined, Air- and Moisture-Stable Palladium–Imidazo[1,5-a]pyridin-3-ylidene Complexes: A Versatile Catalyst Platform for Cross-Coupling Reactions by L-Shaped NHC Ligands

Abstract

We describe the development of [(NHC)Pd(cinnamyl)Cl] complexes of ImPy (ImPy = imidazo[1,5-a]pyridin-3-ylidene) as a versatile class of precatalysts for cross-coupling reactions. These precatalysts feature fast activation to monoligated Pd(0) with 1:1 Pd to ligand ratio in a rigid imidazo[1,5-a]pyridin-3-ylidene template. Steric matching of the C5-substituent and N2-wingtip in the catalytic pocket of the catalyst framework led to the discovery of ImPyMesDipp as a highly reactive imidazo[1,5-a]pyridin-3-ylidene ligand for Pd-catalyzed cross-coupling of nitroarenes by challenging C–NO₂ activation. Kinetic studies demonstrate fast activation and high reactivity of this class of well-defined ImPy–Pd catalysts. Structural studies provide full characteristics of this new class of imidazo[1,5-a]pyridin-3-ylidene ligands. Computational studies establish electronic properties of sterically-restricted imidazo[1,5-a]pyridin-3-ylidene ligands. Finally, a scalable synthesis of C5-substituted imidazo[1,5-a]pyridin-3-ylidene ligands through Ni-catalyzed Kumada cross-coupling is disclosed. The method obviates chromatographic purification at any of the steps, resulting in a facile and modular access to ImPy ligands. We anticipate that well-defined [Pd–ImPy] complexes will find broad utility in organic synthesis and catalysis for activation of unreactive bonds.

Graphical Abstract

Introduction

N-Heterocyclic carbenes (NHCs) have found expansive applications in organic synthesis and catalysis since the first isolation of an archetypal free carbene by Arduengo in 1991.^1–7 The strong σ-donating abilities and easily tunable steric differentiation make them quintessential ligands for transition-metal-catalysis and main group chemistry.^7–12 The invention of new NHC ligands by (1) N-wingtip modification, (2) backbone alteration, or (3) ring remodeling, has continuously been a key way to profoundly change the electronic and steric properties of NHCs. Successful examples in this field include the discovery of cyclic(alkyl)(amino)carbenes (CAACs),¹³ sterically-demanding bioxazoline carbenes (IBiox),¹⁴ abnormal NHCs (aNHCs),^15–16 mesoionic carbenes (MICs),^17–18 among many others.^19–20

Imidazo[1,5-a]pyridin-3-ylidenes (ImPy), first reported in 2005, feature unique architecture leading to a sterically demanding environment around the metal center when the C5 position is substituted by an aryl group (Figure 1A).^21–22 In light of appealing attributes of sterically-restricted NHC ligands,^23–25 we anticipated that highly attractive ligands could be obtained by steric matching of the restricted C5 and flexible N2-substitutes, providing a powerful strategy for the invention of new chemical technologies for bond activation.

Figure 1.

(A) Electronic and structural remodelling of NHC ligands. (B) State-of-the art Pd(II) precatalysts in cross-coupling reactions. (C) Pd(II)–ImPy and steric-matching of [(ImPyMesDipp)Pd(cin)Cl] as air-stable, fast-activating, sterically-restricted Pd(II)–NHC precatalysts.

A major advancement in palladium-based catalyst design is the development of well-defined Pd(II)–precatalysts, where Pd to ligand ratio is 1:1 (Figure 1B).²⁶ This Pd-catalyst design strategy offers a major advantage over in situ formed catalysts due to the stability of Pd–complexes and high price of ancillary ligands (cf. Pd source).^27–36 In this context, Pd(NHC)(cin)Cl,³² Pd–PEPPSI,³³ SingaCycle A1–A3,³⁴ (CAAC)Pd(allyl)Cl,³⁵ and Buchwald palladacycles G1–G4³⁶ have emerged as established benchmarks for palladium-catalyzed cross-coupling reactions, and many of them are now commercially available.

On the other hand, nitroarenes (Ar–NO₂) are highly attractive electrophiles for cross-coupling reactions due to their availability by S_EAr nitration technology complementary to aryl halides, low price as chemical feedstocks and orthogonal reactivty.³⁷ The direct utilization of nitroarenes as coupling partners in cross-coupling reactions by Ar–NO₂ oxidative addition is scarce, and the ligands used for the activation of C–N bond in nitroarenes were initially limited to BrettPhos analogues.^38–50

Subsequently, C5-substituted ImPy were found as active ligands for cross-coupling of nitroarenes.^51–53 However, the protocols relied on large excess of ligands to Pd. Further, long reaction time was required due to slow activation of the putative “Pd(NHC)(acac)Cl” species generated in situ.

We reasoned that the key limitations would be addressed by (1) the development of well-defined, fast-activating [(NHC)Pd(cinnamyl)Cl] complexes of ImPy ligands, which would be bench-stable, yet easily activated to NHC–Pd(0) in the presence of arylboronic acid and base, and (2) steric matching of the C5 restricted and N2 flexible substituents to define catalytic pocket for elementary cross-coupling steps (Figure 1C). We report the development of [(NHC)Pd(cinnamyl)Cl] complexes of ImPy (ImPy = imidazo[1,5-a]pyridin-3-ylidene) as a versatile class of precatalysts for cross-coupling reactions. Most crucially, these air- and bench-stable, well-defined precatalysts feature fast activation to monoligated Pd(0) with 1:1 Pd to ligand ratio in a rigid imidazo[1,5-a]pyridin-3-ylidene framework. Catalyst synthesis, kinetic and catalytic studies, full characterization as well as practical and scalable synthesis by Ni-catalyzed Kumada cross-coupling of this class of imidazo[1,5-a]pyridin-3-ylidene ligands is disclosed.⁵⁴ This study opens the door to activation of unreactive bonds in organic synthesis by well-defined ImPy–Pd catalysts.

Results and discussion

Design Plan.

Based on our previous studies on Pd–NHC catalysis,^{23–25,28–31} we realized that there should be a precise steric match between the rigid C5 and flexible N2 substituents of the catalytic pocket of transition-metal-complexes of imidazo[1,5-a]pyridin-3-ylidenes.

A library of C5-Aryl/N2-Aryl imidazo[1,5-a]pyridinium salts with a systematic variation of the C5 and N2 substitution critical to catalysis, namely ImPyMesMes·HCl (1a), ImPyMesDipp·HCl (1b), ImPyTrippMes·HCl (1c), and ImPyTrippDipp·HCl (1d) were selected for the initial study (Figure 2). The Suzuki-Miyaura cross-coupling of nitroarenes was selected because it is a novel method to activate C–NO₂ bonds, and nitroarenes represent a highly attractive class of orthogonal electrophiles to aryl halides prepared by S_EAr nitration of benzene derivates. At present, very few methods for the biaryl coupling of nitrobenzenes have been reported.^{40, 52–53, 55} It should be noted that (1c) is a privileged ligand for the biaryl coupling of nitroarenes reported previously. NHC ligands without steric hindrance at the ortho-position of the C5-substituent are unreactive (not shown).

Figure 2.

Structures of imidazo[1,5-a]pyridin-3-ylidene ligands used in this study.

The in situ protocol was first used to screen the selected imidazo[1,5-a]pyridin-3-ylidene ligands (Table 1).

Table 1.

Initial Optimization^a

entry	ligand	yield (%)b
1	IMesHCl	0
2	SIMesHCl	0
3	IPrHCl	0
4	SIPrHCl	0
5	ImPyMesMes·HCl	13
6	ImPyMesDipp·HCl	33
7	ImPyTrippMes·HCl	47
8	ImPyTrippDipp·HCl	18

^aConditions: 4-nitroanisole (1.0 equiv), benzeneboronic acid (1.5 equiv), Pd(acac)₂ (5 mol%), ligand (10 mol%), K₃PO₄ (3.0 equiv), H₂O (3.0 equiv), TDA (10 mol%), dioxane (0.2 M), 130 °C, 36 h.

^bGC/NMR yield. TDA = tris(3,6-dioxaheptyl)amine.

4-Nitroanisole and phenylboronic acid were selected as model substrates. Our initial optimization showed that the classical imidazolium salts (IMes·HCl, IPr·HCl) and imidazolinium salts (SIMes·HCl, SIPr·HCl) were totally ineffective for this transformation, resulting in the recovery of nitroarene (Table 1, entries 1–4). Moreover, the screening of imidazo[1,5-a]pyridinium salts revealed that both the least sterically-hindered ImPyMesMes·HCl (13% yield) and the most sterically-demanding ImPyTrippDipp·HCl (18% yield) are less effective than ImPyMesDipp·HCl (33% yield) and ImPyTrippMes·HCl (47% yield) (Table 1, entries 5–8). The yield for ImPyTripp-Mes·HCl was not improved by changing the conditions or prolonging the reaction time (not shown).

Synthesis of Pd(II)–NHC Complexes.

Next, we sought to prepare well-defined Pd(II)–NHC complexes using cinnamyl ancillary ligands. Cinnamyl ancillary ligands were selected due to their well-established capacity to serve as supporting ligands to Pd(II)–NHC and Pd(II)–phosphine complexes and because many different R–allyl ligands are available for tuning the reactivity of Pd(II) complexes.³²

Thus, although the reaction of imidazo[1,5-a]pyridinium salt ImPyMesDipp·HCl with the palladium cinnamyl dimer [{Pd(cin)(μ-Cl)}₂] in the presence of K₂CO₃ in acetone at reflux failed to give the desired product,⁵⁶ we were delighted to find that the desired ImPy–ligated Pd(II) complex [(ImPyMesDipp)Pd(cin)Cl] (2a) was successfully obtained in 73% yield after 3 hours at room temperature when we switched to a stronger potassium tert-butoxide base and THF as a solvent (Scheme 1).

Scheme 1.

Synthesis of Pd(II)–NHC Complexes^a

^aConditions: (a) [Pd(cin)Cl]₂ (0.5 equiv), KOtBu (1.4 equiv), THF (0.1 M), 23 °C, 3 h, 73% (2a), 71% (2b). (b) [Pd(allyl)Cl]₂ (0.5 equiv), KOtBu (1.4 equiv), THF (0.1 M), 23 °C, 3 h, 78%. (c) [Pd(1-tBu-ind)Cl]₂ (0.5 equiv), KOtBu (1.4 equiv), THF (0.1 M), 23 °C, 3 h, 37%.

Following similar procedure, complex [(ImPyTripp-Mes)Pd(cin)Cl] (2b) from the privileged ImPyTrippMes ligand was prepared (Scheme 1). We have also synthesized well-defined complexes [(ImPyMesDipp)Pd(allyl)Cl] (2c) and [(ImPyMesDipp)Pd(1-tBuind)Cl] (2d) bearing allyl and 1-t-Bu-indenyl ancillary ligands utilizing the corresponding Pd dimer precursors. The yield for complex (2d) was comparatively lower, which we attribute to the excessive steric hindrance of the ancillary ligand.⁵⁷ Importantly, all complexes were found to be air- and moisture-stable, with no notable decomposition detected after storing on the bench-top for over 6 months.

Interestingly, in ¹H and ¹³C NMR we observed two isomers namely endo-, and exo- forms in 2a–2c due to the difference in the relative arrangement between the central Cβ-Hβ bond (highlighted in blue) of the allyl ligand and the phenyl ring (red) of the biaryl system (Chart 1). In 2d, the exo- form is not favored because of the steric repulsion between the phenyl ring of the indenyl group and the phenyl ring of the biaryl.

Chart 1.

Endo and exo isomers of Pd(II)–NHC Complexes 2a–2d

With the library of well-defined Pd(II)–NHC complexes of sterically-restricted imidazo[1,5-a]pyridin-3-ylidenes in hand, we next evaluated their reactivity in the Suzuki cross-coupling of nitroarenes (Table 2). We were delighted to find that in addition to the highly desirable 1:1 ratio of ligand to Pd, all of the well-defined complexes showed much higher reactivity than the in situ formed Pd–NHCs in the cross-coupling. As such, the yields using complexes 2a, 2c, 2d were 68%, 81% 57%, respectively, after 10 hours (Table 2, entries 1–3), however, the in situ protocol gave only 33% yield using the same ligand ImPyMesDipp after 36 hours (Table 1, entry 6).

Table 2.

Optimization Using Well-Defined Complexes^a

entry	catalyst	yield (%)b
1c	[(ImPyMesDipp)Pd(allyl)Cl]	68
2c	[(ImPyMesDipp)Pd(cin)Cl]	81
3c	[(ImPyMesDipp)Pd(1-tBu-ind)Cl]	57
4c	[(ImPyTrippMes)Pd(cin)Cl]	74
5	[(ImPyMesDipp)Pd(cin)Cl]	82
6d	[(ImPyMesDipp)Pd(cin)Cl]	67
7e	[(ImPyMesDipp)Pd(cin)Cl]	47
8f	[(ImPyMesDipp)Pd(cin)Cl]	65
9g	[(ImPyMesDipp)Pd(cin)Cl]	31
10h	[(ImPyMesDipp)Pd(cin)Cl]	35

^aConditions: 4-nitroanisole (1.0 equiv), benzeneboronic acid (1.5 equiv), catalyst (5 mol%), K₃PO₄ (3.0 equiv), H₂O (3.0 equiv), TDA (10 mol%), dioxane (0.2 M), 130 °C, 16 h.

^bGC/NMR yield.

^c10 h.

^dw/o water.

^e9 equiv of water.

^f20 mol% TDA.

^gtoluene.

^hCs₂CO₃.

Interestingly, ImPyMesDipp supported precatalyst (2a) outperformed the ImPyTrippMes ligated precatalyst (2b) using the well-defined Pd(II)–NHC complex (81% yield vs. 74% yield, Table 2, entry 2 vs. 4), affirming that a perfect match of sterics at C5 and N2 positions is needed in the catalytic pocket for optimal results. Extending the reaction time did not improve the yield considerably (Table 2, entry 1 vs. entry 5). Further changes of the reaction conditions gave inferior results (Table 2, entries 6–10). We think that TDA serves as a phase-transfer catalyst. Our results showed both 18-crown-6 and TDA can improve the yields (not shown). Similar to 18-crown-6, TDA binds to potassium cations which leads to a higher concentration of phosphate ion (PO₄³⁻) in solution.⁵²

Kinetic Studies.

We next carried out kinetic studies to further compare the catalytic reactivity of the in situ protocol and well-defined Pd(II)–NHC complexes (Figure 3). As shown, well-defined complexes showed much superior kinetics as compared to the in situ protocol, consistent with the fast-activation to catalytically-active monoligated Pd(0)–NHCs by the use of cinnamyl ancillary ligand.⁵⁸ As such, precatalysts (2a) and (2b) gave 51% conversion after 2 hours while the Pd(acac)₂/1b combination only had a conversion of 28%. After 10 hours, 87% of the starting nitroarene was consumed using precatalyst (2a), while the reactions using Pd(acac)₂ and the same imidazo[1,5-a]pyridinium salt ImPyMesDipp, resulted in less than half conversion after the same reaction of time. Thus, the kinetic studies confirm that well-defined Pd(II)–NHC complexes of imidazo[1,5-a]pyridin-3-ylidenes demonstrate much higher reactivity than the in situ formed catalysts. Furthermore, the data also establishes that complex [(ImPyMesDipp)Pd(cin)Cl] (2a) is more reactive than [(ImPyTrippMes)Pd(cin)Cl] (2b). As a further important consideration, ligand ImPyMesDipp is derived from the abundant mesitylene, which is a common feedstock and more readily available than ImPyTrippMes, which is derived from triisopropylbenzene (vide infra).

Figure 3.

Kinetic profile of the Suzuki-Miyaura cross-coupling of 4-MeO-C₆H₄-NO₂ with phenylboronic acid catalyzed by Pd(acac)₂/1b, 2a and 2b at 130 °C.

Characterization of Pd(II)–NHC Complexes.

All of the four novel Pd(II)–NHC complexes were fully characterized by x-ray crystallography (Figure 4). The C(carbene)–palladium bond lengths are compared with known Pd(II)–NHC complexes ligated with imidazolylidenes, such as IPr (Table 3). As shown, the ImPyMesDipp complexes have similar bond lengths to the corresponding [Pd–IPr] complexes.^57–59 However, [(ImPyTrippMes)Pd(cin)Cl] features a shorter C(carbene)–palladium bond length than both [(IPr)Pd(cin)Cl] and [(ImPyMesDipp)Pd(cin)Cl], suggesting a stronger palladium–carbene bond.⁵⁹ Furthermore, the Pd–C(allyl) bond lengths are 2.190 Å (Pd–C3) (2a), 2.221 Å (Pd–C3) (2b), 2.162 Å (Pd–C3) (2c), consistent with the previous studies on [(NHC)Pd(R-allyl)Cl] complexes establishing that the longer Pd–C(allyl)–C(3) bond, the easier the activation to monoligated Pd(0).⁵⁸

Figure 4.

X-ray structure of complex 2a (A), 2b (B), 2c (C), 2d (D). Hydrogen atoms have been omitted for clarity. See SI for selected bond lengths and angles. CCDC 2174493 (2a); CCDC 2174494 (2b); CCDC 2174495 (2c); CCDC 2174496 (2d).

Table 3.

Comparison of C_carbene–Pd Bond Lengths

Pd(II)–NHC	Ccarbene-Pd length (Å)
(IPr)Pd(cin)Cl	2.040(1)
(ImPyMesDipp)Pd(cin)Cl	2.050(2)
(ImPyTripMes)Pd(cin)Cl	2.022(4)
(IPr)Pd(allyl)Cl	2.040(1)
(ImPyMesDipp)Pd(allyl)Cl	2.046(3)
(IPr)Pd(1-tBu-ind)Cl	2.014(3)
(ImPyMesDipp)Pd(1-tBu-ind)Cl	2.030(3)

Topographical steric maps of Pd(II)–NHC complexes 2a–2d determined by the method of Cavallo and co-workers show that [(ImPyMesDipp)Pd(cin)Cl] and (ImPyTripMes)Pd(cin)Cl have %V_bur of 42.1% (Figure 5),⁶⁰ bur which is much higher than that of [(IPr)Pd(cin)Cl] (%V_bur of 34.3%). The % buried volume of [(ImPyMesDipp)Pd(allyl)Cl] and [(ImPyMesDipp)Pd(1-tBuind)Cl] is also much higher (%V_bur of 41.7% and 40.9%, respectively)

Figure 5.

Topographical steric maps of [(ImPyMesDipp)Pd(cin)Cl] (2a) (A), [(ImPyTrippMes)Pd(cin)Cl] (2b) (B), [(ImPyMesDipp) Pd(allyl)Cl] (2c) (C), [(ImPyMesDipp)Pd(1-tBu-ind)Cl] (2d) (D) showing % V_bur per quadrant.

Crucially, the x-ray analysis in combination with the steric topographical maps reveal that the C5-aryl groups in ImPy are positioned closer to the palladium center than the N2-aryl wingtips (cf. IPr). Due to dissymmetry of the ligands, the two quadrants (NE and SE) corresponding to the C5-aryl groups are highly congested, while the other two quadrants (NW and SW) are less sterically-hindered. A closer inspection reveals that the N2-aryl groups have a similar steric effect on the two quadrants (2a: 32.9% vs. 37.4%), while the C5-aryl groups give an unsymmetrical steric pressure on the NE and SE quadrants (2a: 61.4% vs. 36.8%). These features make well-defined Pd(II)–NHC cinnamyl complexes of imidazo[1,5-a]pyridin-3-ylidenes very different from imidazol-2-ylidenes, such as IPr, which have nearly symmetrical steric distribution around the metal center.⁵⁸ This unique catalytic pocket facilitates the approach of substrates to the metal center and reductive elimination.⁶¹

Catalytic Activity.

Next, we examined the scope of the Suzuki cross-coupling of nitroarenes using well-defined Pd(II)–NHC complex (2a) (Table 4). As shown, a wide range of boronic acids could be coupled with electronically-differentiated nitroarenes by C–NO₂ activation. Thus, boronic acids with electron-donating groups were cross-coupled with electron-neutral (5a, 5r), electron-rich (5j) and heterocyclic nitroarenes (5y) in high yields under our standard conditions. Furthermore, electron neutral electrophile/nucleophile combinations (5e, 5f, 5k, 5u) gave the desired products in good yields. Moreover, electronically-deactivated boronic acids (5b, 5c) smoothly reacted with nitroarenes (5n, 5o, 5p, 5s, 5t), delivering the corresponding biaryls in high yields. Furthermore, electronically-deactivated nitroarenes could be cross-coupled with electronically-deactivated boronic acids (5g, 5h) in good yields. In addition, we note that steric hindrance on either nitroarene (5i) or boronic acid component (5q) is tolerated. Finally, heterocyclic boronic acids (5v) and alkyl boronic acids (5w, 5x) also work well using our protocol.

Table 4.

Scope of Suzuki-Miyaura Cross-Coupling of Nitroarenes using Well-Defined [(ImPyMesDipp)Pd(cin)Cl]^a

^aConditions: nitroarene (1.0 equiv), boronic acid (1.5 equiv), [(ImPyMesDipp)Pd(cin)Cl] (5 mol%), K₃PO₄ (3.0 equiv), H₂O (3.0 equiv), TDA (10 mol%), dioxane (0.2 M), 130 °C, 36 h.

^b16 h. c 2 equiv of boronic acid. See SI for details.

Ligand Synthesis.

In light of the intriguing reactivity of sterically-restricted C5-aryl substituted imidazo[1,5-a]pyridin-3-ylidenes, we next developed significantly improved route for ligand synthesis using Ni-catalyzed Kumada cross-coupling and no chromatographic purification (Scheme 2).

Scheme 2.

Modular and Chromatography-Free Synthesis of ImPy Ligands via Ni-Catalyzed Kumada Cross-Coupling

To date, methods to synthesize C5-aryl substituted imidazo[1,5-a]pyridinium precursors to ImPy ligands rely on Pd-catalyzed Suzuki cross-coupling to introduce the aryl group at the C5 position.^53,62–63 However, there are three major shortcomings of this frequently used route: (1) mesitylene-2-boronic acid (100g/$206) and 2,4,6-triisopropylbenzeneboronic acid (100g/$401) are expensive;⁶⁴ (2) Suzuki cross-coupling of sterically-hindered boronic acids is challenging;⁶⁵ (3) chromatographic purification is usually needed due to unsatisfactory yields of the Suzuki cross-coupling.

Having determined the high reactivity of well-defined Pd(II)–NHC cinnamyl complexes of ImPyMesDipp and ImPyTrippMes, we studied new synthetic approach to this class of ligands. As shown in Scheme 2, starting from the commercially available aldehyde (6), acetal (7) could be obtained in near quantitative yield after aqueous work-up. Subsequently, the Ni-catalyzed Kumada cross-coupling between acetal (7) and freshly prepared Grignard reagents (8) delivered the key intermediates (9) in high purities (>95%) and yields (>90%) after extraction, followed by a filtration through a short pad of silica gel. Finally, condensation of (9) directly with the corresponding aniline and paraformaldehyde in the presence of HCl furnished the C5-aryl substituted imidazo[1,5-a]pyridinium salts in excellent yields (82–95%). Acetal deprotection of (9) in a separate step is not required.

There are several noteworthy features of our synthetic route: (1) significantly more efficient Kumada cross-coupling using cheap and abundant 3d transition metal and at lower catalyst loading (1 mol% Ni vs. 5 mol% Pd); (2) feedstock bromoarenes instead of expensive boronic acids as coupling partners (2,4,6-trimethylbromobenzene, 100g/$32; 2,4,6-triisopropylbromobenzene, 100g/$76), which are significantly cheaper than the corresponding boronic acids;⁶⁴ (3) obviating chromatographic purification throughout the synthesis, which permits for a scalable and modular ligand synthesis. We anticipate that this streamlined and economic route will become a major method in expanding the utility of imidazo[1,5-a]pyridin-3-ylidenes in organic synthesis and catalysis.⁵⁴

Orbital Analysis.

To provide detailed electronic characterization of sterically-restricted imidazo[1,5-a]pyridin-3-ylidene ligands, HOMO and LUMO energy levels of 1a–1d in comparison with the classical imidazol-2-ylidenes IMes and IPr were determined (Table 5 and Figure 6). It is now well-established that computation of frontier molecular orbitals gives the most accurate determination of ligand nucleophilicity (more σ-donating, higher HOMO) and electrophilicity (more π-accepting, lower LUMO), while in some cases π-donation should also be considered. The HOMO-1 (σ-bonding orbital) in the series of ImPyMesMes, ImPyMesDipp, ImPyTrippMes, ImPyTrippDipp are -5.86eV, -5.91 eV, -5.87 eV, -5.91 eV, which is in the same range as for the standard imidazol-2-ylidene ligands IMes (-5.90 eV) and IPr (-6.01 eV).

Table 5.

HOMO and LUMO Energy Levels (eV) of Imidazo[1,5-a]pyridin-3-ylidenes Calculated at the B3LYP 6–311++g(d,p) Level^a

entry	NHC	HOMO [eV]	LUMO [eV]
1	1a	−5.86b	−1.24
2	1b	−5.91b	−1.29
3	1c	−5.87b	−1.25
4	1d	−5.91b	−1.30
5	IMes	−5.90	−0.33c
6	IPr	−6.01	−0.48c

^aSee SI for details.

^bHOMO-1, in-plane σ-orbital. HOMO: π-donor orbital, 1a: -5.43 eV, 1b: -5.49 eV, 1c: -5.44 eV; 1d: -5.50 eV.

^cLUMO+2, LUMO+1 due to required symmetry.

Figure 6.

HOMO (π-donating orbital), HOMO-1 (σ-donating orbital) and LUMO (π-accepting orbital) of imidazo[1,5-a]pyridin-3-ylidene 1b. B3LYP 6-311++g(d,p) level. See SI for details.

Furthermore, the LUMO (π-accepting orbital) in the series of ImPyMesMes, ImPyMeDipp, ImPyTrippMes, ImPyTrippDipp are - 1.24 eV, -1.29 eV, -1.25 eV, -1.30 eV, which is much lower than for the standard imidazolylidene ligands IMes (-0.33 eV), IPr (-0.48 eV). The LUMO+1 orbital is located on the N-Ar ring (-0.43 eV,-0.50 eV, -0.45 eV, -0.53 eV) in the series. The results can be compared with the orbital analysis of ImPy carbenes as π-accepting ligands by Shibahara.⁶⁶ Interestingly, the HOMO (π-donating orbital) in the series of ImPyMesMes, ImPyMesDipp, ImPyTrippMes, ImPyTrippDipp are -5.43 eV, -5.49 eV, -5.44 eV, -5.50 eV, which is much higher than the π-donating orbital for the standard imidazol-2-ylidene ligands IMes (-6.44 eV), IPr (-6.55 eV).

The data indicate that (1) sterically-restricted imidazo[1,5-a]pyridin-3-ylidenes are strongly nucleophilic ligands, as expected for sterically-bulky N-Ar NHC ligands, with σ-donation matching those of IPr and IMes and stronger than phosphines, and (2) significantly better π-acceptors than the standard imidazolylidene IPr and IMes ligands. Furthermore, (3) imidazo[1,5-a]pyridin-3-ylidene ligands are characterized by strong π-donating abilities.

Geometry Analysis.

To further determine steric properties of this family of ligands and eliminate effects from the crystal packing, analysis of %V_bur of linear [Cu(NHC)Cl] complexes by DFT computations was performed (Figure 7). Cu(I)–NHC complexes were selected as model linear M–NHC complexes.⁶⁰

Figure 7.

(A-D) Topographical steric maps of linear imidazo[1,5-a]pyridin-3-ylidenes [Cu(NHC)Cl] (1a–1d) showing % Vbur per quadrant. B3LYP 6-311++g(d,p) level. See SI for details.

The %buried volume (%V_bur) of [Cu(NHC)Cl] in the series of ImPyMesMes, ImPyMesDipp, ImPyTrippMes, ImPyTrippDipp are 46.8%, 50.1%, 49.8%, 52.4%. Imidazo[1,5-a]pyridin-3-ylidenes with C5/N2 steric matching combination are sterically similar (%V_bur = 49.8%-50.1%). C5/N2 substitution with sterically-more demanding substituents results in a decrease in size of catalytic pocket (%V_bur = 52.4%), while C5/N2 substitution with less sterically-demanding substituents leads to low %buried volume in this series of ligands (%V_bur = 46.8%). Quadrant distribution of ImPyMesDipp is 55.3%, 55.3%, 45.0%, 45.0%, and for ImPyTrippMes is 60.9%, 61.4%, 38.5%, 38.2%. The %buried volume values can be compared with the reference imidazolylidene [Cu(IMes)Cl], [Cu(IPr)Cl] with (%V_bur) of 36.4%, 42.6%. Furthermore, quadrant distribution can be compared with the reference imidazolylidene [Cu(IPr)Cl] of 41.0%, 44.2%, 41.0%, 44.2% and [Cu(IMes)Cl] of 36.4%, 36.4%, 36.4%, 36.4%.

The data indicate (1) significant increase of steric demand of sterically-restricted imidazo[1,5-a]pyridin-3-ylidene ligands (ImPyIMesIMes of 46.8% vs. IMes of 36.4%); (2) sterically-fixed conformationally-rigid imidazo[1,5-a]pyridin-3-ylidene framework and size-controlled N2-substitution (flexible quadrants). This results in steric features in combination with steric matching between the C5/N2 substituents that are not available in standard imidazolylidenes.

Conclusion

In conclusion, we have reported [(NHC)Pd(cinnamyl)Cl] complexes of ImPy (ImPy = imidazo[1,5-a]pyridin-3-ylidene) as sterically-restricted class of precatalysts for cross-coupling reactions. These precatalysts are fast-activating to monoligated Pd(0), feature 1:1 Pd to ligand ratio, are air- and moisture-stable, and do not require glovebox techniques for the synthesis and application. Steric matching of the fixed C5-aryl and flexible N2-wingtip substituents led to the discovery of ImPyMesDipp as highly effective and practical imidazo[1,5-a]pyridin-3-ylidene ligand for Pd-catalyzed cross-coupling of nitroarenes by challenging C–NO₂ activation. Crystallographic studies demonstrated the unsymmetrical steric encumbrance of the catalytic pocket crucial for catalytic activity. Computational studies established electronic characteristics of this class of ligands. Importantly, we also disclosed a novel synthetic route to the C5-aryl substituted imidazo[1,5-a]pyridin-3-ylidene ligands. Features of our route include efficient and modular Ni-catalyzed Kumada cross-coupling, the use of feedstock starting materials, and chromatography free synthesis, enabling scalable and modular synthesis of conformationally-restricted ImPy ligands. On the basis of high activity of well-defined Pd(II)–NHC cinnamyl complexes of imidazo[1,5-a]pyridin-3-ylidenes, we anticipate that these catalysts will find broad utility in activation of unreactive bonds in organic synthesis and catalysis. Studies to expand the scope of reactions enabled by ImPy ligands are ongoing and will be reported in due course.⁵⁴