An Air-Stable, Single-Component Iridium Precatalyst for the Borylation of C−H Bonds on Large to Miniaturized Scales

Abstract

The functionalization of C−H bonds enables the modification of complex molecules, often with the intention of forming compound libraries. The borylation of aryl C−H bonds is a widely used class of C−H bond functionalization, and conventional catalyst systems for the borylation of C−H bonds consist of an iridium source and an N,N-ligand, in conjunction with pinacolborane, to form the active iridium(III) tris(boryl) catalyst. These multicomponent catalyst systems complicate borylation reactions at large and small scales, due to the air sensitivity of the most common iridium precursor [Ir(cod)OMe]₂, and, particularly on small scale, the challenges associated with dispensing multiple components with differing solubilities or that are air-sensitive. We describe the discovery of an air-stable, single-component iridium precatalyst, [(tmphen)Ir(coe)₂Cl], that generates the same active iridium(III) tris(boryl) catalyst and reacts with higher turnovers, comparable selectivity, and similar scope to those of known catalyst systems for the borylation of aryl and heteroaryl C−H bonds. We show how the development of this precatalyst enables reactions to be run on submicromole scale in a high-throughput experimentation format in conjunction with ChemBead technology, and with a second diversification step that illustrates the potential to diversify structures by chemical sequences involving catalytic reactions, including C−H bond functionalizations, on submicromole scales in the same reaction vessel.

Graphical Abstract

1. INTRODUCTION

The direct functionalization of C−H bonds has become an important strategy for the construction and modification of organic molecules.^1–9 Among the existing methods for the functionalization of aromatic and heteroaromatic C−H bonds, iridium-catalyzed borylation is widely used because the installed boronic ester can undergo a variety of subsequent reactions, including cross-coupling and well-known oxidative functionalization reactions.⁹ The borylation reaction occurs under mild conditions with high regioselectivity and tolerance for functional groups, and it produces highly versatile, organoboronic ester products without the use of prefunctionalized substrates.^10,11 In most cases, the site selectivity^12–14 is orthogonal to the site selectivity of classic methods for the functionalization of C−H bonds, such as electrophilic aromatic substitutions, and it can be controlled by the ligand architecture, often through noncovalent interactions.^15–17

The catalyst composition used most frequently for the borylation of C−H bonds comprises an iridium precursor, often [Ir(cod)OMe]₂ (cod = 1,5-cyclooctadiene), and a phenanthroline- or bipyridine-based ligand, such as 3,4,7,8-tetramethylphenanthroline (tmphen) or 4,4′-di-tert-butylbipyridine (dtbpy) (Scheme 1). This combination of metal precursor and ligand, in conjunction with pinacolborane (HBpin), generates an iridium(III) tris(boryl) complex in situ that undergoes oxidative addition of a C−H bond of the substrate.¹⁸ Subsequent reductive elimination from an iridium-(V) intermediate forms the borylated product.

Scheme 1.

Typical Conditions and Catalyst Systems for the Borylation of Aromatic C−H Bonds

The yield of the iridium(III) tris(boryl) catalyst from [Ir(cod)OMe]₂ is sensitive to the order of addition of components, with high yields requiring the addition of pinacolborane (HBpin) followed by ligand.^18,19 This protocol to generate the active catalyst has been reported to complicate reactions on both large and small scales.^20–22 In one example, researchers at Pfizer reported that [Ir(cod)OMe]₂ was an unreliable catalyst precursor for the borylation of nicotine on 10 kg scale because it is unstable when stored under ambient conditions.²⁰

For such reasons, the utility of transition-metal-catalyzed reactions can be increased by creating precatalysts that are preligated and air-stable, such as the Hoveyda−Grubbs complexes for olefin metathesis and the Buchwald precatalysts for cross-coupling reactions.^23,24 Colacot reported that the combination of 1,10-phenanthroline (phen) and [Ir(cod)Cl]₂ in tetrahydrofuran (THF) forms [(phen)Ir(cod)Cl], which is an air-stable precatalyst for the borylation of aryl and heteroaryl C−H bonds.^25,26 However, this precatalyst was reported to require premixing with isopropanol or HBpin at 80 °C prior to the addition of substrate to generate the active catalyst.²⁶ Moreover, this precatalyst contains 1,10-phenanthroline, which forms an active catalyst with a shorter lifetime than that containing tmphen, due to competing borylation of the aryl C−H bonds of the ligand.²⁷

While the value of a synthetic method is often measured by its ability to be conducted on a gram, decagram, or even kilogram scale, it is often more challenging to conduct catalytic reactions on miniaturized scales. Such scales are valuable for maximizing the amount of experimental data per milligram of material in current and future medicinal chemistry platforms^28–30 by coupling high-throughput experimentation (HTE) campaigns with advanced analytical methods and assays of biological activity.³¹ However, C−H bond functionalization reactions have not been reported on such scales in an array format, despite the opportunity it provides to diversify existing compound libraries of arenes and heteroarenes that are more widely available than pre-functionalized analogs, such as aryl halides or aryl boronates. Prior catalytic reactions on submicromole scales have been limited to cross-coupling reactions, and prior multistep reactions on this scale that include catalytic transformations, even two-step sequences, have not been reported.²⁸

We envisioned that C−H bond functionalization could enable the diversification of molecules on such scales. Small-scale borylations have been reported, but only in the context of reaction optimization with as little as 20 μmol of arene.^19,32

To increase the value of the borylation of C−H bonds from the smallest to largest scales, we sought an air-stable, single-component precatalyst containing tmphen that does not require additives, such as HBpin or an alcohol, to become active under mild conditions and that catalyzes borylations with activity and selectivity that is comparable to commonly used two-component systems. Such a precatalyst would facilitate reactions on a large scale by avoiding air-sensitive precursors or additional activators, thereby enabling controlled formation of the active catalyst. Moreover, such a precatalyst would simplify reactions on miniaturized scales, such as those with nanomoles to micromoles of arene substrate, because combining catalyst components that are unstable or poorly soluble in the reaction media becomes increasingly difficult as reaction volumes are minimized.

We show that [(tmphen)Ir(coe)₂Cl] (coe = cyclooctene) is an air-stable, single-component catalyst for the borylation of aryl C−H bonds that generates the active catalyst with just the two reactants, B₂pin₂ and arene. [(tmphen)Ir(coe)₂Cl] is air-stable and more consistently leads to high yields on typical laboratory or less typical miniaturized scales than prior single-component catalysts or multicomponent systems that include air-sensitive constituents. We demonstrate that borylations with this precatalyst occur with an array of arenes and heteroarenes on multimillimole to decagram scales in a glovebox or on the benchtop, and they occur, in combination with ChemBead technology,^33–35 on submicromole scales. Reactions in a high-throughput experimentation format with a second diversification step in the same reaction vessel illustrate the ability to conduct chemical sequences on this miniaturized scale.

2. RESULTS AND DISCUSSION

2.1. Precatalyst Development.

To develop a single-component precatalyst that is both air-stable and highly active, we began by synthesizing a series of Ir complexes derived from 3,4,7,8-tetramethylphenanthroline (tmphen). We selected tmphen as the ligand for catalyst development, as opposed to 4,4′-di-tert-butylbipyridine (dtbpy), because it forms a catalyst that is substantially longer lived and this ligand is relatively inexpensive.^27,36 We also sought a catalyst that would contain an X-type ligand that would reduce the air and moisture sensitivity relative to that of the methoxide complex and one that would contain a ligand that would dissociate more readily than 1,4-cyclooctadiene (cod). We presume that [Ir(phen)(cod)Cl] required activation with alcohol at higher temperatures because it is an 18-electron complex containing chelating ligands,²⁵ and that the catalyst generated from [Ir(cod)(OMe)]₂ required HBpin to generate the iridium(III) tris(boryl) complex in high yield because the cod ligand binds tightly to the iridium(I) species. Although many iridium(I) complexes are air-sensitive, those containing either alkene or chloride ligands tend to be more stable to air, suggesting that an air-stable but reactive precatalyst could be identified.

With these properties and goals in mind, we prepared a series of tmphen-ligated Ir complexes, [(tmphen)Ir(coe)₂Cl], [(tmphen)Ir(cod)Cl], [(tmphen)Ir(cod)OSiMe₃], and [(tmphen)Ir(cod)OMe], from [Ir(coe)₂Cl]₂ or [Ir(cod)X]₂.³⁷ The activity of these precatalysts was assessed for the model borylation reaction of 2-bromo-1,3-dimethoxybenzene (1) with B₂pin₂ (1 equiv) at 50 °C in THF (0.5 M) (Table 1). The reaction conducted with [(tmphen)Ir(cod)Cl] or [(tmphen)Ir(cod)OSiMe₃] did not lead to any of the boronic ester product 2 (entries 1, 2 and 5, 6) under these conditions, but reactions catalyzed by [(tmphen)Ir(cod)OMe], and [(tmphen)Ir(coe)₂Cl] produced comparable amounts of product 2 when the reaction was stopped at 15 min (entries 3 and 4), suggesting that they form similar amounts of active catalyst. After 1.5 h, reactions catalyzed by either [(tmphen)Ir(cod)OMe] or [(tmphen)Ir(coe)₂Cl] formed the borylated product 2 in high yields (entries 7 and 8).

Table 1.

Comparison of the Yield of Borylation of 2-Bromo-1,3-Dimethoxybenzene for Different Single-Component Precatalysts^a

entry	precatalyst	time	yield (%)
1	[(tmphen)Ir(cod)Cl]	15 min	0
2	[(tmphen)Ir(cod)OSiMe3]	15 min	0
3	[(tmphen)Ir(cod)OMe]	15 min	15
4	[(tmphen)Ir(coe)2Cl]	15 min	17
5	[(tmphen)Ir(cod)Cl]	1.5 h	0
ó	[(tmphen)Ir(cod)OSiMe3]	1.5 h	0
l	[(tmphen)Ir(cod)OMe]	1.5 h	80
8	[(tmphen)Ir(coe)2Cl]	1.5 h	85
9	[(tmphen)Ir(cod)OMe]	1.5 h	10b
10	[(tmphen)Ir(coe)2Cl]	1.5 h	80c

^aConditions: Reactions conducted with 1 equiv (0.25 mmol) of 1. ¹H NMR yields of 2 are reported and were determined using CH₂Br₂ as internal standard.

^bUsing [(tmphen)Ir(cod)OMe] that was allowed to sit in a 20 mL vial on the benchtop open to air for 7 days.

^cUsing [(tmphen)Ir(coe)₂Cl] that was allowed to sit in a 20 mL vial on the benchtop open to air for 1 year.

However, the air stability of the chloride complex was much greater than that of the alkoxide or siloxide complexes. The ¹H NMR spectrum of [(tmphen)Ir(coe)₂Cl] that had been left on the bench exposed to air for 1 year was the same as that of the freshly prepared complex. In contrast, the siloxide and methoxide complexes rapidly decomposed upon exposure to air, as judged by ¹H NMR spectroscopy. Moreover, after storing on the benchtop for 7 days, the methoxide complex lost activity and formed product 2 in only 10% yield after 1.5 h (entry 9), whereas the activity of the chloroiridium cyclooctene complex was the same after a year of storage in air (80% yield, entry 10) as when freshly made, and the synthesis is accomplished by simply mixing the commercially available iridium precursor [Ir(coe)₂Cl]₂ with tmphen in diethyl ether, followed by filtration to afford the compound as a red solid in quantitative yield. Thus, [(tmphen)Ir(coe)₂Cl] was found to be a practical, air-stable precatalyst for the borylation of aryl C−H bonds and was chosen for subsequent studies.

The bonding of [(tmphen)Ir(coe)₂Cl] was investigated by single-crystal X-ray diffraction to study the structural features that contribute to its air stability and high activity as a precatalyst. The structure (Figure 1) consists of a distorted trigonal-bipyramidal geometry about the iridium center. The C−C bond lengths of the bound alkenes of 1.46 and 1.43 Å (versus 1.33 Å in ethylene)³⁸ reflect strong back bonding from iridium and stronger donation by the tmphen ligand than by the trialkyl bisphosphine in (dmpe)Ir(cod)Cl which contains alkenes with bond lengths of 1.44 and 1.40 Å.³⁹ Such back bonding reduces the electron density at iridium and slows the dissociation of cyclooctene in the solid state, thereby suppressing the reaction with oxygen. In solution, however, the alkene dissociates more readily than the chelating cod ligand in [(tmphen)Ir(cod)Cl], and this dissociation facilitates reaction with B₂pin₂ to generate the active iridium(III) tris(boryl) catalyst. The resistance of nitrogen-based ligands to oxidation and the presence of the weakly electron-donating chloride, as opposed to more electron-donating methoxide, also contribute to air stability.

Figure 1.

Structure of [(tmphen)Ir(coe)₂Cl] determined by single-crystal X-ray diffraction. Hydrogen atoms have been omitted for the sake of clarity. Selected bond lengths (Å) and angles: C1−C2:1.432(6), C9−C10:1.459(7), Ir−Cl: 2.143(5), Ir−N1:2.146(4), Ir−N2:2.032(3), N1−Ir−N2:79.3(1)°, Cl−Ir−N1:92.5(1)°.

2.2. Investigation of Catalyst Initiation.

A comparison of the rate at which the air-stable precatalyst and a multicomponent system form the active catalyst was made. [Ir(cod)OMe]₂ forms the tris(boryl) iridium(III) catalyst in higher yield than does [Ir(cod)Cl]₂,^16,20 and Colacot found that the activity of the catalyst generated from [(phen)Ir(cod)Cl] was higher in the presence of alcoholic additives than in their absence.²⁶ Thus, it was not clear if [(tmphen)Ir(coe)₂Cl] would generate the catalyst in yields and rates comparable to those of the conventional system.

We investigated the initiation of the precatalyst by studying the borylation of 2-methylbenzoxazole (3) to form boronic ester 4 over an hour catalyzed by [(tmphen)Ir(coe)₂Cl] and found a similar rate of reaction as with a combination of [Ir(cod)OMe]₂ and tmphen (Figure 2). This result suggests that the single-component system forms the active catalyst with a rate that is similar to that for the formation of the active catalyst from the multicomponent system. For comparison, the reported reaction of 3 catalyzed by the combination of [(phen)Ir(cod)Cl] (0.5 mol %), HBpin (0.5 mol %), and B₂pin₂ (1.1 equiv) in THF (0.3 M) did not give any boronic ester product (4) after 1 h at the higher temperature of 75 °C.²⁶

Figure 2.

Reaction time courses for the borylation of 2-methylbenzoxazole (3, 0.1 M) with [(tmphen)Ir(coe)₂Cl] or a combination of [Ir(cod)OMe]₂ and tmphen as the precatalyst.

2.3. Investigation of Catalyst Speciation.

While both single- and two-component catalyst systems rapidly form active catalysts under typical conditions for the borylation of aryl C−H bonds, we sought to characterize the species that are formed from the precatalysts and to identify potential differences in their modes of activation. To this end, we prepared a phosphite-capped iridium tris(boryl) complex in situ by combining (MesH)Ir(Bpin)₃⁴⁰ and tmphen in THF with excess B₂pin₂, followed by the addition of triphenyl phosphite. While the ³¹P NMR spectrum of this mixture consisted of a single broad singlet centered at +107 ppm (eq 1, Figure S1a), the proton NMR spectrum consisted of at least three related tris(boryl) species resulting from mono- and diborylation of the tmphen ligand (Figure S1b).²⁷ This borylation of the tmphen was suppressed by replacing B₂pin₂ with excess cyclooctene (coe). The combination of (MesH)Ir(Bpin)₃,⁴⁰ tmphen, and coe in THF, followed by the addition of triphenyl phosphite, afforded a pure sample of [Ir(tmphen)(P(OPh)₃)(Bpin)₃], whose structure was unambiguously confirmed by single-crystal X-ray diffraction (Figure S2). The ³¹P NMR spectrum of this complex consisted of a broad singlet with a chemical shift of +106 ppm, which is nearly identical to that obtained with B₂pin₂ (Figure S3).

(1)

The speciation of the single-component catalyst system was then evaluated by mixing the iridium precursor in the presence of B₂pin₂ (as would be the case for a typical catalytic reaction), followed by capping with triphenyl phosphite. We found that the phosphite complex derived from treatment of the single-component precatalyst [(tmphen)Ir(coe)₂Cl] with B₂pin₂ gave rise to a distinct species corresponding to a resonance at +90 ppm in the ³¹P NMR spectrum (eq 2, Figure S4a). The ¹H NMR spectrum, in addition to elemental analysis of the isolated complex, is consistent with its assignment as the bis(boryl) chloride [Ir(tmphen)(P(OPh)₃)(Bpin)₂Cl], formed by the oxidative addition of B₂pin₂ to the iridium(I) precatalyst without subsequent reaction of the chloride ligand (Figure S4b).

(2)

The speciation of the iridium catalyst under the conditions of a typical catalytic reaction was determined with the model arene, julolidine,²⁷ and B₂pin₂ (1 equiv) in THF with the multicomponent and single-component catalyst systems (2 mol % Ir). Heating at 80 °C, followed by the addition of triphenyl phosphite at various reaction times, revealed that both catalyst systems formed the same tris(boryl) iridium complex (δ(³¹P) = +107 ppm), which is analogous to that formed from (MesH)Ir(Bpin)₃ and tmphen in the presence of B₂pin₂, before two catalyst turnovers (eq 3, Figure S5). At an earlier time point (5 min) of the borylation conducted with [(tmphen)Ir(coe)₂Cl], a small (ca. 30% of the total) signal corresponding to [Ir(tmphen)(P(OPh)₃)(Bpin)₂Cl] (+90 ppm) was observed, but this species completely converted to the tris(boryl) complex within 15 min. These results show that both the one- and two-component catalyst systems form the same active tris(boryl) iridium(III) catalyst, even though small differences were observed in the modes of catalyst activation.

(3)

2.4. Scope of the Borylation Reaction Catalyzed by [(tmphen)Ir(coe)₂Cl].

Table 2 shows a series of borylations of arenes and heteroarenes with our single-component catalyst. In every case, the borylation conducted on the benchtop in a nitrogen-purged reaction vessel occurred in yields and regioselectivities that are comparable to those of reactions conducted in a nitrogen-filled glovebox, as judged by ¹H NMR spectroscopy of the crude reaction mixtures with an internal standard (see the SI for detailed comparisons of literature yields with two-component catalyst systems).

Table 2.

Scope of the Borylation of Arenes and Heteroarenes with [(tmphen)Ir(coe)₂Cl]^{a,b,c,d,e,f,g,h,i}

^aConditions: Reactions conducted with 1 equiv (0.25 mmol) of 5a−y. ¹H NMR yields are reported first and isolated yields in parentheses for reactions carried out in an N₂-filled glovebox.

^b1H NMR yields for reactions carried out on the benchtop with air-free technique as outlined in the Supporting Information.

^c1H NMR yield is reported as borylated product which was subsequently transformed to the arylchloride for purification due to the instability of the aryl boronic ester.

^dRun in heptane.

^eRun on 10 g scale.

^fRun with 0.05 mol % catalyst.

^gRun on 10 mmol scale.

^hRun on 10 μmol scale.

ⁱRun with 5 mol % catalyst.

Table 2 includes examples of the activity and selectivity of [(tmphen)Ir(coe)₂Cl] for several previously unreported borylations of bioactive compounds. The borylation of ticlopidine (6v), an antiplatelet drug, occurred in good yield and selectively at the 5-position of the thiophene over borylation on the 1,2-disubstituted benzene. Papaverine, a cardiovascular drug, underwent selective borylation at the 4-position of the quinoline system to afford intermediate aryl boronic ester 6y in moderate yield.

Other cases allowed comparisons to examples conducted previously on a kilogram scale. The high selectivity and yields for the borylation of nicotine (5w) are comparable to those reported by Pfizer on 10 kg scale.¹⁷ We conducted this borylation with [(tmphen)Ir(coe)₂Cl] on 10 mmol and 10 μmol of 5w to illustrate the broad range of scales at which the reactions can be conducted when initiated by the single-component precatalyst. We also evaluated the large-scale borylation of 2-chloro-6-methylpyridine (5u) to form boronic ester 6u, an intermediate in the manufacturing route to a drug candidate that was the subject of recent optimization for scaleup to 27 kg batches.⁴¹ The reaction of 5u on a 10 g scale with 0.05 mol % [(tmphen)Ir(coe)₂Cl] formed boronic ester 6u in quantitative yield and in comparable isolated yield to the one reported on the same decagram scale. Thus, this precatalyst can function as an air-stable, drop-in replacement for the prior catalyst system generated by premixing of [Ir(cod)OMe]₂ and dtbpy at elevated temperatures before addition of the other reaction components.⁴¹

2.5. Borylation of a Pharmaceutical Precursor.

To illustrate the advantages of our new single-component catalyst system further, we compared the activity of several iridium precatalysts for the borylation of 3-chloroiodobenzene with B₂pin₂ outside of a glovebox (Table 3). Researchers at Merck & Co. reported a manufacturing route to the reverse transcriptase inhibitor doravirine that commences with the meta-selective borylation of this arene by a two-component precatalyst system comprising [Ir(cod)(OMe)]₂ and 2,2′-bipyridine.⁴² To better discern differences in turnover numbers (TON) between the precatalyst systems, we employed half the reported loading of precious metal (0.5 mol % Ir) under reaction conditions similar to the process route (50 °C, cyclohexane, 0.33 M). With the cost and sourcing of tmphen in mind, we also prepared and tested the air-stable [(phen)Ir(coe)₂Cl] containing unsubstituted phenanthroline.

Table 3.

Comparison of the Yield of Borylation of 3-Chloroiodobenzene with Different Iridium Precatalysts^a

entry	precatalyst (mol %)	yield (%)	TON
1	[(tmphen)Ir(coe)2Cl] (0.5)	>99.5	200
2	[(tmphen)Ir(coe)2Cl] (0.33)	95.9	290
3	[(phen)Ir(coe)2Cl] (0.5)	75.9	152
4	[(phen)Ir(cod)Cl]b (0.5)	0	0
5	[Ir(cod)OMe]2/bpy (0.25/1)	96.3	193
6	[Ir(cod)OMe]2/bpy (0.25/0.5)	88.0	116
7	[Ir(cod)OMe]2/tmphen (0.25/0.5)	96.3	193
8	[Ir(cod)OMe]2/tmphenc (0.25/0.5)	0	0
9	[Ir(cod)Cl]2/tmphen (0.25/0.5)	13.1	26
10	[(tmphen)Ir(coe)2Cl] (0.1)	82.0	820
11	[Ir(cod)OMe]2/bpy (0.05/0.2)	32.7	327
12	[Ir(cod)OMe]2/bpy (0.05/0.1)	14.8	148

^aConditions: Reactions (entries 1−9) were conducted with 1 equiv (1 mmol) of 7 on the benchtop with air-free technique. Reactions (entries 10−12) were conducted with 1 equiv (4.5 mmol) of 7 in an N₂-filled glovebox. ¹H NMR yields of 8 are reported and were determined with CH₂Br₂ as internal standard.

^bIsopropanol (0.5 mol %) was added as activator.

^cUsing [Ir(cod)OMe]₂ that had been stored outside the glovebox.

TON, turnover number.

Compared to previously reported multicomponent catalyst systems, only our single-component precatalyst [(tmphen)Ir(coe)₂Cl] quantitively produced the boronic ester product 8 under these conditions. Reactions with a catalyst loading of only 0.33 mol % (entry 2) occurred in high yield with 290 turnovers. This turnover number is nearly 50% greater than that of the combination of tmphen and [Ir(cod)OMe]₂ (entry 7).

Furthermore, our new precatalyst was superior to the two-component system used under the reported process conditions (entries 5−6), which include [Ir(cod)OMe]₂ as the iridium source. This difference is evident from reactions with only 0.1 mol % catalyst that were assembled in an N₂-filled glovebox instead of the benchtop (entries 10−12). An almost 3-fold greater turnover number was measured for the reaction catalyzed by [(tmphen)Ir(coe)₂Cl] (entry 10) than for the reaction catalyzed by the combination of bpy and [Ir(cod)OMe]₂ (entry 11).

Entries 7 and 8 illustrate the liabilities of using [Ir(cod)OMe]₂ as the iridium precursor in air. While the combination of tmphen and [Ir(cod)OMe]₂ that had been stored in the glovebox produced the boronic ester product 8 in high yield (entry 7), the reaction conducted with tmphen and [Ir(cod)OMe]₂ that had been stored in a sealed vessel outside the glovebox gave no measurable product (entry 8). Researchers at Pfizer specifically noted that the air sensitivity of [Ir(cod)OMe]₂ led them to use [Ir(cod)Cl]₂ for the large-scale borylation of nicotine;²⁰ however, reactions of 3-chloroiodobenzene conducted with [Ir(cod)Cl]₂ as the precursor under similar conditions occurred in a substantially lower yield of the desired product (entry 9).

In the same vein, the use of the single-component precatalyst previously described by Colacot,^25,26 which contains an unsubstituted phenanthroline ligand and is reported to require isopropanol or HBpin as an additional activating reagent, did not produce any boronic ester product under these conditions (entry 4). In contrast, the phenanthroline analogue of our new single-component catalyst, [(phen)Ir(coe)₂Cl], gave the product in 76% yield without any additional activator (entry 3).

Further studies were conducted to evaluate how different preactivation conditions affect the yields of the borylation of 3-chloroiodobenzene and other (hetero)arenes catalyzed by Colacot’s system (see SI Figure S13). Under the conditions used to generate the active catalyst from our single-component catalyst (stirring [Ir] and B₂pin₂ at room temperature), borylation reactions conducted with Colacot’s system give little to no yield. Even when using the preactivation conditions described by Colacot (heating [Ir], B₂pin₂, and additive at 80 °C for 1 h), the yield was shown to vary, depending on whether the borylation reaction was run in heptane with iPrOH as the additive or in THF with HBpin as the additive.

Taken together, these results demonstrate that the yields and turnover numbers from reactions initiated with our single-component catalyst were both higher than those of the prior multicomponent, state-of-the-art precatalyst while addressing the liability from air-sensitive precursors. In addition, our single-component precatalyst can catalyze borylation without additional activator and thereby relieves the need to vary reaction conditions to obtain acceptable yields of the boronic ester products.

2.6. Miniaturization of the C−H Borylation Reaction for HTE.

On a small scale, automation of a catalytic process comprising multiple catalyst components and reagents can be difficult. Preparing and combining stock solutions of multiple, sensitive catalyst components, such as [Ir(cod)OMe]₂ and HBpin, or with different solubilities, such as [Ir(cod)OMe]₂ and tmphen, leads to varying results when reactions are converted from bench scale to miniaturized scale. These characteristics have been a barrier to the further miniaturization of C−H borylation to submicromole scales.

We envisioned the single-component, air-stable catalyst [(tmphen)Ir(coe)₂Cl] could be delivered in submicromole amounts and address these concerns. Moreover, multistep reactions on this scale, even two-step sequences, have not included catalytic processes and have involved one bond construction and one deprotection.³⁰ Thus, we sought to conduct the combination of borylation and derivatization with this single-component precatalyst on submicromole scales.

To conduct the catalytic reactions on this scale, we loaded the precatalyst on ChemBeads, which are glass beads dry coated with solid catalysts or reagents by using acoustic mixing. They can be handled like ordinary glass beads rather than the catalysts adhered to them. Milligram quantities of bulk material can then deliver micrograms of a given catalyst.^43,44 While many applications use a large number of 200−300 μm diameter ChemBeads (>2 mg or >100 beads) to achieve uniform dosing within the acceptable range of ±10%,^{33,34,45–47} the desire to conduct experiments on a further miniaturized (10- to 50-fold smaller) scale would require a small number (<10 beads) of nonuniform 200−300 μm diameter ChemBeads per reaction, leading to inaccurate delivery of the precatalyst. To address this issue, we coated the precatalyst on larger diameter but uniformly sized 1 mm glass beads. ChemBeads loaded with [(tmphen)Ir(coe)₂Cl] were thus prepared by coating uniformly sized 1 mm diameter glass beads with the iridium precatalyst by using a resonant acoustic mixer to achieve a loading of 39 μg/bead. The amount of solid coated on each glass bead was previously found to be controlled within ±10%.³⁵

Figure 3 shows the results from conducting a sequence of C−H borylation and Suzuki coupling with 24 arenes and heteroarenes in a 24-well HTE format on submicromole scale. Two ChemBeads coated with the iridium precatalyst (0.11 μmol, 27 mol %) were dispensed into each well using a parallel bead dispenser, which we have described previously.³³ B₂pin₂ (0.44 μmol, 1.1 equiv) was then added as a stock solution in CPME using a liquid handling robot, and the mixture was stirred for 10 min, after which time the (hetero)arene (0.05−0.09 mg, 0.4 μmol, 1 equiv) was added as a stock solution in CPME. The reaction was then heated at 80 °C for 20 h with orbital stirring. The total solvent volume and substrate concentration in CPME were 11 μL and 0.04 M, respectively. Because some of the aryl boronic esters decomposed upon liquid chromatography-mass spectrometry (LCMS) analysis, they were converted in a subsequent Suzuki coupling in the same vessel to the corresponding biaryl compounds derived from 1-iodo-3,5-dimethylbenzene. To do so, (1,4-bis(di-tert-butylphosphino)butane)PdCl₂ (0.24 μmol, 60 mol %) coated on ChemBeads was used as the catalyst in conjunction with potassium trimethylsiloxide as a soluble base (see the SI for optimization).^48,49 The amount of product obtained after the Suzuki coupling was measured by integration of the peak corresponding to the product in the LCMS trace, relative to that of an internal standard. The amount of product was normalized to the amount of product in the well containing the most product (C1). For 22 of the 24 arenes evaluated, the expected biaryl product was identified, indicating that the borylation-Suzuki sequence occurred. The borylated products in wells H1 and G2 did not undergo Suzuki coupling under the reported conditions; however, analysis by LCMS after the borylation reaction showed that high conversion to the arylboronates ester did occur in these wells.

Figure 3.

High-throughput evaluation of the borylation of arenes and heteroarenes with [(tmphen)Ir(coe)₂Cl] on single ChemBeads. The color of the cell depicts the normalized amount of product formed after borylation and Suzuki coupling. The highest ratio of product versus an internal standard (salicylic acid) observed by LCMS in the crude reaction mixture was assigned the value of 1 (C1). The ratio of LCMS area of product versus the internal standard obtained in other cells was normalized with respect to the value observed in well C1. ^a The borylated substrates in wells H1 and G2 did not undergo Suzuki coupling under these conditions; however, analysis after the borylation reaction showed that the borylation reaction occurred to high conversion in these wells.

2.7. Scope of the C−H Borylation-Suzuki Coupling Sequence in HTE.

We then evaluated the one-pot sequence comprising the borylation of 12 (hetero)arenes followed by Suzuki coupling of 8 (hetero)aryl halides to generate 96 potential products on a 400 nmol scale (Figure 4, Condition A). The heteroarenes chosen to undergo borylation were selected on the basis of their yields in the one-pot borylation-Suzuki coupling sequence in Figure 3. In addition, the substrates represent a broad array of heteroaryl scaffolds, including those containing N, O, and S atoms in both saturated and unsaturated rings.

Figure 4.

High-throughput evaluation of the one-pot borylation of heteroarenes with [(tmphen)Ir(coe)₂Cl] followed by Pd- or Ni-catalyzed Suzuki coupling with aryl (Condition A) or alkyl (Condition B) halides. The color of the cell depicts the normalized amount of product formed after the Suzuki coupling. The highest ratio of product versus an internal standard (salicylic acid) observed by LCMS in the crude reaction mixture was assigned the value of 1 (D9 and G7 for conditions A and B, respectively). The ratio of LCMS area of product versus the internal standard obtained in other cells were normalized with respect to the value observed in these wells.

Products from the one-pot borylation-Suzuki sequence with a combination of 12 (hetero)arenes and 8 (hetero)aryl halides were observed in 81 of 96 wells. Products were observed with intermediates relevant to medicinal chemistry, including the monomers from Merck’s informer library (F and G). The sequential reactions of aryl halides F and G gave product at a lower frequency than did those of the less complex coupling partners, but they did form product in 15 of 24 reactions.

Similarly, nicotine and ticlopidine (11 and 12), two biologically active substrates, formed the coupled product with a lower frequency than did those with the less complex heteroarenes, but the product was formed in 9 of 16 reactions. Moreover, the high yield observed for the borylation of 11 and 12 (Figure 3, cells G3 and H3) shows that the borylated products did not undergo Suzuki coupling with some of the electrophiles. Overall, this set of experiments establishes the utility of our approach to conducting two transformations in one pot with less than 0.1 mg of limiting reagent.

We also conducted a one-pot sequence comprising the borylation of 12 (hetero)arenes followed by Ni-catalyzed Suzuki coupling of 8 alkyl iodides (Figure 4, Condition B). We previously reported a one-pot sequence of C−H borylation and Ni-catalyzed alkylation of the corresponding boronic ester.⁵⁰ However, the scope of substrates that underwent borylation and coupling with alkyl iodides was limited to substituted arenes and benzofuran as a representative heteroarene. Thus, we evaluated 96 ligands for the Ni-catalyzed reaction of 1-benzyl-3-pincaolboronic ester-1H-pyrazole with 2-iodopropane and identified 3,8-dimesityl-1,10-phenanthroline (3,8-Mes₂phen) as the ligand that gave the highest yield of alkylated product when combined with NiBr₂(DME), 2-propanol as solvent, and sodium tert-butoxide as base (see the Supporting Information).

We found that 81 out of 96 potential products from the one-pot sequence of borylation and alkylation were observed by LCMS. Like the sequence of borylation and Suzuki coupling of aryl halides, a significant product was observed for reactions of the two biologically active substrates, nicotine and ticlopidine (11 and 12). The fewest number of coupled products was observed for the sequence with indole (8). The instability of the 2-boryl indole intermediate and the participation of the indole N−H in side reactions with the alkyl iodide likely led to the low frequency of coupled products formed. Among the reactions of alkyl iodides, those of secondary alkyl iodides (A−D) gave lower yields of the coupled product than did those of primary alkyl iodides (E−H).

2.8. Scope of the C−H Borylation-Chan-Lam Coupling Sequence in HTE.

We also conducted a one-pot sequence comprising C−H borylation of the same 12 (hetero)arenes used in the previous experiment, followed by Chan-Lam coupling with 8 different amines to generate 96 different potential products (Figure 5). Initial investigations revealed that commonly used conditions for Chan-Lam couplings with Cu(OAc)₂ as the oxidative coupling agent and acetonitrile as solvent were heterogeneous and did not yield a product. Thus, we modified the conditions by using DBU as a soluble base and dimethyl sulfoxide (DMSO) as a solvent because copper(II) acetate dissolves in DMSO.

Figure 5.

High-throughput evaluation of the one-pot borylation of (hetero)arenes with [(tmphen)Ir(coe)₂Cl] on single ChemBeads followed by Chan-Lam coupling with amines. The color of the cell depicts the normalized amount of product formed after the Chan-Lam coupling. The highest ratio of product versus an internal standard (salicylic acid) observed by LCMS in the crude reaction mixture was assigned a value of 1 (G7). The ratio of LCMS area of product versus the internal standard obtained in other cells was normalized with respect to the value observed in well G7.

Under these homogeneous conditions, 43 of 96 potential arylamine products were formed. Some of the arenes and heteroarenes giving the products from the two steps in the highest frequency were those that contained a pyridine nitrogen (2−4). The amines that gave the coupled product from the Chan-Lam coupling at the highest frequency included amino-indoline and amino-benzothiophene (F and G). These two heteroarylamines gave the coupled product in 75% or greater of the cases. In addition, 7 of 12 potential arylamine products were observed from the reactions with the least sterically hindered alkylamine (A), and this rate was significantly higher than that for any other alkylamine tested.

2.9. Scope of the C−H Borylation-Bromination Sequence in HTE.

Finally, we evaluated the 24 arenes and heteroarenes from our previous panel in Figure 3 in a one-pot sequence comprising borylation and bromination to afford 24 potential aryl bromide products (Figure 6). We used conditions with CuBr₂ as the bromide source.⁵¹ Of the 24 reactions, 17 yielded brominated product, as observed by LCMS. This experiment included an expanded scope of heteroarenes for this one-pot transformation over that of the published combination of borylation and bromination, which included just three pyridines.

Figure 6.

High-throughput evaluation of one-pot borylation of heteroarenes with [(tmphen)Ir(coe)₂Cl] on single ChemBeads followed by copper-mediated bromination. The color of the cell depicts the normalized amount of product formed after the Suzuki coupling. The highest ratio of product versus an internal standard (salicylic acid) observed by LCMS in the crude reaction mixture was assigned the value of 1 (C1). The ratio of LCMS area of product versus the internal standard obtained in other cells was normalized with respect to the value observed in well C1.

Given the importance of such heterocycles, it is valuable to determine the origin of the absence of the seven aryl bromides that did not form. The sequence to form bromides A1, C2, and D2, presumably, did not occur because the oxidative conditions of the bromination are sufficiently strong to oxidize the heteroarene. The sequence to form E2 led to the deprotection of the pyrazole nitrogen during the bromination step. The formation of heteroarenes B2 and A3 was likely inhibited by the known instability of the corresponding aryl bromide products under ambient conditions. The basis of the lack of formation of B1 is not known.

3. CONCLUSIONS

We have shown that [(tmphen)Ir(coe)₂Cl] is a highly active precatalyst for the iridium-catalyzed borylation of aromatic C−H bonds. This single-component catalyst can be prepared in a single step that is operationally simple and high-yielding from commercially available precursors, affording the complex as an easily stored, air-stable powder. The discovery and development of this preligated complex enable the borylation of a wide variety of arenes and heteroarenes, including medicinally relevant compounds and intermediates, with improved turnovers, reaction rates, selectivity, and scope to conventional catalyst systems. The reliable dispensing of the precatalyst as an air-stable solid without the need for the air-sensitive pinacolborane (HBpin) as an activating reagent simplifies assembly of the borylation reaction from large to small scales.

Using ChemBead technology, the solid iridium catalyst can be consistently dispensed in microgram quantities for high-throughput experimentation methods, enabling the miniaturization of aryl C−H borylation to submicromole scales. Through downstream transformations, we show that the intermediate boronic esters can be diversified on this scale in the same reaction vessel. A wide variety of products, including those derived from medicinally relevant intermediates, were accessed by cross-coupling and halogenation transformations that install further functionality or handles for synthesis.

More broadly, this study provides new insight into the design of new catalysts for the borylation of C−H bonds. Specifically, pairing a ligand that readily dissociates in solution (coe) with an anionic ligand that is weakly electron-donating (Cl) produced an air-stable complex that readily forms the catalytically active iridium(III) tris(boryl) complex under conventional reaction conditions. We anticipate that additional single-component iridium catalysts for the functionalization of C−H bonds, such as those composed of other N,N-ligands that offer distinct reactivity or selectivity to tmphen for the borylation of C−H bonds, can be prepared from the blueprint described herein.