C(sp²)–H Activation with Bis(silylene)pyridine Cobalt(III) Complexes: Catalytic Hydrogen Isotope Exchange of Sterically-Hindered C–H Bonds

Abstract

The bis(silylene)pyridine cobalt(III) dihydride boryl, trans-[^ptolSiNSi]Co(H)₂BPin (^ptolSiNSi = 2,6-[EtNSi(N^tBu)₂CAr]₂ C₅H₃N, where Ar = C₆H₅CH₃, and Pin =pinacolato) has been used as a precatalyst for the hydrogen isotope exchange (HIE) of arenes and heteroarenes using benzene-d₆ as the deuterium source. Use of D₂ as the source of the isotope produced modest levels of deuterium incorporation and stoichiometric studies established modification of the pincer ligand through irreversible addition of H₂ across the silylene leading to catalyst deactivation. High levels of deuterium incorporation were observed with benzene-d₆ as the isotope source and enabled low (0.5 – 5 mol%) loadings of the cobalt precursor. The resulting high activity for C–H activation enabled deuterium incorporation at sterically encumbered sites previously inaccessible with first-row metal HIE catalysts. The cobalt-catalyzed method was also compatible with aryl halides, demonstrating a kinetic preference for chemoselective C(sp²)–H activation over C(sp²)–X (X = Cl, Br) bonds. Monitoring the catalytic reaction by NMR spectroscopy established cobalt(III) resting states at both low and high conversions of substrate and the overall performance was inhibited by the addition of HBPin. Studies on precatalyst activation with cis-[^ptolSiNSi]Co(Bf)₂H and cis-[^ptolSiNSi]Co(H)₂Bf (where Bf = 2-benzofuranyl), support the intermediacy of bis(hydride)aryl cobalt intermediates as opposed to bis(aryl)hydride cobalt complexes in the catalytic HIE method. Mechanistic insights resulted in an improved protocol using [^ptolSiNSi]Co(H)₃ NaBHEt₃ as the precatalyst, ultimately translating onto higher levels of isotopic incorporation.

Graphical Abstract

INTRODUCTION

The deuteration of organic molecules by metal-catalyzed C–H activation has applications ranging from mechanistic studies in site-selective functionalization to drug synthesis where labeled compounds are used throughout the discovery and development process to elucidate a drug molecules’ metabolic profile and to assess its toxicity ^{1, 2, 3} Deuterated molecules also have the potential to serve as pharmaceutical candidates with distinct properties as evidenced by several deuterated small molecules that are currently in clinical trials. For example, deutetrabenazine (Austedo) was recently approved by the FDA for the treatment of tardive dyskinesia caused by Huntington’s disease.^4,5 The slower rate of metabolism at deuterated positions, resulting from kinetic and equilibrium isotope effects, increases bioavailability enabling lower drug dosages. Accordingly, the development of direct catalytic hydrogen isotope exchange methods is of paramount importance given the numerous applications of deuterated and tritiated molecules.

Historically, developments in transition metal-catalyzed hydrogen isotope exchange (HIE) have relied on precious metal catalysts given their established precedent in C–H bond activation.⁶ Increasing attention has been devoted to first-row transition metal alternatives due to their unique selectivity as compared to more commonly used precious metals. Iron compounds bearing electron-donating pincer ligands, for example, operate with predictable, sterically-driven site selectivity, in contrast to iridium catalysts that rely on directing groups for activation of proximal C–H bonds. ⁷ Examples include pyridine(N-heterocyclic) carbene iron bis(dinitrogen) and dialkyl complexes reported by our group and more recently, a bis(phosphine) carbene iron dihydride described by de Ruiter and co-workers (Figure 1A).^8,9 Mechanistic experiments on the former class of compounds and DFT studies on the latter are consistent with iron(II) intermediates responsible for activation of the C(sp²)–H bonds. Nickel hydride complexes supported by α-diimine ligands are effective precatalysts at high catalyst loadings (25 mol%) for the deuteration of heteroarenes and pharmaceuticals using D₂ or T₂ as the source of the isotopic label.¹⁰ In some cases such as MK-6096, a dual orexin receptor antagonist for insomnia, extremely high specific activity was observed. By contrast, C(sp²)–H HIE catalysts based on cobalt¹¹ and manganese¹² are underdeveloped and are currently limited to examples that rely on directed ortho H/D exchange where D₂O serves as the deuterium source. Catalysts based on these metals that operate without directing groups have not yet been reported. Accordingly, next-generation, first-row transition metal catalysts that are more robust and more active are desirable to expand the scope and applications of catalytic deuteration reactions. A significant challenge in catalytic HIE is the deuteration of sterically encumbered C(sp²)–H sites (e.g., ortho to a methyl group). While highly active precious metal catalysts have recently been developed to address this challenge, methods relying on Earth-abundant first-row metal catalysts are lacking.¹³

Figure 1.

A. Hydrogen isotope exchange with iron compounds bearing electron-donating pincer ligands. B. Cobalt HIE catalysts reported in this work.

The continued development of first-row metal complexes supported by more-electron-rich pincers is of interest given the marked difference in reactivity and selectivity arising from the varied electronic properties of the tridentate chelate.² Ligands containing N-heterocyclic silylenes are attractive for this purpose as reports from Driess and coworkers have demonstrated that replacement of more commonly used phosphine and nitrogen donors with N-heterocyclic silylenes results in more electron-rich metal centers.¹⁴ For example, Driess and Cui reported that activation of a bis(silylene)pyridine cobalt(II) dichloride complex with excess NaBEt₃H in the presence of cyclohexene generated an active catalyst for C(sp²)–H borylation.^15a Subsequent studies from our laboratory identified cobalt(III) dihydride boryl and trihydride sodium triethylborohydride complexes as well-defined, single-component precatalysts.^15b Notably, these catalysts are thermally robust and only cobalt(III) complexes were observed as precatalysts or resting states. Because the number of experimental observables are limited, open questions remain as to whether these catalysts undergo redox cycling between cobalt(I) and cobalt(III) oxidation states as well as the nature of the cobalt intermediate responsible for C–H bond activation. It is also possible that metal-ligand cooperative processes are operative during catalytic borylation.

The unique electronic properties of pincer ligands supported by N-heterocyclic silylenes coupled with their thermal robustness prompted exploration of the corresponding cobalt compounds for catalytic hydrogen isotope exchange. Here we describe the application of pyridine bis(silylene) cobalt(III) precatalysts to catalytic HIE in arenes and heteroarenes using benzene-d₆ as the isotope source (Figure 1B). High levels of deuterium incorporation were observed across a range of substrates including sterically encumbered C(sp²)–H bonds that are typically inaccessible with transition metal catalysts were also labeled. A kinetic preference for C(sp²)–H bond activation over C(sp²)–X was observed which contrasts the general reactivity of first-row metals^16,17 and expands the functional group compatibility of the method.

RESULTS AND DISCUSSION

The HIE activity of previously reported pyridine bis(silylene) cobalt (III) precatalysts using D₂ as the deuterium source was evaluated. The catalytic HIE of 1a with D₂ gas in THF using 5 mol% of 1-(H)₂BPin at 80 °C resulted in 50% deuterium incorporation at the 4- and 5-position. Similar results were obtained using 1-(H)₃·NaBHEt₃ (Scheme 1). N,N-dimethylbenzamide (1b) was also evaluated as a substrate. Prior results for the catalytic HIE of 1b in THF and 4 atm of D₂ using 1 mol% pyridine(N-heterocyclic) carbene iron bis(dinitrogen) at 45 °C resulted in 70% and 64% deuterium incorporation at the para- and meta-sites, respectively. When using 5 mol% of 1-(H)₂BPin, 22% deuterium incorporation was observed at the para- and meta-position (Scheme 1). No C(sp³)–H deuterium incorporation was observed at the bonds α to nitrogen. Varying the pressure of D_2, solvent, and temperature did not significantly improve the level of deuterium incorporation.

Scheme 1.

Cobalt-catalyzed HIE using D2 as the Deuterium Source.

^a17 h instead of 24 h.

Previously, our laboratory reported the synthesis of [^ptolSiNSi(H)]CoH₂(H₂) from addition of 4 atm of H₂ to 1-(H)₂BPin or at 80 °C in THF for 16 hours. To gain additional insight into the observed HIE activity and reactivity of bis(silylene) cobalt (III) precatalysts with H₂, stoichiometric studies with H₂ were conducted. Addition of 1 atm of H₂ at room temperature resulted in formation of new diamagnetic compound, identified as (^ptolSiNSi(H)₂)Co(H)₂BPin, and previously reported [^ptolSiNSi(H)]CoH₂(H₂).^15b This same diamagnetic cobalt compound was observed alongside other unidentified cobalt products when 1-(H)₂BPin was treated with 3.05 equivalents of PMe₃. This product arises from generation of [^ptolSiNSi(H)]CoH₂(H₂) followed by capture of H₂ from another equivalent of 1-(H)₂BPin. (See Supporting Information for Details). Cooling a saturated hexanes solution of the unpurified mixture at −35 °C produced crystals suitable for X-ray diffraction and a representation of the solid-state structure of (^ptolSiNSi(H)₂)Co(H)₂BPin is presented in Figure 2. The synthesis of (^ptolSiNSi(H)₂)Co(H)₂BPin arises from addition of dihydrogen to one of the silylene sidearms in 1-(H)₂BPin. Notably, solid state structures resulting from cooperative silicon(II) mediated H₂ cleavage with cobalt complexes have not been reported. Driess and co-workers reported opening of one of the N-heterocyclic silylenes to generate a silyl ligand upon exposure of a is(N-heterocyclicsilylene)xanthene nickel(0) complex to 1 bar of dihydrogen.^14b In the cobalt example, one of the hydrogen atoms (H18 in Figure 2) has transferred to the amidinate, generating a new C(sp³)–H bond and a silicon hydride (Si1–H48 in Figure 2). This process is related to a report by Roesky and co-workers describing cooperative B–H bond activation whereby coordination of silicon to HBPin results in hydride transfer and Si–B bond formation.^14j So and co-workers have also reported a hydride transfer from silicon hydride to an amidinate ligand to generate a diaminosilylene.^14k

Figure 2.

Molecular structure of [(^ptolSiNSi)(H)₂]CoH₂BPin of at 30% probably ellipsoids with hydrogen atoms except H18 and cobalt hydrides omitted for clarity.

To probe the reversibility of the addition of H₂ across the silylene portion of the chelate, a THF-d₈ solution of [^ptolSiNSi(H)₂]Co(H)₂BPin and [^ptolSiNSi(H)]Co(H₂)H₂ was freeze-pump-thawed three times and left to stand under static vacuum (Scheme 2). Monitoring the reaction by NMR spectroscopy established no conversion of the mixture of [^ptolSiNSi(H)₂]Co(H)₂BPin and [^ptolSiNSi(H)]Co(H₂)H₂ to 1-(H)₂BPin, supporting irreversible H₂ addition. Heating the solution to 60 °C led to the conversion of (^ptolSiNSi(H)₂)Co(H)₂BPin to [^ptolSiNSi(H)]CoH₂(H₂).

Scheme 2.

Pincer modification upon addition of H2 to 1-(H)2BPin and Evaluation of the Reversibility of H2 Addition.

Despite observing H/D exchange using D₂ gas as the deuterium source, the overall levels of isotopic incorporation were lower compared to previously reported iron-catalyzed methods. Previously, an inverse pressure dependence was observed with (H₄-^iPrCNC)Fe(N₂)₂, where higher activity was observed at lower pressure. In the case of bis(silylene)pyridine cobalt complexes, rapid oxidation of cobalt(I) takes place to yield cobalt(III) complexes and competitive D₂ versus arene C–H activation from a cobalt(I) deuteride result in overall low levels of isotopic incorporation as the rate of D₂ activation is expected to be significantly faster. It was postulated that switching to a different source of deuterium where activation of the deuterium source does not significantly outcompete substrate C(sp²)–H activation might result in higher levels of deuterium incorporation. Benzene-d₆, a common NMR solvent, was selected due its convenience and method of activation (e.g., C(sp²)–D) and should enable deuterium transfer to arene substrates with overall high levels of isotopic incorporation.

To probe if benzene-d₆ would be a viable deuterium source, the reactivity of 1-(H)₂BPin in C₆D₆ was explored. Heating a benzene-d₆ solution of trans-(^ptolSiNSiCo(H)₂BPin (1-(H)₂BPin) to 60 °C resulted in deuterium incorporation into the cobalt-hydrides as well as in the 4-position of the pyridine (Scheme 3). Continued heating at 60 °C resulted in additional deuterium incorporation at the sterically encumbered C(sp²–H) sites on the p-tolyl substituent of the N-heterocyclic silylene sidearms.

Scheme 3.

Reactivity of 1-(H)₂BPin in Benzene-d₆.^a

^aDeuterium incorporation was determined by ¹H NMR spectroscopy. See Supporting Information for experimental details.

Presumably, deuterium incorporation in 1-(H)₂BPin occurs through a bimolecular process. The observation of deuterium incorporation into the ortho-to-methyl sites of the tolyl substituents demonstrated the unique ability of [(SiNSi)Co] catalysts to activate and ultimately deuterate sterically hindered positions in an arene ring and prompted investigation into the generality of catalytic HIE using benzene-d₆ as the deuterium source (Table 1). Standard conditions for catalytic experiments employed 1 mol% of 1-(H)₂Bpin in a 0.25 M solution of substrate in benzene-d₆ at 80 °C. In many cases, the cobalt-catalyzed reaction proceeded efficiently at lower temperatures; 80 ºC was selected for consistency across a range of arenes and heteroarenes. The progress of each catalytic reaction was monitored by ¹H NMR spectroscopy.

Table 1.

Scope of the Catalytic Deuteration of Arenes with 1-(H)₂BPin using Benzene-d₆ as the Deuterium Source. ^a,b,c,d

^aReactions conducted in J Young NMR tubes. Deuterium incorporation determined by a combination of ¹H, ¹³C, and ¹⁹F NMR spectroscopies. No C(sp³)–H deuterium incorporation was observed for all substrates examined.

^b5.0 mol%

^c3.0 mol%.

^d2.0 mol% and 50 °C.

Complete deuteration (>98% D incorporation) of 2-fluoro-m-xylene (3a) was observed over the course of 24 hours. Notably, high levels of deuterium incorporation were also observed with more electron-rich aromatics lacking fluorine substituents (3b- 3d). In each case examined, high levels of deuterium incorporation were obtained for the C(sp²)–H sites ortho to the alkyl substituents. These sites are inaccessible with previously reported iron and nickel catalysts or in selected cases, exhibit low levels of isotopic incorporation. To further highlight the high activity of 1-(H)₂Bpin, naphthalene (3e) was examined as a substrate. Prior work with pyridine(bis)carbene iron and bis(phosphine) carbene iron dihydride led to high levels of isotopic incorporation at the sterically accessible β-C–H bonds. The α-C–H bonds were unaffected. By contrast, employing 1 mol% of 1-(H)₂Bpin in benzene-d₆ at 80 °C resulted in complete deuteration (>98%) of the β-C–H bonds along with the high levels of isotopic incorporation (77%) at the sterically encumbered α-C–H bonds.

The cobalt-catalyzed method was also effective with N-heteroarenes. Both lutidine (3f) and the related pyridine derivative, 3g were deuterated with >98% deuterium incorporation at all C(sp²)–H sites. State-of-the-art iron-catalyzed methods resulted in exclusive and sole deuteriation of the 4-position of the lutidine ring.^8,9 Heteroarenes lacking substitution in the α-position were also well tolerated. For example, 4-picoline (3h) and 3-picoline (3i) underwent complete deuteration and distinct site selectivity was also observed for nicotine (3j) compared to prior methods.¹⁰ For example, complete deuteration (>98%) was obtained at the 5-position of the pyridine ring. The cobalt-catalyzed HIE method was also compatible with other N-heteroarenes such as N-methyl pyrrole (3k), which underwent H/D exchange with 96% deuterium incorporation at all four sites. Functional groups such as amides, ethers, tertiary amines, and oxazolines were well tolerated levels and resulted in high levels of isotopic incorporation (3l-3o).

Aryl halides were also evaluated for their compatibility with the cobalt-catalyzed HIE method. A challenge in many C–H functionalization reactions, including catalytic hydrogen isotope exchange, is the chemoselective activation of C(sp²)–H bonds in the presence of a C(sp²)–X (X = Cl, Br, I) functionality. Studies with cationic cobalt [(PNP)Co] and iridium [(PNP)Ir] complexes highlight the general trend of third-row metals exhibiting a preference for C(sp²)–H bonds while first-row congeners are more reactive toward C(sp²)–X.^16,17,18 Chlorobenzene (3p) was compatible with the cobalt-catalyzed method as deuterium incorporation was observed at the meta and para positions. By comparison, pyridine(bis)carbene iron complexes reported by our group are incompatible with aryl chlorides and the bis(phosphine) carbene iron dihydride catalyst described by de Ruiter and co-workers led to deuteration strictly at the para-position.^8,9 More complex molecules were also well tolerated in the reaction. For example, NMe-paroxetine (3q) and papaverine (3r) underwent cobalt-catalyzed HIE. Lastly, to probe the utility of the cobalt-catalyzed HIE method, a large-scale deuteration was performed. Heating a solution of 2.03 mmol of 1a in a 0.70 M benzene-d₆ solution containing 0.5 mol% of 1-(H)₂BPin resulted in a 97% isolated yield of 1a-d₂.

Selectivity of Cobalt-Catalyzed C(sp²)–H Activation.

Catalysts for C–H functionalization with first-row metals have demonstrated complementary site selectivity to those based on precious metals. For example, the [(^iPrPNP)Co]-catalyzed C(sp²)–H borylation of fluorinated arenes exhibits ortho-selectivity arising from reversible oxidative addition of the C(sp²)–H bond under thermodynamic control. ^19,20,21 More recently, less electron-donating terpyridine ligands have been shown to alter the selectivity to the meta-sites due to kinetically-controlled oxidative addition.²² With in-situ activated bis(silylene)pyridine-supported cobalt(II) precursors, Driess, Cui, and co-workers demonstrated the borylation of fluoroarenes with bis(silylene)pyridine cobalt(II) complexes proceeds with meta-to-fluorine selectivity.^15a To gain additional insight into the site-selectivity preferences of the C(sp²)–H activation step, the reaction of fluorinated arene (1a) and 1-(H)₂BPin was studied. Monitoring the reaction by ²H NMR spectroscopy revealed a constant ratio of ~55:45 (5-position/4-position) slightly in favor of the 5-position (meta-to-fluorine) throughout time course of the HIE reaction until high conversion was achieved, where the ratio becomes 50:50 and both sites are completely deuterated. The observed deuterium incorporation as a function of time supports a slight kinetic preference for C(sp²)–H activation at the meta-to-fluorine position.

Determination of the Catalyst Resting State as a Function of Time.

The catalytic HIE of 1a with benzene-d₆ and 15 mol% of 1-(H)₂BPin at 80 °C was monitored by ¹H and ¹⁹F NMR spectroscopy at 23 °C to determine the identity of the catalyst resting state (Scheme 4A). A higher loading of 15 mol% was used to facilitate the observation of the cobalt compounds present in solution. The dihydride boryl 1-(H)₂BPin was observed throughout the reaction (Scheme 4B). The cobalt dihydride signals diminished in intensity as the C(sp²)–H activation begins to have increased isotopic incorporation. For example, at higher conversions, the cobalt hydride signal at −10.91 ppm diminished in intensity, consistent with H/D exchange to generate 1-(D)₂BPin and 1-(H)(D)BPin and is consistent with observations from ²H NMR spectroscopy (See Figure S13-S16) A proposed pathway to account for the observed resting state is presented in Figure 3. From 1-(H)₂BPin, loss of HBPin generates a putative cobalt(I) hydride which reacts with either 1a or benzene-d₆. The reaction of the cobalt(I) hydride with benzene-d₆ leads to productive H/D exchange generating a cobalt(I) deuteride that reacts with 1a to transfer deuterium and regenerate the cobalt (I) hydride. Presumably, after catalytic turnover, HBPin reacts with a cobalt(I) hydride to generate the dihydride boryl resting state. At higher conversions, as the concentration of 1a-d₂ and DBPin increases, the formation of 1-(H)(D)BPin and 1-(D)₂BPin was observed by ¹H NMR and ²H NMR spectroscopy, respectively. The slow rate of H/D exchange of the cobalt hydride signals in 1-(H)₂BPin relative to that of the arene substrate suggests H–B activation to regenerate Co(III) is fast relative to C–D activation. It is worth noting that it is unclear if any of the elementary steps involving redox cycling at cobalt are silicon assisted. These processes have been investigated computationally with other transition metal complexes. ¹⁴

Scheme 4.

A. Determination of Catalyst Resting State as a Function of Time. B. ¹H NMR Spectrum (Obtained at 23°C) of the Reaction Mixture of 1a with C₆D₆ in the Presence of 15 mol% 1-(H)₂BPin at Different Time Points. Aromatic Region (Top) and Hydride Region (Bottom) Depicted.

Figure 3.

Proposed pathway accounting for the experimentally observed resting state.

Another pathway to account for the formation of a cobalt(I) hydride is H₂ loss followed by activation of fluoroarene 1a and C(sp²)–B reductive elimination. The borylation of 1a and generation of cobalt (I) or cobalt (III) fluoraryl complexes was not detected by ¹⁹F NMR spectroscopy during the time course of the reaction. Furthermore, stirring a THF solution of 1a and 1 equivalent of B₂Pin₂ with 10 mol% of 1-(H)₂Bpin at 80 °C for 24 hours resulted in no conversion.

Effect of HBPin and Atmosphere on Cobalt-Catalyzed HIE.

Based on the observed catalytic resting states and electron-donating ability of [SiNSi] pincer ligands, precatalyst activation by reductive elimination of HBPin was expected to limit access to the active catalyst as the equilibrium strongly favors cobalt(III). To further experimentally support this hypothesis, parallel catalytic HIE experiments were conducted with 1-(H)₂BPin and 1a in the presence and absence of added HBPin. The progress of both catalytic reactions was monitored by ¹⁹F NMR spectroscopy. The sample containing HBPin exhibited a dramatically slower rate of HIE as compared to the sample without (Scheme 5). For example, after 22 minutes, the sample lacking HBPin reached 42% conversion, while the sample with 0.5 equivalents of HBPin reached 33% conversion at 150 minutes (Scheme 5). To further explore the inhibitory effect of HBPin, parallel experiments monitoring the HIE activity of 1-(H)₂BPin with 1-fluoro-3-(trifluoromethyl)benzene in the presence and absence of one equivalent of DBPin were conducted. The progress of each reaction was monitored by ¹⁹F NMR spectroscopy, and after 15 minutes, the sample without DBPin reached one half-life, while the sample with one equivalent of DBPin reached <20% conversion (See Figure S19). These results demonstrate that DBPin is also an inhibitor of the reaction. The competition between the arene substrate or HBPin/DBPin for the cobalt(I) deuteride resulted in an overall decreased rate of catalytic HIE. Notably, the formation of HBPin was detected by ¹H NMR spectroscopy under the reaction conditions at low conversions. The effect of dihydrogen was also evaluated and shown to have an inhibitory role (see Supporting Information for details).

Scheme 5.

Effect of HBPin on Cobalt-Catalyzed HIE with 4.

^aDeuterium incorporation determined by ¹⁹F NMR spectroscopy.

Pincer ligands containing N-heterocyclic silylene side arms have been reported to promote hydride and other ligand migration reactions. For example, cobalt-to-silicon and palladium-to-silicon hydride migrations have been previously identified.^{14g, 15b} Notably, this transfer from the metal to the silylene or vice-versa, changes the coordination number of the metal complex and opens a coordination site on the metal. To rule out any inhibition by dinitrogen coordination, parallel experiments were conducted. Reaction monitoring revealed no correlation between the pressure of N₂ and deuterium incorporation as a function of time (See Figure S24).

Evaluation of Other Well-defined Co(III) Complexes.

Given the HBPin inhibition and observed resting states with 1-(H)₂BPin, other bis(silylene)pyridine cobalt (III) complexes were evaluated for catalytic HIE performance. It was postulated that 1-(H)₃·NaBHEt₃ may result in increased HIE activity given that the active cobalt(I)-hydride is generated from H₂ loss followed by accumulation of H₂ generated in the headspace. The catalytic HIE of 1a with benzene-d₆ and 15 mol% of 1-(H)₂BPin or 1-(H)₃·NaBHEt₃ at 80 °C was monitored by ¹H and ¹⁹F NMR spectroscopy. With 1-(H)₂BPin as the precatalyst, complete deuteration of 1a was observed after 60 minutes (Scheme 6). By comparison, use of 1-(H)₃·NaBHEt₃ instead resulted in complete deuteration within 12 minutes (Scheme 6).

Scheme 6.

Evaluation of 1-(H)₃ NaBHEt₃ for HIE.

To further evaluate the HIE activity of 1-(H)₃·NaBHEt₃, a series of arenes was evaluated in the cobalt-catalyzed H/D exchange method (table 2). Chlorobenzene (3p) underwent complete deuteration (>98%) of the meta and para sites with 1-(H)₃·NaBHEt₃ resulting in overall higher levels of isotopic incorporation compared to when 1-(H)₂Bpin was used as the precatalyst. A notable feature of bis(silylene)pyridine cobalt(III) complexes is the ability to perform C(sp²)–H activation in the presence of C(sp²–X) bonds. For example, 4-fluorochlorobenzene (3s) and 4-fluorobromobenzene (3t) underwent cobalt-catalyzed HIE at room temperature employing 5 mol% of 1-(H)₃·NaBHEt₃. At room temperature, 1-(H)₃·NaBHEt₃ resulted in higher levels of deuterium incorporation than 1-(H)₂Bpin. Performing the catalytic reactions at higher temperatures resulted in lower deuterium incorporation and fluorobenzene was detected by ¹⁹F NMR spectroscopy, the amount of which was identical to the catalyst loading indicating C(sp²)–X reduction. The ability to perform the catalytic HIE at lower temperatures with 1-(H)₃·NaBHEt₃ is critical for substrates which pose chemoselectivity issues such as 3s and 3t. To probe if 1-(H)₃·NaBHEt₃ and 1-(H)₂Bpin have similar functional group tolerances, the catalytic HIE of NMe-Paroxetine (3q) was evaluated. Overall, higher levels of deuterium incorporation were observed with 1-(H)₃·NaBHEt₃ (3.8D/molecule) compared to 1-(H)₂Bpin (2.6D/molecule). No deuterium incorporation was observed at the meta-to-fluorine positions in 3q with 1-(H)₃·NaBHEt₃ or 1-(H)₂Bpin, which is likely due to the sterics of the piperidine ring attached to the fluoroarene. Prior work with nickel hydride complexes supported by α-diimine ligands evaluated for the HIE of NH-Paroxetine resulted in a total of 3.85D/molecule with 25 mol% of precatalyst.¹⁰ Furthermore, the improved protocol resulted in complete deuteration (>98%) of etoxazole (3o) at the ortho and meta position. Presumably, the lack of deuteration in the 1,2,4-ring system of 3o with 1-(H)₃·NaBHEt₃ or 1-(H)₂Bpin is a result of the steric environments of the tert-butyl group and oxazoline.

Table 2.

Scope of the Catalytic Deuteration of Arenes with 1-(H)₃·NaBHEt₃ using Benzene-d₆ as the Deuterium Source.^a,b

Reactions conducted in J Young NMR tubes. Deuterium incorporation determined by a combination of ¹H, ¹³C, and ¹⁹F NMR spectroscopies.

^c2.0 mol% and 50 °C

^d3.0 mol%.

To gain additional insight into precatalyst activation and the proposed mechanism of catalytic HIE, bis(silylene)pyridine cobalt(III) complexes with different modes of activation were evaluated for HIE performance. A different mode of precatalyst activation is likely operative with cis-1-(Bf)₂H, whereby C(sp²)–H reductive elimination of benzofuran is required to access a cobalt(I) complex (Scheme 7a). From the cobalt(I)-benzofuranyl, activation of benzene-d₆ and C(sp²)–D reductive elimination leads to H/D exchange. With the previously reported bis(aryl)hydride complex, cis-(^ptolSiNSi)Co(Bf)₂H (cis-1-(Bf)₂H), no significant deuterium incorporation was observed with 1a or 4 (Scheme 7b). To gain additional insight into precatalyst activation, a benzene-d₆ solution of cis-1-(Bf)₂H was heated to 80 °C for 20 hours, resulting in the formation of benzofuran along with an unidentified paramagnetic cobalt product (See Figure S154–S156).The observed lack of catalytic HIE activity likely arises from a slower C– HH H/D oxidative addition with the cobalt(I)-benzofuranyl and/or facile deactivation of cobalt(I)-benzofuranyl complex as compared to the putative cobalt(I)-hydride generated from 1-(H)₂BPin or 1-(H)₃ Na BHEt₃. To further probe this hypothesis, cis-(^ptolSiNSi)Co(H)₂Bf (cis-1-(H)₂Bf) was evaluated as a precatalyst to explore whether reductive elimination of benzofuran would generate the same cobalt hydride as derived from 1-(H)₂BPin or 1-(H)₃·NaBHEt₃ (Scheme 7c). Heating a benzene-d₆ solution containing either 1a or 4 to 80 °C in the presence of 5 mol% of cis-1-(H)₂Bf resulted in higher isotopic incorporation compared to cis-1-(Bf)₂H (Scheme 7d). For example, use of cis-1-(H)₂Bf resulted in 78% deuterium incorporation into the 4- and 5-position of 1a after 43 minutes. With 4 as the substrate, deuterium incorporation (>60%) was observed at the C-2 and C-3 positions of benzofuran as compared to cis-1-(Bf)₂H, where no significant deuterium (<10%) incorporation was observed.

Scheme 7.

Evaluation of Well-Defined Cobalt(III) Complexes for HIE Activity.

CONCLUSIONS

A cobalt-catalyzed method for hydrogen isotope exchange of (hetero)arenes has been developed using low loadings of a bis(silylene)pyridine cobalt (III) complex and benzene-d₆ as the deuterium source. The introduction of strongly σ-donating N-heterocyclic silylenes resulted in active complexes that enabled high levels of isotopic incorporation. Despite the tert-butyl groups present in the amidinate side arm, HIE at sterically encumbered sites was viable. Stoichiometric studies with dihydrogen and X-ray diffraction revealed previously unknown pincer modifications relevant to catalysis. N-heterocyclic silylenes represent a unique ligand class due to their electronic properties and propensity to engage in metal-ligand cooperativity. The insights from this study can aid the development of further transformations catalyzed by bis(silylene)pyridine cobalt complexes. These results demonstrate that while the donating properties of pincer ligands are critical for HIE activity, other factors including chemical metal-ligand cooperativity must also be considered when using small molecules like D₂ as the isotope source. Given that H₂ is a common byproduct in C–H functionalization reactions, these insights may be applied to other catalytic C–H functionalization reactions.