C(sp²)–H Hydroxylation via Catalytic 1,4-Ni Migration with N₂O

Abstract

Herein, we report a Ni-catalyzed C(sp²)–H hydroxylation of aryl bromides with N₂O as an oxygen-atom donor. The reaction is enabled by a 1,4-Ni translocation that results in ipso/ortho difunctionalized products. Regioselectivity and stereocontrol are dictated by a judicious choice of the ligand backbone, thus giving access to either carbonyl or phenol derivatives and offering an opportunity to repurpose hazardous substances en route to valuable oxygen-containing building blocks.

Nitrous oxide (N₂O) ranks as the third most emitted greenhouse gas and is currently the largest ozone-depleting substance emitted through human activities, with an atmospheric lifetime of 116 ± 9 years.¹ Not surprisingly, chemists have recently been challenged to repurpose N₂O en route to added-value chemicals by means of catalytic transformations.² Among various scenarios,³ the utilization of N₂O as an oxygen-transfer reagent is particularly attractive given that N₂ is released as a benign byproduct. While the poor σ-donor and π-acceptor properties of N₂O might constitute a drawback when combined with transition metals,⁴ seminal stoichiometric work by Hillhouse (Scheme 1, top) set the basis for designing catalytic oxygen-transfer events.⁵ Despite the elegant advances realized, particularly from Cornella (path a) that showed the viability for enabling ipso oxygen-transfer reactions with aryl iodides as substrates,⁶ a catalytic oxygen transfer at sp² C–H sites still constitutes an uncharted cartography in both the N₂O and C–H functionalization arenas (path b).⁷

Scheme 1. Oxygen-Transfer Reactions with N₂O and Ni.

As part of our ongoing interest in the catalytic valorization of greenhouse gases⁸ and the ability of nickel catalysts to enable translocation events at previously unfunctionalized C–H reaction sites,⁹ we wondered whether it would be possible to design an oxygen-transfer event at sp² C–H linkages with N₂O and readily available aryl bromides via 1,4-Ni translocation.¹⁰ At the outset of our investigations, however, it was unclear whether such a strategy could be designed given (a) the low binding affinity of N₂O to transition metals,⁴ (b) the proclivity of alkynes to promote multiple insertion events at C–Ni linkages,¹¹ (c) the likelihood of promoting homodimerization pathways of either alkynes or aryl halides, and (d) the subtleties required for effecting a 1,4-Ni translocation prior to N₂O insertion. If successful, however, we anticipated that such a strategy might constitute a worthwhile endeavor for chemical invention, offering an unrecognized opportunity to promote oxygen transfer at previously unfunctionalized sp² C–H sites with N₂O as a coupling partner. In addition, such a technique would also constitute a new entry point to promote the ipso/ortho-difunctionalization of aryl halides by incorporating both C(sp²)–C(sp²) and C(sp²)–O linkages in a concerted, site-selective manner with readily available precursors. Herein, we report the successful realization of this goal, resulting in a regiodivergent strategy that promotes oxygen transfer at either vinyl or aryl sp² C–H sites with an excellent site-selectivity profile from simple and commercially available building blocks (Scheme 2).

Scheme 2. sp² C–H Hydroxylation with N₂O via a 1,4-Ni Shift.

We began our study by evaluating the reactivity of 1-bromo-4-(trifluoromethyl) benzene (1a) with 1-phenyl propyne (2a) under an N₂O atmosphere (2 bar). The choice of 1a was not arbitrary, as it allows the identification of byproducts arising from N₂O insertion prior to 1,4-Ni shift or parasitic homodimerization events via ¹⁹F NMR spectroscopy. After considerable optimization (Table 1),¹² a combination of NiBr₂·diglyme (10 mol %), L1 (20 mol %), and Mn (3.0 equiv) in DMA at 50 °C afforded the best results, giving rise to 3a in 80% assay yield with a >20:1 E/Z ratio and an exquisite 40:1 (3a:4a) selectivity pattern (entry 1). Under these reaction conditions, it is worth noting that not even traces of 4-(trifluoromethyl) phenol arising from N₂O insertion at the initial aryl sp² C–Ni bond were found in the crude mixtures.^6a As anticipated, a subtle interplay of electronic and steric effects at the ligand backbone is critical for success. While similar yields were found for L2 or L3, the lack of substituents adjacent to the nitrogen atom (L4) or the inclusion of substituents at both C2 and C9 (L5) had a deleterious effect on reactivity and site selectivity (entries 2–5). A similar trend was observed for otherwise related 2,2′-bipyridines L6 and L7 (entries 6 and 7). Strikingly, a close inspection of Table 1 revealed that both L4 and L7 delivered exclusively 4a, albeit in low yields, thus suggesting the possibility for developing a regiodivergent oxygen-transfer scenario (entries 4 and 7). On the other hand, the utilization of reductants or nickel precatalysts other than Mn or NiBr₂·diglyme led to lower yields and selectivities (entries 8–11). Control experiments revealed that all of the reaction parameters were critical for success (entry 12).

Table 1. Optimization of the Reaction Conditions^a.

entry	deviation from standard conditions	3a (%)b	E/Z	4a (%)b
1	none	80 (79c)	>20:1(12:1)	2
2	L2 instead of L1	74	>20:1	3
3	L3 instead of L1	68	>20:1	3
4	L4 instead of L1	0	–	18
5	L5 instead of L1	30	>20:1	0
6	L6 instead of L1	67	>20:1	3
7	L7 instead of L1	0	–	18
8	Zn instead of Mn	49	>20:1	3
9	Nil2 instead of NiBbr2· diglyme	51	>20:1	13
10	DMF instead of DMA	66	>20:1	7
11	N2O (1 atm)	58	>20:1	7
12	no Ni, Mn or N2O	0	–	0

^aConditions: 1a (0.20 mmol), 2a (0.24 mmol), NiBr₂·diglyme (10 mol %), L1 (20 mol %), Mn (0.60 mmol), and DMA [0.2 M] at 50 °C under N₂O (2 atm) for 48 h.

^b1H NMR yield using 1,1,2,2-tetrachloroethane as an internal standard.

^cIsolated yield.

With the optimized conditions in hand, we next examined the generality of our protocol. As is evident from the results compiled in Table 2, the catalytic oxygen-transfer event via 1,4-Ni translocation with N₂O could be performed independently of whether electron-rich or electron-poor aryl bromides were utilized as substrates, giving rise to the desired phenol derivatives in good yields (3a–3e). Interestingly, the presence of additional halogen atoms (3f) or organometallic partners (3g–3h) did not interfere with productive oxygen transfer at the ortho sp² C–H site, leaving ample room for derivatization via conventional cross-coupling reactions.¹¹ Although 3i–3l were obtained in good yields and excellent stereochemistry, a non-negligible erosion in regioselectivity was observed, thus evidencing the subtle electronic interplay dictating the targeted 1,4-Ni migration. This observation is also illustrated by the different reactivities found for meta- substituted aryl bromides. While good yields and selectivities were found with electron-withdrawing backbones (3m–3o), the inclusion of electron-donating moieties (3p) or π-extended systems (3r) had a deleterious effect on regioselectivity.¹³ While one might argue that the inclusion of meta- substituents might compromise the site-selective functionalization at different sp² C–H bonds, this was not the case, and 3m–3p were all obtained with an exclusive 1,3,4-substitution pattern. Albeit in lower yields, 1-bromo-2-fluorobenzene (1q) could be employed as a substrate, resulting in the 1,2,3-functionalized phenol derivative 3q.¹² The presence of heterocycles with nitrogen donors did not interfere with productive oxygen transfer via 1,4-Ni translocation, delivering the targeted phenol with C–O bond formation occurring exclusively at C7 (3s). Finally, a selection of aryl bromides decorated with fluorine atoms (1t, 1u) delivered the corresponding phenol derivatives 3t and 3u with an exclusively 1,2,4,5-tetrasubstitution pattern, thus arguing against the intervention of concerted-metalation deprotonation pathways.^14,15

Table 2. C(sp²)–H Hydroxylation of Aryl Bromides with N₂O via Catalytic 1,4-Ni Migration^a,b.

^aConditions: 1 (0.20 mmol), 2 (0.24 mmol), NiBr₂·diglyme (10 mol %), L1 (20 mol %), Mn (0.60 mmol), NaI (0.40 mmol), and DMA [0.2 M] at 50 °C under N₂O (2 atm) for 48 h. Isolated yields, average of two independent runs; E/Z corresponds to the isolated E/Z stereoisomeric ratio; rr corresponds to the regioisomeric ratio of 3 vs 4 by ¹H NMR analysis of the crude.

^b3af: 4-Octyne (2.0 equiv).

Encouraged by our initial findings, we focused our attention on evaluating the influence of the alkyne counterpart on both the reactivity and selectivity (Table 2). As shown, similar yields and selectivities were found independently on the electronics at the alkyne terminus. Among these, the combination of an electron-rich alkyne with an electron-deficient aryl bromide was particularly efficient, leading to an exquisite regio- and stereoselective oxygen-transfer event with N₂O at the sp² C–H bond (3w, 3y). Likewise, the utilization of meta- or ortho-substituted alkynes resulted in 3z and 3aa with high regiocontrol. The latter is particularly illustrative given the erosion in yield found when the oxygen-transfer event is attempted with meta-substituted aryl bromides (3m–3p). Equally interesting was the ability to incorporate longer alkyl chains (3ab, 3ac) or sulfur-containing heteroaromatics (3ad) at the alkyne terminus without significant erosion in yield or selectivity. Furthermore, the successful preparation of 3ae and 3af stands as a testament to the generality of our oxygen-transfer reaction with diaryl- or dialkyl-substituted olefins, delivering the targeted phenols in high regio- and stereoselectivity.

As judged by the data provided in Table 1 (entries 4 and 7), the lack of substituents adjacent to the nitrogen atom in both the 1,10-phenanthroline and 2,2′-bipyridine ligand series under otherwise identical reaction conditions resulted in a selectivity switch. In these cases, the oxygen-transfer reaction occurred via a formal 1,2-carboalkoxylation, leading to the exclusive formation of 4a arising from a N₂O transfer at the vinyl sp² C–Ni bond prior to 1,4-Ni translocation. This result is particularly noteworthy, as it advocates the notion that the 1,4-Ni shift could be controlled by the nature of the ligand backbone, resulting in a regiodivergent scenario for accessing oxygen-containing compounds. After some optimization,¹²L7 turned out to be the most competent ligand for enabling the 1,2-carboalkoxylation event with N₂O as the oxygen-transfer reagent, leading to the corresponding ketone derivatives with exquisite site-selectivity (Table 3). Albeit in moderate yields, a diverse range of aryl ketones displaying different electronic effects at the arene backbone could be obtained with high regiocontrol. Overall, these results should be assessed against the challenge that is addressed, suggesting that the development of regiodivergent oxygen-transfer reactions might constitute a new gateway for rapidly and reliably accessing valuable oxygen-containing products with N₂O as the coupling partner.

Table 3. Oxygen Transfer via 1,2-Carboalkoxylation^a,b.

^aConditions: 1 (0.20 mmol), 2 (0.20 mmol), NiBr₂·diglyme (10 mol %), L7 (20 mol %), Mn (0.60 mmol), NaI (0.40 mmol), and DMA [0.2 M] at 60 °C under N₂O (2 atm) for 48 h. Isolated yields, average of two independent runs; rr = regioisomeric ratio of 4 vs 3 by ¹H NMR analysis of the crude.

^b1a (1.20 equiv).

Although unraveling the mechanistic underpinnings of this reaction should await further investigations, we decided to gather indirect evidence about the mechanism of the reaction by studying the reaction of 1x-d₅ with N₂O under our Ni/L1 regime (Scheme 3, top). As anticipated, full deuterium incorporation was observed at the vinyl position (>99% D) in 3x-d₅. In addition, no significant intermolecular kinetic isotope effect was observed in the competitive reaction of 1x and 1x-d₅ (k_H/k_D= 1.13).¹² This result suggests that the 1,4-Ni translocation at the arene sp² C–H bond might not be rate-determining.¹⁶ Aiming at gaining a deeper understanding of whether the key 1,4-Ni shift and/or N₂O insertion was catalyzed by either Ni(I) or Ni(II) centers, we turned our attention to investigate the reactivity of some of the putative reaction intermediates. To this end, we prepared Ni–I and Ni–II complexes by simple exposure of 5 or 6 to Ni(cod)₂ and TMEDA as ligands (Scheme 3, bottom).¹⁷ As expected, X-ray crystallography evidenced a canonical square-planar geometry for both complexes. Although Ni–I rapidly undergoes E/Z-isomerization in solution,¹⁸ oxygen transfer from N₂O cleanly delivered 3i or 4i in the presence of Mn with either L1 or L7. Notably, these reactions occurred only in the presence of L1/L7, thus showing the importance of 1,10-phenanthroline or the 2,2′-bipyridine backbone on reactivity. More interestingly, Ni–II gave rise to 3i regardless of whether L1 or L7 was utilized as a ligand in the presence of Mn. Altogether, these results suggest that (a) 1,4-Ni translocation is uniquely enabled by a Ni/L1 regime, whereas (b) N₂O insertion can be effected at the targeted sp² C–Ni bond with both L1 or L7. While 1,4-Ni shifts have been proposed to operate via Ni(II) species,^9c,9d the experiments compiled in Scheme 3 suggest the intermediacy of Ni(I) species prior to reaction with N₂O.^19,20 While the latter can be obtained via single-electron transfer with Mn as a reductant, our available data does not allow us to rule out the intervention of comproportionation events as a means to access Ni(I) complexes.²¹

Scheme 3. Preliminary Mechanistic Studies.

In summary, we report the means to enable the C(sp²)–H hydroxylation of aryl bromides via a 1,4-Ni shift with N₂O as an oxygen-atom transfer reagent, resulting in a formal ipso/ortho-difunctionalization event that forges C(sp²)–C(sp²) and C(sp²)–O linkages in a tandem fashion. Good yields and excellent regio- and stereoselectivities are obtained regardless of whether electron-rich or electron-poor arenes are utilized. A judicious choice of ligand allows for dictating the selectivity of the process, resulting in regiodivergent scenarios en route to either phenols or ketones. Further studies into the exploitation of N₂O as an oxygen-transfer reagent are currently underway in our laboratories.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c07018.

• Experimental procedures and spectral and crystallographic data (PDF)

Author Contributions

^∥ H.Z. and J.R. contributed equally to this work.

The authors declare no competing financial interest.