Catalytic Activation of N2O at a Low-Valent Bismuth Redox Platform

Herein we present the catalytic activation of N2O at a BiI⇄BiIII redox platform. The activation of such a kinetically inert molecule was achieved by the use of bismuthinidene catalysts, aided by HBpin as reducing agent. The protocol features remarkably mild conditions (25 °C, 1 bar N2O), together with high turnover numbers (TON, up to 6700) and turnover frequencies (TOF). Analysis of the elementary steps enabled structural characterization of catalytically relevant intermediates after O-insertion, namely a rare arylbismuth oxo dimer and a unique monomeric arylbismuth hydroxide. This protocol represents a distinctive example of a main-group redox cycling for the catalytic activation of N2O.

* sı Supporting Information ABSTRACT: Herein we present the catalytic activation of N 2 O at a Bi I ⇄Bi III redox platform. The activation of such a kinetically inert molecule was achieved by the use of bismuthinidene catalysts, aided by HBpin as reducing agent. The protocol features remarkably mild conditions (25°C, 1 bar N 2 O), together with high turnover numbers (TON, up to 6700) and turnover frequencies (TOF). Analysis of the elementary steps enabled structural characterization of catalytically relevant intermediates after O-insertion, namely a rare arylbismuth oxo dimer and a unique monomeric arylbismuth hydroxide. This protocol represents a distinctive example of a main-group redox cycling for the catalytic activation of N 2 O. N itrous oxide (N 2 O) is known to be a potent greenhouse gas, with much greater warming potential than CO 2 . 1 The concentration of this gaseous molecule in the atmosphere has significantly increased in the modern era as a result of human activities, thus generating an environmental threat. 1b In this sense, the development of catalytic strategies for the decomposition of N 2 O has recently drawn much attention. 2 From the chemical standpoint, N 2 O is a thermodynamically powerful O atom transfer reagent, and indeed, Nature has evolved a powerful enzymatic pathway for converting N 2 O into N 2 in microbial denitrification processes. 3 Yet, N 2 O is kinetically inert, which poses challenges for its catalytic activation with artificial systems. 2c In contrast to heterogeneous catalysts, 2a,b,d a handful of examples are known to be capable of activating N 2 O based on homogeneous catalysts ( Figure 1A), 2e mainly consisting of transition metals (Ru, 4 Rh, 5 Co, 6 and others 7 ).
With the aim of emulating transition-metal-like reactivity, a growing number of organo-main-group compounds have been shown to activate small molecules. 8 Among them, many lowvalent p-block compounds circumvented the kinetic barrier for N 2 O activation and formed structurally unique O-containing compounds. 9 Additionally, frustrated Lewis pairs (FLPs) 10 and N-heterocyclic carbenes (NHCs) 11 have also been shown to capture N 2 O and eventually cleave the N−O bond. Recently, degradation of N 2 O with disilanes initiated by a fluoride anion has also been achieved, 12 representing alternative pathways based on main-group systems. However, despite these precedents, catalytic redox processes for the activation of N 2 O by a main group compound still remain elusive.
Based on earlier precedents, activation of N 2 O could be facilitated by low-valent main group species, through the formation of an M−O bond with release of N 2 . 9 If catalytic turnover is to be achieved, such M−O species should be reduced satisfactorily back to the starting low-valent main group catalyst. Although this process is well-established for high-valent oxo compounds, 13 access to low-valent counterparts is nontrivial. Nevertheless, bismuth presents itself as an interesting candidate to address such a challenging idea, due to its demonstrated ability to access various oxidation states. 14 One-electron Bi II ⇄Bi III has been described by Coles 15 and recently by Lichtenberg, 16 whereas, two-electron Bi redox catalysis (Bi I ⇄Bi III and Bi III ⇄Bi V ) has recently been reported by us. 17−19 We have previously demonstrated that the N,C,Nchelated bismuthinidene complexes, originally reported by Dostaĺ, 20 provide a privileged platform for Bi I ⇄Bi III redox cycling. Herein, we demonstrate that N,C,N-chelated bismuthinidenes are able to catalyze N 2 O deoxygenation in the presence of pinacolborane (HBpin) ( Figure 1B). The catalytic system features the activation of N 2 O at remarkably mild conditions (25°C, 1 bar) with high TON (up to 6700) and TOF. Ligand design and structural analysis on the bismuthinidene catalysts enabled the full characterization of key Bi III −oxo intermediates by NMR, X-ray, and HRMS.
To interrogate the reactivity of bismuthinidenes with N 2 O, we initially subjected complex 1 to a N 2 O atmosphere (1 bar) in THF-d 8 at −78°C (Scheme 1). The green solution slowly turned pale yellow with concomitant evolution of gas. Analysis of the head space by GC-TCD identified the formation of N 2 during the reaction. 21 1 H NMR analysis at −40°C revealed complete consumption of 1 after 45 min and the formation of a major species containing intact N,C,N-ligand scaffold and a C−Bi bond. 21 ESI-HRMS analysis of this mixture clearly suggested the formation of dimeric arylbismuth oxides [(ArBiO) 2 +H + , calcd 937.33012, found 937.33070]. Such species was found to be dynamic in solution and thermally unstable, preventing its characterization by crystallographic techniques. The observed behavior for this species is consistent with other related Ar−Bi III oxo or sulfido dimers. 22−24 In order to shed light on the possible structure of this species, the tBu on imines was replaced with m-terphenyl (m-Tp, 4 and 5, Figure 2A), which has been previously utilized to stabilize reactive organobismuth compounds such as Bi−H, 25 BiBi, 26 and Bi−OH. 27 Due to the sensitivity of aryl-(ket)imines to metal hydrides, 28 we developed a facile and scalable procedure to obtain 4 and 5: reduction of the parent arylbismuth dichlorides 2 and 3 with Cp 2 Co afforded 4 and 5 after simple filtration as dark purple and red-purple solids respectively in very high yields ( Figure 2A). 21 X-ray crystallography revealed that, in spite of the steric bulkiness of the m-Tp groups, the bond lengths and angles resemble those reported for 1 and related ketimine-N,C,N-complexes of bismuth ( Figure 2B). 20 When 4 was exposed to a N 2 O atmosphere at room temperature, the color slowly changed from dark purple to pale yellow and evolution of N 2 was observed ( Figure 3A). 1 H NMR analysis at −50°C confirmed the full consumption of 4 after 40 min and indicated the formation of a single species with an asymmetric N,C,N-pincer backbone. Crystals suitable for X-ray crystallography were obtained by slow diffusion of npentane into a concentrated toluene solution of the reaction mixture at −78°C. The crystal structure unequivocally determined the presence of the dimeric mono-organobismuth oxide 6, which features two μ-oxo bridge moieties ( Figure 3B).
Examples of mono-organobismuth(III) oxides are rare. 22,23,29 Due to the high polarity of the Bi−O bond and the large difference in orbital size between Bi and O, these oxides readily undergo dimerization or polymerization. 30 To the best of our knowledge, only two crystal structures of dimeric mono-organobismuth oxides have been reported: a  syn-23 and an anti-isomer. 22,31 Here, 6 represents an antiisomer with a slightly asymmetric Bi 2 O 2 core ( Figure 3B). As a result of the weak coordination of N1 to Bi1 [Bi1−N1, 2.672(4) Å], the Bi1−O2 distance [2.120(3) Å] is marginally longer than Bi1−O1 [2.103(3) Å]. Interestingly, one m-Tp group in each half of the complex points away from the central Bi. Although H27 could not be refined unambiguously, the short C27−O2 distance (3.104 Å) strongly indicated a hydrogen bonding between H27 (and its symmetric H) and O2. 1 H NMR at −50°C reveals dramatically different chemical shifts for both imines (8.11 and 10.04 ppm), thus endorsing the hydrogen-bonding proposed. Yet, the dynamic imine coordination was indicated by the exchange peaks in ROESY-NMR and convergence of these imine peaks at higher temperatures as shown in VT-NMR data ( Figure 3C). On the other hand, DOSY-NMR experiments suggested that the Bi 2 O 2 ring of 6 was preserved in solution and no dissociation occurred. 21 The structure of 6 suggests that similar species are formed when 1 is oxidized with N 2 O (Scheme 1).
At this point, we speculated that the dimeric nature of 6 could be the result of a rapid dimerization of a monomeric terminal Ar−BiO compound. Based on previous examples, 32 we speculated that replacement of the imines with ketimines would favor the isolation of a monomeric species via tautomerization processes. To entertain this hypothesis, 5 was subjected to N 2 O in THF-d 8 ( Figure 4A). Similar to 6, 1 H NMR of the resulting orange-red solution indicated the formation of one single species. X-ray crystallography unequivocally determined that 7 was a monomeric organobismuth hydroxide ( Figure 4B). The high quality of the crystals allowed the unambiguous assignment of the positions of the H1, H8a, H8b, and H10 (3 H). One of the Me groups in the ketimines converted into a CH 2 , resulting in a reduction of the C−C length in C7−C8 [1.3552 (16) Å], consistent with a double bond. 33 The longer C7−N1 and Bi1−N2 distances [1.3832 (15) and 2.6117(9) Å] compared to C9−N2 and Bi1− N1 [1.2848(14) and 2.2319(9) Å] also manifest the presence of an amido bond in one of the arms of the pincer. 34 Interestingly, the OH points to a phenyl group of a m-Tp with a short H−phenyl centroid distance of 2.622 Å, indicative of a weak OH···π interaction. 27 Due to the high tendency to form oxides or clusters through dehydration, reports on well-defined organobismuth hydroxides are limited; 24a,27,35 yet, 7 is noticeably stable.
The monomeric compound 7 represents a tautomeric form of a monomeric Ar−BiO, an elusive species which has yet to be reported. In the same way, 6 can be conceived as the result of a fast dimerization process of two molecules of monomeric Ar−BiO. The formation of hydroxide 7 highlights the high basicity of the O atom in Ar−BiO, which could be better described as a polarized BiO bond: Ar−Bi + −O − . Therefore, it is reasonable to assume that both 6 and 7 are fingerprints for the transient generation of such elusive species (int-I, both resonance structures depicted; Scheme 2), which rapidly dimerizes or tautomerizes to the more stable compounds 6 and 7.
Having identified the intermediacy of Bi−O bonds after N 2 O activation, we explored the reduction of 6 and 7 to Bi(I) to sustain a putative catalytic cycle. Among other uses, 36 HBpin has been utilized as a deoxygenation agent for amine and phosphine oxides 37 as well as for the catalytic reduction of CO 2 . 38 Inspired by this reactivity, we treated 6 with 2.0 equiv of HBpin, which resulted in immediate formation of a dark purple solution (Scheme 3). Bismuthinidene 4 formed in 79% yield judging by 1 H NMR. Similarly, the reduction of 7 gave 78% of 5. Meanwhile, ca. 1 equiv of HBpin was converted to a mixture of HO−Bpin (8) and (pinB) 2 O (9). 39 At this point, we decided to merge this reactivity to unfold a catalytic system for the activation of N 2 O with Bi(I) compounds. Blank experiments demonstrated that no reaction occurs in the absence of Bi(I) ( Table 1, entry 1). Catalytic N 2 O deoxygenation with HBpin proceeded smoothly at room temperature in the presence of 1 mol % of 4 or 5, with the TON reaching 54 and 89, respectively (entries 2 and 3). The higher efficiency of 5 over 4 could be ascribed to the higher stability of oxobismuth species 7 compared to 6. To our

Scheme 2. Postulated Intermediates during Oxidation of Bi(I) with N 2 O
Journal of the American Chemical Society pubs.acs.org/JACS Communication delight, when 1 was revisited as catalyst, a dramatic rate enhancement was observed; the reaction was complete in 3 min with vigorous release of N 2 gas (entry 4). The high reactivity of 1 permitted lowering the catalyst loading to 0.01 mol % (entries 4 to 7). Unprecedentedly, the TON reached to 6700 at 0.01 mol % catalyst loading (entry 7). In addition, the TOF was estimated to be 52 min −1 at 0.1 mol % catalyst loading (entry 5). 21 Since dimerization and tautomerization have already been shown to proceed really fast, we believe that species similar to 6 and 7 could probably be involved in the catalytic cycle. The mild conditions and high catalytic efficiency contrast with the elevated temperatures, high catalyst loadings, and/or prolonged reaction times usually required for transition metals.
In conclusion, this work demonstrates the capacity of bismuthinidenes to catalytically activate N 2 O in a Bi(I)⇄ Bi(III) redox platform. The synthesis of sterically congested bismuthinidenes using m-Tp substituents on the imines permitted isolation and characterization of catalytically relevant species such as bismuth oxide dimer 6 and bismuth hydroxide 7. Bis-imine and bis-ketimine N,C,N-chelated bismuthinidenes provide the first main-group redox platform for catalytic N 2 O decomposition. The ambient conditions and the very high catalytic efficiency make this system akin to transition-metal   Journal of the American Chemical Society pubs.acs.org/JACS Communication