Manganese(I)‐Catalyzed C−H Activation: The Key Role of a 7‐Membered Manganacycle in H‐Transfer and Reductive Elimination

Abstract Manganese‐catalyzed C−H bond activation chemistry is emerging as a powerful and complementary method for molecular functionalization. A highly reactive seven‐membered MnI intermediate is detected and characterized that is effective for H‐transfer or reductive elimination to deliver alkenylated or pyridinium products, respectively. The two pathways are determined at MnI by judicious choice of an electron‐deficient 2‐pyrone substrate containing a 2‐pyridyl directing group, which undergoes regioselective C−H bond activation, serving as a valuable system for probing the mechanistic features of Mn C−H bond activation chemistry.

Processes utilizing Mn I ,p articularly [Mn(C^N)(CO) 4 ] 2, [5,6] have been of broad interest. Them echanistic features of the remarkable synthetic work of Ackermann and Wang, [3,4] where intermediates 4a-c have been proposed, prompted us to examine whether they could be detected and characterized and then subsequently be shown to deliver organic products such as 5-7.C omplexes 4d-f,f ormed by insertion of internal alkynes are known, [6,7] but their competence in terms of af ully connected reaction system, affording organic products,h as not been examined. As 18electron species containing four CO ligands,p ossessing high thermodynamic stability,t hey are unlikely to be directly involved in the catalytic cycle. [8] Herein we describe asuitable reaction system (1g!4g! 5g or 6g,S cheme 1) that takes advantage of the exquisite reactivity of an electron-deficient 2-pyrone ring system containing a2 -pyridyl directing group (1g). We recognized that the 2-pyrone could act as ah emilabile ligand in 7membered manganacycle 4g,p otentially providing sufficient stabilisation for observation of this key intermediate.O ur findings demonstrate that 4g acts as ac entral manifold to reductive elimination and H-transfer, giving products 5g and 6g,respectively,w ith details described herein.
Our study began with the reaction of 2-pyrone 1g with BnMn(CO) 5 in hexane at 75 8 8C, which gave cyclometalated 2g cleanly and in quantitative yield (Scheme 2). Complex 2g was fully characterized (see the Supporting Information);asingle crystal X-ray structure confirmed that regioselective CÀH activation occurred at C3, in keeping with Pd II -direct arylations of 2-pyrones, [9] albeit most likely by a s-CAM-type process. [10] We hypothesized that UV irradiation [11] of 2g would lead to solvated intermediate I Pyr (Scheme 2, middle inset). [12] Subsequent alkyne trapping via intermediate II Pyr ,w ould then convert into the alkyne insertion manganacycle 4g.UV irradiation (Hg/Xe Arc lamp,2 00-2500 nm) of am ixture of 2g and 3 (1.1 equiv) in [D 8 ]THF at 240 K(at 5min intervals), and reaction monitoring by 1 HNMR spectroscopy between intervals,r evealed the formation of an ew intermediate that grows up to 9.6 %conversion. Further irradiation resulted in spectral broadening (paramagnetic species), but crucially,full NMR analysis of manganacycle 4gwas possible,with HMQC/ HMBC correlation methods/n.O.e.e xperiments.A nalysis shows that 4g formed regioselectively at C3 (Scheme 2, bottom inset). MS analysis also confirmed the presence of 4g (LIFDI m/z 427 for [M] + C and ESI m/z 428 for [MH] + )i n solution.
Experimentally there is evidence in 4g of an interaction between the 2-pyrone olefinic bond (C6-C11) and the Mn I center at d = 159.7 ppm (C6) and d = 145.9 ppm (C11), which stabilizes the tricarbonyl complex. Computational studies (DFT methods) confirm that HOMOÀ4w ithin 4g has 2pyrone-Mn bonding character (see the Supporting Information), confirming 4g as af easible structure.T he small coordination shifts in the 13 C{ 1 H} NMR spectrum imply this interaction is weak, although generation of avacant site at Mn (4g' ')a nd subsequent alkyne coordination (4g' '' ')o ught to be feasible.T he DFT studies for III Pyr (4g)a nd III Ph (4a) indicate no low-lying vacant orbitals (HOMO-LUMO gap = 1.70154-1.97588 eV), consistent with Mn having an 18electron count.
Warming of the [D 8 ]THF solution of 4g to room temperature led to the formation of the reductive elimination product 5g (Scheme 3). Complex 5g was fully characterized (see the Supporting Information) and confirmed by X-ray analysis to possess aM n(CO) 3 anion. 5g was also formed in 87 %y ield on treatment of 2g with 3 (1.1 equiv.) at 80 8 8C, Et 2 O, 18 h( sealed tube). Thus,t he same reaction pathway (2g+ 3!5g)r esults from either UV irradiation or thermal heating, validating our approach in utilizing UV irradiation to enable detection and characterization of intermediate 4g.
Interestingly,c atalytic reactions of 1g with 3,u nder the reaction conditions reported by Wang et al. [4] for2 -phenyl-pyridine 1a (conditions:B rMn(CO) 5 ,C y 2 NH, Et 2 O, 100 8 8C for 6-24 h), do not lead to formation of alkenylated products (for example, 6g). This indicates that the rate of reductive elimination from 4g to give 5g is faster than the rate for alkyne H-transfer to give 6g(see above). We rationalized that reaction of 2g in neat phenylacetylene 3 would enable Htransfer to become the dominant pathway (Scheme 4), but the reaction afforded three new products.F irstly,t he H-transfer product 6g was formed in 28 %yield;anexcess of 3 favors Htransfer over reductive elimination. Central to the success of the reaction is coordination of asecond molecule of alkyne 3 and subsequent alkyne H-transfer of intermediate 4g.T he other products 8 and 9 were unexpected, resulting from an oteworthy Diels-Alder reaction (DAR) of 3 with the 2pyridine ring, [13] followed by ring fragmentation (single-Scheme 1. Manganese(I)-catalyzed CÀHa ctivation, and potentialp roducts and intermediates.
Scheme 2. Cyclomanganation of 1g gives 2g,w hich upon photolysis with phenylacetylene 3 gives 4g.T he X-ray structure of 2g is given (top right, ellipsoidss et at 50 %probability;H-atoms omitted and Mn atom labeled only for clarity). Insets:proposed transienti ntermediates on route to 4g and the key NMR data for 4g. Scheme 3. Thermally controlled reductive elimination from either 2g or 4g to give 5g.A nX-ray structure of asingle crystal of 5g is also shown (ellipsoids set to 50 %probability;H-atoms omitted and Mn atom labeled only,for clarity). crystal X-ray structures of 8 and 9 confirmed the molecular connectivity,c orrelating with NMR spectroscopy,s ee the Supporting Information). Compound 9 shows that the 2pyrone participated in as econdary inverse electron demand DAR. [14] Along with 6g,b oth 8 and 9 derive from 4g,w here the DARs and 2-pyridyl fragmentation are secondary reactions.
To understand the steps leading to the formation of 5g DFT methods were used (Scheme 5, see the Supporting Information for details of DFT calculations). Starting from II Pyr ,f ormed via loss of CO from 2g and coordination of 3, insertion of coordinated alkyne into the MnÀC(pyrone) bond proceeds through alow-energy transition state (TS IIPyr-IIIPyr )to give III Pyr .T he latter intermediate is equivalent to characterized 4g.C À Nr eductive elimination from III Pyr ,v ia transition state TS IIIPyr-5g-iso ,r esults in the formation of the 2methyl-4-oxo-6-phenyl-4H-3,7l 5 -pyrano[4,3-a]quinolizin-7ylium ring system (5g). AD RC analysis of TS IIIPyr-5g-iso revealed that the imaginary eigenvector led to 5g-iso (the coordination isomer of 5g); a p-slip then gives 5g.
Thec orresponding potential energy surface for the phenyl-substituted system (giving the Chen and Wang product 5a)r evealed that the same reaction pathway was viable (pathway shown in gray in Scheme 5). Thebarrier to insertion of 3 (TS II-III )was slightly greater (Gibbs energies at 298.15 K relative to the respective compound II + 25 kJ mol À1 for 2pyrone versus + 34 kJ mol À1 for phenyl) and that III Pyr was relatively higher in energy than III Ph (À76 kJ mol À1 versus À95 kJ mol À1 ). To explain the different outcome from the phenyl and 2-pyrone substituents it is informative to consider the higher energy of TS IIIPh-5a-iso (+ 26 kJ mol À1 )a gainst TS IIIPyr-5g-iso (À16 kJ mol À1 ). Therefore,t he energetic spans for reductive elimination are 60 kJ mol À1 (2-pyrone) and 121 kJ mol À1 (phenyl). When compared with the formation of IV Pyr and IV Ph ,w hich is the next step in forming H-transfer products 5g and 5a,r espectively,i ti se vident that the reductive elimination to form 5g is competitive,b ut in the case of 5a the much larger energetic span to reductive elimination allows for productive catalysis via alkyne coordination to give IV Ph . [4] While no double alkyne insertion products were detected in reactions of 2g with phenylacetylene 3,t he reaction of related derivative 2h with 3 resulted in exclusive formation of double alkyne insertion product 10 (Scheme 6; the structure of 6h is shown as an expected alkenylated product). This remarkable result shows the impact that as ubtle change to the pyridyl directing group has on the barriers to these steps.
We rationalized the experimental observations by DFT calculations,which enabled amechanism for this reaction and the differences between the phenyl-and 2-pyrone-substituted complexes to be proposed (Scheme 7). In the case of the pyrone derivative,c oordination of alkyne to III Pyr results in formation of IV Pyr having two energetically accessible fates. H-transfer through TS IVPyr-VPyr (+ 3kJmol À1 )r esults in the formation of alkynyl complex V Pyr which would liberate 6h, however insertion of the alkyne into the MnÀCbond of IV Pyr through TS IVPyr-VIPyr (+ 4kJmol À1 )affords more energetically favourable VI Pyr .T he process seen for reactions of 2h has resulted in the formation of two CÀCb onds.Preliminary investigations indicate that this proceeds through a"two-steps no intermediate" pathway [15] with the initial insertion into the Mn À Cb ond, followed by cyclization giving as ix-membered ring without an intermediate.However,inVI Pyr the Mn is h 3coordinated to the pendant pyridyl group and newly formed ring. To form 10 Pyr ,which is the lowest point on the potential energy surface at À320 kJ mol À1 ,t he Mn needs to migrate to the alternative ring-face.W ep ostulate that this involves migration onto one of the phenyl rings in the ligand, for example, VIIa Pyr .The ring rotates allowing the Mn to migrate to the other face of the pentadienyl system, giving VIIb Pyr .Itis reasonable to presume that this proceeds via al ow energy ring-walking process. [16] In the case of the phenyl derivative,a ll of the states predicted for the 2-pyrone system are viable;however, TS IVPh-VIPh is far higher in energy than TS IVPh-VPh (+ 41 kJ mol À1 versus À1kJmol À1 ). Therefore,i nsertion of the second alkyne is non-competitive,w ith the H-transfer pathway leading to the alkenylated product, consistent with experimental observations.
In conclusion, we have detected and characterized ac ommonly proposed 7-membered manganacycle 4g (of direct relevance to generic structure 4,Scheme 1). Manganacycle 4gsits at the selectivity junction to reductive elimination or H-transfer steps.Depending on the reaction conditions, 5g or 6g products form that correspond to reductive elimination and protonation pathways,respectively.Double alkyne insertion to give 10 has also been revealed in these studies.O ur observations provide the first clear cut evidence that manganacycles such as 4 are key intermediates in Mn I -mediated CÀHb ond activation processes involving substrates containing directing groups. [3,4,7] More generally,s uch intermediates may be considered as leading to side reactions,b ut here we have shown that it presents an opportunity to control product selectivity.Serendipitously we have uncovered arare example of aD AR of ap yridine derivative,w here the intermediate fragments to form products such as 8 and 9.T aken together, our findings provide aunique insight into Mn I -mediated CÀH bond activation processes,e specially how relatively minor changes in substrate structure influence product selection; Mn I -based metallocycles clearly offer rich chemistry, [3] much potential, and warrant further study more generally in organic and organometallic chemistry.