Iron-Catalyzed Reactions via C(sp3)-H and C-O Bond Cleavage

概要

1. Introduction
Directed functionalization of inert C–X (X = H, O, F, etc.) has attracted much attention in recent years because it streamlines organic synthesis by directly transforming an inert C–X bond into a valuable functional group and it has great potential to afford innovative materials. These reactions have been developed with various kinds of transition metals since a significant discovery of a catalytic C–H activation. However, most of them require expensive and toxic metals under harsh conditions or highly reactive reagents. Therefore replacement of such metals, conditions, and reagents has been demanded.

Iron is ubiquitous, inexpensive and less toxic and it has unique reactivity. Therefore, the development of iron catalyst has attracted much attention instead of expensive or toxic metals. However, reactions of iron are difficult to control and they have limitations mostly because strong reductants are required and they cause the generation of highly reactive iron species through over-reduction of iron. If milder reagents are used with iron-catalyzed transformations of C–X bonds, the reaction scope will be broadened, and the improvement of such reactions can lead to a synthesis of new molecules. Herein I report iron-catalyzed reactions via C–H and C–O cleavage under mild conditions. The first reaction transforms aliphatic C–H bond into aryl or alkenyl group with organoboron reagents. The second reaction utilizes C–O bond cleavage and carbometalation to afford organomagnesium. It can access many kinds of indene derivatives.

2. C(sp3 )–H bond Functionalization with boron reagents
The directed functionalization of a C(sp3 )–H bond is more difficult than that of a C(sp2 )–H bond because of the lack of stabilizing interactions in the organometallic intermediate. Thus, these reactions are typically catalyzed by a precious metal under harsh reaction conditions. Our group has achieved the iron-catalyzed arylation of C(sp3 )–H bonds with zinc reagents. However, the zinc reagent reduced the iron to a low-valent species, resulting in narrow scope and undesired homocoupling.

Organoboron are advantageous coupling partners due to their availability. However, there have been only a few examples of iron catalysis using boron reagents because the transmetallation is difficult. As previously discovered, a catalytic amount of zinc salt can facilitate the transmetalation, and I found that the reaction with phenylborate proceeds in the presence of an iron, diphenylphosphinoethylene (dppen) as a ligand and 1,2-dichloroisobutane (DCIB) as an oxidant (Table 1). However, initially a large amount of catalyst was required, and decreasing the amount resulted in low yield. I hypothesized that the turnover of the catalyst becomes slower, and the active species decompose before being re-oxidized by DCIB. Therefore, I investigated the effect of concentration, to find that the product yield increased with the concentration. The arylation reaction showed the broad scope and good functional group tolerance.

Alkenylation proceeded in moderate yield under the same conditions. According to the hypothesis described, I envisioned that a stronger donating diphosphine ligand, MeO-dppen (Scheme 1) might stabilize the organoiron species. Accordingly, the alkenylation reaction proceeded well when this ligand was used.

I performed the arylation reaction using a stoichiometric amount of iron in the absence of oxidant (Scheme 2). Notably, an equimolar amount of zinc salt was required for the reaction to proceed, suggesting that zinc and iron cooperatively catalyze this reaction. A tiny small amount of the homocoupling product (biphenyl) was observed, confirming that iron is not reduced to a low-valent species and therefore an organoiron(III) species cleaves the C–H bond (Scheme 3).

In summary, I have developed an iron-catalyzed directed arylation of aliphatic C–H bonds with aryl-, heteroaryl- and alkenylboron reagents. The chemoselectivity of organoboron reagents, combined with organoiron(III) catalysis resulted in broad reaction scope, functional group tolerance, and product selectivity, as compared with the reactions using an organozinc reagent. This catalytic system also enabled the unprecedented directed olefination of aliphatic C–H bonds with simple alkenyl reagents.

3. Iron-catalyzed Cyclization via C–O Cleavage and Carbomatallation for Synthesis of 1H-Indene derivatives
Indene structure has interesting optical and electronic properties. Among of them, carbon-bridged oligo(phenylenevinylene)s shows remarkable photophysical properties and stability because of their rigid and planner structure, and they are applied to organic electronic devices such as solar cells and laser. To access these carbon-bridged compounds, our group has developed the reductive cyclization to give 3-lithioindene as synthetic module using lithium naphthalenide. This reaction has good reactivity but a limitation of substrates due to the use of highly reactive lithium naphthalenide. Transition-metal-catalyzed C–O cleavage and carbometalation of alkyne to overcome the problem can provide efficiently the generation of organometallic compounds as a nucleophile in organic synthesis. Iron has the potential to cleave C–O bond and bring carbometalation. Here, the cyclization of arylakyne possessing methoxy group via C–O cleavage and carbometalation to form alkenyl magnesium intermediate with an iron catalyst is described (Scheme 4).

An arylalkyne substrate possessing methoxy group (1) was successfully cyclized using magnesium as a reductant, lithium chloride as an additive and a catalytic amount of FeBr2 and triphenylphosphine to afford the desired indene 2 in 90% yield upon quenching with D2O (Table 2). The investigation showed the key parameters of the cyclization. The screening of transition metal and reductant indicated that the combination of iron (II) bromide and magnesium gave the desired product 2 though other transition metals (Ni, Co) and reductants did not work well. The most effective factor was additive (Table2). In particular lithium chloride and bromide salt gave the desired product in good yield (entry 2, 3) while the yield is low using fluoride salt (entry 1). Magnesium and zinc salt do not affect the reaction (entry 4, 5). Ligand was also effective. Triphenylphosphine only improved the reaction (entry 6). On the other hand, the reaction was suppressed by other ligands (entry 7, 8, 9, 10).

Indene derivatives trapped with various electrophiles is shown (Table 3). Alkyl group was incorporated in good yield. This organomagnesium intermediate can react with alkyl chloride, bromine and iodide even if it has ester moiety. Tin, boron, and iodine were also incorporated in moderate yield and they can be utilized for further synthetic reagents. Homocoupling reaction took place when dichloro propane was used as an oxidant. Carbonyl group was incorporated in moderate yield. The corresponding secondary alcohols were oxidized under air to be a ketone. Double cyclized reaction and gram-scale reaction succeeded in high yield. A proposed mechanism is that iron(II) is reduced first to low-valent iron species by metal magnesium. The iron species affords one electron to arylalkyne substrate to generate radical anion intermediate and cleave C–O bond. Carbometalation takes place and the alkenyl group on iron is exchanged to magnesium to afford the alkenyl magnesium. The iron is reduced by magnesium again. In summary, I have developed the iron-catalyzed cyclization of arylalkynes via C–O bond cleavage and carbometalation to form the alkenyl organomagnesium reagent. This reaction could be applied to further reaction (Scheme 5).

4. Conclusion
During the Ph.D. course, I have developed iron-catalyzed reactions under milder conditions by tuning the catalytic system. The first reaction was directed aliphatic C–H bond functionalization with organoboron reagents. This reaction affords arylated products in good yield with broad functional group tolerance under the suppression of the undesired homocoupling reaction. By using new bidentate phosphine ligand, the stereoselective olefination of C(sp 3 )–H bonds was also achieved. The second reaction was reductive cyclization to afford organomagnesium as a synthetic module. Lithium chloride dramatically improved the reaction. The organomagnesium was able to be trapped with various electrophiles to give indene derivatives. This reaction has good potential to synthesize π-extended molecules.