CO2 Fixation with Epoxides under Mild Conditions with a Cooperative Metal Corrole/Quaternary Ammonium Salt Catalyst System

Abstract The cooperative catalytic activity of several metal corrole complexes in combination with tetrabutyl‐ammonium bromide (TBAB) has been investigated for the reaction of epoxides with CO2 leading to cyclic carbonates. It was found that the use of just 0.05 mol % of a manganese(III)corrole with 2 mol % TBAB exhibits excellent catalytic activity under an atmosphere of CO2.

shown relatively high catalytic activity under CO 2 autoclave conditions (> 5bar) and at elevated temperatures (usually > 100 8C). [4,[8][9][10] In contrastt op orphyrin-based systems, the closelyr elated corrolem acrocycle can stabilize metal ions in highero xidations tates, [12][13][14][15][16] making them unique reagents with extraordinary catalytic properties. [17,18] In this study,w ef ocused on manganese,i ron, cobalt, copper,a ntimony and bismuth 5,10,15-tris(pentafluorophenyl) corrole (MTpFPC) complexes 1a-f ( Figure 1) for CO 2 fixation reactions. The center metal ionsa re complexed as Bi III , [13] Co IV , [14] Cu II , [15] Fe IV , [14,16] and Mn III , [14] accordingly,a nd the three C 6 F 5 -groups in the meso po-sitions5 ,1 0, and1 5o ft he macrocycle withdraw electron density from the 18-p-electron system. Ac onsequence of this effect is the improved stabilityo fs uch high-valent metal corroles. Nozaki and co-workersr ecently reported the use of Fe corrolesa nd bis(triphenylphosphine)iminium chloride as an additive for the copolymerization of epoxidesw ith CO 2 under high pressure conditions. [18] Interestingly,i nt his case study,t he formation of the cyclic carbonates was more or lesstotally suppressed. Based on our recenti nteresti nt he use of metal corroles as catalysts [17a] and the recent progress in the use of cooperative CO 2 -fixation catalyst systems based on metal complexesi nc ombination with simple nucleophilic halide sources such as tetrabutylammoniumb romide (TBAB), [5] we reasoned that the use of alternative metal-based corroles together with TBAB may result in av ery powerful cooperative catalyst system for the CO 2 fixation with epoxides. Such ac atalysts ystem may even operate under an atmospheric pressure of CO 2 at low temperature. Because of this lower CO 2 pressure, we argued that this synergistic catalystc ombinationm ay allow us to selectively access cyclic carbonates instead of polymerization products,w hich would thus result in ah ighly complementary   approacht oN ozaki's impressive polymerization protocol [18] by relying on as imilar corrole system. Ta ble 1g ives an overview of the most significant resultso btained in ad etailed screening of different metal corroles 1 in combination with TBAB as ac heap nucleophilic organic halide source for the solvent-free CO 2 fixation of styrene oxide (2a) under an atmosphere of CO 2 (using ab alloon). As expected, only the synergistic combinationo fc orroles 1 and TBAB allows for ar easonable conversion within ar elatively short reaction time (at slightly elevated temperatures). In contrast, the absence of either 1 or TBAB resultedi nn oo ro nly very slow formation of 3a only (entries 1-4). Te sting of the different metal corroles 1a-f next showed that Mn-and Fe-basedo nes clearly outperformed the otherm etal complexes tested herein (entries 1a nd 5-9). Owing to the superior catalytic performance of Mn-corrole 1a we further fine-tuned the reaction conditions with this system (entries [10][11][12][13]. Hereby,i tw as found that reducingt he catalyst loading below 0.01 mol % 1a resultedi n a reduced conversion rate (entries 10 and1 1). On the other hand, carrying out the reaction with 0.05 mol % 1a and 2mol %T BAB( entry 12) leads to ah igherc onversion,a nd as lightly longer reaction time of 8hresultsi na lmost full conversion of 2a under relativelymild conditions with low Mn-corrole loadings (entry 13).
Havingi dentified the best-suited cooperativec atalyst combination and reaction conditions for the solvent-free CO 2 -fixation of epoxide 2a,w en ext investigated the scope of this protocol by using other simple epoxides 2 ( Table 2). Most of the epoxides reacteda tasimilar rate to the parents tyreneoxide 2a at 60 8C. Only the diphenylmethylether-based startingm aterial (entry 6) ande pichlorhydrine (entry 7) showedas lightly slower conversion of less than 90 %under standard conditions. In contrast, some aliphatic epoxides even showed good conversion at room temperature (entries10-12), thus proving the generality of this method for the CO 2 fixation with epoxides 2.
The synergistic effect of an organic nucleophilic halide source and aL ewis acidic metal complex for these CO 2 -fixation reactions has been the subject of detailed recent mechanistic studies. [5,7,8] For example the groups of North et al. and Ren and Lu have done systematic investigations of salen-and salphen-based systems, [5,7] while Hasegawa et al. have carriedo ut very detailed DFT investigations for porphyrin-based catalyst systemsa te levated CO 2 pressure. [8] Based on these comprehensives tudies we propose that the herein-reported Mn-corrole 1a/TBABs ystem operates through an analogousm echanism (Scheme1). To corroborate the mechanism,w ep erformed DFT calculations of the three main proposed intermediates A-C.T he corrole macrocycle exhibits ad ome-shaped structure after coordinationw ith the ring-opened substrate (intermediate A)a nd the manganese atom lies slightly above the plane defined by the four nitrogen atomso ft he corrole ring. The Mn atom is coordinated by the four nitrogen atoms and axially by the oxygen atom of the ring-opened epoxide. After the insertion reaction of CO 2 ,t he axial pyramidal conformation is distorted (see structure B). Herein, one Oa tom originating from CO 2 is axially coordinating with ad istance of 1.8 to the manganese ion, and the other oxygen atom is 2.86 apart from the Br methyleneg roup and can easily perform,i nt he final step, the ring-closure to the cyclic carbonate 3.
To obtain further mechanistic details, we investigated the time course UV/Vis spectralc hanges occurring to the catalyst 1a during the reaction. The typical UV/Vis absorption spectrum of 1a and propylene oxide (Figure 2, solid black line, 1) changes significantly after addition of TBABu nder aC O 2 atmosphere (dotted green line, 2). While the Soret band maxi-  [19] n.r. = no reaction. mum at 410-420 nm remained unaffected, as trong increase of the absorption band and ah ypsochromic shift from 485 to 472 nm was immediately observed (Figure 2, transition 1!2). The latter absorption band is known to be very sensitive towards axial ligation and the change in absorption wavelength can be attributed to the binding of an axially ligatedo xygen atom either as alkoxide intermediate A and/orc arbonate intermediate B (compare with Scheme 1). [20,21] To get further information regarding the expected intermediates (Scheme1)a nd the apparent suggestion that the electronic spectrap resented in Figure 2r eflect them, we compared the spectralc hanges upon coordination of simple model compoundst oc orrole 1a (i.e.,a na lkoxide for intermediate A,acarboxylate fori ntermediate B,F igures S1 and S2). The hereby obtainedU V/Vis spectra are identicalt ot he UV/Vis spectrum of 1a in the presence of propylene oxide,T BAB and CO 2 (dotted green line, Figure 2) andc learly support the presence of an axially ligated oxygen. [20,21] However, no noticeable differences between the spectra obtained upon coordination of an alkoxide or acarboxylate to 1a could be detected, still making it impossible to unambiguously assign the illustrated UV/Vis spectrumi nF igure 2 (dotted green line, 2) to either intermediate A or B. Finally,a fter full conversion of propylene oxide to carbonate 3 the UV/Vis absorption spectrum ( Figure 2, dashedr ed line, 3) changes back, comparably to the one observed in the beginning for the non-reacted catalyst species 1a ( Figure S3).
After recycling the manganese corrole species (column chromatography of the reaction mixtures first with heptanes/ EtOAc = 10:1-3:1 to isolate the carbonates and then with heptanes/EtOAc = 1:1t oe lute the catalyst), analysiso fE SI-MS and 19 FNMR spectra ( Figures S4 andS 5) revealed that the manganese corrole remained intact and was reusablef or further transformations.
To conclude, we have identified that TBAB/Mn III corrole 1a is the best-suited cooperative catalyst combination so far.O wing to the superior catalytic performance of Mn-corrole 1a we further fine-tuned the reaction conditions with this system. Hereby, it was found that reducing the catalystl oadingb elow 0.01 mol % 1a resulted in ar educed conversion rate. On the other hand, carrying out the reaction with 0.05 mol % 1a and 2mol %T BAB led to an increased conversion, and finally as lightly longerr eaction time (8 h) resultsi n> 95 %c onversion of 2 under relativelym ild conditions with low Mn-corrole loadings.
The dramatic changes observed in the time course UV/Vis spectra for 1a during the reactionc ould be attributed to the effect of axial binding of the oxygen atom of the ring-opened epoxide, and to the intermediate B after CO 2 insertion/fixation. Finally,w eh ave shown that recycling of the Mn-corrole is possible and makes the Lewis-acidic manganese corrolec omplex reusablefor furthertransformations.