μ-Oxo-bis[(octacosafluoro-meso-tetraphenylporphyrinato)iron(iii)] – synthesis, crystal structure, and catalytic activity in oxidation reactions

We describe the synthesis and X-ray crystal structure of μ-oxo-bis[(octacosafluoro-meso-tetraphenylporphyrinato)iron(iii)] [(FeTPPF28)2O]. This novel iron complex is an efficient catalyst for oxidative biaryl coupling reactions of diarylamines and carbazoles. The asymmetric oxidative coupling in the presence of an axially chiral biaryl phosphoric acid as co-catalyst provides the 2,2′-bis(arylamino)-1,1′-biaryl in 96% ee. The Wacker-type oxidation of alkenes to the corresponding ketones with (FeTPPF28)2O as catalyst in the presence of phenylsilane proceeds at room temperature with air as the terminal oxidant. For internal and aliphatic alkenes increased ketone/alcohol product ratios were obtained.


Introduction
Limited resources and environmental issues have promoted the development of sustainable chemistry. In organometallic catalysis the classical noble metals like palladium and iridium, which are expensive and toxic, are being replaced by rst row transition metals. Among those, iron is the prime candidate for environmentally benign organometallic catalysis because of its high abundance and low toxicity. 1,2 In nature, porphyrin-iron complexes are essential biocatalysts with cytochrome P450 enzymes as the most important class. These oxidoreductases occur in nearly all organisms. 3 Moreover, they also catalyze uncommon transformations like rearrangements, cyclizations, and intramolecular C-C and C-heteroatom coupling reactions. [4][5][6][7][8][9][10] These reactions generally proceed either via a hydrogen atom transfer (HAT) or single-electron transfer (SET) process. 11,12 Therefore, the design of novel electron-decient porphyrin-iron complexes could open up the way to unprecedented biomimetic reactions for organic synthesis. In this respect, the properties of m-oxo-bridged binuclear porphyrinoid complexes have recently attracted a lot of attention. 13,14 Results and discussion In the present study, ‡ we describe the synthesis, structural characterization, and applications in catalysis of the strongly electron-decient complex m-oxo-bis[(octacosauoro-meso-tetraphenylporphyrinato)iron(III)] [(FeTPPF 28 ) 2 O] (5c) ( Table 1). We focused our efforts on the uorinated porphyrin ligands, since in addition to the electron-withdrawing effect of the uorine atoms they improve considerably the solubility of the complexes. 15,16 Homogeneous catalysis sometimes suffers from low solubility. However, moderately uorinated organometallic complexes generally allow a broader spectrum of solvents that can be used. The rst syntheses of b-octauoro-substituted meso-tetraphenylporphyrins and their zinc complexes were reported independently by two different groups in 1997. 17,18 The direct introduction of uorine substituents at the porphyrin ring is not possible and thus b-octauoro-meso-tetraphenylporphyrins (3a and 3c) are synthesized by condensation of 3,4-diuoropyrrole (2a) with the corresponding benzaldehydes. The b-octauorinated porphyrins 3a and 3c are accessible from 3,3,4,4-tetrauoropyrrolidinium chloride (1) via a three-step sequence reported by DiMagno et al. 18,19 According to 1 H and 19 F NMR analysis (see SI), the elimination of hydrogen uoride proceeds nearly quantitatively. However, DiMagno et al. isolated 3,4-diuoropyrrole (2a) in only 53% yield due to the extremely high volatility of this compound. 19 We found that the overall yield of the porphyrins 3a and 3c is considerably improved by avoiding the isolation of 2a and submitting the crude product directly to the cyclocondensation step. The formation of the uorinated tetraphenylporphyrin-iron complexes 4a-4c from the corresponding porphyrins using the classical conditions described by Adler (DMF at reux) 20 led to a complex reaction mixture. This mixture is resulting from nucleophilic aromatic substitution at the uorinated porphyrins by dimethylamine, formed by decarbonylation of the solvent. Adapting the conditions reported by Freire et al., complexation of 3a-3c was achieved by reaction with iron(II) chloride in acetonitrile at 120°C. 21 Several preparations of m-oxo-porphyrinoid-iron complexes have been reported. 13, 14,22,23 The b-octauoro-substituted m-oxoiron complex (FeTPPF 8 ) 2 O (5a) could not be prepared due to the extremely low solubility of the corresponding chloro-iron complex 4a. Complex 5b was described previously. 24 Elution of the chloro complex 4c over activated alumina (CH 2 Cl 2 /MeOH, 95 : 5) provided quantitatively the m-oxo complex (FeTPPF 28 ) 2 O (5c).
Deep red cubic crystals of complex 5c suitable for X-ray crystallography were obtained by recrystallization from dichloromethane ( Fig. 1). 25 26 and m-oxo-bis[(octapropylporphyrazinato)iron(III)] (1.7601(12), 1.7501(12)Å). 13a Leroy et al. investigated the catalytic activity of the porphyrinato-iron(III) chloride complexes 4a and 4c for epoxidation and   into the corresponding hypochlorite complexes and studied their catalytic reactivity in epoxidation and chlorination reactions. 28b Previously, we described iron-catalyzed oxidative C-C and Cheteroatom coupling reactions using hexadecauorophthalocyanine-iron(II) (FePcF 16 ) as well as the corresponding m-oxoiron(III) complex ([FePcF 16 ] 2 O) as catalysts and air as terminal oxidant. 29 We have now studied in detail the catalytic activity of the uorinated porphyrin-iron(III) complexes 4a-4c, 5b, and 5c in oxidative coupling reactions. We postulated that the strong electron-withdrawing effect of the uorine atom should increase the catalytic activity of (FeTPPF 28 ) 2 O (5c) as compared to the unsubstituted FeTPP system in analogy to our observations with the peruorinated phthalocyanine-iron complexes. The oxidative C-C homocoupling of N-phenyl-2-naphthylamine (6) was selected as model system. Using the previously reported catalyst FePcF 16 and methanesulfonic acid (MsOH) as additive, the biaryl compound 7 was isolated in 62% yield along with 11% of the carbazole 8 (Table 2, entry 1; Fig. 2). 30 Initial attempts with the unsubstituted meso-tetraphenylporphyrin-iron complex FeTPPCl and the corresponding b-octauorinated complex 4a in the presence of methanesulfonic acid as additive gave no turnover (Table 2, entries 2 and 3). The peruorinated complex FeTPPF 28 Cl (4c) gave only traces of the product 7 (entry 4). In conclusion, none of the chloro complexes 4a-4c showed signicant catalytic activity in the C-C coupling of 6.
The m-oxo-iron complexes are assumed to be intermediates in the catalytic cycle of oxidations with porphyrin and phthalocyanine-iron complexes. 13,14,29,31 Thus, we tested the m-oxo-iron complex (FeTPPF 28 ) 2 O (5c) as catalyst under the same conditions used above for the complexes 4a-4c and obtained the biaryl 7 in 6% yield (entry 5). Performing the reaction under an atmosphere of pure oxygen improved the yield only slightly (entry 6). Variation of the additive improved the yield signicantly and revealed that strong Brønsted acids (TFA and TfOH, entries 8 and 9) and the Lewis acids tris(pentauorophenyl)borane (entry 10) and boron triuoride diethyl etherate (entry 11) gave the best results. Control experiments conrmed that both the iron catalyst (entry 12) and the Lewis acid (entry 13) are required for the reaction to proceed. The iron-catalyzed oxidative coupling of 6 was generally performed using non-dried solvents under an ambient atmosphere. Finally, we have demonstrated that water-free conditions with dried air and anhydrous solvents led to a further slight increase of the yield of 7 and a decrease of the reaction time (entry 14).   Under the optimized reaction conditions identied above (Table 2, entry 9), the effect of the uorine substitution and of the axial ligand at the iron atom was investigated using the porphyrin complexes FeTPPCl, 4a-4c, 5b, and 5c as catalysts (Table 3). Two general trends have been observed. The complex FeTPPCl had no catalytic activity at all and the complexes 4a-4c exhibited a very low catalytic activity providing 7 in yields below 10% (Table 3, entries 1-4). However, the m-oxo-iron complexes (FeTPPF 20 ) 2 O (5b) and (FeTPPF 28 ) 2 O (5c) led to much higher turnover numbers and provided 7 in yields of 57 and 78%, respectively (entries 5 and 6). We concluded that the signicant difference in catalytic activity is caused by the different axial ligand. Thus using the chloro-iron complex 4c, we added 3 mol% of silver triate to the reaction mixture in order to generate in situ FeTPPF 28 OTf, which led to much shorter reaction times (4 h instead of 48 h) and afforded the biaryl compound 7 in 89% yield (entry 7).
Based on our experimental ndings, we postulate the following mechanism for the (FeTPPF 28 ) 2 O-catalyzed oxidative coupling considering the strong inuence of the additive (Scheme 1). The reaction is believed to be initiated by an SET oxidation followed by coupling and proton loss. 32 The key step is an SET from the substrate to the iron(III) complex, as previously observed by Baciocchi et al. for the oxidation of N,N-dimethylanilines with FeTPPF 20 Cl (4b). 12b The SET process is much more efficient with strongly electron-decient iron(III) complexes. Chen et al. found that in situ exchange of the axial ligand from chloride to triate enhances the reactivity of porphyrin-iron(III) complexes for oxidation reactions signicantly since triate is a weaker donor than chloride. 33 In order to rationalize our experimental ndings described above, we followed the iron-catalyzed oxidative coupling by UV-vis experiments ( Fig. 3 and S1 †).
We observed that the m-oxo complex (FeTPPF 28 ) 2 O (5c) is rapidly hydrolyzed by strong Brønsted acids. Aer addition of TfOH to a solution of 5c, the characteristic peak at 553 nm disappeared whereas two peaks at 497 and 618 nm emerged, which are assigned to the complex FeTPPF 28 OTf. Alternatively, the latter complex can be generated in situ by reaction of FeTPPF 28 Cl (4c) with silver triuoromethanesulfonate. The UVvis spectrum of the reaction mixture with 5c as catalyst and TfOH as additive showed aer 1 hour all three peaks, indicating that both porphyrin-iron(III) triate and the m-oxo complex 5c are present. Thus, using the m-oxo complex (FeTPPF 28 ) 2 O (5c) in combination with a strong acid (TfOH) as catalyst generates a catalytic system more reactive than FeTPPF 28 Cl (4c) ( Table 3, entries 4 versus 6). Additional support derives from the increase in catalytic activity observed by exchange of the chloro against the triato ligand (Table 3, entries 4 and 7).
We have studied the catalytic activity of (FeTPPF 28 ) 2 O (5c) for the oxidative coupling of a selection of diarylamines 10a-10c ( Table 4). The variation of the additive showed that the Lewis acid BF 3 $OEt 2 was more efficient than the previously used   Brønsted acids (Table S1 †). Using these modied conditions, the coupling of 10a-10c proceeded more slowly but gave improved yields compared to our previous results using FePcF 16 as catalyst. 29a We then explored the possibility to achieve an asymmetric catalytic oxidative coupling of 10d to the atropisomeric biaryl compound 11d using (FeTPPF 28 ) 2 O (5c) as catalyst (Table 5). 1,1 ′ -Biaryl-2,2 ′ -phosphoric acids have been established as efficient chiral catalysts for asymmetric catalysis by Akiyama, Terada, and List. 34 Recently, we have shown that oxidation of 10d with FePcF 16 as catalyst in the presence of 10 mol% of the chiral phosphoric acid (R)-12 as co-catalyst afforded 11d in 71% yield and 90% ee ( Table 5, entry 1). 30 The chiral phosphate counter-ion was believed to direct the asymmetric coupling of the radical cation generated from 10d by an initial single-electron transfer. Using 5c as catalyst in the presence of 20 mol% of (R)-12 led to the biaryl compound 11d in 96% ee (Table 5, entry 3).
Due to their physical properties, 3,3 ′ -bicarbazoles represent promising candidates for hole-transporting materials in organic light-emitting diodes (OLEDs). 46 Previous procedures for the synthesis of 3,3 ′ -bicarbazoles by oxidative homocoupling required stochiometric amounts of iron(III) chloride, 47 DDQ, 48 or rhodium as noble metal catalyst. 49 Our method using oxygen as terminal oxidant in the presence of (FeTPPF 28 ) 2 O (5c) as catalyst  enables the rst iron-catalyzed oxidative coupling of the carbazoles 15a-15c to the 3,3 ′ -bicarbazoles 16a-16c (Table 6). Another iron-catalyzed oxidation process recently investigated by our group is the Wacker-type oxidation of olens to ketones. [50][51][52][53] The oxidation of 2-vinylnaphthalene (17a) to 2acetylnaphthalene (18a) served as a model system in order to test different phthalocyanine-and porphyrin-iron complexes under an atmosphere of pure oxygen (Table 7). Our previous results showed a much higher catalytic activity of the uorinated phthalocyanine-and porphyrin-iron complexes as compared to their unsubstituted analogs ( Table 7, entries 1-8). 50,51 Moreover, the importance to use the appropriate silane reducing agent in combination with the corresponding iron catalyst was emphasized. [50][51][52][53] The present results conrm this strong inuence of the silane when using the iron complexes FeTPPF 20 Cl (4b) and (FeTPPF 20 ) 2 O (5b) as catalysts for the Wacker-type reaction ( Table 7, entries [8][9][10][11]. Based on these previous results, we expected a high catalytic activity for the iron complexes 4c and 5c with the peruorinated porphyrinato ligand octacosauoro-meso-tetraphenylporphyrin. To our surprise, we had no turnover at all in the oxidation of the olen 17a using FeTPPF 28 Cl (4c) as catalyst and either triethylsilane or triphenylsilane as reducing agent under otherwise identical reaction conditions (Table 7, entries 12 and 13). Phenylsilane was proven to be the best reducing agent for the iron-catalyzed Wacker-type reaction with tris(1,3-diketonato)iron(III) complexes as catalysts. 53 Indeed, using complex 4c as catalyst in combination with phenylsilane provided the ketone 18a in 76% yield along with 9% of the corresponding alcohol 19a (Table 7, entry 14). Basically the same results were obtained for the oxidation of 17a to 18a using the three reducing agents Et 3 SiH, Ph 3 SiH, and PhSiH 3 in combination with (FeTPPF 28 ) 2 O (5c) as catalyst ( Table 7, entries [15][16][17]. These results provide further evidence for our mechanistic hypothesis with m-oxo[diiron(III)] complexes as intermediates in the catalytic cycle of the Wackertype oxidation which also applies to the present reaction using complex 5c as catalyst. 51 Finally, we tested air instead of an atmosphere of pure oxygen as re-oxidant for our iron complex ( Table 7, entry 18). Although the reaction time was prolonged, we were delighted that the yield of the desired product 18a increased to 87%, whereas only traces of the alcohol 19a could be detected.
Using (FeTPPF 28 ) 2 O (5c) as catalyst under the optimized reaction conditions, we have tested the Wacker-type oxidation for a range of different olens 17a-17h (Table 8). A special focus was on those olens which gave poor results in our previous study with FePcF 16 or [FePcF 16 ] 2 O as catalysts, the cyclic olens 17e-17g and the aliphatic olen 17h. [50][51][52] For the simple styrene derivatives 17a-17d, the oxidation with 5c as catalyst proceeded smoothly affording the corresponding ketones 18a-18d in yields as high or even slightly better compared to those obtained with FePcF 16 as catalyst, 50 albeit longer reaction times were required. The differences between the results with (FeTPPF 28 ) 2 O (5c) and the peruorophthalocyanine-iron complex as catalyst were most pronounced for the oxidation of the more challenging substrates (cyclic olens and aliphatic olens). The results with the substrates 17e-17h show that the selectivity of the reaction is shied signicantly towards the ketone at the expense of the alcohol by-product. For example, the Wackertype oxidation of the nitrochromene 17e catalyzed by (FeTPPF 28 ) 2 O (5c) afforded the chroman-4-one 18e in 79% yield along with only 11% of the corresponding alcohol 19e (previous result with FePcF 16 as catalyst under O 2 : 62% of 18e and 35% of 19e). 52 Compound 18e represents a synthetic precursor for the pyrano[3,2-a]carbazole alkaloid euchrestifoline. [52][53][54] Also for the (FeTPPF 28 ) 2 O-catalyzed oxidation of the cyanochromene 17f and the dihydronaphthalene 17g to the ketones 18f (93% yield)  and 18g (89% yield), we observed a much higher selectivity in favor of the ketones (previous yields with FePcF 16 as catalyst under O 2 : 65% for 18f and 68% for 18g). 50 Most strikingly, oxidation of the aliphatic alkene 1-octadecene (17h) using (FeTPPF 28 ) 2 O (5c) as catalyst provided 2-octadecanone (18h) in 53% yield, 55 whereas the corresponding reaction with FePcF 16 required more of the catalyst (10 mol%), pure oxygen as reoxidant, and elevated temperature (78°C) but still led preferentially to the alcohol 19h (42% yield) along with 18h (30% yield). 50 Thus, we have shown that the Wacker-type oxidation of olens using the new catalyst (FeTPPF 28 ) 2 O (5c) proceeds smoothly with ambient air as nal oxidant. The present reaction gives higher selectivities in favor of the desired ketones as compared to the corresponding reaction with FePcF 16 as catalyst which needs pure oxygen as reoxidant to achieve the best turnover numbers.

Conclusions
We have described the synthesis of the novel peruorinated porphyrin-iron complex m-oxo-bis[(octacosauoro-meso-tetraphenylporphyrinato)iron(III)] [(FeTPPF 28 ) 2 O]. The high activity of this catalyst in oxidation reactions has been demonstrated for the biaryl coupling and the Wacker-type reaction. The twofold aryl C-H bond activation was exploited for the oxidative coupling of diarylamines leading to 2,2 ′ -bis(arylamino)-1,1 ′biaryls. In the presence of an axially chiral biaryl phosphoric acid as co-catalyst this coupling proceeds in up to 96% ee. The (FeTPPF 28 ) 2 O-catalyzed oxidative coupling of 2-hydroxy-, 9-alkyl/ aryl-, and 3-hydroxycarbazoles affords regioselectively 1,1 ′ -, 3,3 ′ -, and 4,4 ′ -bicarbazoles and has been applied to the synthesis of a variety of bicarbazole natural products including the rst synthesis of integerrine B. The Wacker-type oxidation of alkenes, including internal and aliphatic alkenes, previously considered as difficult substrates, with (FeTPPF 28 ) 2 O as catalyst in the presence of phenylsilane proceeds at room temperature with air as terminal oxidant and provides the corresponding ketones in high yields. The present ndings are paving the way for the development of mild and selective oxidation reactions under biomimetic conditions resembling those of the enzymatic oxidative processes in nature dependent on cytochrome P450 heme proteins.

Data availability
The data supporting this article have been uploaded as part of the ESI. †

Conflicts of interest
There are no conicts to declare.