Surprising Route to a Monoazaporphyrin and Full Characterization of Its Complexes with Five Different 3d Metals

In the search for mild agents for the oxidative cyclization of tetrapyrromethane to the corresponding corrole, we discovered a route that leads to a monoazaporphyrin with three meso-CF3 groups. Optimization studies that allowed access to appreciable amounts of this new macrocycle paved the way for the preparation of its cobalt, copper, nickel, zinc, and iron complexes. All complexes were fully characterized by various spectroscopic methods and X-ray crystallography. Their photophysical and electrochemical properties were determined and compared to those of analogous porphyrins in order to deduce the effect of the peripheral N atom. Considering the global efforts for designing efficient alternatives to platinum group metal (PGM) catalysts, they were also absorbed onto a porous carbon electrode material and studied as electrocatalysts for the oxygen reduction reaction (ORR). The cobalt complex was found to be operative at a quite positive catalytic onset potential and with good selectivity for the desirable 4-electrons/4-protons pathway.


■ INTRODUCTION
The ever-increasing interest in modified porphyrins may be traced back to the synthesis of the N 4 macrocycles that serve as prosthetic groups in heme, chlorophylls, Vitamin B 12 , and other natural systems.These tremendous efforts were led by prominent scientists including R. Willstatter, H. Fisher, and R. B. Woodward, whose contributions were recently outlined by Senge et al. in an excellent review. 1 Research activity on nonnatural porphyrins and related macrocycles continues to flourish because of their relevance to fundamental science, including the study of aromaticity and redox noninnocence, as well as for diverse practical applications. 2−18 This also holds true for derivatives in which the metal-coordinating core is modified, being composed of less or more than four N atoms and/or by non-nitrogen elements, 6,10,13 as well as for those bearing a different macrocyclic skeleton. 5,14,16,19Regarding the latter class, the largest advances are arguably in corroles, N 4 macrocycles with one direct bond between two pyrrole subunits like in the prosthetic group of Vitamin B12 but fully conjugated and aromatic like in classic porphyrins. 19,20ne major reason for the interest in corroles is that the seemingly small differences relative to porphyrins, at about 10% reduction in the size of the coordination core and the presence of three rather than two protic nitrogen atoms therein, induce very significant changes in the reactivity/ stability of the corresponding metal complexes. 3,5,8The advantageous outcomes of these features have been outlined in quite a few contemporary reviews, which also include recent synthetic breakthroughs, such as access to minimally substituted corroles. 9,12,21,22ne additional and less emphasized possible difference between corroles and porphyrins is the largest possible symmetry axis in the corresponding metal complexes, C 2 and C 4 , respectively.This affects not only spectroscopic features but also allows for selective substitution of the macrocyclic CH protons, as in the case of the 2,17-bis-sulfonated derivative which continues to be the most intensively studied corrole for medicinal applications. 23,24An interesting hybrid case is presented by monoazaporphyrins (MAPs), with nitrogen instead of one meso-C-R position (R = H, alkyl, or aryl) of porphyrins since they have the coordination core of porphyrins and the symmetry of corroles (Scheme 1).−32 Despite these reports, MAPs have been quite neglected in the field as compared not only to porphyrins, phthalocyanines, and the related tetraazaporphyrins but also relative to corroles.−35 Recent research has focused mainly on spectroscopic features, 33 but there is one example regarding outstanding reactivity: Neya et al. reported a 50 times higher oxygen affinity of a MAP-reconstituted myoglobin relative to the native protein. 36e now report how an incidental finding during attempted improvement of a corrole synthesis led to facile access to a novel MAP derivative with three CF 3 substituents on the meso-C positions.This was followed by the preparation and characterization of zinc(II), nickel(II), copper(II), cobalt(II), and several iron(III) complexes of the macrocycle.The spectroscopic and redox properties of these complexes were compared to those of analogous porphyrins with either H or CF 3 substituents on C5.Also included is the first investigation of the utility of MAP transition metal complexes as electrocatalysts, which revealed that the cobalt(II) complex is quite a potent and selective catalyst for the oxygen reduction reaction (ORR).

■ RESULTS AND DISCUSSION
One conclusion from research that focused on gaining efficient access to the tris-CF 3 -substituted corrole H 3 (tfc) (also for the much more sensitive parent corrole) was that better results are obtained when the oxidant used in the last step is milder than the commonly employed DDQ. 12,23Using PIFA, the oxidative cyclization of the CF 3 -substituted tetrapyrrane CF 3 -TP afforded H 3 (tfc) in 12−20% yield (Scheme 2, pathway a), which initiated our search for even milder oxidants.One of the examined options was an aqueous ammonia/methanol solution of K 3 Fe(CN) 6 , which has previously been used to affect the oxidative cyclization of a biladiene to a corrole. 37Much to our initial surprise, the sole isolable product was not the expected corrole but monazaporphyrin 1-H 2 (Scheme 2, pathway b).A search of the corrole literature revealed that the first-ever corrole synthesis also reported a (not fully characterized) MAP as a minor byproduct. 38The new 1-H 2 is novel by means of having C1 substituents on the meso-C atoms and unsubstituted β-pyrrole positions.There is only one similar MAP derivative, the iridium complex with C 6 F 5 groups on the meso-C positions, which was reported by Palmer et al. but not further studied. 39he initially employed reaction conditions provided 1-H 2 in 4% yield and no other isolable macrocycle (entry 1, Table 1).
Trying to gear the reaction toward H 3 (tfc), 300 mL of aq.NaOH or K 2 CO 3 were employed as alternatives for NH 4 OH as bases (entries 2 and 3, Table 1), but neither corrole nor porphyrin was formed in detectable amounts.Reducing the volume of methanol, required for dissolving precursor CF 3 -TP, from 50 to 15 mL and thus increasing the amount of dissolved ammonia, raised the yield of the MAP from 4 to 8% (Table 1, entry 4).A further jump in the yield of 1-H 2 to 15% was achieved by replacing regular methanol with a 7N NH 3 methanol solution (Table 1, entry 5).Experiments in NH 3 / MeOH only, with no aqueous ammonia, did not yield any detectable product.This clearly shows that NH 4 OH is a critical cosolvent for dissolving the inorganic oxidant K 3 Fe(CN) 6 , but it is also oxidized and provides nitrogen for MAP formation.The best results were obtained by reducing the temperature of the reaction to 60 °C (Table 1, entry 4), yielding 1-H 2 in a respectable 33% yield.Considering the price and long-term stability of 7N NH 3 /CH 3 OH solutions, an alternative ammonia precursor was sought for.The choice was ammonium carbamate, the solid and cheap material that is very soluble in methanol and was previously reported for different nitrogen insertion processes. 40With 26 equiv of ammonium carbamate as an additive, 1-H 2 was obtained in 26% chemical yield (Table 1, entry 7) Free-base 1-H 2 and the corresponding metal complexes (vide infra) were fully characterized by the combination of high-resolution mass spectrometry (HRMS) (Figures S1−S7   S8−S9).The NH protons of 1-H 2 appear at a chemical shift of −2.5 ppm and are as sharp as in regular porphyrins and not very broad as in corroles (Figure S8). 41nsights into the effect of the meso-N atom on the photophysical and redox properties of the new MAP required the synthesis of related porphyrins, with either CH or CCF 3 at the position occupied by a nitrogen atom in 1-H 2 .Compounds 2-H 2 and 3-H 2 , which have the same symmetry as 1-H 2 and regular porphyrins, respectively, were prepared by slight modifications of previously reported procedures (Scheme 3). 42,43The electronic spectra revealed typical near-UV (Soret) and four visible (Q) bands for all compounds (Figure 1a).But 1-H 2 differs from the two regular porphyrins by (a) a significantly broader Soret band and (b) by alternating intensity of the Q bands rather than a gradual decrease in their absorbance. 43−46 The 400 nm excitation of optically matched dichloromethane solutions revealed that the emission intensity of 1-H 2 is 1.5 and 3.3 times larger than those of 2-H 2 and 3-H 3 , respectively (Figure 1b).This phenomenon likely reflects greater rigidity and fewer degrees of freedom when the position that differs among these three macrocycles bears a nitrogen atom rather than a CH or CCF 3 unit.The combination of more pronounced Q bands and stronger emission of 1-H 2 suggests potential photocatalytic applications for this compound and the corresponding metal complexes.
X-ray quality crystals were obtained for all of the new MAP complexes as well as for the analogous porphyrin 2-H 2 (Figure 2).One interesting observation is that not only the d 8 nickel(II) complex but also the d 7 cobalt(II) and the d 9 copper(II) complexes are 4-coordinate.The average metal− N bond lengths are 1.9284, 1.9102, and 1.9650 Å for 1-Co, 1-Ni, and 1-Cu, and the metal ion deviations from the macrocycle plane are 0.089, 0.116, and 0.097 Å, respectively.Comparison with octaethylporphyrinato metal complexes reveals an identical trend in the M−N bond lengths.This has been explained by the difference in metal ion size, which is 0.58, 0.55, and 0.57 Å for Co II , Ni II , and Cu II , respectively, affecting interporphyrin π−π interactions and deviation from planarity. 47,48It can be deduced that the MAP cobalt and nickel complexes follow the same trend as their ionic radii, with Co II bigger than Ni II , while the difference between the  A feature common to all complexes is that the macrocycle is not planar.They adopt the ruffled geometry, with meso-C atoms alternately located below and above the macrocycle plane, while all of the pyrroles are tilted to the same direction. 15,49he electronic spectra of the complexes (Figure 3) were recorded in CH 2 Cl 2 , except for 1-Zn which was insufficiently soluble and was hence examined in THF.The divalent cobalt, copper, nickel, and zinc complexes displayed "normal" spectra: 50 a single near-UV (Soret) band with maxima at 389 ± 3 nm and one dominant visible (Q) band at 576 ± 5 nm with a shoulder at about 40 nm shorter wavelengths.(Figure 3a−b).The spectra of the trivalent iron complexes are quite different: mononuclear 1-FeCl has a split Soret band, a charge transfer band at 488 nm, and three Q bands, while the Soret band of binuclear 1-Fe 2 O is even more split and it has only one Q band and no charge transfer band (Figure 3c).HRMS and extensive NMR data are provided in the Supporting Information (SI) (S1−S7 and S8−S18).
Investigation of the redox properties of the new MAP started with a comparison to the analogous porphyrins, which share with 1-H 2 the presence of three meso-C−CF 3 but have either CH (2-H 2 ) or another C−CF 3 (3-H 2 ) group rather than an N atom.Circular voltammograms of these metal-free compounds exposed one quasi-reversible process for each (Figure 4a) that corresponds to the reduction and reoxidation of the corresponding macrocycle.The half-wave potentials (E 1/2 ) were deduced to be −1.07,−0.92, and −0.86 V for 2-H 2 , 1-H 2 , and 3-H 2 , respectively.Considering the differences in the redox potentials relative to 2-H 2 uncovers that the N atom is more electron-withdrawing than CH (ΔE 1/2 = 0.15 V) and only slightly less than C−CF 3 (ΔE 1/2 = 0.21 V).These studies were followed by investigating 1-Zn(OH 2 ), whose reversible redox event at E 1/2 = −1.21V (Figure 4b) is safely attributed to ring reduction.
The complexes with potentially redox-active metal ions displayed reversible/semireversible reduction waves with E 1/2 values of −0.88 V for 1-Co and −1.03 V for both 1-Cu and 1-Ni (Figure 4c).This trend is consistent with the reduction of all complexes being macrocycle-centered but influenced by the electronegativity of the chelated metal ion.This variable was deduced by Kadish et al. in the study of many tetraphenylporphyrin metal complexes to be in the order of Zn 2+ > Cu 2+ ∼ Ni 2+ > Co 2+ . 51,52Mononuclear iron(III) complex 1-FeCl displays two reversible redox events (Figure 4d) of which the first, with E 1/2 = −0.29 V, may safely be attributed to an Fe II / Fe III couple because it is so much less negative than that of free-base 1-H 2 and zinc complex 1-Zn. 53,54The second reduction (E 1/2 = −0.91V) is already at potentials where macrocycle reduction must be considered possible.The case of binuclear oxo-bridged iron(III) complex 1-Fe 2 O is more  complex since the first and second events appear at E 1/2 = −0.88V and −1.1 V (Figure 4e), at which both the macrocycle and the metal ion could be redox active.All of the complexes exhibited only irreversible transformations at more negative potentials, which is common for porphyrins not substituted by large aryl groups that protect against decomposition. 34,42aving deduced that the emission intensity of the new MAP is larger than that of porphyrins with identical substituents, the fluorescence quantum yields of 1-H 2 and 1-Zn(OH 2 ) were determined.This was done relative to free-base 5,10,15,20tetraphenylporphyrin (H 2 TPP) (Figure 5), whose quantum yield is 0.11 in toluene under aerobic conditions. 55The quantum yields of 1-H 2 and 1-Zn(OH 2 ) were deduced to be 0.075 and 0.036, respectively.The latter value is very similar to the quantum yield of ZnTPP (0.033), 55,56 which is commonly used for photocatalysis and photoredox processes. 55he transition metal complexes of the new MAP were briefly examined for exploring their potential utility as electrocatalysts for the heterogeneous oxygen reduction reaction (ORR), frequently considered as most crucial for the future of fuel cells. 57−64 Vulcan XC-72R (Vulcan) and Black Pearls 2000 (BP2000) were employed here, noting that the latter has a much bigger surface area (1537 versus 235 m 2 /g) and ten times more mesopores. 65The process started by dissolving the molecular catalyst in 2propanol and mixing it with the insoluble carbon.Both iron complexes were excluded from this study because 1-FeCl was not stable in isopropanol and 1-Fe 2 O was not soluble in it.In contrast with meso-aryl substituted porphyrins and corroles, whose adsorption onto BP2000 is much more significant than onto Vulcan, 57 1-Co, 1-Ni, and 1-Cu were fully adsorbed on both carbons (Figure 6) when mixed in a 0.8/10 mg ratio.This phenomenon may be confidently attributed to the small size of meso-CF 3 relative to meso-aryls: the latter are perpendicular to the macrocycle and hence interfere with its π-stacking interactions, which are responsible for the spontaneous physisorption onto the carbon surface.
Studying the azaporphyrin complexes for ORR at pH 13 exposed the nickel and copper complexes as being irrelevant for catalyzing this reaction.They even harmed the performance of the nonmodified carbon electrode, presumably by blocking active sites.1-Co was however active when absorbed on either Vulcan or BP2000 (Figure 7a,b), with an onset potential of 0.84 V on both carbons, a half-wave potential (E 1/2 ) of 0.72 V on BP2000 and 0.74 V on Vulcan, and producing only 12− 16% H 2 O 2 at 0.5 V and 19% at 0.7 V (Figure 7c, percentages calculated by eqs S1).This clearly shows that the metal plays a significant role in this catalysis, as all other parameters (macrocycle size, meso-substituents, lack of axial ligands, and adsorption on carbon) are identical and that even the carbon type does not play a significant role (Figure 7d).
The results obtained with electrode modifications by 1-Co were briefly compared to those with the cobalt complexes of analogous porphyrins 2-H 2 and 3-H 2 .Both porphyrin complexes (2-Co and 3-Co) were fully adsorbed when treated  with either Vulcan or BP2000 (Figure S19), similar to the observation with 1-Co.Despite the small differences� occupation of the C20 position by C−H, CCF 3 , and N, respectively�the catalytic activity of electrodes modified by them varied significantly (Figure 8).This was revealed by deducing that the superiority of 1-Co@BP2000 is its onset potential: more positive by 50 and 40 mV relative to 2-Co@ BP2000 and 3-Co@BP2000, respectively (Figure 8a).
On the other hand, significantly less undesired H 2 O 2 is formed with the two latter when adsorbed onto BP2000   (Figure 8b).By this criterion, the best results were obtained for catalysis by 3-Co@BP2000, producing only 10 to 20% H 2 O 2 .Full characterization and further studies of the new cobalt complex, including other electrocatalytic processes of clean energy relevance, will be reported in the near future.

■ CONCLUSIONS
Introduced is the facile synthesis of a novel free-base monoazaporphyrin with no substituents on the β-pyrrolic atoms and electron-withdrawing trifluoromethyl groups on the meso-C positions.The divalent zinc, cobalt, copper, and nickel complexes and both mononuclear and binuclear iron(III) complexes were fully characterized in terms of structures and photophysical and redox properties.Carbon electrodes modified by the cobalt complex were found to be worthy electrocatalysts for heterogeneous oxygen reduction with a quite positive catalytic onset potential and good selectivity for the desirable reduction to water rather than to hydrogen peroxide.

S y n t h e s i s o f [ 1 0 , 1 5 , 2 0 -T r i s ( t r i fl u o r o m e t h y l ) -5azaporphyrinato]iron(III) Chloride (1-FeCl).
A methanol solution (20 mL) of FeCl 2 •4H 2 O (650 mg, 3.2 mmol) was added to a solution of 1-H 2 (65 mg, 0.13 mmol) in CHCl 3 (20 mL), and the mixture was stirred and heated to reflux for 24 h.After cooling to room temperature and removing the solvent, the crude material was purified by silica gel column chromatography (20% MeOH/CHCl 3 , red fractions collected).The resulting solid was dissolved in CH 2 Cl 2 , washed 3 times with 10% aq.HCl and once with water, and dried over sodium sulfate, and the solvent was evaporated.The compound was crystallized from a CH 2 Cl 2 /hexane (1:1) solution to afford 1-FeCl as a dark solid (25 mg, 0.016 mmol, 13% yield).X-ray quality crystals were grown by slow evaporation of a concentrated CH 2 Cl 2 /benzene (1:3) solution.Paramagnetic  (20 mL), and the mixture was stirred and refluxed for 24 h.After the mixture was cooled to room temperature and the solvent removed, the crude material was purified by silica gel column chromatography (20% MeOH/CHCl 3 , red fractions collected).The resulting solid was dissolved in CH 2 Cl 2 , washed three times with aq.NaOH and once with water, and dried over sodium sulfate, and the solvent was evaporated.The compound was crystallized from a CH 2 Cl 2 /hexane (1:1) solution to afford 1-Fe 2 O as a dark solid (25 mg, 0.02 mmol, 5% yield).X-ray quality crystals were grown by slow evaporation of a concentrated CH 2 Cl 2 /benzene (1:3) solution.Paramagnetic 1 H NMR (400 MHz, CDCl 3 ) δ (ppm): 14.50 (4H, br s), 15.34 (8H, br s), 15.67 (4H, br.s). 19  .Synthesis was performed according to a published procedure 42 while noting that a mistake therein.5-trifluoromethyldipyrromethane that is listed as being used in the last step of the synthesis leads to 5,10,15,20-tetrakis(trifluoromethyl)porphyrin (3-H 2 ).By using unsubstituted dipyrromethane, prepared by a previously reported procedure, 9 the desired product was obtained after purification on a silica column with 10% ethyl acetate/hexane eluent. 42,432-H 2 and 3-H 2 were synthesized in 9 and 5.5% yield, respectively.The 1 H and 19 F NMR data were identical to those reported in the published procedures. 42,43ASSOCIATED CONTENT * sı Supporting Information The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.4c00436.Methods, experimental details, crystal data, adsorption on both carbon supports, and comparison of ORR activity of each porphyrin on Vulcan and BP2000 (PDF) Scheme 1. Fully NH-Deprotonated Porphyrin, Monoazaporphyrin (MAP), and Corrole for Emphasizing the Analogy of MAP to (a) the Former in Terms of Charge and (b) the Latter by Virtue of Identical Symmetry Scheme 1. Fully NH-Deprotonated Porphyrin, Monoazaporphyrin (MAP), and Corrole for Emphasizing the Analogy of MAP to (a) the Former in Terms of Charge and (b) the Latter by Virtue of Identical Symmetry

Figure 2 .
Figure 2. H-omitted presentation of the X-ray structures of the different complexes.Color legend: red�oxygen, yellow�fluorine, blue�nitrogen, green�chlorine, gray�carbon, purple�cobalt, orange�iron, light blue�nickel, black�copper, and light green� zinc.The numbers below the crystals indicate the deviation of the metal center from the 24 atom plane of the macrocycle (ΔM 24 ) in Å.

Table 1 .
Yields of 1-H 2 Obtained by Variation of the Reaction Conditions a d Neither 1-H 2 nor H 3 (tfc) is formed.