Ligand noninnocence in FeNO corroles: insights from β -octabromocorrole complexes †

The ﬁ rst FeNO octabromocorroles have been synthesized including four β -octabromo- meso -tris( p -X- phenyl)corrole derivatives Fe[Br 8 T p XPC](NO) (X = CF 3 , H, CH 3 , OCH 3 ) and the β -octabromo- meso -tris- (penta ﬂ uorophenyl)corrole complex, Fe[Br 8 TPFPC](NO). The last complex, which proved amenable to single-crystal X-ray structure determination, exhibits the geometry parameters: Fe – N(O) 1.643(8) Å, N – O 1.158(9) Å, and a FeNO angle of 176.4(6)°. The more electron-de ﬁ cient complexes exhibit increased instability with respect to NO loss and also higher infrared NO stretching frequencies ( ν NO ). Interestingly, DFT calculations and IR marker bands indicate a noninnocent {FeNO} 7 -(corrole (cid:129) 2 − ) formulation for all FeNO corroles, both β -H 8 and β -Br 8 , with essentially the same degree of corrole radical character. Instead, an electron-de ﬁ cient corrole appears to exert a ﬁ eld e ﬀ ect resulting in reduced Fe-to-NO back-donation, which accounts for both the increased instability with respect to NO loss and the higher ν NO ’ s.


Introduction
6][7][8][9] The question thus arises, to what extent is such a description contingent upon the exact electronic character, i.e., the substitution pattern, of the corrole?Our earlier study focused on FeNO complexes of mesotris( p-X-phenyl)corrole ligands, i.e., the Fe[TpXPC](NO) series.In this work, we have synthesized and characterized the corresponding β-octabromo complexes, i.e., Fe[Br 8 TpXPC](NO), where (Fig. 1; X = CF 3 , H, CH 3 , OCH 3 ), as well as Fe[Br 8 TPFPC]-(NO), where Br 8 TPFPC is the highly electron-deficient β-octabromo-meso-tris( pentafluorophenyl)corrole ligand.Qualitatively, the stability of the complexes appears to decrease as the corrole ligand becomes increasingly electron-deficient.Thus, it was gratifying that a single crystal X-ray structure could be obtained for Fe[Br 8 TPFPC](NO), the most electron-deficient complex studied.Electron-poor FeNO corroles also exhibit higher infrared NO stretching frequencies (ν NO ), which vary over a range of some 40 cm −1 .It is tempting to interpret this trend as reflecting a slightly higher degree of {FeNO} 6 character of the metal center, along with reduced radical character on the corrole.We shall see, however, that broken-symmetry DFT calculations indicate a different rationale.

Synthesis and chemical stability
The FeNO octabromocorroles were prepared in much the same manner as the corresponding β-H 8 complexes, i.e., via iron insertion in pyridine-methanol, followed by addition of aqueous NaNO 2 .The syntheses were greatly facilitated by the recent improvements in the syntheses of the H 3 [Br 8 TpXPC] (X = CF 3 , H, CH 3 , OCH 3 ) 10,11 and H 3 [Br 8 TPFPC] ligands. 12,13he characteristic blood-red color of the FeNO complexes provided a simple visual test of their stability, which, as mentioned above, varied significantly across the compounds studied.Thus,  ](NO), turned brown in about an hour or so, indicating decomposition.These complexes also decomposed, with NO loss, in the solid state, but only over days.Spectroscopic and electrochemical characterization was therefore carried out on freshly prepared samples, although certain measurements ( particularly molar extinction coefficients and elemental analyses) could not be carried out for the more unstable complexes.
Molecular structure of Fe[Br 8 TPFPC](NO) and general structural considerations 14 Attempted crystallization and single-crystal X-ray analysis of Fe[Br 8 TpCH 3 PC](NO) led, to our surprise, to a rather poorquality structure for a μ-oxodiiron 10-methoxyisocorrole complex.Fortunately, X-ray quality crystals of Fe[Br 8 TPFPC]-(NO) could be obtained by diffusion of benzene into a heptane solution of the complex, providing the first X-ray structure of an FeNO octabromocorrole.The geometrical parameters of the FeNO units (listed under Fig. 2) were found to be nearly identical to those found for {FeNO} 6 porphyrin-type complexes, 15 i.e., Fe-N(O) ∼1.65 Å, N-O ∼1.16 Å, and a nearly linear FeNO angle >170°.We shall see, however, that other probes, notably infrared spectroscopy and broken-symmetry DFT calculations, favor a {FeNO} 7 -(corrole •2− ) description, for both β-H 8 and β-Br 8 complexes. 16espite the apparent similarity with {FeNO} 6 complexes, careful examination of the skeletal bond distances of Fe-[Br 8 TPFPC](NO) (Fig. 3), as well as those in DFT (B3LYP/ STO-TZ2P) optimized structures, does afford evidence of corrole radical character.The bipyrrole part of the corrole macrocycle of Fe[Br 8 TPFPC](NO) exhibits characteristic bond length alternations that are not seen in an innocent [Br 8 TPFPC] complex such as Ir[Br 8 TPFPC](Me 3 N) 2 . 17Analogous bond length alternations have also been noted for noninnocent β-H 8 complexes such as Fe[TPC](NO) and Fe[TPC]Cl. 5hese bond length alternations all correspond to removal of an electron from the corrole HOMO depicted in Fig. 4.

DFT calculations
Broken-symmetry DFT (B3LYP) calculations played a major role in establishing the noninnocent, radical nature of FeNO corroles. 5An examination of the broken-symmetry spin density profiles (Fig. 5) and Mulliken spin populations (Table 2) now also provide an attractive explanation for the variation of ν NO across the complexes studied.Table 2 shows that both the Fe and NO spin populations decrease (in absolute value) with increasing electron-deficient character of the corrole ligand.Because the latter spin density may be viewed as a measure of occupation of the NO π* orbitals, we may conclude that a strongly electron-deficient corrole ligand leads to a lesser degree of backbonding and hence to a stronger NO bond and a higher ν NO .Thus, between Fe[TpOMePC](NO) and Fe[Br 8 TPFPC](NO), the magnitude of the broken-symmetry   spin population on the NO decreases by 0.12, which correlates with an experimentally observed upshift of 42 cm −1 in the ν NO .Interestingly, there does not appear to be a clear correlation between the amount of broken-symmetry spin density on the macrocycle and its electron-deficient character.Indeed, to a first approximation, all FeNO corroles exhibit the same degree of corrole radical character (as measured by the amount of broken-symmetry spin density on the corrole).
In our earlier study of FeNO meso-triarylcorroles, 5 we identified 3-4 IR skeletal modes that are indicative of a noninnocent corrole macrocycle.Using IR spectroscopy and DFT calculations in concert, we have now identified the same modes for FeNO octabromocorroles (Fig. 6 and 7).Fig. 7 depicts the significant shifts in these bands between noninnocent Fe[Br 8 TPC](NO) and the known, innocent analogue Au-[Br 8 TPC].These frequency shifts may be qualitatively interpreted in terms of changes in skeletal bond order as a result of removal of an electron from the corrole HOMO depicted in Fig. 4.

Redox potentials
As shown in Table 1, both the meso aryl and β-Br substituents exert a significant effect on the redox potentials of FeNO cor-roles.Thus, β-octabromination upshifts both the oxidation and reduction potentials by about 400 mV (Fig. 8).As shown in Fig. 5, the calculated spin densities of both the cationic and anionic states of a typical FeNO corrole are almost exclusively localized on the FeNO unit in a manner that is characteristic of a S = 1/2 {FeNO} 7 group.[20][21]

Electronic absorption spectroscopy
The Soret maxima of several but not all families of metallotriarylcorroles redshift markedly in response to increasing electron-donating character of the para substituent on the meso-aryl groups.Over a lengthy series of studies, 5,6,18,[22][23][24][25][26] we have shown that this effect is specific to noninnocent corroles and the substituent effects arise from aryl-to-corrole charge transfer character in certain transitions in the Soret region.Indeed, the strongly substituent-sensitive Soret maxima of the Fe[TpXPC](NO) series provided the first clue that FeNO corroles are noninnocent, i.e. not {FeNO} 6 . 5The Fe[Br 8 TpXPC]-(NO) series reported herein provides the first example of a series of noninnocent metallotriarylcorroles whose Soret Fig. 7 Selected IR marker bands, including visual depiction of the eigenvectors and experimental and broken-symmetry B3LYP/STO-TZP frequencies (cm −1 ).

Dalton Transactions Paper
This maxima are essentially unaffected by para substituents on the meso-aryl groups (Table 1 and Fig. 9).Based on an examination of the valence MOs of the complexes in question, this unique behavior appears to be clearly attributable to steric inhibition of resonance, i.e., the bulky β-Br groups preclude any significant mesomeric interactions between the aryl groups and the corrole macrocycle in these compounds.This scenario may be contrasted with the Cu[Br 8 TpXPC] series, where strong saddling leads to enhanced aryl-corrole mesomeric interactions and hence also to strongly substituent-sensitive Soret maxima.18a,d,e

Conclusions
A detailed, multitechnique study of FeNO octabromocorroles has led to considerable deepening of our appreciation of FeNO corroles as strongly noninnocent complexes.Broken-symmetry DFT calculations indicate that the amount of radical character of the corrole macrocycle is essentially constant for all the FeNO corroles studied and does not vary significantly with increasing electron-deficient character of the corrole macrocycle.Instead, an electron-deficient corrole ligand leads to decreased π-backdonation from the iron to the NO, which manifests itself experimentally in a higher NO stretching frequency.The relative instability of the more electron-deficient FeNO corroles with respect to NO loss is also reasonably ascribed to decreased π-backdonation in these complexes.A combined experimental-DFT approach has also led to the identification of structural and IR spectroscopic signatures for ligand noninnocence in FeNO corroles.Uniquely, the Soret maxima of FeNO octabromocorroles do not shift as a function of para substituents on the meso-aryl groups, as they do for other families of noninnocent metallotriarylcorroles.Steric inhibition of aryl-corrole mesomeric interactions in these sterically crowded complexes appears to be responsible for this unusual phenomenon.

Experimental section
Instrumentation Ultraviolet-visible (UV-vis) spectra were recorded on an HP 8454 spectrophotometer in CH 2 Cl 2 .Cyclic voltammetry was carried out with an EG&G model 263A potentiostat equipped with a three-electrode system: a glassy carbon working electrode, a platinum wire counter electrode, and a saturated calomel reference electrode (SCE).Tetrabutylammonium perchlorate, recrystallized twice from absolute ethanol and dried in a desiccator for at least 2 weeks, was used as the supporting electrolyte.The reference electrode was separated from the bulk solution by a fritted-glass bridge filled with a solvent/supporting electrolyte mixture.All potentials were referenced to the SCE.A scan rate of 100 mV s −1 was used.The anhydrous CH 2 Cl 2 solutions were purged with argon for at least 5 min prior to the electrochemical measurements and an argon blanket was maintained over the solutions during the measurements. 1H NMR spectra (400 MHz) were recorded in CDCl 3 (referenced to residual CHCl 3 at 7.26 ppm) at 298 K on a Varian Inova 400 spectrometer.High-resolution electrosprayionization (HR-ESI) mass spectra were recorded on an LTQ Orbitrap XL mass spectrometer.IR spectra were acquired as an average of 32 scans with a 1 cm −1 resolution on a Varian 7000e FT-IR spectrometer.

General synthetic details
All reagents and solvents were obtained from Sigma-Aldrich and used as purchased unless otherwise noted.CHROMA-SOLV® HPLC grade n-hexane and dichloromethane were used for chromatography.FeCl 2 •4H 2 O was obtained from Merck.NaNO 2 was obtained from Fluka.Chloroform-d was obtained from Euriso-top.Pyrrole was passed through a short column of basic alumina (Merck, 5 cm height) before use to remove brownish impurities and stored at −20°C.Silica gel 60 (0.04-0.063 mm particle size; 230-400 mesh, Merck) was used for flash chromatography.
Free-base β-octabromo-meso-triarylcorroles H 3 [Br 8 TpXPC] 11 and H 3 [Br 8 TPFPC] 12 were prepared via the reductive demetalation of the copper and manganese complexes, respectively, as described earlier.For H 3 [Br 8 TpXPC] (X = CH 3 and OCH 3 ), iron insertion and nitrosylation could be achieved in good yields via slight modification of a one-pot literature procedure.2a The procedure, however, did not work for H 3 [Br 8 TPFPC] and led to relatively poor yields for H 3 [Br 8 TPC] and H 3 [Br 8 TpCF 3 PC].In the latter two cases, the major byproduct could be readily converted to the FeCl corrole, 27 and subsequently nitrosylated via another literature procedure.For H 3 [Br 8 TPFPC], a two-step procedure via Fe[TPFPC]Cl and subsequent nitrosylation was employed.All of the iron nitrosyl complexes were stored in sealed vials at 4 °C.Because of the limited stability of the compounds, especially with respect to denitrosylation, satisfactory elemental analyses did not prove possible.Full spectroscopic characterization was carried out on all the compounds and, in one case, also single-crystal X-ray structure determination.Detailed procedures for the preparation and purification of the various complexes are as follows.

Synthesis of Fe[Br 8 TpOMePC](NO)
Free-base β-octabromo-meso-tris(4-methoxyphenyl)corrole (100 mg, 80 µmol) was added to a 50 mL two-necked roundbottomed flask equipped with a magnetic stirrer and a reflux condenser and dissolved in pyridine (4.0 mL) and methanol (8.0 mL).The system was the closed off with rubber septa and vented via an oil bubbler.Argon was bubbled through the solution while it was virgorously stirred for 20 min.Iron(II) chloride tetrahydrate (318 mg, 1.6 mmol) was added to the solution under an argon blanket.The solution was then heated to 69 °C under a slow flow of argon for 1 h after which heating was discontinued and a saturated aqueous solution of NaNO 2 (2.0 mL) was added dropwise via syringe, followed by stirring for an additional 30 min.The solution was then placed in an ice bath, quenched with cold water (40 mL), and filtered.The dry, filtered precipitate was washed with CH 2 Cl 2 and the filtrate was rotary-evaporated to dryness.Flash chromatography (silica, 1 : 2 CH 2 Cl 2 : hexane, subsequently 1 : 1) afforded Fe[Br 8 TpOCH 3 PC](NO) as the first, deep red band.Yield: 98 mg (92%).UV-vis (CH 2 Cl 2 ): λ max (ε × 10 −4 (M −1 cm −1 )) (nm); 394 (7.1), 549 (1.6).

Synthesis of Fe[Br 8 TPC](NO)
Free-base β-octabromo-meso-triphenylcorrole (42 mg, 36 µmol) and iron(II) chloride tetrahydrate (144 mg, 0.72 mmol) were subjected to the same procedure as described above.Flash chromatography (silica, 1 : 4 CH 2 Cl 2 : hexane) afforded Fe-[Br 8 TPC](NO) as the first deep red band and also yielded second, darker reddish-brown band.The latter band was collected, evaporated to dryness, and dissolved in 20 mL CH 2 Cl 2 .The solution was washed twice with ∼20 mL 7% HCl (aq.) and once with water, dried over MgSO 4 , and filtered.A saturated aqueous solution of NaNO 2 (2.0 mL) was added to the solution, and the solution was stirred for 4 h at room temperature.The solution was washed with water, dried over MgSO 4 , filtered, and evaporated to dryness.

Synthesis of Fe[Br 8 TPFPC]Cl
Free-base β-octabromo-meso-tris( pentafluorophenyl)corrole (30 mg, 21 µmol) was added to a 50 mL two-necked round-bottomed flask equipped with a magnetic stirrer and a reflux condenser and dissolved in dry DMF (25 mL).The system was the closed off with rubber septa and vented via an oil bubbler.Argon was bubbled through the solution under vigorous stirring for 20 min.Anhydrous iron(II) chloride (53 mg, 0.42 mmol) was added to the solution under an argon blanket and the solution was heated to reflux.The reaction was monitored by TLC (silica, CH 2 Cl 2 : hexane 1 : 2).After 40 min of refluxing, no starting material could be detected by TLC and heating was discontinued.The solution was cooled to room temperature and the solvent removed via rotary evaporation.The solid material was dissolved in diethyl ether and passed through a short silica column with diethyl ether as eluent.After evaporation of the diethyl ether, the resulting solid was dissolved in CH 2 Cl 2 and washed twice with 2 M HCl (aq.) and once with water.The organic layer was dried over Na 2 SO 4 and evaporated to dryness.Flash chromatography (silica, CH 2 Cl 2 ) afforded Fe[Br 8 TPFPC]Cl as a reddish brown band.Yield: 29 mg (91%).UV-vis (CH 2 Cl 2 ): λ max , nm (log ε, M −1 cm −1 ) 418 (4.66), 563 (0.79). 19

Synthesis of Fe[Br 8 TPFPC](NO)
Chloroiron β-octabromo-meso-tris( pentafluorophenyl)corrole (20 mg, 13 µmol) was dissolved in CH 2 Cl 2 (20 mL) and a saturated aqueous solution of NaNO 2 (5.0 mL) was added to the solution.The solution was stirred at room temperature for 6 h, washed with water, dried over Na 2 SO 4 , and filtered, and the filtrate was evaporated to dryness.Flash chromatography (silica, 1 : 2 CH 2 Cl 2 : hexane) of the residue afforded Fe[Br 8 TPFPC](NO) as a deep red band.Yield: 13 mg (68%).X-ray quality crystals were obtained by vapor diffusion of n-heptane into a benzene solution of the title compound within a sealed vial.UV-vis (CH 2 Cl 2 ): λ max , nm (log ε, M −1 cm −1 ) 395 (5.52), 560 (1.28). 19 The structure was solved by dual-space methods (SHELXT) and refined by full-matrix least-squares on F 2 (SHELXL-97) using 1282 parameters and 0 restraints.The hydrogen atoms on carbon atoms were generated geometrically and refined as riding atoms with C-H = 0.95-0.99Å and U iso (H) = 1.2 times U eq (C) for aromatic and CH 2 groups and U iso (H) = 1.5 times U eq (C) for CH 3 groups.The maximum and minimum peaks in the final difference Fourier map were 4.

Computational details
All calculations were carried out with the ADF 2014 28 program system, the B3LYP 29,30 exchange-correlation functional, and appropriately fine integration grids and tight SCF and geometry optimization criteria.Scalar relativistic effects were taken into account with the ZORA Hamiltonian 31 and ZORA STO-TZP basis sets, except for the structures shown in Fig. 3, for which ZORA STO-TZ2P basis sets were employed. 32Grimme's D3 dispersion correction was used throughout. 33All optimized geometries were verified as true minima via numerical frequency calculations.

Fig. 5
Fig. 5 Spin density plots for the neutral, anionic, and cationic states of Fe[Br 8 TPC](NO).The α and β spin densities are shown in cyan and burgundy, respectively, The contour has been set at 0.009 e Å −3 .