Electronic Structure of Metallophlorins: Lessons from Iridium and Gold Phlorin Derivatives

Phlorins have long remained underexplored relative to their fully conjugated counterparts, such as porphyrins, hydroporphyrins, and corroles. Herein, we have attempted to bridge that knowledge gap with a scalar-relativistic density functional theory (DFT) study of unsubstituted iridium and gold phlorin derivatives and a multitechnique experimental study of iridium-bispyridine and gold complexes of 5,5-dimethyl-10,15,20-tris(pentafluorophenyl)phlorin. Theory and experiments concur that the phlorin derivatives exhibit substantially smaller HOMO–LUMO gaps, as reflected in a variety of observable properties. Thus, the experimentally studied Ir and Au complexes absorb strongly in the near-infrared (NIR), with absorption maxima at 806 and 770 nm, respectively. The two complexes are also weakly phosphorescent with emission maxima at 950 and 967 nm, respectively. They were also found to photosensitize singlet oxygen formation, with quantum yields of 40 and 28%, respectively. The near-infrared (NIR) absorption and emission are consonants with smaller electrochemical HOMO–LUMO gaps of ∼1.6 V, compared to values of ∼2.1 V, for electronically innocent porphyrins and corroles. Interestingly, both the first oxidation and reduction potentials of the Ir complex are some 600 mV shifted to more negative potentials relative to those of the Au complex, indicating an exceptionally electron-rich macrocycle in the case of the Ir complex.


■ INTRODUCTION
Since the beginnings of photodynamic therapy in the early part of the twentieth century, porphyrins have been a cornerstone of the approach. 1,2−7 Very recently, we have begun to survey a wider range of 4d and 5d metalloporphyrin analogues, with an emphasis on near-infrared (NIR) absorbing systems.Examples include recent studies of isocorroles 8,9 and 6-azahemiporphycenes. 10 In the present study, we have attempted to extend our survey to phlorins, 11−14 a class of reduced porphyrins with a storied history.
In 1960, Woodward and co-workers discovered a dihydroporphyrin intermediate, which readily oxidized to a porphyrin, in the course of their celebrated total synthesis of chlorophyll a. 15−17 Woodward named this intermediate phlorin and speculated that a similar isomer of porphyrin should also exist.Ten years later, Dolphin and co-workers realized Woodward's prediction by isolating a zinc isoporphyrin. 18Over the subsequent half-century, several examples of phlorins 11−14 and isoporphyrins 19−22 have appeared in the literature.−26 However, our general appreciation of the electronic structure of phlorins has languished, a key knowledge gap that we have also sought to bridge in this study.
In free-base form, the distinction between phlorin and isoporphyrin is rather obvious (Scheme 1).Phlorin is a twoelectron reduced derivative of porphyrin, in fact an isomer of chlorin.Phlorin also resembles corrole by virtue of its triprotic N 4 core.Isoporphyrin, in contrast, is a porphyrin isomer with a monoprotic N 4 core.For complexes involving redox-active metals, however, the distinction between the two macrocycles can be tricky; indeed, an Au(III) phlorin and an Au(I) isoporphyrin are only valence tautomers.Metal−ligand bond distances provide an important clue, but quantum chemistry, with careful control of group theory, can also provide a definitive distinction, as we show in this study.
Against that backdrop, we report here a combined experimental and theoretical study of two 5d metallophlorins, including wide-ranging density functional theory (DFT) calculations, temperature-dependent NMR studies of structure and conformation, cyclic voltammetry, optical and photophysical studies, and singlet oxygen sensitization measurements.The experimental studies were carried out on gold 5,5dimethyl-10,15,20-tris(pentafluorophenyl)phlorin, Au-[DMTPFPhl], a known compound, 27,28 and on its iridiumbispyridine analogue Ir[DMTPFPhl](py) 2 , a new compound and the first iridium phlorin (Scheme 2).The results, in our view, significantly deepen our understanding of key geometric and electronic structural aspects of metallophlorins.
■ RESULTS AND DISCUSSION DFT Calculations.Phlorin may be viewed as formally arising via a nucleophilic attack by a hydride on the meso carbon of a porphyrin.This formal transformation has profound structural and electronic consequences.Structurally, phlorins are strongly ruffled, with meso positions alternately above and below the mean plane of the central nitrogens.The ruffling destroys the σ/π distinction among molecular orbitals, resulting in extensive metal(d)-macrocycle(π) orbital mixing.−31 Thus, a porphyrin's a 2u HOMO, with large amplitudes at meso positions, may be expected to be strongly destabilized in a phlorin, whereas the a 1u HOMO should remain relatively unperturbed.These expectations are confirmed by scalar-relativistic DFT (B3LYP 32−34 -D3 35 /STO-TZ2P) calculations on the model compounds Au[Phl] and Ir[Phl](NH 3 ) 2 , where Phl 3− is the trianion of unsubstituted phlorin.Both molecules were found to exhibit strongly ruffled, C s optimized geometries, with the planar C 2v geometries about a quarter of an eV higher in energy.To place their electronic properties in context, we also performed analogous calculations on unsubstituted platinum porphyrin, Pt[Por], and gold corrole, Au [Cor]. 36The main results are presented in Table 1 and Figures 1 and 2 and may be summarized as follows.
• For the two reference compounds Pt[Por] and Au[Cor], 9][30][31]37 The values for Pt [Por] are in excellent agreement with those determined for free-base Scheme 1. Proticity and Isomeric Relationships Among Free-Base Porphyrinoids

Inorganic Chemistry
−43 The EAs of about 1.2 eV are also in excellent agreement with experimental gas-phase values 44 and previous DFT calculations. 45,46reflecting smaller HOMO−LUMO gaps, as expected for nonaromatic compounds.Finally, an examination of the occupied MOs allowed us to rule out any potential isoporphyrin character for the compounds studied.Taking the case of Au[Phl], note that the purely macrocycle-based HOMO transforms as a′ under C s point group symmetry, whereas the LUMO and LUMO + 1, each with partial Au(5d x 2 −y 2 ) character, transform as a″.Accordingly, using simple group-theoretical manipulations (moving two electrons from a′ to a′′), we could computationally study the Au(I) isoporphyrin state.The optimized geometry revealed a relatively planar macrocycle, as indeed observed experimentally for an authentic Zn-isoporphyrin complex. 47However, B3LYP-D3 calculations (as well as DFT calculations with several other common exchange−correlation functionals) revealed an energy ∼40 kcal/mol above the Au(III) phlorin ground state, conclusively ruling out an isoporphyrin ground state.Whether axial ligands, such as phosphines and isocyanides, might stabilize an Au(I) isoporphyrin remains an intriguing question for the future.
The above theoretical findings are nicely confirmed by Ultraviolet−visible−near-infrared (UV−vis-NIR) absorption, phosphorescence, and electrochemical measurements, as outlined below.
Synthesis and Characterization.We chose an iridiumbispyridine and a gold phlorin derivative for a general physicochemical study of metallophlorins.These complexes are not only expected to be electronically innocent but were also of interest as potential candidates for NIR phosphorescence and singlet oxygen-sensitizing properties.The ligand used in our experimental studies, 5,5-dimethyl-10,15,20tris(pentafluorophenyl)phlorin, DMTPFPhl, was chosen in light of the superior photophysical properties found for pentafluorophenyl-appended corroles. 48Of the two complexes, Au[DMTPFPhl] is a known compound; 28 its crystal structure (Figure 3), however, had not been previously reported.The complex, like other phlorins, exhibits a strongly ruffled macrocycle.The 1 H and 19 F NMR spectra of the Ir complex from ambient temperature down to −50 °C (Figures S1 and  S2), however, are consistent with time-averaged C 2v , as opposed to C s , symmetry.Thus, the two geminal methyl groups at the 5-position appear as a single signal.Also, in the 19 F NMR spectra, the ortho and meta fluorines on a given mesopentafluorophenyl group do not split into diastereotopic pairs.B3LYP-D3 calculations suggest a plausible explanation for the conundrum: rapid ruffling inversion of the phlorin macrocycle on the nuclear magnetic resonance (NMR) time scale.Thus, the calculations predict an energy of only 0.18 eV for the planar C 2v conformation of Ir[Phl](NH 3 ) 2 relative to the ruffled C s ground state (the corresponding value for Au[Phl] is 0.24 eV).Frequency analyses established these planar forms as true transition states with a single imaginary frequency.We also calculated the barriers for the actual complexes studied, including all peripheral substituents, and found the barriers to be only slightly higher, ∼0.33 eV for either metal; such a value is also consistent with time-averaged C 2v geometry on the NMR time scale.
As noted before for Au phlorins, 28 both complexes exhibit strong NIR absorption in the 700−900 nm region, with λ max at 806 and 770 nm for Ir[DMTPFPhl](py) 2 and Au-[DMTPFPhl], respectively (Figure 4).These absorption maxima are some 200 nm red-shifted relative to those observed for Ir 49−53 and Au 54−58 corroles, respectively, indicating a significantly smaller HOMO−LUMO gap in the case of the metallophlorins.The same picture also emerges from cyclic voltammetry measurements, as described below.
Au[DMTPFPhl] exhibits four oxidations, of which the first and third are essentially reversible and the second and fourth are partially irreversible. 28Following oxidation over the full positive potential range, the reduction sweep exhibits a certain hysteresis in the form of a peak at −0.02 V vs saturated calomel electrode (SCE).Performing separate scans for the first For each value, the symmetry of the higher-energy state is indicated within parentheses (the lower-energy state always being the fully symmetric ground state).b These states were only optimized under C s , with the symmetry plane identified with the mean molecular plane, so as to allow for Jahn−Teller distortions.c Because the LUMOs of Pt[Por] are degenerate (e g ) under D 4h , the vertical electron affinity was computed by running the anion calculation under the Abelian subgroup D 2h (which has no degenerate representations).oxidation and the first and second oxidation reveals that the peak at −0.02 V vs SCE is due to the quasi-reversibility of the second oxidation.The complex, however, reaches its normal neutral state before reduction to the anion commences.The first oxidation and reduction potential of Au[DMTPFPhl] at 0.62 and −0.94 V vs SCE, respectively, translate to a rather low electrochemical HOMO−LUMO gap of 1.56 V.A similarly low electrochemical HOMO−LUMO gap is also observed for Ir[DMTPFPhl](py) 2 (Figure 5).These HOMO−LUMO gaps are well over half a volt lower than those observed for Ir and Au corroles.
As expected on the basis DFT calculations described above (Table 1), the redox potentials of the two metallophlorins examined differ substantially: both the first oxidation and reduction potentials of Ir[DMTPFPhl](py) 2 are shifted to more negative potentials by some 600 mV relative to those of Au[DMTPFPhl], an indication that the phlorin macrocycle in the Ir complex is exceptionally electron-rich.For both metals, the redox-active MOs (the HOMO and the LUMO) are similar in character, except that, as discussed above, ruffling lifts the orthogonality between the phlorin's π-LUMO and Au 5d x 2 -y 2 orbital (Figures 1 and 2), so the reduction in the Au case occurs in a partially metal-centered manner.The electronrichness of Ir phlorins relative to their Au counterparts nicely mirrors the previously noted electron-richness of Ir corroles relative to Au corroles.Indeed, in the absence of strongly electron-withdrawing substituents, the reduction potentials of Ir corroles are generally not observable within the potential window of common electrochemical solvents. 52Finally, the origin of the small feature at −0.93 V for Ir[DMTPFPhl](py) 2 remains uncertain, but a plausible explanation involves loss of NIR Phosphorescence of Phlorins.The phlorins exhibited weak near-infrared phosphorescence in the emission region of 850−1150 nm in anoxic toluene when excited with a 405 nm LED excitation source at room temperature.Ir[DMTPFPhl](py) 2 , with a phosphorescence quantum yield of 0.01%, was observed to exhibit a stronger phosphorescence relative to Au[DMTPFPhl], whose phosphorescence quantum yield was too low to be reliably determined by the relative method (Table 2).Although the overall phosphorescence profiles of both phlorins are similar, Au[DMTPFPhl] exhibits a red-shifted phosphorescence maximum (967 nm) compared to Ir[DMTPFPhl](py) 2 (950 nm, Figure 6), emphasizing the importance of the transition metal center in determining the exact energetics of the luminescence.Their phosphorescence lifetimes, however, are similar: 22 μs for Au[DMTPFPhl] and 20 μs for Ir[DMTPFPhl](py) 2 (see Figures S3 and S4).

■ CONCLUSIONS
We have presented a combined theoretical and experimental study of iridium and gold phlorin derivatives that, in our view, has significantly deepened our understanding of key electronic structural characteristics of metallophlorins.DFT calculations, NIR absorption phosphorescence, and electrochemical measurements all indicate a substantially smaller HOMO−LUMO gap relative to typical, electronically innocent metalloporphyrins and metallocorroles.Interestingly, the Ir complex was found to exhibit dramatically lower redox potentials (by a margin of some 600 mV) relative to its Au counterpart, indicating a much more electron-rich macrocycle in the former.Both the Ir and Au complexes were found to photosensitize singlet oxygen formation, with quantum yields of 40 and 28%, respectively.
■ EXPERIMENTAL SECTION Materials and Instruments.Reagents and solvents were used as purchased, unless noted otherwise.Pyridine was predried over and distilled from NaOH and stored over 4 Å molecular sieves.Dichloromethane was predried over and distilled from P 4 O 10 and stored over activated 4 Å molecular sieves.5,5-Dimethyl-10,15,20tris(pentafluorophenyl)phlorin was prepared as previously described. 28Ultraviolet−visible (UV−vis) spectra were recorded on an HP 8454 spectrophotometer in CH 2 Cl 2 . 1 H (400 MHz) and 19 F (376 MHz) NMR spectra were acquired on a 400 MHz Bruker Avance III HD spectrometer equipped with a 5 mm SmartProbe BB/ 1H (BB = 19  Cyclic voltammetry was carried out at ambient temperature with a Gamry Reference 620 potentiostat equipped with a three-electrode system: a 3 mm disk glassy carbon working electrode, a platinum wire counter electrode, and a saturated calomel reference electrode (SCE).Tetra(n-butyl)ammonium hexafluorophosphate was used as the supporting electrolyte.Anhydrous CH 2 Cl 2 (Aldrich) was used as the solvent.The electrolyte solution was purged with argon for at least 2 min prior to all measurements, which were carried out under an argon blanket.The glassy carbon working electrode was polished using a polishing pad and 0.05 μm polishing alumina from ALS, Japan.All potentials were referenced to the SCE.

S y n t h e s i s o f G o l d 5 , 5 -D i m e t h y l -1 0 , 1 5 , 2 0 -t r i s -(pentafluorophenyl)phlorin.
Into a Schlenk tube charged with free-base phlorin (24 mg, 0.0286 mmol) and gold(III) acetate (5 equiv, 54 mg, 0.144 mmol) and preflushed with argon, dry pyridine (4 mL) was introduced under argon flow.After 5 min of degassing, the mixture was stirred at 50 °C for 50 min, at which point TLC (with 7:3 n-hexane/CH 2 Cl 2 as eluent) on neutral alumina indicated that the free-base phlorin had been completely consumed.The suspension obtained was filtered through Celite and the filtrate was evaporated.The residue obtained was dissolved in a minimum volume of CH 2 Cl 2 and placed on a silica gel column (18 cm × 3 cm).Eluting with 17:3 hexane/CH 2 Cl 2 yielded a yellowish-brown band identified as the desired gold(III) phlorin ( 12 Synthesis of Iridium-bispyridine 5,5-Dimethyl-10,15,20tris(pentafluorophenyl)phlorin.A round-bottom flask was charged with free-base phlorin (26 mg, 0.031 mmol), [Ir(cod)Cl] 2 (20 equiv, 416 mg, 0.62 mmol), and K 2 CO 3 (163 equiv, 0.7 g, 5.06 mmol).Upon flushing with argon, dry THF (10 mL) was introduced under argon flow.After 5 min of degassing, the green suspension was stirred at reflux for 1.5 h, at which point TLC (with 1:1 n-hexane/ CH 2 Cl 2 ) on neutral alumina indicated the continued presence of freebase phlorin.An additional quantity of K 2 CO 3 (226 equiv, 0.97 g, 7.02 mmol) was added to the reaction mixture, and refluxing was continued for an additional 2.5 h.By that time, the reaction mixture turned brown, indicating that the free-base phlorin had been consumed.Heating was discontinued, and pyridine (1 mL) was added to the reaction mixture while stirring was continued.After 0.5 h, the mixture was filtered through a pad of Celite, and the filtrate was evaporated.The residue obtained was dissolved in a minimum volume of CH 2 Cl 2 and chromatographed on a basic alumina column (7 cm × 2 cm) with CH 2 Cl 2 as eluent, whereupon the desired complex was obtained as a brown solid (14 mg, 0.0118 mmol, 38.1%) after washing with n-pentane.Inorganic Chemistry concentrated solution of the complex in pyridine.A suitable crystal with dimensions 0.38 × 0.10 × 0.08 mm 3 was mounted on MiTeGen holder oil on a Nonius Kappa Apex II diffractometer.The crystal was kept at a steady T = 110.0(1)K during data collection.The structure was solved with ShelXT 67 using dual methods and Olex2 1.5 as the graphical interface. 68The model was refined with ShelXL 69 2018/3 using full-matrix least-squares minimization on F 2 using the ShelXle GUI.All non-hydrogen atoms were refined anisotropically.Hydrogen atom positions were calculated geometrically and refined using the riding model.Photophysical Measurements.All photophysical studies (phosphorescence and lifetime) were performed in sealed cuvettes under dry N 2 using degassed toluene.NIR phosphorescence and lifetimes were measured on an OLIS NIR CPL Solo spectrofluorometer using ∼10 −4 M toluene solutions and a 405 nm LED light source.Lifetime spectra were collected using pulsed excitation at 405 nm and time-resolved emission measurements fixed at the peak of the strongest emission.A first-order exponential decay curve was fitted to the collected data to estimate the fluorescence lifetime (τ obs ).Values are reported as measured lifetimes (observation wavelength).The NIR PMT detector is a HAMAMATSU model H10330C-75-C2 cooled at −60 °C using an internal air-cooled thermoelectric cooler.The detector uses an InGaAs photocathode material.The sample is illuminated with LEDs (Everlight) driven to an output of 1000 mW.
For phosphorescence quantum yields, the dye solution in a screwcapped cuvette was unscrewed prior to each measurement, diluted to a new concentration, and then purged with nitrogen for 3 min.Yb(tta) 3 (H 2 O) 2 was used as the reference compound (QY = 0.35% in toluene).Singlet oxygen quantum yields were calculated by the relative method using 9,10-diphenylanthracene as the singlet oxygen acceptor.The metallophlorins were dissolved in 9:1 v/v EtOH/THF to a concentration of approximately 15 μM, adjusted for identical absorption at the excitation wavelength.A 0.28 mM solution of 9,10diphenylanthracene in 9:1 v/v EtOH/THF was also prepared.Equal volumes of the 9,10-diphenylanthracene solution and each photosensitizer solution were then mixed in a 1 cm path-length cuvette.The mixture was saturated with oxygen by bubbling air through it for 3 min.The cuvette was sealed and the absorbance was measured before it was irradiated with light (two 405 nm Everlight LEDs each driven to an output of 1000 mW, 10 nm slit) for 3 min.The cuvette was then shaken and an absorbance spectrum was remeasured.The experiment was repeated 4 times.Singlet oxygen quantum yields were calculated from the slope of the curve (absorbance at 372 nm vs. time) using methylene blue as the reference (φ = 0.48).
Computational Methods.The DFT calculations described above employed the scalar-relativistic ZORA 70−72 (zeroth-order regular approximation to the Dirac equation) Hamiltonian, the dispersion-corrected B3LYP-D3 method, and all-electron ZORA STO/TZ2P basis sets, all as implemented in the ADF program system. 73,74A number of additional exchange−correlation functionals were also examined, but both the optimized geometries and calculated IPs showed minimal variations across the different methods; 39−43 accordingly, only the B3LYP-D3 results have been reported here.The following point group symmetries were used: Pt[Por] (D 4h ), Au[Cor] (C 2v ), Au[Phl] (C s ), and Ir[Phl](NH 3 ) 2 (C s ).All energies were calculated in the gas phase with the ΔSCF method, i.e., as differences in total electronic energy between initial and final states.When needed, different ionized and excited states were calculated by specifying the number of electrons in each irrep; these occupations are all explicitly specified in the Supporting Information.Note also that the optimized geometry of {Pt[Por]} + (used in the calculation of the adiabatic IP of the neutral compound) is only C 4h , as a result of skeletal bond length alternations arising from a pseudo-Jahn−Teller distortion. 75

■ ASSOCIATED CONTENT Data Availability Statement
All data generated or analyzed in this study are included in this published article and its Supporting Information.

Figure 5 .
Figure 5. Cyclic voltammograms in dry dichloromethane with 0.1 M TBAPF 6 at ambient temperature.Scan rate = 100 mV/s.The inset of Au[DMTPFPhl] shows the first oxidation and the first and second oxidation as separate scans.

Table 2 .
Summary of the Phosphorescence and Lifetime of the PhlorinsToo weak to be reliably determined by the relative method.b "−" Not reported. a