β,β-Directly Linked Porphyrin Rings: Synthesis, Photophysical Properties, and Fullerene Binding

Cyclic porphyrin oligomers have been studied as models for photosynthetic light-harvesting antenna complexes and as potential receptors for supramolecular chemistry. Here, we report the synthesis of unprecedented β,β-directly linked cyclic zinc porphyrin oligomers, the trimer (CP3) and tetramer (CP4), by Yamamoto coupling of a 2,3-dibromoporphyrin precursor. Their three-dimensional structures were confirmed by nuclear magnetic resonance (NMR) spectroscopy, mass spectrometry, and single-crystal X-ray diffraction analyses. The minimum-energy geometries of CP3 and CP4 have propeller and saddle shapes, respectively, as calculated using density functional theory. Their different geometries result in distinct photophysical and electrochemical properties. The smaller dihedral angles between the porphyrin units in CP3, compared with CP4, result in stronger π-conjugation, splitting the ultraviolet–vis absorption bands and shifting them to longer wavelengths. Analysis of the crystallographic bond lengths indicates that the central benzene ring of the CP3 is partially aromatic [harmonic oscillator model of aromaticity (HOMA) 0.52], whereas the central cyclooctatetraene ring of the CP4 is non-aromatic (HOMA –0.02). The saddle-shaped structure of CP4 makes it a ditopic receptor for fullerenes, with affinity constants of (1.1 ± 0.4) × 105 M–1 for C70 and (2.2 ± 0.1) × 104 M–1 for C60, respectively, in toluene solution at 298 K. The formation of a 1:2 complex with C60 is confirmed by NMR titration and single-crystal X-ray diffraction.


UV-vis Absorption, Fluorescence and Excitation Spectra
All measurements were carried out in non-deaerated toluene solution at 298 K. UV-vis absorption spectra of P1, CP3 and CP4 were measured at concentration of 10 -6 M; for fluorescence measurement, excitation wavelengths are 417 nm (P1), 424 nm (CP3) and 420 nm (CP4). Fluorescence lifetimes were measured in timecorrelated single photon counting (TCSPC) mode using a picosecond pulsed diode laser (EPL-475, λ = 473.5 nm) as the excitation source and detection at 631 nm, 680 nm and 656 nm for P1, CP3 and CP4, respectively. Fluorescence quantum yields were measured using tetraphenylporphyrin(Zn) (ZnTPP) (Φ = 0.029 in nondeaerated toluene) as standard. 6 Fluorescence quantum yields were calculated using the formula: where subscripts (s) and (r) refer to sample and reference, Φ is the fluorescence quantum yield, F is the integral photon flux, 1 -10 -A(λEx) is the absorption factor at the wavelength of excitation. 7 Table S1. Summary of photophysical properties of P1, CP3 and CP4.  S8 Figure S3. Comparison of UV-vis absorption and fluorescence excitation spectra of CP4 measured in toluene solution at 25 °C. Figure S4. Plots of integrated emission peak (560-750 nm) against 1-10 -A(λEx) value of reference compound TPP(Zn) and P1 in toluene at 298 K. Two independent measurements were conducted with the excitation wavelength at a) 417 nm and b) 543 nm, respectively. The average fluorescence quantum yield of P1 is calculated to be 0.031. Figure S5. Plots of integrated emission peak (560-750 nm) against 1-10 -A(λEx) value of reference compound TPP(Zn) and CP3 in toluene at 298 K. Two independent measurements were conducted with the excitation wavelength at a) 421 nm and b) 424 nm, respectively. The average fluorescence quantum yield of CP3 is calculated to be 0.040.
S9 Figure S6. Plots of integrated emission peak (560-750 nm) against 1-10 -A(λEx) value of reference compound TPP(Zn) and CP4 in toluene at 298 K. Two independent measurements were conducted with the excitation wavelength at a) 420 nm and b) 426 nm, respectively. The average fluorescence quantum yield of CP4 is calculated to be 0.068. a) b) S10 DFT Calculations DFT calculations were performed using Gaussian 16/A.03 software package. 8 Geometries were optimized for each conformation of CP3 and CP4 using B3LYP level of theory and 6-31G(d,p) basis set. Nucleus independent chemical shifts (NICS) were calculated using the gauge invariant atomic orbital (GIAO) approach, as implemented in Gaussian 16/A.03, at the GIAO-B3LYP/6-31G(d,p) level. 9 Figure S7. The lowest-energy geometries of a) CP3 (only PPP conformer was shown) and b) CP4 calculated at the B3LYP/6-31G(d,p) level of theory and the relative energy of the other conformation of c) CP3 and d) CP4 calculated using the same method. Figure S8. Frontier molecular orbitals and energy levels of P1 calculated by DFT at the B3LYP/6-31G(d,p) level and it has doubly degenerate LUMO and nearly degenerate HOMO. S11 Figure S9. Frontier molecular orbitals and energy levels of CP3 calculated by DFT at the B3LYP/6-31G(d,p) level and it has doubly degenerate HOMO and LUMO. where Ri and Ropt are the i th bond length of the C-C bond in the analyzed ring and the bond length of benzene ring (Ropt = 1.388 Å), respectively. n is the number of C-C bonds in the analyzed ring and α = 257.7 Å -2 is a normalization factor that gives HOMA value of 1 for perfect aromatic benzene ring and a HOMA value of 0 for an alternating nonaromatic Kekulé cyclohexatriene ring.
The uncertainty (standard deviation) in the HOMA value (sH) was calculated from the uncertainty of the bond lengths (sR,i) using the equation: The bond length data and calculated HOMA indexes of the central six-and eight-membered rings for CP3, CP4 and CP4·2C 60 are shown in the following table: Typical procedure for UV-vis titrations: Fullerene (c = 1.65 × 10 -3 M for C60 and 3.34 × 10 -4 M for C70) dissolved in a toluene solution containing CP4 (c = 1.65 × 10 -6 M) was added to the toluene solution of CP4 (c = 1.65 × 10 -6 M) and the UV-vis absorption spectra were recorded at 298 K. The change in the absorbance of CP4 caused by addition of C60 was calculated by subtracting the absorption intensity at 421 nm to 605 nm, at which wavelengths C60 has the same absorbance. For C70, the absorption intensity at 421 nm was subtracted to that at 518 nm to calculate the change of UV-vis absorption caused by adding C70. Binding curves were obtained by plotting y = A421 nm -A605 nm or A421 nm -A518 nm against the concentration of C60 and C70. Association constants Ka were evaluated by applying a nonlinear curve fitting of y observed for CP4 upon titration with C60 or C70 using the following equation: where, ymax indicates the maximum change of UV-vis absorption intensity at complete complexation of CP4, Ka, x and H indicate the binding constant, concentration of fullerenes and concentration of binding sites of host CP4, respectively. A is a constant. Figure S11. UV-vis absorption spectra change of the solution of CP4 upon addition of C60 at room temperature. The spectra were corrected by subtracting C60 absorption background. Inset shows change of (A421 nm -A605 nm) with the addition of C60 and the red line is the fitting curve using the 1:1 binding equation. Figure S12. UV-vis absorption spectra change of the solution of CP4 upon addition of C60 at room temperature. The spectra were corrected by subtracting C60 absorption background. Inset shows change of (A421 nm -A605 nm) with the addition of equivalents of C60 and the red line is the fitting curve using the 1:1 binding equation. Figure S13. UV-vis absorption spectra change of the solution of CP4 upon addition of C70 at room temperature. The spectra were corrected by subtracting C70 absorption background. Inset shows change of (A413 nm -A518 nm) with the addition of C70 and the red line is the fitting curve using the 1:1 binding equation. Figure S14. UV-vis absorption spectra change of the solution of CP4 upon addition of C70 at room temperature. The spectra were corrected by subtracting C70 absorption background. Inset shows change of (A413 nm -A518 nm) with the addition of C70 and the red line is the fitting curve using the 1:1 binding equation.
Typical procedure for fluorescence titrations: Fullerene (c = 1.68 × 10 -3 M for C60 and 3.34 × 10 -4 M for C70) dissolved in a toluene solution containing CP4 (c = 1.65 × 10 -7 M) was added to the toluene solution of CP4 (c = 1.65 × 10 -7 M) and the fluorescence spectra were recorded at 298 K with excitation wavelength of 415 nm. Binding curves were obtained by plotting the fluorescence intensity at 600 nm against the concentration of fullerene. Association constants Ka were evaluated by applying a nonlinear curve fitting of y observed for CP4 upon titration with C60 or C70 using the flowing equation: where, ymax indicates the maximum change of fluorescence intensity at complete complexation of CP4 with fullerene, Ka, x and H indicates the binding constant, concentration of fullerenes and concentration of binding sites of CP4, respectively. A is a constant. The average binding constants of two independent measurements are (1.13 ± 0.02) × 10 5 M -1 and (7.83 ± 0.06) × 10 5 M -1 for C60 and C70, respectively.     Typical procedure for 1 H NMR titrations: Equivalents of fullerene dissolved in toluene-d8 was added to a solution of CP4 in toluene-d8 (0.7 mL) and the excess solvent was evaporated to sustain the whole solvent volume of 0.7 mL. The resulting solution was subjected to 1 H NMR spectroscopy (400 MHz) measurement at 298 K. Binding curves were obtained by plotting the chemical shift of protons (y) against the concentration of fullerene. Association constants Ka were evaluated by applying a nonlinear curve fitting of y observed for CP4 upon titration with C60 or C70 using the flowing equation: where, ymax indicates the maximum change of proton chemical shift at complete complexation of CP4 with fullerene, Ka, x and H indicates the binding constant, concentration of fullerenes and concentration of binding sites of CP4, respectively. A is a constant. S17 Figure S19. 1 H NMR spectra change of CP4 upon addition of C60 in toluene-d8 at 298 K (400 MHz).  Single crystal X-ray diffraction data for structure 1 were collected using a (Rigaku) Oxford Diffraction SuperNova A diffractometer and reduced using CrysAlisPro. The structure was solved using SuperFlip 12 and refined using CRYSTALS 13,14 as detailed in the CIF. Figure S24. X-ray Single-crystal structure of 1 (Hydrogen atoms and solvent molecules were omitted for clarity). The thermal ellipsoids are 50% probability level. The crystal used was modulated, but sufficient information could be obtained from using the standard cell and a disordered structure along with various restraints (DFIX, DANG, FLAT, SADI, BUMP, SIMU, RIGU). Solvent masking was employed, suggesting nine solvent methanol molecules per asymmetric unit. Figure S25. X-ray Single-crystal structure of CP3 (hydrogen atoms and solvent molecules omitted for clarity). The thermal ellipsoids are 50% probability level.  Data for crystals of CP4·2C 60 were initially collected using synchrotron radiation at Diamond Light Source 18 and processed using the XIA2 software. 19 The data for this structure (included herein as CP4·2C 60 -II) were of poor quality, and exhibited a structure with two CP4 molecules and four fullerenes in the asymmetric unit. The structure solved well using SuperFlip, 12 but the refinement, carried out on F 2 within the CRYSTALS suite 13,14,20 was poor due to the quality of the data, which was suboptimal partly as a result of radiation damage, but also because of the complexity of the problem.
Examination of the structure suggested the presence of pseudo symmetry which was thought to potentially be caused by a phase transition. For this reason, the sample was re-examined to see if the presence of a phase transition could be confirmed and, whether data collected on a high temperature phase would be better.
In all cases, the structure solved readily, however in the main cyclic porphyrin oligomer there were prolate ellipsoids associated with the tertiary butyl groups, indicative of disorder. This was modelled using a split-site model with same-distance, thermal similarity and vibrational restraints to ensure the distances, angles and displacements remained sensible. This made a relatively marginal improvement to the refinement statistics so attention turned to the void which was found to include diffuse residual electron density believed to be due to disordered solvent. The application of SQUEEZE 21,22 to leave a void from which the electron density was removed improved the refinement again. However, there remained a considerable amount of residual electron density around the fullerenes. This coupled with the need for extensive restraints to maintain the geometry and large displacement ellipsoids was strongly suggestive of disorder. Initial attempts to model this suggested that more than two positions would be required and the complexity of this model was less than ideal given the already low data to parameter ratio. For this reason, a hollow sphere 23 was used coupled with the atomic model, and this improved the refinement considerably.
This hollow sphere is represented by a single atom in the center of each fullerene with an occupancy of approximately thirty. The fullerene is then completed by the atomic model where each atom has an S24 occupancy of approximately half. These occupancies were then appropriately weighted and refined competitively. This was implemented for both fullerenes in the orthorhombic phase, but also all four in the triclinic polymorph. In both cases, although it improved the result. Once the refinement had been stabilized, the hollow sphere was then removed and SQUEEZE used iteratively to remove the electron density from the void before reinstating the hollow spheres.
Although final structure for CP4·2C 60 -I is one of poor resolution, this is likely the best currently achievable result for this material and confirms the gross structure and the 2:1 ratio of fullerene to each cyclic porphyrin oligomer, which was seen in both phases.
The result for the triclinic polymorph, CP4·2C 60 -II is of even lower resolution, so are not discussed in detail though the structure is included as supplementary material for completeness. However, the results are in keeping with those seen for the better orthorhombic polymorph (CP4·2C 60 -I). Together, the structures make a compelling case for the conclusions reported in the manuscript, namely the gross structure, conformation and connectivity of the CP4 species and the 2:1 ratio of fullerene to each cyclic porphyrin oligomer. Figure S27. X-ray single crystal structure of CP4·2C60 (hydrogen atoms and solvent molecules were omitted for clarity).