Synergistic Effect of Tetraaryl Porphyrins Containing Corannulene and Other Polycyclic Aromatic Fragments as Hosts for Fullerenes. Impact of C60 in a Statistically Distributed Mixture of Atropisomers

Synergistic effect of tetraaryl porphyrins containing corannulene and other polycyclic aromatic fragments as host for fullerenes. Impact of C60 in a statistically distributed mixture of atropisomers. Celedonio M. Álvarez,* Héctor Barbero, Sergio Ferrero and Daniel Miguel GIR MIOMeT, IU CINQUIMA/Química Inorgnica, Facultad de Ciencias, Universidad de Valladolid, E-47011, Valladolid, Spain ABSTRACT: Symmetric meso-tetraaryl porphyrins bearing phenanthrene, pyrene and corannulene moieties in meta positions have been synthesized in a straightforward procedure under microwave irradiation by quadruple Suzuki-Miyaura reactions. Their 1H NMR spectra showed the typical pattern of 4 atropisomers distributed accordingly to their statistical ratio not properly separable due to their fast isomerization. Their ability to bind Buckminsterfullerene has been tested with the whole mixture and different behaviors have been found, being  isomer corannulene-substituted porphyrins the best hosts in the family. Porphyrins are a vast family of naturally occurring macrocycles having multiple functions in living beings. Many applications have been developed 1 by using these astounding architectures, being one of them their use as hosts for fullerenes, especially for C60, which is considered a very promising and potentially useful building block for new electronic devices. 2 The main approach in order to bind fullerenes by means of supramolecular forces has consisted of building structures with more than one porphyrin yielding molecular tweezers, trimers or strapped porphyrins leading to very high affinities. 3 However, there are not examples in which a single porphyrin shows binding abilities to C60 beyond cocrystallization, 4 except for the work of Ishizuma and Kojima et al. 5 whose curved porphyrins were able to do such task. Under our research line regarding corannulene-functionalized molecules to act as fullerenes hosts, 6 we decided to link this molecule to a porphyrin scaffold in order to sum association abilities of corannulene 7 to the porphyrin core. Additionally, other planar aromatic fragments were linked to the porphyrin core as well and tested for comparison purposes. Compounds 2HP1 and ZnP1 could be synthesized with the help of a microwave reactor with a modified method reported elsewhere. 8 From that point we reasoned that Suzuki reaction from borylated polyaromatic fragments with known Schlenk techniques would lead to 2HP2a,b,c and ZnP2a,b,c families (see Schemes S1-S3 for a description of all routes attempted), however a great amount of homocoupling by-products were obtained as well as non-negligible amount of porphyrins with Pd in their cavity (Figures S57-S59). This mixture leads to a very cumbersome chromatographic separations. So we turned to a synthetic route in which pinacol boronate fragment was located in the porphyrin core and reacted with brominated polyaromatic fragments as depicted in Scheme 1. In order to get free base porphyrins, removal of Zn atom was preferred. Additionally, all coupling reactions were carried out with microwave irradiation because reaction times were significantly improved. 9 SCHEME 1 Best synthetic route towards porphyrin targets Reagents and conditions: a) Propionic acid, nitrobenzene, MW, 200 ºC; b) Zn(OAc)2·2H2O, CHCl3, MW, 120 ºC; c) B2pin2, [PdCl2(dppf)], AcOK, dioxane, MW. d) [PdCl2(dppf)], tBuONa, toluene, MW; e) CF3CO2H. New porphyrins were fully characterized by spectroscopic methods. All of them have expected Soret bands around 430 nm and Q bands between 500 and 600 nm (Figures S67-S70) in their absorption spectra. 1H NMR spectra show clearly the presence of statistically distributed atropisomers, 10 especially for H6 singlets (see Supporting Information for numbering of the system), as depicted in Figure 1. Attempts to separate each atropisomer individually, both by preparative methods or HPLC, were unfruitful, even though HPLC chromatogram showed clearly four separated peaks. NMR tests of each fraction indicated the presence of the mixture showing that isomerization occurred even at low temperature. 11 FIGURE 1 Schematic representation of the atropisomers and their statistical distribution. Notwithstanding such limitation, we decided to analyze the ability of the whole mixture to bind C60 and, at the same time, to observe potential differences among all individual atropisomers. 1H NMR was chosen to be the technique because, unlike UV/Vis experiments, it is able to distinguish all isomers. This was complemented by Mass Spectrometry experiments. Before performing any titration, we firstly attempted to know the stoichiometry of the adduct. Job plots were carried out for all compounds, however, only H1 was observed as an average signal due to its inherent broadness, so only 2HP2 set could be properly analyzed (Figures S77-S79). Surprisingly, all porphyrins showed 1:1 adduct formation and this was confirmed by MS (Figures S60-66). We assumed that this stoichiometry was maintained in ZnP2 derivatives because only 1:1 adducts were located in all mass spectra. Titrations of phenanthrene and pyrene-substituted porphyrins ZnP2a,b and 2HP2a,b provided several little chemical shifts changes, indicating supramolecular association for some atropisomers. This was not expected because it is known that compounds bearing this type of PAHs do not establish adducts with C60 due to their low shape complementarity, 12 this suggests that the host-guest formation is taking place due mainly to the porphyrin core. 13 On the other hand, corannulene derivatives, 2HP2c and Zn2P2c, showed great chemical shift changes, except for one atropisomer. Interestingly, we observed the same pattern in all chemical shift changes, and especially for H1, H2 and H6 hydrogens in all compounds. This suggests a similar behavior. In order to assign signals for each atropisomer, we carried out computational studies 14 for 2HP2 set of isomers and the optimized structures provided interesting features. Intramolecular association of PAHs units at the same side is so preferred that leads to deformation of the porphyrin reaching a saddle-like conformation for the most stable isomers, . On the other hand,  isomers are the least stable and PAHs moieties are not interacting among them. This lack of interaction could be the reason for the low stability. Thus, a quasi-planar porphyrin core as well as vague preorganization as host are observed, as seen in Figure 2 for 2HP2c. The other atropisomers present intermediate situations and stabilities (Tables S1-S3 and Figures S80-S91). FIGURE 2 Optimized structures of: (a)2HP2c; (b) 42HP2c. Theoretical NMR predictions for H6 proton signal in 2HP2c gave, from lowest to highest field, this order: < (bottom) < (outer) < < inner) <  as depicted in Figure 3. FIGURE 3 Simulated H6 resonance in toluene of 2HP2c pattern and theoretical assignments with an integral ratio of 1:1:2:2:1:1. From that starting point, we followed the evolution of singlet H6 during titration with C60. Simultaneously, the evolution of singlet H2 proton (-pyrrole) could be traced due to a great disruption observed upon addition of initial aliquots. It must be noted that these H2 protons are all equivalent for  and  isomers, but there are two nonequivalent protons for  and  isomers. Additionally, this procedure was applied for ZnP2c compound to check similar behaviors and reproducibility. First of all, (bottom) and (outer) predicted H6 chemical shifts changes were followed and their curves fitted yielding equal constants. However, (inner) H6 could not be followed due to spectral overlap with H6 during titration. Once located and determined those proton signals, another atropisomer could be directly assigned, . Its H6 signal is the one left with a ratio of 2, (Figure 3). The titration for this atropisomer gave a very different constant (three-fold increase for 2HP2c and one order of magnitude in ZnP2c). So, at this point, just two minor signals, corresponding to  and  isomers, were left to be matched with computational results. It was assumed that theoretical assignment was correct due to the accuracy observed for the other predictions. Unfortunately, H6 proton for  atropisomer could not be followed because it got broadened after the first addition. These findings are all gathered in Figure 4. Figure 4: 1H NMR showing H6 chemical shifts in deuterated toluene of ZnP2c (up) and 2HP2c (bottom) upon addition of aliquots of C60. Colored domains enclose the assignments according to the pattern established in Figure 3. Note the overlap of H6 (inner) proton in  isomer with H6 proton in  isomer and the 'disappearance' of H6 proton in  isomer after the first addition. On the other hand, H2 protons (β-pyrrole) evolution was followed as well in order to identify each signal in the course of titration experiments. Each chemical shift change observed was systematically subjected to a fitting process covering all possibilities until association constants matched those previously estimated for H6. Fortunately, signals corresponding to and  isomers, especially for ZnP2c, were located easily due to the lack of spectral overlap during titration. Spectral changes of  atropisomer (which, depending on FID processing, can be seen as one big broad signal or two resolved shifts) were also easily followed. A crossover with a broad chemical shift change that unexpectedly raised in the beginning of the titration was observed in the last additions. This new signal was accepted to come from the remaining  atropisomer, so we carried out the same fitting procedure, resulting in a new association constant. This discussion is shown in Figure 5. 15 Figure 5: 1H NMR showing H2 chemical shifts in deuterated toluene of ZnP2c (up) and 2HP2c (bottom) upon addition of aliquots of C60.Colored domains enclose the assignments according to the pattern established in Figure 3. Note the crossover between green and yellow domains. This protocol was applied for phenanthrene and pyrene-substituted porphyrins (ZnP2a, ZnP2b, 2HP2a and 2HP2b) resulting in very similar findings, but less pronounced changes of chemical shifts were obserbed. These results were not consistent with our initial assumptions, because we thought that 2HP2c ideally had a good preorganization in a double tweezer mode and could lead to moderate association. However, we did not find chemical shift changes big enough to be measured in any of all atropisomers. Furthermore, as stated previously, structures of optimized minima clearly showed that preorganization of this atropisomer to adapt fullerene shape is not efficient, while other atropisomers preorganizations are much more efficient. The best results were obtained for corannulene-functionalized porphyrins, especially ZnP2c. ZnP2c and ZnP2c showed the best association constants in the mixture of atropisomers, which are 22330 and 12140 M-1, respectively. The presence of the metal favors supramolecular adduct formation. As commented above, 2HP2 (free-base) porphyrins possess average signals coming from internal hydrogens (H1) and could be followed during titration experiments., so an average association constant could be given, concluding that our corannulene-based porphyrins are very good host for C60 because their association constants are one order of magnitude better than those bearing planar polycyclic hydrocarbons. In light of these findings, synergy between a porphyrin core and well adapted corannulene groups to bind Buckminsterfullerene is suggested. All constants estimated are gathered in Table 1. TABLE 1. Summary of Ka of new porphyrins vs C60 estimated in toluene-d8. a Could not be estimated due to spectral overlap. b Determined from  change of H2 -pyrrole protons only. c Not measured (no average signal). d  very low to estimate a constant reasonably. In summary, a set of six tetrasubstituted porphyrins in meso positions with polycyclic aromatic hydrocarbons have been prepared very easily by using microwave irradiation. The existence of a statistically distributed mixture of atropisomers was identified by NMR and assigned with the help of a combination of titration experiments of the whole mixture with C60 and computational studies. Complexation studies were carried out and a great difference was observed for each individual atropisomer, being the best hosts,  and  porphyrins bearing corannulene as substituent whose crude estimated association constants were among the highest of those reported for single-porphyrin hosts. Theoretical calculations have suggested that preorganization of corannulene moieties in a tweezer-like fashion are crucial. This initial research has opened new possibilities to design more complex 'picket fence' ( atropisomer) porphyrins having multiple corannulene fragments with the objective to increase their affinity towards fullerenes thanks to a synergy between porphyrin and non-planar polycyclic hydrocarbon. EXPERIMENTAL SECTION All reagents were purchased from regular vendors and used without further purification. Solvents were either used as purchased or dried according to procedures described elsewhere. 16 Microwave reactions were carried out with an Anton Paar Monowave 300 Reactor. Column chromatography was carried out using Silica gel 60 (particle size 0.040-0.063 mm; 230-400 mesh) as the stationary phase and TLC were performed on precoated silica gel plates (0.25 mm thick, 60 F254) and observed under UV light and/or dipping in anisaldehyde. NMR spectra were recorded in 400 MHz and 500 MHz instruments. 1H and 13C NMR chemical shifts (δ) are reported in parts per million (ppm) and are referenced to TMS, using solvents as an internal reference. Coupling constants (J) are reported in hertz (Hz). Standard abbreviations used to indicate multiplicity: s = singlet, d = doublet, t = triplet, m = multiplet, br = broad. 13C{1H} spectra for all porphyrin derivatives had to be recorded with the following acquisition parameters due to their fast relaxation: 90º pulse, relaxation delay of 5ms and acquisition time of 250ms. 1H and 13C assignments were performed by utilizing ZnP2a 2HP2a ZnP2b 2HP2b ZnP2c 2HP2c  180 N/Aa 790 1180 22330b 7700b  0d 0d 0d 540 1730 1820  240 130 690 1240 12140 5890  0 d 0 d 0 d 0 d 0 d 0 d Average N/Ac (4.5±0.05)102 N/Ac (4.6±0.05)102 N/Ac (5.4±0.2) 103 2D NMR methods band selective HSQC and band selective HMBC. High resolution mass spectra were recorded with a MALDITOF system with N2 laser (337 nm, pulse energy 100 J, 1 ns). Acceleration voltage of 19 kV and reflector positive mode were used. UV/Vis absorption spectra wavelengths (λ) are reported in nanometers (nm) and molar absorption coefficient (ε) are reported in M-1·cm-1 units. Corannulene and bromocorannulene were prepared according to literature procedures. 17 General method for polyaromatic boronate esters preparation: Bromoarene (0.4 mmol), bis(pinacolato)diboron (152 mg, 0.6 mmol), [PdCl2(dppf)] (15 mg, 0.02 mmol) and AcOK (118 mg, 1.2 mmol) were mixed in a Schlenk flask under inert atmosphere. 2 mL of dry dioxane were added and the mixture was degassed. Then, it was refluxed overnight. Solvent was removed under vacuum before a purification by column chromatography with SiO2 gel, hexane/AcOEt (5:1 for 9-( pinacolatoboron)phenanthrene) (1), (20:1 for 1-(pinacolatoboron)pyrene (2) and 1-(pinacolatoboron) corannulene) (3). Yields over 85%. Spectral data are in agreement to those published. 18 2HP1: pyrrole (277 μl, 4 mmol), 3-bromobenzaldehyde (468.0 μl, 4 mmol), propionic acid (12 ml, 160 mmol) and nitrobenzene (7 ml, 68 mmol) were mixed in a sealed reactor specifically designed for microwave irradiation. The mixture was stirred inside microwave reactor at 200 ºC for 15 minutes (a pressure of 5 bar was achieved). The resulting black crude obtained after cooling at room temperature was taken up and MeOH was added (50 mL), precipitating a purple solid. Filtered in a Büchner funnel and washed thoroughly with more MeOH. The solid was collected, placed in an oven and kept under reduced pressure (200 ºC, 24 hours, 50 mbar) to finally get pure 2HP1 whose spectroscopic data agree with those reported 19 (0.295 g, 32% yield). ZnP1: 2HP1 (0.25 0g, 0.268 mmol) and Zn(AcO)2·2H2O (0.294 g, 1.34 mmol) were mixed in a sealed reactor specifically designed for microwave irradiation and dissolved in 12 ml of CHCl3. The solution was stirred at 130 ºC for 2.5 hours (a pressure of 6 bar was achieved). Once finished, solvent was removed in vacuo. The resulting solid was washed several times with water and dried to get a purple solid corresponding to ZnP1 (0.203 g, 76 % yield). 1H NMR (500 MHz, CDCl3) δ 8.96 (s, 2H, H2), 8.38 (s, 1H, H6), 8.16 (br, 1H, H10), 7.95 (d, J = 8.2 Hz, 1H, H8), 7.64 (d, J = 8.2 Hz, 1H, H9). 13C{1H} NMR (101 MHz, CDCl3) δ 150.0 (C3), 144.5 (C5), 137.1 (C6), 132.9 (C10), 132.1 (C2), 130.9 (C8), 128.0 (C9), 110.0 (C4). HRMS (MALDI-TOF): m/z = 987.8010 [M]+ (calcd. 987.8026 for C44H24Br4N4Zn). UV/Vis (toluene) λ 404 (ε: 17033), 424 (ε: 195165), 550 (ε: 10220), 588 (ε: 2308). ZnP2: ZnP1 (80 mg, 0.08 mmol), bis(pinacolato)diboron (122 mg, 0.48 mmol), [PdCl2(dppf)] (12 mg, 0.016 mmol) and AcOK (94 mg, 0.96 mmol) were mixed in a reactor specifically designed for microwave irradiation inside a two-necked round-bottom flask in order to put the mixture under inert atmosphere. 1.8 mL of dry dioxane and 116 L of dry pyridine were added and the mixture was degassed before sealing the vial. Then, it was stirred inside microwave reactor at 150 ºC for 2 hours (a pressure of 3 bar was achieved). Solvent was removed under vacuum and the resulting dark residue was redissolved in a mixture of CHCl3/MeOH 95:5 before quickly passing it through a plug of SiO2 gel. The deep red filtrate was concentrated to afford a purple solid (85 mg, 90% yield). 1H NMR (500 MHz, CDCl3) δ 8.91 (s, 2H, H2), 8.66 (br, 1H, H6), 8.31 (d, J = 7.5 Hz, 1H, H10), 8.21 (d, J = 7.5 Hz, 1H, H8), 7.76 (t, J = 7.5 Hz, 1H, H9), 1.38 (s, 12H, H12). 13C{1H} NMR (126 MHz, CDCl3) δ 150.3 (C3), 142.5 (C5), 140.5 (C6), 137.1 (C10), 133.9 (C8), 132.1 (C2), 127.8 (C7), 126.1 (C9), 121.1 (C4), 84.1 (C11), 25.1 (C12). HRMS (MALDI-TOF): m/z = 1180.5043 [M]+ (calcd. 1180.5047 for C68H72B4N4O8Zn). UV/Vis (toluene) λ 402 (ε: 17927), 425 (ε: 324268), 550 (ε: 16098), 590 (ε: 3049). General procedure for quadruple C-C cross coupling: Porphyrin derivative (2HP1, ZnP1 or ZnP2) (0.02 mmol), PAH derivative (9-Bromophenanthrene, 1-Bromopyrene, Bromocorannulene, 1, 2 or 3) (0.08 mmol), [PdCl2(dppf)] (11.7 mg, 0.016 mmol) and tBuONa (23.1 mg, 0.24 mmol) were mixed in a reactor specifically designed for microwave irradiation inside a two-necked round-bottom flask in order to put the mixture under inert atmosphere. 2.2 ml of dry and degassed toluene were added as well as 30 L of dry pyridine before sealing the vial. The mixture was then sonicated thoroughly and stirred vigorously inside microwave reactor at 130 ºC for 1 hour (a pressure of 3 bar was achieved). After such time, solvent was removed under vacuum and the black residue was subjected to a purification by column chromatography yielding the pure purple solid corresponding with the expected product. ZnP2a: chromatography conditions: SiO2 gel, hexane/AcOEt gradient elution (10:1 - 5:1 - 1:1 - AcOEt). 98% yield. 1H NMR (500 MHz, CDCl3) δ: 9.20 (s, 2H, H2), 8.85 – 8.75 (m, 1H, H17), 8.75 – 8.65 (m, 1H, H20), 8.50 – 8.41 (m, 1H, H6), 8.41 – 8.35 (m, 1H, H23), 8.35 – 8.27 (m, 1H, H10), 8.05 – 7.96 (m, 2H, H12+H8), 7.96 – 7.86 (m, 2H, H14+H9), 7.73 – 7.51 (m, 4H, H16+H21+H15+H22). 13C{1H} NMR (126 MHz, CDCl3) δ: 150.4 (C3), 143.1 (C5), 139.2 (C7), 138.7 (C11), 136.4 (C6), 133.7 (C10), 132.3 (C2), 131.7 (C24), 131.4 (C19), 130.9 (C13), 130.2 (C18), 129.4 (C8), 128.8 (C14), 128.2 (C12), 127.1 – 126.7 (C9+C15+C16+C21+C22+C23), 123.1 (C17), 122.6 (C20), 121.1 (C4). HRMS (MALDI): m/z = 1380.4075 [M]+ (calcd. 1380.4104 for C100H60N4Zn). UV/Vis (toluene) λ 300 (ε: 38806), 400 (ε: 14179), 426 (ε: 211940), 552 (ε: 11642), 590 (ε: 4627). ZnP2b: chromatography conditions: SiO2 gel, hexane/AcOEt gradient elution (5:1 - 1:1 - AcOEt). 92% yield. 1H NMR (500 MHz, CDCl3) δ: 9.26 – 9.21 (m, 2H, H2), 8.68 – 8.59 (m, 1H, H23), 8.59 – 8.47 (m, 1H, H6), 8.45 – 8.35 (m, 1H, H10), 8.35 – 8.22 (m, 2H, H12+H13), 8.22 – 7.86 (m, 8H, H22+H8+H9+H19+H16+H18+H20+H15). 13C{1H} NMR (126 MHz, CDCl3) δ: 150.4 (C3), 143.3 (C5), 139.5 (C7), 137.6 (C11), 137.0 (C6), 133.6 (C10), 132.2 (C2), 131.6 – 130.7 (C26+C17+C21+C14), 129.8 (C8), 128.9 (C24), 128.1 (C12), 127.8 (C22 or C16), 127.5 (C22 or C16), 126.7 (C9), 126.0 (C19), 125.4 (C23), 125.2 – 124.8 (C13+C15+C18+C20+C25), 121.0 (C4). HRMS (MALDI): m/z = 1476.4104 [M]+ (calcd. 1476.4104 for C108H60N4Zn). UV/Vis (toluene) λ 329 (ε: 36774), 347 (ε: 48548), 405 (ε: 20806), 428 (ε: 195161), 551 (ε: 12903), 591 (ε: 4839). ZnP2c: chromatography conditions: SiO2 gel, hexane/AcOEt gradient elution (5:1 - 1:1 – AcOEt – CHCl3) 85% yield. 1H NMR (500 MHz, CDCl3) δ: 9.18 (s, 2H, H2), 8.74 – 8.65 (m, 1H, H6), 8.42 – 8.31 (m, 1H, H10), 8.28 – 8.21 (m, 1H, H8), 8.21 – 8.11 (m, 2H, H24+H12), 8.00 – 7.90 (m, 1H, H9), 7.90 – 7.72 (m, 7H, H14+H15+H17+H18+H20+H21+H23).13C{1H} NMR (126 MHz, CDCl3) δ: 150.3 (C3), 145.2 (C25 or C27), 143.5 (C5), 137.7 (C7), 136.2 (C6), 133.9 (C10), 132.1 (C2), 123.0 (C25 or C27), 129.0 (C8), 127.5 (C24), 127.2 – 126.7 (C9 + C14 + C15 + C17 + C18 + C20 + C21 + C23), 126.2 (C12), 120.9 (C4). HRMS (MALDI): m/z = 1668,4133 [M]+ (calcd. 1668,4104 for C124H60N4Zn). UV/Vis (toluene) λ 296 (ε: 146667), 400 (ε: 27407), 427 (ε: 438889), 551 (ε: 22037), 590 (ε: 4630). General procedure for Zn removal: Zinc porphyrin derivatives (ZnP2a, ZnP2b or ZnP2c) (0.02 mmol) were dissolved in 2 mL of DCM. To the deep red solution, 2 ml of CF3CO2H were added at once and the color turned immediately to green. The mix-