Synergy of Electrostatic and π–π Interactions in the Realization of Nanoscale Artificial Photosynthetic Model Systems

Abstract In the scientific race to build up photoactive electron donor‐acceptor systems with increasing efficiencies, little is known about the interplay of their building blocks when integrated into supramolecular nanoscale arrays, particularly in aqueous environments. Here, we describe an aqueous donor‐acceptor ensemble whose emergence as a nanoscale material renders it remarkably stable and efficient. We have focused on a tetracationic zinc phthalocyanine (ZnPc) featuring pyrenes, which shows an unprecedented mode of aggregation, driven by subtle cooperation between electrostatic and π–π interactions. Our studies demonstrate monocrystalline growth in solution and a symmetry‐breaking intermolecular charge transfer between adjacent ZnPcs upon photoexcitation. Immobilizing a negatively charged fullerene (C60) as electron acceptor onto the monocrystalline ZnPc assemblies was found to enhance the overall stability, and to suppress the energy‐wasting charge recombination found in the absence of C60. Overall, the resulting artificial photosynthetic model system exhibits a high degree of preorganization, which facilitates efficient charge separation and subsequent charge transport.

Transmission electron microscopy. Transmission electron microscopy was performed with a LEO EM912-Omega (Zeiss) with an acceleration voltage of 80 kV. Selected area electron diffraction (SAED) patterns have been recorded using a Phillips CM30 S/TEM at an accelerating voltage of 300 kV at a camera length of nominally 175 mm with a calibration using gold nanoparticles with known lattice constant. For SAED an aperture with diameter of 300 nm was imposed on the area of interest. Samples were prepared by drop-casting solution onto a carbon coated copper grid. The grids were dried for 1h in a desiccator. Sample preparation. Stock solutions for all ZnPcs were prepared in DMSO. Mixed water/DMSO solutions were obtained by injecting the corresponding amounts of stock solution into milliQ ultrapure water.
Pyrenyl-1-boronic acid pinacol ester [1] In a round-bottom flask, pyrene-1-boronic acid (1.94 g, 7.88 mmol) and pinacol (2.84 g, 24.0 mmol) were dissolved in 10 mL of THF and 40 mL of diethyl ether, and warmed up to reflux. After 3 h, the reaction mixture was dried under reduced pressure and purified by column chromatography, using heptane/ethyl acetate (20:1) as eluent. The product was obtained as a white solid. Yield: 2.39 g, 92%. Mp: 120 o C. 1  3-hydroxy-5-(1-pyrenyl)pyridine (5) In a round-bottom flask, pyrenyl-1-boronic acid pinacol ester (2.39 g, 7.27 mmol) and 3bromo-5-hydroxypyridine (1.24 g, 7.10 mmol) were dissolved in 90 mL of 1,4-dioxane under argon atmosphere. Pd(dppf)Cl2·CH2Cl2 (340 mg, 0.42 mmol) was then dissolved and 80 mL of previously deoxygenated aqueous sodium carbonate (2.4 M) was added. The reaction mixture was heated at 100 o C for 4 hours. After cooling down, both phases were separated and the aqueous phase was washed with THF (3 x 10 mL). The combined organic phases were dried over anhydrous Mg2SO4 and dried under reduced pressure. The crude residue was purified by column chromatography, using hexane/ethyl acetate (3:2) as eluent. The product was obtained as a white solid. Yield: 1.28 g, 61%. Mp: >250 o C. 1  4-(5-(1-pyrenyl)pyridyl-3-oxy) phthalonitrile (6) In a round-bottom flask, compound 5 (493.8 mg, 1.67 mmol) and 4nitrophthalonitrile (323.1 mg, 1.87 mmol) were dissolved in DMF (20 mL) under argon atmosphere. Anhydrous potassium carbonate (1.0 g, 7.24 mmol) was then added in portions and the mixture was stirred at room temperature for 8 hours. The reaction was poured over a water-ice mixture and the precipitate was filtered off and purified by recrystallization from hot ethanol. The product was obtained as a white solid. Importantly, because the assignment of aromatic signals corresponding to the pyrene moiety is extremely difficult once incorporated into the ZnPc macrocycle (i.e., for compounds 1 and 7), a full characterization by NMR spectroscopy (including DEPT-135, 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-13 C HSQC spectra) was performed at this stage (see below), helping a lot for the characterization of subsequent products. Yield: 612.4 mg, 87%. Mp: 92 o C. 1 (7) In a sealed tube, phthalonitrile 6 (346 mg, 0.82 mmol) and anhydrous zinc acetate (75 mg, 0.41 mmol) were dissolved in DMAE (5 mL) under argon atmosphere. The reaction was heated to 150 o C for 24 hours. The reaction mixture was then poured over water/methanol (1:1) mixture and the blue precipitate was filtered off. The green solid was purified by column chromatography using toluene/1,4dioxane/pyridine (40:10:1) as eluent. The product was obtained as a green solid (162 mg, 45%). Mp: > 250 o C. 1 13 [M] +4 (100). *Due the presence of a regioisomeric mixture and the overlapping of signals, the complete assignment of signals in the 1 H-and 13 C-NMR spectrum was not possible for this compound.
The synthesis of compounds 2 and 3 (trans3 isomer) was achieved by adapting a previously described methodology (Scheme S2). Scheme S2. Synthesis of fullerenes 3 and 4. Note that 10 was obtained as a mixture of regioisomers, purified by column chromatography (eluent gradient, from chloroform to chloroform/THF 1:1). The trans3 isomer was the majoritarian species, therefore being the only one hydrolysed, leading to the compound (3) that has been utilized in the subsequent complexation studies with the ZnPc 1.
Tetraethyl bromomethylenediphosphonate (8) [3] In a 100 mL round-bottomed flask, n-buthyllithium 1.6 M in hexane (6.9 mL, 11.0 mmol) was dissolved into 20 mL of freshly distilled THF at -78 o C. Subsequently, a solution of diisopropylamine (1.6 mL, 11.4 mmol) in 10 mL of freshly distilled THF was added and stirred for 10 min. Thereafter, a solution of tetraethyl methylenediphosphonate (1.2 mL, 4.8 mmol) in 10 mL of freshly distilled THF was added dropwise. When the addition was completed, the reaction was left to warm up to room temperature, followed by the addition of a solution of dibromotetrachloroethane (1.8 g, 5.5 mmol) in 10 mL of freshly distilled THF. After 30 min, the reaction was poured over an H2SO4 2M/DCM (1:1) mixture (60 mL) and the phases were separated. The aqueous phase was washed with DCM (2 x 25 mL) and the combined organic phases were dried over anhydrous Mg2SO4 and, the solvent was removed in vacuo. The dark yellow oil crude was purified by column chromatography, employing ethylacetate as eluent, resulting into a yellowish oil. Yield: 1.34 g, 75%. General procedure for the Bingel-Hirsch reaction [4] In a 1 L round bottom flask, dry toluene (400 mL) were deaereated, followed by the addition of C60 (510 mg, 0.74 mmol) and NaH 60% in mineral oil (200 mg, 5 mmol), previously washed with toluene. The resulting solution was warmed up to 80 o C, followed by the dropwise addition of 8 (659 mg, 1.79 mmol) in anhydrous toluene (5 mL), resulting in gas evolution (H2). The reaction mixture was left stirring for 90 min, filtered off afterwards and the resulting solution was dried under reduced pressure.

Aggregation studies of the ZnPcs
Organic solvent effect Figure S5. Absorption and emission spectra of 2 (a, b) and 6 (c, d) at 2.0 x 10 -6 M recorded for different solvent ratios of DMSO and water, from 100 vol% DMSO to 98 vol% water. Emission recorded at λexc = 615 nm.

S21
Light scattering Figure S6. Distribution function of relaxation times for 1 at increasing water content (curves from θ = 90°). Hydrodynamic radii extracted from extrapolation q against 0.

Molecular Modelling
Two general approaches were used to study the monomeric 1 and 2 as well as the resulting aggregates. Dreiding forcefield calculations [5] utilizing Forcite Forcefield Calculation Package included in Materials Studio software 2018 [6] were used for preoptimization, conformer screening, simulated annealing and molecular dynamics. For this purpose, the Dreiding Zn parameter were parametrized against wB97xd/def2-SVP DFToptimized structures of 1 to reproduce the Zn-N bond distance (Zn-N ≈ 1.9 Å) and angles (∢(N1-Zn-N3) ≈ 180°, ∢(N1-Zn-N2) ≈ 90°). All other parameters, including the parameters for the (12,6)-Lennard Jones potential were kept from the original Dreiding Forcefield. First the Monomer of 1 and 2 were optimized, followed by a conformation scan including 10000 conformers of a unique regioisomer, varying all dihedral angles of the single bonds at once, followed by an optimization of all conformers. A series of 5 monomers, including the energetic most favourable conformations with significant difference in structure, was chosen to be then further optimized with DFT in Gaussian 16. [7] Optimization was performed with dispersion and long range corrected wB97xd functional, including PCM solvation of water. [8][9][10][11][12][13][14] Basis set for optimization and single point calculation was def2-SVP. To elucidate the size of the basis set single point calculations were also performed on the wB97xd/def2-TZVP//wB97xd/def2-SVP and wB97xd/def2-TZVPPD//wB97xd/def2-SVP level of theory ( Figure S21 and Table S3). [15][16] Dimers of 1 were constructed under the estimation, that the preferable monomer geometry stays intact. In total 12 input structures were generated featuring four pyrene-pyrene-, four pyrene-ZnPc-, and four ZnPc-ZnPcinteractions, respectively. Dreiding forcefield optimizations, followed by 1000 simulated annealing cycles and reoptimization of each structure was performed. The best stabilized dimer was then used as starting geometry to restart this cycle a second and a third time. The energetically best four conformers were chosen to study on the wB97xd/def2-SVP level of theory. Optimization was performed with water PCM solvation and single point calculations including counterpoise correction were performed in gas phase, toluene and DMSO, as the basis set is incomplete. Thus, the interaction energies, basis set superposition errors, deformation energy of the monomer and total dimerization energies could be estimated according to the following equations: Next, the first 12 vertical excitations involving singlet states were studied for the dimer and monomer. TD-DFT with the same method and basis was used. DMSO was chosen as PCM solvent, as this offers better comparability of the monomer to experimental data. The same search algorithm of Dreiding optimization and simulated annealing was performed to find dimers and trimers of 1 with 3 and 4. Similar, DFT optimizations were performedhowever we exchanged the basis set to 3-21G [17][18] for the pre-opzimizations and Lan2DZ (Lan2DZdp for phosphorus) [19][20][21][22][23][24][25] for the final iterationsas they showed faster SCF convergence than def2-SVP on this larger complexes. The information of the dimer orientation and binding geometry was then used to form larger aggregates of 1 with Dreiding optimizations and simulated annealing.       To extrapolate the dimerization energies towards the energy gain per monomer in the aggregates, we applied a linear superposition model. This is useful, as the number of interaction motifs differs for the aggregates. Please note, that this is only a coarse guess for multiple reasons: First, only the first layer of aggregation was t aken in account. Secondno change in electronic structure or reorganization is included, as the energies are extrapolated from the dimer calculations. We choose to separate the attractive Van der Waals interactions from the repulsive Coulomb part, to account for the short-and longer-range character, respectively (typically ~ −6 vs. ~ −1 ). The number of π-interactions can be directly extrapolated from the dimer structure. For the Coulomb part we counted all pyridinium cations within a Cutoff of 10 Å next to the monomer pyridinium moieties. π-π-interactions in aggregates per monomer π-π-interactions in the dimer Coulomb close contacts in aggregate (10 Å cutoff-distance)

Coulomb interaction in dimers
e r E mon kcal mol -1 dimer A dimer B dimer C dimer D Figure S24. wB97xd/def2-SVP predicted dimerization energy (ED, top) and possible interaction energy in the aggregates per monomer (Emon), assuming equal interaction energy per binding motif and usage of all possible interaction motifs (bottom) as function of the Kirkwood solvent parameter, fitted with an monoexponential decay.  Table S9. Vertical transition energies for 1 calculated with wB97xd/def2-SVP in DMSO PCM solvation. Contribution of initial and final molecular orbitals and the amplitude of contribution for the given transition.