Strategic Synthesis of ‘Picket Fence’ Porphyrins Based on Nonplanar Macrocycles**

Abstract Traditional ‘picket fence’ porphyrin systems have been a topic of interest for their capacity to direct steric shielding effects selectively to one side of the macrocycle. Sterically overcrowded porphyrin systems that adopt macrocycle deformations have recently drawn attention for their applications in organocatalysis and sensing. Here we explore the combined benefits of nonplanar porphyrins and the old molecular design to bring new concepts to the playing field. The challenging ortho‐positions of meso‐phenyl residues in dodecasubstituted porphyrin systems led us to transition to less hindered para‐ and meta‐sites and develop selective demethylation based on the steric interplay. Isolation of the symmetrical target compound [2,3,7,8,12,13,17,18‐octaethyl‐5,10,15,20‐tetrakis(3,5‐dipivaloyloxyphenyl)porphyrin] was investigated under two synthetic pathways. The obtained insight was used to isolate unsymmetrical [2,3,7,8,12,13,17,18‐octaethyl‐5,10,15,20‐tetrakis(2‐nitro‐5‐pivaloyloxyphenyl)porphyrin]. Upon separation of the atropisomers, a detailed single‐crystal X‐ray crystallographic analysis highlighted intrinsic intermolecular interactions. The nonplanarity of these systems in combination with ‘picket fence’ motifs provides an important feature in the design of supramolecular ensembles.


Synthesis and Characterization of Compounds
The general procedure of porphyrin condensation to synthesize 3, 5, 7, 9 and 10.
To a pre-dried 2L round-bottomed flask, anhydrous dichloromethane (500 mL -1 L) was added and purged with argon. Benzaldehyde 3A, 5A, 7A, 9A or 10A (0.9 − 1.1 eq.) and 3,4-diethyl-1H-pyrrole 11 (1.00 eq.) were added and the reaction mixture was stirred at room temperature under a slow steady flow of argon. After 15 min, BF3×Et2O (0.1 eq.) was added and the reaction flask was shielded from ambient light. After stirring for 16 h DDQ was added and 3 h later triethylamine (0.1 eq.) was added to neutralize BF3×Et2O. The solution was concentrated in vacuo to yield the crude product mixture, washed with NaOH (1 M), water, and dried over anhydrous magnesium sulfate. The organic extract was concentrated in vacuo for further purification by silica gel flash column chromatography. Scheme S1. Synthetic scheme for the preparation of the target compounds.
The mixture was dissolved in a small volume of dichloromethane and transferred to a silica gel column (CH2Cl2). Three green fractions (CH2Cl2) were collected and dried under reduced pressure to yield three products as dichromatic (green/purple) crystalline solids:  Figure S42. UV-vis spectrum of α,β,α,β-12, α2β2-12 and α3β-12 in chloroform.

Structural Determination of Isolated Structures
Crystals were grown following the protocol developed by Hope, liquid-liquid diffusion in CHCl3 and methanol or oversaturated solutions in DMSO. [7] Using Olex2, the structure was solved with the XT structure solution program, using the intrinsic phasing solution method and refined against │F 2 │ with XL using least squares minimization. [8] The C and N bound H atoms were placed in their expected calculated positions and refined as riding model: N-H = 0.88 Å, C-H = 0.95-0.98 Å, with Uiso (H) = 1.5Ueq (C) for methyl H atoms and 1.2Ueq (C, N) for all other atoms other H atoms. Details of data refinements can be found in Table S1. All images were prepared by using Olex2. [8a] In the structure of 9A one t-Bu acetate group modelled as disordered over two positions with almost equal occupancy (53/47%) using restraints (SADI and ISOR) and constraints (EADP/EXYZ for C17C17a and EADP C19/C19a).
In the structure of 9 a pivaloyl group at O114 was modelled over two positions using SIMU restraint in a 72:28% occupancy. Two of the trifluoroacetate molecules was modelled over two positions using rigid models in a 50:50% occupancies.
In the structure of 8 the phenyl moiety at C10_2 was modelled over two positions using restraints (SADI, SIMU, ISOR, AFIX 66) in a 54:46 % occupancy. The ethyl groups at C18_2, C7_2, C12_1 were modelled over two positions using restraints (SADI, SIMU) in 58:42 %, 54:46 %, 75:25 % occupancies respectively. DMSO molecules were modelled using rigid models, moreover, units at S26S, S7S and S1S were modelled over two positions in 25:75, 40:60, 40:60 % occupancies. In the structure there were solvent access ible voids that contained large amounts of solvent molecules, however, due to high disorder these could not be modelled reliably and were omitted using OLEX2 maps.
In the structure of α3β-12 one ethyl group at C18 was disordered and modelled in two locations (36:54% occupied) with restraints (SADI, SIMU). Three pivaloate groups at C10, C15, C20 disordered and modelled in two locations (57:33%; 26:74%; 85:15% correspondingly) with restraints (SADI, SIMU). The solvent molecules in the lattice void were modelled with rigid groups and consist of DMSO In the structure of α2β2-12 one ethyl group at C2 was disordered and modelled in two locations (51:49% occupied) with restraints (SADI, SIMU). One pyrrole ring carbon and ethyl group at C13 was disordered and modelled in two locations (64:36% occupied) with res traints (SADI, SIMU). One pivaloate group at C15 disordered and modelled in two locations (52:48%) with restraints (SADI, SIMU). The solvents in the lattice void were modelled with rigid groups and consist of DMSO and H2O.
The structure α,β,α,β-12 was refined as a 2-component twin. In the first unit, two pivaloate groups at C15, C20 disordered and modelled in two locations (57:43%; 60:40% correspondingly) with restraints (SADI, SIMU). Two nitro groups at C5, C15 disordered and modelled in two locations (61:49%; 47:53% correspondingly) with restraints (SADI, SIMU). In the second unit, two pivaloate groups at C10B, C15B disordered and modelled in two locations (52:48%; 56:44% correspondingly) with restraints (SADI, SIMU). The solvent molecules in the lattice void were modelled with rigid groups and consist of chloroform and dichloromethane. Figure S59. Molecular structures of isolated and analyzed compounds by X-ray crystallography. Hydrogen atoms and solvent molecules omitted for clarity; thermal ellipsoids give 50% probability.    Figure S62. Stability evaluation of α,β,α,β-12, α2β2-12, and α3β-12, by 1 H NMR recorded in CDCl3. Top: over a period of 22h at 25 C and 18h at 50C (note, no spectroscopic changes was observed at room 25 C). Bottom: enlargement of the aromatic region signals of the samples a) α,β,α,β-12, b) α2β2-12, and c) α3β-12 recorded after 18h at 50C. In blue, the percentage value of the integrated characteristic signal corresponding to the domain atropisomer. In red, the predominant observable signal corresponding to the equilibrated atropisomer. Note, due to the very broad profile of α2β2-12 characteristic signal, the tracing the exact atropisomeric composition can be unattainable. Highlighted with star symbol is unidentified signal that could potentially correspond to the formation of α4-12.