Crystallographic evidence for global aromaticity in the di-anion and tetra-anion of a cyclophane hydrocarbon

[24]Paracyclophanetetraene is a classic example of a macrocyclic hydrocarbon that becomes globally aromatic on reduction to the di-anion, and switches to globally anti-aromatic in the tetra-anion. This redox activity makes it promising as an electrode material for batteries. Here, we report the solid-state structures of the di- and tetra-anions of this cyclophane, in several coordination environments. The changes in bond length on reduction yield insights into the global aromaticity of the di-anion (26π electrons), and anti-aromaticity of the tetra-anion (28π electrons), that were previously deduced from NMR spectra of species generated in situ. The experimental geometries of the aromatic di-anion and anti-aromatic tetra-anion from X-ray crystallographic data match well with gas-phase calculated structures, and reproduce the low symmetry expected in the anti-aromatic ring. Comparison of coordinated and naked anions confirms that metal coordination has little effect on the bond lengths. The UV-vis-NIR absorption spectra show a sharp intense peak at 878 nm for the di-anion, whereas the tetra-anion gives a broad spectrum typical of an anti-aromatic system.


General Methods
Tetrahydrofuran (THF) and hexanes were purchased as dry from Sigma Aldrich and then freshly distilled over sodium with benzophenone as indicator before use.THF-d8 for NMR purposes was purchased from Sigma Aldrich and distilled over sodium and potassium alloy in the presence of benzophenone under reduced pressure and then stored inside a glovebox.Alkali metals were purchased from Sigma Aldrich, washed with hexanes to remove mineral oil and stored inside a glovebox under argon.All reactions and follow-up procedures were performed in custom-made glass systems under an atmosphere of argon, using the 'break-and-seal' technique. [1] 1H and 7 Li NMR spectra were recorded on Bruker Ascend 500 MHz NMR spectrometer. 1 H NMR spectra were referenced against the residual solvent peak (THF-d8 δH = 3.58 ppm); for 7 Li NMR, 0.1 M LiCl in THF-d8 (δLi = 0.00 ppm) was used as a reference.Samples were prepared in the NMR tubes inside a glovebox, wrapped with Parafilm and then quickly flame-sealed after removing from glovebox.UV-vis-NIR absorption measurements were carried out in a 1 cm path length quartz cuvette with PTFE cap at 298 K using Jasco V770 spectrophotometer.THF for UV-vis-NIR measurements was stored over lithium metal to remove residual oxygen/moisture.Samples were prepared inside a glovebox, the cuvette was closed tightly with a PTFE cap, wrapped with Parafilm, and removed from the glovebox.Semipreparative GPC was carried out on a Shimadzu recycling GPC system equipped with a LC-20 AD pump, SPD20A UV detector and a set of JAIGEL 3H (20 × 600 mm) and JAIGEL 4H (20 × 600 mm) columns in THF as the eluent at a flow rate of 3.5 mL/min.

Synthetic Procedures
[24]Paracyclophanetetraene was synthesized as described in the literature. [2]We found that during separation using GPC, parts of the tubing that were exposed to laboratory light got permanently yellow-colored.It is therefore recommended to wrap up visible parts of tubing with aluminum foil or perform the separation in the dark.

2.1.
[Li + (THF)4][{Li + (THF)3}C32H24 2-] (Li2-1 2-) THF (1.0 mL) was added to a customized glass system (Fig. S1, left) containing excess Li metal (6.0 mg, 0.86 mmol, 171 equiv.)and PCT (1, 2.0 mg, 0.005 mmol).The mixture was stirred at 25 °C under argon for 2 hours.The initial yellow color (neutral ligand) changed to orange in 20 min.The mixture was filtered after 2 hours while still orange (Fig. S1, right), and the orange filtrate was layered with anhydrous hexanes (3.0 mL).The ampule was sealed under argon and stored at 5 °C.Dark brown block-like crystals were present after 7 days.Yield: 3.5 mg, 70%.It is important to not keep the reaction for too long, otherwise the color changes to brown/red due to formation of tetra-reduced product.Figure S1.Left: Customized reaction vessel, [1] and right: Reaction mixture with Li at the dianion stage.THF (1.0 mL) was added to a customized glass system containing excess Na metal (6.0 mg, 0.26 mmol, 52 equiv.)and PCT (1, 2.0 mg, 0.005 mmol).The mixture was stirred at 25 °C under argon.The initial yellow color (neutral ligand) changed to orange in 5 min.The mixture was filtered after another 20 minutes when the solution became intense orange, and the filtrate was layered with anhydrous hexanes (2.0 mL).The ampule was sealed under argon and stored at 5 °C.Brownish long plates with blue shine started forming after 30 min, crystallizing fully from the solution after 4 days.Diffraction experiments indicated formation of two types of solvates with different numbers of coordinated THF molecules.Yield: 4.5 mg, 95% calculated for Na2-7THF-1 2-.

[Na
THF (1.0 mL) was added to a customized glass system containing excess Na metal (6.0 mg, 0.26 mmol, 52 equiv.)PCT (1, 2.0 mg, 0.005 mmol) and 18-crown-6 (2.72 mg, 2.1 equiv., 0.01 mmol).The mixture was stirred at 25 °C under argon.The initial yellow color (neutral ligand) changed to orange in 20 min.The mixture was filtered after another 40 minutes when the solution became intense orange, and the filtrate was layered with anhydrous hexanes (2.0 mL).The ampule was sealed under argon and stored at 5 °C.Crystals started forming quickly and the layers were fully mixed after 5 days.Yield: 5.9 mg, 90%.

[K
THF (1.0 mL) was added to a customized glass system containing excess K metal (6.0 mg, 0.15 mmol, 31 equiv.)PCT (1, 2.0 mg, 0.005 mmol) and [2.2.2]cryptand (3.7 mg, 2.1 equiv., 0.10 mmol).The mixture was stirred at 25 °C under argon for 1.5 hour.The initial yellow color (neutral ligand) changed to orange in 10 minutes.The mixture was filtered after 1.5 hours, and the orange filtrate was layered with anhydrous hexanes (2.0 mL).The ampule was sealed under argon and stored at 5 °C.Dark brown blocks started appearing after one hour and they were fully precipitated after 3 days.Yield: 5.7 mg, 85%.

[Li
THF (1.0 mL) was added to a customized glass system containing excess Li metal (6.0 mg, 0.86 mmol) and PCT (1, 2.0 mg, 0.005 mmol).The mixture was stirred at 25 °C under argon for 5 hours.The initial yellow color (neutral ligand) has changed to orange in 1 hour and to reddish brown in 4 hours (Fig. S2).The mixture was filtered after 5 hours, and the reddish-brown filtrate was layered with anhydrous hexanes (4.0 mL).The ampule was sealed under argon and stored at 5 °C.Dark brown blocks were present after 7 days.Yield: 1.5 mg, 30%.The yield is relatively low due to good solubility of the tetra-reduced product in the mixture of THF and hexanes.

Measurement and refinement details
Data collection of Li2-1 2-, Na2-7THF-1 2-, Na2-6THF-1 2-, Na2-crown-1 2-, K2-crypt-1 2-, and Li4-1 4-were performed at 100(2) K on a Huber 4-circle system with a DECTRIS PILATUS3 X 2M(CdTe) pixel array detector using ϕ scans (synchrotron radiation at λ = 0.49594 Å) located at the Advanced Photon Source, Argonne National Laboratory (NSF's ChemMatCARS, Sector 15, Beamline 15-ID-D).The dataset reduction and integration were performed with the Bruker software package SAINT (version 8.38A). [3]Data were corrected for absorption effects using the empirical methods as implemented in SADABS (version 2016/2). [4]The structures were solved by SHELXT (version 2018/2) [5] and refined by full-matrix least-squares procedures using the Bruker SHELXTL (version 2019/2) [6] software package through the OLEX2 graphical interface. [7]All non-hydrogen atoms, including those in disordered parts, were refined anisotropically.Hydrogen atoms were included in idealized positions for structure factor calculations with Uiso(H) = 1.2 Ueq(C).In the structure model of Li2-1 2-, five THF molecules were found to be disordered.In Na2-7THF-1 2-, five THF molecules were found to be disordered.In Na2-6THF-1 2-, two THF molecules were found to be disordered.In Na2-crown-1 2-, the whole structure, except two sodium cations, was found to be disordered.In K2-crypt-1 2-, one of the cryptand molecules, the THF molecule, the nhexane molecule, and the 1 2-were found to be disordered.In Li4-1 4-, six THF molecules were found to be disordered.The disordered molecules were modeled with two orientations with their relative occupancies refined.The geometries of the disordered parts were restrained to be similar by using restraints SADI and DFIX.The anisotropic displacement parameters of the disordered molecules were restrained to have the same Uij components by using the restraint SIMU, with a standard uncertainty of 0.01 Å 2 .In each unit cell of Na2-crown-1 2-, two THF solvent molecules were found to be severely disordered and removed by the Olex2's solvent mask subroutine. [7]The total void volume was 230.4 Å 3 , equivalent to 6.31% of the unit cell's total volume.Further crystal and data collection details are listed in Table S1.

Crystal structures and structural parameters
Bond lengths, torsion and dihedral angles of neutral 1 are as in previously reported crystal structure. [8]We used these values rather than those from more recent work as in the latter an AFIX 66 command was used during refinement on all phenyl rings which made all of them regular hexagons with bond length of 1.39 Å. [2] Another polymorph described in the more recent study features eight independent molecules of 1 and this dataset has low completeness, therefore we decided to not use it as a reference point during our comparisons.Table S2.Selected bond distances (Å) in 1 and Li2-1 2-along with a labelling scheme.The Li cation binds to the ethylene bridge at the (C31/C32) site.
Table S3.Selected dihedral and torsion angles (°) in 1 and Li2-1 2-along with a labelling scheme.The Li cation binds to the ethylene bridge at the (C31/C32) site.

[K
. ORTEP drawing of the asymmetric unit of K2-crypt-1 2-, drawn with thermal ellipsoids at the 40% probability level.Color key: C gray, H white, O red, N spring-green, and K dark-orchid.Interstitial solvent molecules are omitted for clarity.

Theoretical calculations
Geometries were optimized using Gaussian 16 software [9] on a CAM-B3LYP def2-TZVP level of theory in the gas phase (no solvation) with unconstrained C1 symmetry. [10,11]We calculated all the vibrational frequencies and none of them were imaginary.TD-DFT calculations were performed similarly and using 64 states in the calculations, visualizing calculated UV-vis spectra using GaussView 6.0.16 software.

Calculated bond lengths and angles
Figure S39 shows the bond lengths in the linkers (black, outside the macro ring), in para-phenylenes (black, inside phenyls), torsion angles in linkers (blue and green) and dihedral angles between planes of para-phenylene rings (red).Dihedral and torsion angles were calculated using Mercury software.Torsion angles in the linkers have the same sign on opposite sides of the molecule (same torsion direction), the other two have the same sign but opposite.There are very small differences between calculated bond lengths around the molecule (second decimal place), the figure shows mean values for simplicity.

NICS calculations
Positions of dummy atoms (Bq) were calculated at the of the macrocycle, phenyl rings and at the middle of the C=C bonds in the linkers.They were also calculated 1 Å above and below these points (Fig. S45).The Z axis (direction of the magnetic field) is perpendicular to the macrocycle plane.NICS analysis was reported for the neutral 1 and dianion 1 2-but has never been reported before for the tetraanion 1 4-. [12]In some cases, the NICS(1) and NICS(-1) values on the opposite sides of the ring are the same (equivalent due to symmetry) and were indicated as such in the Table S2.In phenyl rings, NICS(1) / NICS(-1) were indicated as in/out to distinguish between Bq positions pointing towards the inside the macroring (in) or towards outside of the macroring (out).

HOMA Calculations
HOMA (Harmonic Oscillator Model of Aromaticity) is one of the simplest and most widely used indices for describing aromaticity based on molecular geometry.It uses the C-C bond length in benzene as a standard of perfect aromaticity.The HOMA index can be calculated using the following equation: 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 = 1.388Å), 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 bond length data and calculated HOMA indexes of the vinylene bridges in 1, 1 2-and 1 4-are listed in the Table S17: Table S17.Bond length data used for calculating HOMA values of vinylene bridges in 1, 1 2-and 1 4-.

Figure S2 .
Figure S2.Reaction mixture with Li at the tetraanion stage.

S4
Figure S3.ORTEP drawing of the asymmetric unit of Li2-1 2-, drawn with thermal ellipsoids at the 40% probability level.Color key: C gray, H white, O red, and Li sky-blue.

Figure S7 .
Figure S7.ORTEP drawing of the asymmetric unit of Na2-6THF-1 2-, drawn with thermal ellipsoids at the 40% probability level.Color key: C gray, H white, O red, and Na blue.

4. 2 1 Figure
Figure S19.UV-vis-NIR spectra of dianion generated during in-situ reaction of 1 with sodium compared with the final stage of the reduction with Na and Li metals.All spectra recorded in THF at 25 °C.

Figure S20 .
Figure S20.UV-vis-NIR spectra of potassium-generated dianion and final stage of reduction, indicating formation of the tetraanion (very low solubility).All spectra recorded in THF at 25 °C.

Figure S21 .
Figure S21.Black surface of potassium indicating precipitation during the tetraanion formation stage.The solution becomes almost colorless.

Figure S23 .
Figure S23.Reaction of 1 with Li monitored by 1 H NMR over time.Orange areas indicate region where dianion resonates.In this case, the reaction reached the stage of the dianion even before measurement was performed and tetraanion started to appear.After 50 min, there are almost no signs of the dianion in the mixture.Some decomposition/reactivity of impurity products started appearing after a week.THF-d8, −30 °C, 500 MHz.Asterisk indicates signals of THF and TMS.

Figure S24 .
Figure S24.Temperature dependence of the 1 H NMR spectrum of the in-situ generated Li4-1 4− .THF-d8, 500 MHz.Asterisk indicates signals of THF.Intensity was normalized to the residual signal of THF.Note: different sample than in Fig. S23.

Figure S25 .
Figure S25.Reaction of 1 with Na monitored by 1 H NMR over time.Orange areas indicate region where dianion resonates.Mixture of dianion and tetraanion is formed.THF-d8, −30 °C, 500 MHz.Asterisk indicates residual solvent signals of THF and signal of TMS.

Figure S32 . 5 . 6 2 -Figure S33. 1 H
Figure S32.Temperature dependence of 1 H NMR spectrum of Na2-crown-1 2-, THF-d8, 500 MHz.Solubility drops drastically at low temperatures and the sample crystallizes readily, that is why any impurity signals appear as similar in intensity to the product.Impurities come from reduced by-products after synthesis of 1 which could not be removed even by sublimation.Asterisk indicates residual solvent signals of THF and signals of crown and TMS.

Figure S34 . 4 -Figure S35. 1 H
Figure S34.Temperature dependence of 1 H NMR spectrum of K2-crypt-1 2-, THF-d8, 500 MHz.Solubility drops drastically at low temperatures and the sample crystallizes readily, that is why any impurity signals appear as similar in intensity to the product.Impurities come from reduced by-products after synthesis of 1 which could not be removed even by sublimation.Asterisk indicates residual solvent signals of THF and signals of cryptand and TMS.

Figure S46 .
Figure S46.Location of dummy atoms (violet) used for NICS calculations (for 1 as an example).