Redox-Switchable Aromaticity in a Helically Extended Indeno[2,1-c]fluorene

Molecular switches have received major attention to enable the reversible modulation of various molecular properties and have been extensively used as trigger elements in diverse fields, including molecular machines, responsive materials, and photopharmacology. Antiaromaticity is a fascinating property that has attracted not only significant fundamental interest but is also increasingly relevant in different applications, in particular organic (opto)electronics. However, designing systems in which (anti)aromaticity can be judiciously and reversibly switched ON and OFF remains challenging. Herein, we report a helicene featuring an indenofluorene-bridged bisthioxanthylidene as a novel switch wherein a simultaneous two-electron (electro)chemical redox process allows highly reversible modulation of its (anti)aromatic character. Specifically, the two thioxanthylidene rotors, attached to the initially aromatic indenofluorene scaffold via overcrowded alkenes, adopt an anti-folded structure, which upon oxidation convert to singly bonded, twisted conformations. This is not only associated with significant (chir)optical changes but importantly also results in formation of the fully conjugated, formally antiaromatic as-indacene motif in the helical core of the switch. This process proceeds without the buildup of radical cation intermediates and thus enables highly reversible switching of molecular geometry, aromaticity, absorbance, and chiral expression under ambient conditions, as evidenced by NMR, UV–vis, CD, and (spectro)electrochemical analyses, supported by DFT calculations. We expect this concept to be extendable to a wide range of robust antiaromatic–aromatic switches and to provide a basis for modulation of the structure and properties of these fascinating inherently chiral polycyclic π-scaffolds.


General Remarks
All reagents were obtained from commercial sources and used as received without further purification.Dry solvents were obtained from a MBraun solvent purification system.Progress of the reactions was determined by TLC: silica gel 60, Merck, 0.25 mm.The TLC plates were visualized with ultraviolet (UV) light (λ = 254 nm or 355 nm).
Microwave reactions were carried out using a CEM Discover SP synthesis system.
Column chromatography was performed on a Biotage Selekt System.High Resolution Mass Spectrometry (HRMS) measurements were performed using an LTQ Orbitrap XL.NMR spectra were recorded on a Bruker Avance Neo with Cryoprobe Prodigy BBO ( 1 H: 600 MHz, 13 C: 151 MHz), a Varian Mercury Plus ( 1 H: 400 MHz) or an Agilent MR ( 1 H: 400 MHz) instrument.Chemical shifts (δ) are in parts per million (ppm) relative to TMS.For 1 H NMR spectroscopy, the splitting pattern of peaks is designated as follows: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd (doublet of doublets), td (triplet of doublets), dq (quartet of doublets), and qt (quartet of triplets).Single-crystal X-ray diffraction measurements were performed on a Bruker-AXS D8 Venture diffractometer.UV/Vis absorption spectra were recorded on a Agilent Cary 8454 spectrophotometer in a 1 cm quartz cuvette.CD spectra were obtained on a Jasco J-815 spectropolarimeter.Electrochemical measurements were carried out with a Palmsense 4 potentiostat in a three-electrode setup comprising a Pt wire counter electrode, a non-aqueous Ag/AgNO3 reference electrode (10 mM AgNO3 in CH3CN, 100 mM TBAPF6) and a glassy carbon disk working electrode (3 mm diameter).

Electrochemical
grade TBAPF6 was obtained from Sigma Aldrich.
Spectroelectrochemistry was performed using a quartz spectroelectrochemical cell with 1 mm pathlength (ALS Japan).Solvents used for spectroscopic studies were of spectroscopic grade (UVASOL, Merck).Geometry optimizations and TD-DFT calculations were performed using the Orca 5.0.1.package. 1 The NICS-XY scan was calculated using the Gaussian 16 Rev.B.01 software package. 2

Synthesis of rac-3:
The synthesis of rac-3 was adapted from literature in order to obtain a racemic mixture. 3microwave vial was charged with 2 4 (40 mg, 104 µmol, 1 eq.) followed by the addition of THF (3 mL).The solution was degassed with nitrogen for 5 min before adding a spatula tip of Ag2CO3 and [RhCl(PPh3)3] and heating the mixture to 180 °C for 1.5 h in a microwave reactor.After cooling the reaction mixture to room temperature, the solvent was evaporated under reduced pressure followed by the addition of CH2Cl2 (10 mL), pyridinium chlorochromate (67 mg, 311 µmol, 3 eq.)and celite (120 mg).The reaction mixture was stirred at room temperature for 3 h before filtering it over SiO2/celite (1:4), followed by washing with CH2Cl2 until a colorless elution was obtained.Evaporation of the solvent under reduced pressure yielded pure racemic rac-3 (23 mg, 58%) as a red solid.
The analytical data of the obtained compound matched with literature reported data. 4 flame-dried round-bottom flask was charged with rac-3 (30 mg, 78 µmol, 1 eq.) and Lawesson's reagent (95 mg, 235 µmol, 3 eq.)followed by the addition of dry toluene (12 mL).The flask was equipped with a condenser and the mixture was heated to reflux for 1.5 h.After cooling the reaction mixture to room temperature, the mixture was filtered over a SiO2 plug into a flame-dried Schlenk flask.The plug was further eluted with toluene until an almost colorless solution eluted (~10 mL).The resulting black solution containing the dithioketone 4 was degassed with nitrogen for 10 min and used without purification for the subsequent step.
In the meantime, a flame-dried microwave vial was charged with 5 5 (177 mg, 784 µmol, 1 eq.) followed by the addition of dry diethyl ether (10 mL), MgSO4 (640 mg) and a saturated solution of KOH in methanol (1 mL).The mixture was cooled to 0 °C before adding Ag2O (727 mg, 3.14 mmol, 4 eq.) followed by stirring the reaction mixture vigorously for 1 h at the same temperature.After stopping the stirring, 5 mL of the resulting purple supernatant containing the diazo compound 6 (4.6 eq.wrt.rac-3) was added to the previously prepared solution of 4. The resulting reaction mixture was stirred at room temperature for 45 min before adding hexamethylphosphorous triamide (HMPT, 100 µL, 6.6 eq.).After stirring for another 15 min, the mixture was diluted with water and toluene and the phases were separated.The aqueous phase was extracted with toluene and the combined organic layers were washed with brine and dried with MgSO4 before filtration and evaporation of the solvent under reduced pressure.The resulting crude product was purified by automated flash column chromatography (SiO2, pentane/CH2Cl2 1:0 to 8:2 to 0:1) followed by precipitation from a CH2Cl2/methanol mixture to yield pure rac-1 (9 mg, 15%) as an orange solid.

X-Ray Crystallography (CCDC 2341064)
Switch rac-1 was crystallized by slow diffusion of a layer of methanol on top of a layer of the compound dissolved in CD2Cl2.A single-crystal was mounted on a cryoloop and placed in the nitrogen stream (100 K) of a Bruker-AXS D8 Venture diffractometer with a Cu Kα (λ = 1.54178Å) source.Data collection and processing was carried out using the APEX4 software suite from Bruker. 6The structure was solved using SHELXT 7 and refinement performed using SHELXL 8 in the OLEX2 software package. 9The cif file of the structure is provided as additional file.Contributions from disordered solvent were removed using the PLATON/SQUEEZE routine. 10No A-or B-level alerts were raised by CheckCIF for the fully refined structure of switch rac-1.Table S1.Crystal data and structure refinement for switch rac-1.

UV-Vis Dilution Series
The absorption spectra of a series of diluted samples of rac-1 were measured (Figure S6).The absorption value at 408 nm was then plotted vs. the concentration of the solution and a linear fit was applied (Figure S7).The linear relationship validates the adherence to the Lambert-Beer-Law and thus confirms absence of any aggregation.
Furthermore, it enabled exact determination of the concentration of diluted samples when weighing was troublesome (e.g.pure enantiomers, where only minor amounts of samples were at hand).

HPLC Separation & g-Factor Plot
Separation of enantiomers of rac-1 was performed using a Shimadzu Prominence HPLC system equipped with a chiral CHIRALPAK IE column from Daicel Corporation and a photo diode array detector.
Analytical separation was achieved by using n-heptane/CH2Cl2 (7:3) as the mobile phase (flowrate 0.8 mL/min) on a 4.6x250 mm column (5 µm particle size) and injecting 10 µL of the sample (0.8 mg/mL in the mobile phase).
Assignment of the absolute configuration to the peaks in the chromatogram was based on DFT calculations (vide infra).In the case of (P)-1 2+ (generated by oxidation with 2 equivalents of magic blue), a decay curve of the CD signal at 325 nm was observed at 75 °C in 1,2-dichloroethane.
However, the absorption spectrum of the sample at the end of the experiment differed considerably from the initial absorption spectrum, which indicates decomposition/sidereaction rather than a clean racemization process.It should be noted that at room temperature no racemization was observed during the time course of the spectroelectrochemical measurements, indicating that the dication is also configurationally stable.

H NMR Chemical Redox Studies
To a solution of rac-1 in CD2Cl2 (≈0.5 mg/mL) was added an excess of solid Fe(ClO4)3•xH2O (~20 mg).The resulting suspension was briefly sonicated whereupon its colour changed to dark purple.The solid oxidant was then removed by filtration through a syringe filter (0.45 µm) and the NMR spectrum of the so-generated 1 2+ recorded.Re-reduction was performed analogously using solid Zn powder.Sonication  Alternatively, chemical oxidation was also performed using trifluoroacetic acid (TFA).
A suspension of rac-1 in TFA-d was stirred at room temperature over the course of two months, which slowly oxidized rac-1 to rac-1 + , yielding a deeply purple colored solution.
A 1 H NMR spectrum was recorded after filtering off remaining starting material and adding CD2Cl2 as lock reference.In the oxidation with TFA, no over-oxidized species are obtained and the dication rac-1 2+ is stable in solution over months.
DFT calculations using the r 2 SCAN-3c functional have proven to be reliable and cost-efficient for the geometry optimization of various overcrowded alkenes.Further conformational analysis: A conformational analysis was performed for (P)-1 at the r 2 SCAN-3c/CPCM(CH2Cl2) 12,13 level of theory.Potential energy minima were confirmed by subsequent frequency calculations.Additionally, the transition state from the ground state to the second most stable conformer (TS1) was calculated at the r 2 SCAN-3c/CPCM(CH2Cl2) 12,13 level of theory.The transition state was confirmed by the presence of a single imaginary harmonic frequency.The side view of the optimized structures as well as their relative energies are displayed in Figure S17.
Full TD-DFT: Using the optimized structures, full TD-DFT calculations of (P)-1 and (P)-1 2+ were performed at the B3LYP/6-311g*/CPCM(CH2Cl2) 13,[18][19][20] level of theory and considering 100 singlet transitions.The calculated CD spectrum of (P)-1 (with an energy shift of -83 eV) matches well with the measured spectrum (Figure S18).To get an understanding of the involved molecular orbitals in the redshifted region of the measured spectrum of 1 2+ , the full TD-DFT calculations were analyzed.The first calculated 8 states are summarized in Table S2 and the involved molecular orbitals (HOMO-2 to LUMO+2) are plotted in Figure S20.The visualization of the molecular orbitals reveals occupied π orbitals on the central helicene core and unoccupied π* orbitals on the thioxanthylium rotors and the as-indacene core.Thus, the redshifted region is partially governed by charge-transfer characteristics from the central helicene core to the cationic thioxanthylium rotors.

Electrochemical Measurements and DYREX Mechanism
Prior to each measurement the GC working electrode was mechanically polished with 0.05 µm alumina slurry and briefly sonicated in H2O/EtOH (1:1).TBAPF6 was used as electrolyte in all cases (100 mM for standard voltammetric experiments and 200 mM for all spectroelectrochemical experiments).All measurements were performed in DCM.
-0.6 -0.As shown in Figure S23, we initially hypothesized that the one-electron oxidation of the initially neutral 1af,af, wherein both rotors are in the anti-folded state, first transiently generates the radical cation with the same geometric arrangement (1 +• af,af, EOx 1 ).This oxidation is associated with conversion of the C-C double bond to a single bond of the oxidized thioxanthylium motif, inducing immediate geometric rearrangement of this rotor to the more favorable twisted state (1 +• tw,af, k1).For this state, various mesomeric structures can be considered, including the one shown in Figure S23, wherein the radical resides on the second rotor and the fully conjugated antiaromatic as-indacene is formed.In analogy to the first step, this species can quickly geometrically rearrange to the "doubly twisted" state (1 +• tw,tw, k2).One-electron oxidation of this species to the target dication 1 2+ tw,tw (EOx 2 ) is at least as facile, or presumably even more facile than the initial oxidation (EOx 2 ≤ EOx 1 , potential inversion/compression) such that in effect both oxidations happen simultaneously.In an alternative mechanism, a conformational rearrangement precedes electron transfer. 21,22To gain further insight into which mechanism is most likely operating in this system further studies were conducted as detailed in the following.Both the smaller currents and larger peak separation arise, at least in part from the significant change in temperature, i.e. significantly slowed down diffusion and significant IR drop.However, upon removal of the cooling bath and slow warming to room temperature, continuous CV scans showed significant changes that cannot be solely attributed to these effects, in particular for the oxidation wave.Specifically, it is clearly observable that the oxidation peak observed at low temperature disappears upon warming and gives rise to a new (i.e. the "original", room temperature) oxidation wave at lower potentials.This behavior can, for example, arise from a thermal equilibration between different (conformational) species. 21,22Specifically, the more cathodic oxidation observed at room temperature might arise from oxidation of a minor species of a different geometry with a lower oxidation potential (for example containing twisted rotors).If both the energy difference and the activation energy barrier between this species and the most stable doubly anti-folded state is sufficiently low for them to quickly equilibrate at a given temperature, then only oxidation of the conformer with a lower oxidation potential is observed.At lower temperature the population of this higher energy, lower oxidation potential species would be prevented, such that the CVs would only reflect the redox properties of the lowest energy conformer.We attempted to further investigate this by fast-scan CV studies with scan rates up to 20 V/s (Figure S25).At higher scan rates the oxidation peak appears to broaden, which is potentially the result of appearance of a new peak at higher potential (analogous to the experiments at low temperature and reflective of an equilibrium between different species).However, due to significant distortions at these high scan rates this cannot be clearly resolved.To further probe whether different conformers of 1 are accessible, we subjected rac-1 to VT-NMR in TCE-d2 over a wide temperature range between -28 °C to 80 °C (Figure S26).Across this whole temperature range only small shift differences in the proton signals were observed, which are unlikely to arise from (significant) geometric changes.This was also corroborated by computational conformational analysis, which revealed that numerous other conformers of neutral 1 can in principle exist, whereby one or two of the rotors adopt differently folded or twisted structures (Figure S17).
However, these species are all either much higher in energy or thermally inaccessible due to high activation energy barriers, which would suggest that conformational rearrangements do not precede electron transfer.
Taken together, these results do not allow an unambiguous identification of the specific dynamic redox mechanism.However, regardless of temperature, scan rate and the specific mechanism, it appears that all oxidations and reductions occur via virtually simultaneous 2-electron transfers, such that in no cases (the build-up of) the intermediate monoradical cation is observable, as also clearly evidenced by the spectroelectrochemical studies (Figure 7).

Figure S5 .
Figure S5.ORTEP plot (ellipsoid at 50% probability) of the unit cell of switch rac-1.Hydrogens are omitted for clarity.

Figure S6 .
Figure S6.Absorption spectra of solutions of rac-1 in CH2Cl2 with different concentrations.

Figure S7 .
Figure S7.The absorbance at 408 nm of differently concentrated solutions of rac-1 in CH2Cl2 is plotted versus its concentration and linearly fitted.
was carried out until the original orange colour of the solution was recovered.If the oxidation with Fe(ClO4)3•xH2O was carried out for too long, unknown over-oxidized species were obtained.

Figure S17 .
Figure S17.DFT-optimized structures of different conformers and the first transition state (TS1) of (P)-1 from the first to the second most stable conformer at the r 2 SCAN-3c/CPCM(CH2Cl2) level of theory.

Low temperature electrochemical studies of 1 Figure S24 .
Figure S24.CVs of 0.5 mM 1 in CH2Cl2, 100 mM TBAPF6 at ν = 100 mV/s at different temperatures between ~-40 °C (black line) and room temperature (red line).The black arrows indicate the changes in the peaks upon warming.

Figure S25 .
Figure S25.CVs of 0.5 mM 1 in CH2Cl2, 100 mM TBAPF6 at high scan rates up to 20 V/s.

Table S2 .
2ummary of the calculated first 8 states of (P)-12.The orbitals with the largest individual weight within the respective transition states are written in bold.