Highly Cooperative Photoswitching in Dihydropyrene Dimers

Abstract We present a strategy to achieve highly cooperative photoswitching, where the initial switching event greatly facilitates subsequent switching of the neighboring unit. By linking donor/acceptor substituted dihydropyrenes via suitable π‐conjugated bridges, the quantum yield of the second photochemical ring‐opening process could be enhanced by more than two orders of magnitude as compared to the first ring‐opening. As a result, the intermediate mixed switching state is not detected during photoisomerization although it is formed during the thermal back reaction. Comparing the switching behavior of various dimers, both experimentally and computationally, helped to unravel the crucial role of the bridging moiety connecting both photochromic units. The presented dihydropyrene dimer serves as model system for longer cooperative switching chains, which, in principle, should enable efficient and directional transfer of information along a molecularly defined path. Moreover, our concept allows to enhance the photosensitivity in oligomeric and polymeric systems and materials thereof.


PyFm-dimer
A Schlenk tube was charged with 2-formyl-7-[N-methylpyrrole]-DHP (31 mg, 0.09 mmol), 2-formyl-7-bromo-DHP (46 mg, 0.14 mmol, 1.5 equiv.), potassium acetate (18 mg, 0.18 mmol, 2.0 equiv.) and 1.0 mL of dry DMAc. The mixture was degassed and put under argon atmosphere three times. Palladium(II) acetate (1 mg, 5 µmol, 5 mol%) was added and the reaction stirred at 100 °C overnight. After allowing the mixture to cool to room temperature, water was added and the mixture was extracted with DCM several times. The combined organic phases were washed with brine and dried over magnesium sulfate. Removal of the solvent under reduced pressure afforded the crude product which was purified by column chromatography (silica, petroleum ether/chloroform = 2:1 to chloroform). The product was obtained as a dark greenish blue solid (18 mg, 0.03 mmol, 33%).

2-Formyl-DHP
According to literature procedure [6] DHP (200 mg, 0.86 mmol) was dissolved in 20.0 mL of dry DCM and cooled to 0 °C under argon atmosphere. First, titanium tetrachloride (142 µL, 1.29 mmol, 1.5 equiv.) was added, whereas the solution turned from green to red. Afterwards (dichloromethyl)methylether (101 µl, 1.12 mmol, 1.3 equiv.) was added dropwise, followed by a color change to turquoise. The reaction mixture was stirred at 0 °C for one hour and additional three hours at room temperature. Then the mixture was carefully poured on ice and the aqueous phase extracted with DCM several times. The combined organic phases were washed with brine and dried over magnesium sulfate. Column chromatography (silica, petroleum ether/DCM = 2:1 to DCM) afforded the product as a dark pink solid (100 mg, 0.39 mmol, 45%).

E-fumaro-nitrile-DHP
In a flame-dried Schlenk tube 2-formyl-DHP (52 mg, 0.20 mmol) and fumaronitrile (17 mg, 0.22 mmol, 1.1 equiv.) were dissolved in 1.5 mL of dry DCM under argon atmosphere. The solution was cooled down to -78 °C and tri-nbutylphosphine (59 µl, 24 mmol, 1.2 equiv.) was added dropwise. The reaction mixture was stirred at -78 °C for 30 min and then overnight at room temperature. Water was added and the aqueous phase extracted with ethyl acetate, followed by washing with brine and drying over magnesium sulfate. After the solvent was removed, the product was purified by column chromatography (silica, petroleum ether/ethyl acetate = 2:1) which was obtained as a pink solid (34 mg, 0.11 mg, 52%).

CN-dimer
In a Schlenk flask 2-formyl-DHP (10 mg, 0.04 mmol) was dissolved in 1.0 mL of dry DCM under argon atmosphere and one drop DBU was added. 2-Fumaronitrile-DHP (13 mg, 0.04 mmol, 1.0 equiv.) dissolved in 2 mL of dry DCM was added dropwise to the aldehyde solution. The reaction mixture was stirred overnight at room temperature. Afterwards the solvent was removed under reduced pressure and the residue purified by column chromatography (silica, petroleum ether/DCM = 1:1 to DCM). Two turquoise bands were isolated corresponding to two different isomers. One of them was identified as the symmetric E,E-dimer (5 mg, 9 µmol, 22%). The synthesized amount was not sufficient for 13 C-NMR analysis.

PhCN-dimer
Para-xylenedicyanide (17 mg, 0.11 mmol) and 2-formyl-DHP (72 mg, 0.28 mmol, 2.5 equiv.) were suspended in 5.0 mL of ethanol. Afterwards cesium hydroxide (5 mg, 0.03 mmol) was dissolved in 0.5 mL of ethanol and added dropwise. The reaction mixture was stirred overnight at room temperature, whereas the precipitation of the product was observed. After dilution with chloroform it was washed three times with water, followed by brine. The organic phase was dried over magnesium sulfate and the solvent removed under reduced pressure. The residue was purified by column chromatography (silica, chloroform). The obtained crude product was further purified by precipitation into petroleum ether to remove any remaining starting material. The dimer was obtained as a dark blue solid (30 mg, 0.05 mmol, 43 %).         Figure S27. NMR-spectra of PyFm-dimer (CD2Cl2) during 660 nm irradiation at 15 °C. Figure S28. NMR-spectra of PyFm-dimer (CD2Cl2) after irradiation in the dark at 15 °C.

PyFm-dimer
For the determination of both ring-opening quantum yields the PyFm-dimer was first pre-irradiated at 0 °C with 579 nm (intense 500 W mercury lamp) for 180 min. Then, irradiation with a 660 nm LED was performed to reach the photothermal equilibrium (PTE). The back reaction was monitored for 60 min in the dark. The cuvette was warmed up to room temperature and cooled back to 0 °C to verify the complete recovery of the cc isomer. The spectra of the thermal back reaction were divided by the pure cc spectra ( Figure S30). In the wavelength region 700-730 nm the ratio appears constant, thus the cc isomer is the only isomer absorbing in that range. Figure S30. Spectra of thermal back reaction divided by spectra of pure cc isomer. In the region of 700-730 nm the ratio is constant and equals the amount of cc present at that time.
The average ratio equals the amount of cc isomer being built up with time during the thermal back reaction and the corresponding concentrations were fitted according to integrated rate laws for an irreversible consecutive reaction (Equation I, Figure S31). Because the reaction does not start with 100% oo isomer present, the remaining amount of cc was included in the fit, assuming the amount of co is negligible at the PTE. The values for k1 and k2 allowed the determination of the concentrations of oo and co at any point of time during the back reaction using equation II and III, respectively ( Figure S32). Figure S32. Concentrations of all three isomers during the thermal back reaction at 0 °C. The concentrations of cc were derived from the experimental spectra in order to calculate the rate constants. The concentrations of the other two isomers were calculated according to the kinetic rate laws for irreversible consecutive reactions.
Next, the spectra of pure oo and co isomer were calculated (see MS Figure 4a). At the photothermal equilibrium (t = 0 for the thermal back reaction) the oo isomer is the major isomer and the pure oo spectra was obtained by subtracting the amount of cc still present at the PTE (Equation IV). For the co isomer the spectra between 10 and 30 min were considered because in that time frame the co isomer is the predominating isomer and the amount of cc and oo at each time t were subtracted according to equation V.

Aλ (oo)
Absorbance at wavelength λ of pure oo isomer Aλ (co) Absorbance at wavelength λ of pure co isomer

Aλ (cc)
Absorbance at wavelength λ of pure cc isomer A λ (t) Absorbance at wavelength λ at time t

A λ (PTE)
Absorbance at wavelength λ at t=0 oo (t) percentage of oo isomer present at time t

cc (PTE)
percentage of cc isomer present at t=0 cc (t) percentage of cc isomer present at time t With the calculated spectra, it was possible to determine the concentrations of each isomer during the irradiation. For the cc isomer the irradiation spectra are simply divided by the pure cc spectra and the obtained ratio in the range of 700-730 nm is used for determining the concentrations at the corresponding point of time. Subtraction of the amount of cc from the irradiation spectra and division by the pure co spectra identifies a constant region at 560-590 nm where the ratio equals the amount of co isomer present ( Figure S33). The amount of oo isomer can then be obtained by the relationship oo (t) = 1 -cc (t) -co (t). Figure S33. Division of irradiation spectra by the pure cc spectra shows a constant region at 700-730 nm (left). Irradiation spectra minus the amount of cc isomer, divided by pure co spectra (right) gives a constant region at 560-590 nm.
The differential equations for the concentrations of each isomer were numerically integrated and the previously determined concentrations were then fitted according to the following Equations (VI-VIII), providing values for ϕcc→co and ϕco→oo based on the least-square method (see Figure S34).

PyFm-monomer
The determination of the quantum yield for the PyFm-monomer was conducted under the same conditions and with the same experimental setup as for the dimer. Irradiation with 579 nm (500 W mercury lamp) until the photothermal equilibrium, was followed by monitoring the thermal back reaction in the dark at 0 °C ( Figure S35).

Methodology
For photochemical reactions involving higher excited states and their crossings, the methods of the choice are the Complete Active Space SCF (CASSCF) and CASPT2 that includes dynamic correlation energy in a perturbative approach. Previously Boggio-Pasqua et al. [8] studied the DHP to CPD transformation of the single unsubstituted chromophore with this methodology using active space of 16 electrons in 16 orbitals. Such calculations are still on the edge of current computational resources and would clearly yield a totally non-tractable active space for the dimers. Therefore, for the present investigation, we used Time-Dependent Density Functional Theory (TD-DFT) as computationally affordable choice. As was shown recently by Boggio-Pasqua and Garavelli [9] the TD-CAM-B3LYP [10] shapes of the potential energy surface of excited states (ES) are comparable with the one obtained with CAS/CASPT2. Moreover CAM-B3LYP is well-suited for charge transfer excitations which may occur as result of push-pull character of some of the investigated dimers. We first performed ground state (GS) optimizations with the CAM-B3LYP functional and the 6-31G(d) basis set, followed by vibrational frequency calculation. Optimizations of excited state were performed at the same level of theory. The influence of the solvent (chloroform) was included by means of Polarizable Continuum Model (PCM). For ES calculations, the linear response formalism (LR) [11] of PCM was used. For all of our calculations the Gaussian16 program [12] was used with the so-called ultrafine DFT integration grid and a tight SCF-KS convergence criterion.

Solvent effects on molecular structure
In the experiment different solvents were used for different molecules to allow good solubility at given conditions (THF for PyFm-dimer and DCM for PhCN-dimer, again THF for Ester-dimer). In the prediction phase of the experimental work we used gas phase geometry optimizations to avoid any confusion regarding the choice of solvent. We observed several situations with oscillator strength borrowing between states in gas phase and thus the ring-opening state was identified mainly based on orbital picture and not on the basis of the computed f. The optimized geometries are usually very similar in gas phase and the solvent (differences of ca. 0.002 Å) with one interesting exception. We have found that the twisting between DHP units of the PyFm-dimer responds significantly to solvent polarity. In the gas phase we observed q = 1.566 Å and in chloroform q = 1.549 Å. Smaller q relates to more conjugated (and parallel) DHP units in more polar solvents. Interestingly, this was not found for the structurally similar PyCN-dimer which attains planar DHP-bridge-DHP geometry already in the gas phase (with q very similar to chloroform). Both molecules have polar acceptor groups on the opposite side of the dimer. Repolarization of the formyl group in more polar solvents is likely the cause of the observed effect. Figure S38. Electron density differences upon the excitation to the S1 state of the Isoin-dimer. The red and blue lobes indicate regions of increase and decrease of electron density upon excitation. Side alkyl chains were omitted for clarity. Isosurface value 0.001 au. Figure S39. Electron density differences upon the excitation to the S1 state of the PyCN-dimer. The red and blue lobes indicate regions of increase and decrease of electron density upon excitation. Isosurface value 0.001 au. Isoin-dimer S1 1.5867