Push‐Pull Stiff‐Stilbene: Proton‐Gated Visible‐Light Photoswitching and Acid‐Catalyzed Isomerization

Abstract Donor‐acceptor substituted stiff‐stilbene is shown to undergo isomerization induced by visible light avoiding the need for harmful UV light. This visible‐light photoswitching is inhibited by protonation of the dimethylamino‐donor unit, disrupting the push‐pull character and thus, gating of the photochromic properties is allowed by acid/base addition. Remarkably, the addition of a mild acid also triggers fast thermal back‐isomerization, which is unprecedented for stiff‐stilbene photoswitches usually having a very high energy barrier for this process. These combined features offer unique orthogonal control over switching behavior by light and protonation, which is investigated in detail by 1H NMR and UV/Vis spectroscopy. In addition, TD‐DFT calculations are used to gain further insight into the absorption properties. Our results will help elevating the level of control over dynamic behavior in stiff‐stilbene applications.


Experimental section General methods and materials:
THF, MeCN, and toluene were dried using a Pure Solve 400 solvent purification system from Innovative Technology. Dry DMF was purchased from Acros Organics and DMSO-d6 , MeCN-d3 and CDCl3 were purchased from Eurisotop. DMSO-d6 was stored under N2 over molecular sieves (4Å) and CDCl3 was filtered over basic Al2O3 to remove DCl. The degassing of the solvents was carried out by purging with N2 for 15 min, unless noted otherwise. (E)-2,2',3,3'tetrahydro-1,1'-biindenylidene [(E)-6] was prepared according to a procedure reported in the literature. [1] All other chemicals were commercial products and were used without further purification. Column chromatography was performed using silica gel (SiO2) purchased from Screening Devices BV (Pore diameter 55-70 Å, surface area 500 m 2 g -1 ) and neutral aluminum oxide (Al2O3) purchased from Fluka Analytical. Thin-layer chromatography (TLC) was carried out on aluminum sheets coated with silica 60 F254 and neutral aluminum oxide obtained from Merck. Compounds were visualized with UV light (254 nm) or by staining with potassium permanganate. 1 H and 13 C NMR spectra were recorded on Bruker AV 400 and Bruker 500 Ultra Shield instruments at 298 K unless indicated otherwise. Chemical shifts () are denoted in parts per million (ppm) relative to residual protiated solvent (DMSO-d6: for 1 H detection,  = 2.50 ppm; for 13 C detection,  = 39.52 ppm; CDCl3: for 1 H detection,  = 7.26 ppm; for 13 C detection,  = 77.16 ppm; MeCN-d3: for 1 H detection,  = 1.94 ppm; for 13 C detection,  = 1.32 and 118.26 ppm). The splitting pattern of peaks is designated as follows: s (singlet), d (doublet), t (triplet), q (quartet), p (quintet), h (septet), m (multiplet), br (broad). High-resolution mass spectrometry (ESI-MS) was performed on a Thermo Scientific Q Exactive HF spectrometer with ESI ionization. IR spectra were recorded on a Perkin Elmer Spectrum Two FT-IR spectrometer. The wavenumber ( is in units of reciprocal centimeters (cm 1 ) and the intensity (v = cm -1 ) is designated as follows: s (strong), m (medium), w (weak), very w (very weak), br (broad), and sh (shoulder). Melting points were determined, with a Büchi M560 apparatus. UV-Vis spectra were recorded on an Agilent Cary 8454 spectrometer using 1 cm or 1 mm quartz cuvettes. Irradiation of UV-Vis and NMR samples was carried out using Thorlab model LEDs (17 mW) positioned at a distance of 1 cm to the sample. Longpass filters with cut-on wavelengths (cut-on) of 280±5 nm and 360±5 nm were acquired from Newport Corporation. [2,3] First, TiCl4 (4.18 mL, 37.9 mmol) was slowly added to a vigorously stirred suspension of Zn powder (4.95 g, 75.7 mmol) in THF (38 mL) under N2 atmosphere. The resulting mixture was stirred at reflux for 2 h and cooled to rt. Subsequently, 5-bromo-1-indanone (4.01 g, 19.0 mmol) was added to the black suspension and the mixture was stirred at reflux for another 16 h, cooled to rt and treated with a saturated aqueous NH4Cl solution (50 mL) and extracted with CHCl3 (4 × 80 mL). The volume of the combined organic layers was reduced to 30 mL and the resulting precipitate was filtered off and air-dried to obtain (E)-2 (2.21 g, 59%) as a yellow solid. Rf = 0.48 (SiO2, pentane); m.p. 273 -275 °C ( [3] : 269  271 °C); 1 H NMR (400 MHz, CDCl3, (E)isomer assignment based on the NOESY spectrum in Figure S4 [3] ). The filtrate was concentrated and purification by column chromatography

Attempted mixed McMurry reaction of 5-bromo-1-indanone and compound 5:
First, TiCl4 (0.18 mL, 1.7 mmol) was slowly added to a vigorously stirred suspension of Zn powder (0.22 g, 3.4 mmol) in THF (1 mL) under a N2 atmosphere. The resulting mixture was stirred at reflux for 2 h and cooled to rt. Then, 5-bromo-1-indanone (94 mg, 0.45 mmol) and compound 5 (66 mg, 0.38 mmol) were added to the black suspension and the mixture was stirred at reflux for another 16 h, cooled to rt and treated with a saturated aqueous NH4Cl (10 mL) solution and extracted with CHCl3 (3 × 20 mL). The combined organic layers were dried over Na2SO4 and concentrated, to afford a crude residue (see Figure S13 for 1 H NMR analysis). Figure S1. 400 MHz 1 H NMR spectrum of (E)-2 measured at 298 K in CDCl3.

H and 13 C NMR Spectra of new compounds
S8 Figure S2. 400 MHz 1 H NMR spectrum of (Z)-2 measured at 298 K in CDCl3.

UV-Vis photoisomerization studies
Note: When a sample of (E)-1 in MeCN was irradiated and a spectrum recorded, the absorption increased upon recording additional spectra. As thermal isomerization is negligible at rt (see Figure S19), this increase was ascribed to Z→E back isomerization caused by the deuterium light source (UV range) of the used diode array UV-Vis spectrophotometer. This process could be avoided by placing a longpass filter (cut-on = 280±5 nm) in in front of the light source.  irradiation (for 8 s) was virtually the same as that using 365 nm, revealing similar PSS ratios.
A cut-on filter ( = 280±5 nm) was placed in front of the light source of the spectrophotometer.

Quantum yield determination
The photon flux of the used Thorlab model M405F1 high-power LED (max = 405 nm) was determined at 1/3 of the maximum power by measuring the production of ferrous ions from potassium ferrioxalate. [4] This production was determined at different irradiation times and the slope of a plot of the concentration of Fe 2+ ions against time ( Figure S16) corresponds to the rate (r) of Fe 2+ ion formation, which was found to be: 1.50  10 -6 M s -1 . Hence, the moles of photons absorbed per time unit (Nh/t = moles of Fe 2+ / t) in a 2 mL solution can be calculated using the reported quantum yield of ferrioxalate ( = 1.14), [5] which gives: 2.63  10 -9 mol s -1 . Figure S16. Concentration of Fe 2+ ions obtained after different 405 nm irradiation periods of a stirred (1200 rpm) 2 mL solution of potassium ferrioxalate (0.15 M) in a 1 cm quartz cuvette. [4] Next, a solution of (E)-1 in MeCN (1.0  10 -4 M) in a 1 cm quartz cuvette was irradiated with the same light source in the same cuvette holder and the absorbance increase at  = 450 nm was monitored over time. The molar absorptivity of (Z)-1 at this wavelength was estimated by deconvolution of the PSS405 absorption spectrum using the E/Z ratio determined by 1 H NMR spectroscopy (see Figure S18) and using the known absorption spectrum of (E)-1.

S26
For the time-based measurements, 1 L of TFA was added to 2 mL of a stirred (1200 rpm) solution of (E)-1 (2.5 × 10 5 M in MeCN) at 20 °C. The solution was irradiated with 340 nm light for 3 min and thereafter a UV-Vis spectrum was recorded every 5 s. A cut-on filter ( = 360±5 nm) was placed in front of the light source of the spectrophotometer to minimize undesired photoisomerization. The rate constants (k) were obtained by fitting to the equation A = A0e -kt using OriginPro 9.1 software. The average of this constant was used to calculate halflife (t1/2) and Gibbs free energy of activation ( ‡ G°) using the following equations:

Time-dependent DFT calculations
Using the DFT geometry-optimized structures, TD-DFT calculations (solving for 30 singlet excited states) were performed at the same B3LYP/6-31G(d,p) level of theory and IEFPCM acetonitrile solvation model using Gaussian 09. [6] The output data was again visualized using Avogadro 1.2.0. [7]