Photophysics of Fluorescent Contact Sensors Based on the Dicyanodihydrofuran Motif

Abstract Fluorescent molecular rotors have been used for measurements of local mobility on molecular length scales, for example to determine viscosity, and for the visualization of contact between two surfaces. In the present work, we deepen our insight into the excited‐state deactivation kinetics and mechanics of dicyanodihydrofuran‐based molecular rotors. We extend the scope of the use of this class of rotors for contact sensing with a red‐shifted member of the family. This allows for contact detection with a range of excitation wavelengths up to ∼600 nm. Steady‐state fluorescence shows that the fluorescence quantum yield of these rotors depends not only on the rigidity of their environment, but – under certain conditions – also on its polarity. While excited state decay via rotation about the exocyclic double bond is rapid in nonpolar solvents and twisting of a single bond allows for fast decay in polar solvents, the barriers for both processes are significant in solvents of intermediate polarity. This effect may also occur in other molecular rotors, and it should be considered when applying such molecules as local mobility probes.


Synthesis and Characterization
Preparation of compounds 1a [1] and 1b [2] has been described previously. Compounds 2a and 2b were prepared by a procedure analogous to the one reported by Han et al. [3] , using the appropriate dialkylaminobenzaldehyde. All reactions were conducted under N2.
Compound 2a: 4-dimethylaminobenzaldehyde (675 mg, 4.53 mmol) and 2-(3-cyano-4,5,5trimethylfuran-2(5H)-ylidene)malononitrile (388 mg, 2.06 mmol) were dissolved in 18 mL of pyridine, and 3 drops of acetic acid were added. The mixture turned dark blue after approximately 10 minutes of stirring at room temperature. After 48 hours, the mixture was poured into 150 mL of ice-water and filtered. The precipitate was dissolved in CH2Cl2, washed thoroughly with water/brine and the organic solvent was evaporated. The obtained dark blue powder was reprecipitated from CH2Cl2/MeOH to give 211 mg of 2a (31 %). A similar preparation of this compound can be found in reference [4]. 1  Compound 2b: A mixture of 4-(bis(2-hydroxyethyl)amino)benzaldehyde (778 mg, 3.72 mmol) and 2-(3-cyano-4,5,5-trimethylfuran-2(5H)-ylidene)malononitrile (340 mg, 1.69 mmol) in pyridine (200 mL) was stirred at room temperature until all reagents dissolved. After that, 3 drops of acetic acid were added to this mixture, which turned dark after 10 minutes of stirring at room temperature. The reaction mixture was left to stir for 24 hours, after which it was poured into 150 mL of ice-water. 0.1 M HCl was added until a precipitate could be observed. This precipitate was filtered, and 140 mg of blue powder with a reddish shine was obtained (22 %). Preparation of this compound is also described in ref. [5] . 1

Glass silanization
Cleaning: A Teflon rack with 4 borosilicate glass cover slips (2 cm × 2 cm × 170 μm) was sonicated in aqueous Extran solution (0.3 % w/w) for 30 mins, then sonicated in deionized water (10 min) and in ethanol (30 mins), dried under an air current, and placed in an ozone photoreactor for 2 hours.
Silanization with 3-aminopropyltrimethoxysilane (APTES): 80 mL of aqueous ethanol (96 %, v/v) was mixed with a small amount of acetic acid (so that pH = 5). To this mixture, APTES (2 mL) was added. Cover slips were placed in this solution and were left to react (with stirring) for 25-30 minutes. After this time, the cover slips were sonicated in ethanol 2 times for 30 minutes, dried in air, and annealed at 130 °C for 3 days.

Immobilization of 1b and 2b
Monolayer M1: 1b (15 mg, 0.037 mmol) was dissolved in 80 mg of dry dimethylformamide (DMF) under N2. To this solution, (benzotriazol-1-yloxy)tris(dimethylamino)phosphonium hexafluorophosphate (BOP; 100 mg, 0.230 mmol) and hydroxybenzotriazole (HOBt; 35 mg, 0.23 mmol) were added. 120 μL of N,N-diisopropylethylamine (DIPEA) were added to this solution, and a rack with cover slips (previously functionalized with APTES) was immersed in the solution. This was left to stir overnight. After that, cover slips were sonicated in ethanol twice for 15 minutes, and dried under an air current. Cover slips were stored in plastic containers until use.
Monolayer M2: Compound 2b (10 mg, 0.025 mmol) was dissolved in dry dichloromethane under N 2 atmosphere. To this solution, 15 % phosgene solution (0.50 mL, 0.66 mmol) in toluene was added. This solution was stirred for 5 minutes in order to provide enough time for the acylation reaction of phosgene with the alcohol groups to take place, after which a rack with APTES silanized cover slips was added. This was stirred for approximately 20 minutes after which the cover slips were taken out and thoroughly washed in dichloromethane, toluene and ethanol by sonication for 15 minutes in each solvent. Samples were stored in sealed plastic containers under Ar until use.

Surface density of molecular rotors
In order to estimate the grafting density of chromophores on our functionalized surfaces, we measured UV-Vis absorption spectra ( Figure S1). Absorbance is very low, as expected from a single dye layer attached to the surface. Compared to spectra of 1 and 2 in solutions, absorption spectra of M1 and M2 appear to be broader. If we nonetheless assume that the optical cross sections of immobilized molecules are the same as in toluene solutions ( = 62200 M -1 cm -1 and  = 42300 M -1 cm -1 for 1a and 2a, respectively), Lambert-Beer law can be used to estimate the number of molecules per unit area to be ~A/(2) (since both sides of cover slips are functionalized). The grafting density obtained in this way is ~1 molecule per 10 nm 2 for both M1 and M2. Such a low grafting density should, in principle, not allow for a large degree of homo FRET and self quenching. Figure S1. Electronic absorption spectra of cover slips used in contact imaging.

Polystyrene beads
Polystyrene spheres (microbeads) were purchased from the Precision Plastic Ball Company (Addingham, Ilkley, UK) and roughened by placing them over 240 grit sand paper on an orbital shaker (1200 rpm) for 2 days.

Photophysical measurements.
Almost all equipment and methods used for spectroscopic experiments and microscopy has been described in earlier work. [2,[6][7][8][9] For analysis of the transient absorption data we used Glotaran 1.5.1. [10] Fluorescence lifetime images were measured with a Olympus IX-71 microscope equipped with a MicroTime 200 TCSPC unit (PicoQuant GmbH) and a 100x 1.4 NA objective (UplanSApo, Olympus), mounted on a piezo-scanning stage (Physik Instruments GmbH). A detection pinhole with a diameter of 75 m was used. An NKT Photonics SuperK Extreme Supercontinuum white laser (80 MHz) was used as the excitation source. Because of the nonexponential nature of the fluorescence decays, the average fluorescence lifetimes are defined as the average times that emitted photons take to reach the detector after the excitation pulse.
Fluorescence spectra of M1 and M2 (figure 9 in main text) were obtained by passing the emitted light in the confocal microscope to a spectograph (Spectra Pro-150, Acton Research Corp.) equipped with an EM-CCD camera (PhotonMax 512B, Roper Scientific). Figure S2. Transient absorption spectra of 1a in toluene. a) transient spectra at different delay times; b) selected time traces (black markers) and fits (colored lines) produced by compartmental global analysis; c) decay-associated difference spectra; d) species-associated spectra. Figure S3. Vis-pump / vis-probe probe measurements for 1a in MeCN: a) selected transient spectra; b) selected time traces (black markers) and fits (colored lines) produced by compartmental global analysis; c) decay-associated difference spectra (with time constants); d) species-associated difference spectra. Figure S4. Förster-Hoffmann plot for 1a and 2a in glycerol. The viscosity was varied by varying the temperature. [11]