Coumarin Dimer Is an Effective Photomechanochemical AND Gate for Small-Molecule Release

Stimulus-responsive gating of chemical reactions is of considerable practical and conceptual interest. For example, photocleavable protective groups and gating mechanophores allow the kinetics of purely thermally activated reactions to be controlled optically or by mechanical load by inducing the release of small-molecule reactants. Such release only in response to a sequential application of both stimuli (photomechanochemical gating) has not been demonstrated despite its unique expected benefits. Here, we describe computational and experimental evidence that coumarin dimers are highly promising moieties for realizing photomechanochemical control of small-molecule release. Such dimers are transparent and photochemically inert at wavelengths >300 nm but can be made to dissociate rapidly under tensile force. The resulting coumarins are mechanochemically and thermally stable, but rapidly release their payload upon irradiation. Our DFT calculations reveal that both strain-free and mechanochemical kinetics of dimer dissociation are highly tunable over an unusually broad range of rates by simple substitution. In head-to-head dimers, the phenyl groups act as molecular levers to allow systematic and predictable variation in the force sensitivity of the dissociation barriers by choice of the pulling axis. As a proof-of-concept, we synthesized and characterized the reactivity of one such dimer for photomechanochemically controlled release of aniline and its application for controlling bulk gelation.


I. DFT calculations
All calculations were performed with the Gaussian 16 suite of software in vacuum.The UHF formalism was applied to all open-shell singlets (TS1, Int and TS2) and the wavefunction stability checks were performed on all converged geometries of TS1, Int and TS2; on all initial guess structures of the transition states and on all converged geometries of the dimer coupled to force >4 nN.The Berny algorithm was applied to locate stationary points.The nature of each converged strain-free geometry and select force-coupled geometries was confirmed by the presence of 0 (minima) or 1 (saddle points) imaginary frequencies obtained by analytical frequency calculations.Tight convergence criteria and ultrafine integration grids were used in optimisations and frequency calculations.All force-coupled stationary geometries were optimized as previously described. 1Force dependent electronic energies, constrained distances and thermodynamic corrections (see below) were obtained by interpolation of the results of relaxed potential energy scans of either the terminal MeOC … COMe (series 2, 4-6) or terminal HC…CH (series 3) distance of the dimer.
The initial (guess) geometry of the 1 st dissociation transition state of each unsubstituted coumarin dimer was generated by scanning, at uMPW1K/6-31+G(d), the proximate scissile bond (hh isomers) or any scissile bond (ht isomers) of the cyclobutene core, identifying the converged geometry corresponding to the highest energy, performing wavefunction stability test, frequency calculation, unconstrained optimization to a saddle point, another wavefunction stability test and final frequency.The initial guess geometry of TS2 was derived for each conformer of Int by scanning the remaining scissile bond.Conformers of Int were generated systematically as previously described. 2Guess geometries of transition states of the alternative dissociation path (distal bond scission followed by proximate bond scission, Fig. 2b main text) were generated by manually elongating the scissile bond of the dimer (TS1a) or of each conformer of Int (TS2a) to the value of the scissile bond in the corresponding converged conformer of TS1 or TS2.The initial guess geometries of the concerted TSs (ht isomers) were obtained by manually elongating both scissile bonds of the cyclobutane core to 1.85 Å.
Thermodynamic corrections (TCs) to electronic energies of individual converged geometries were calculated statistical-mechanically in the pseudo-harmonic oscillator/rigid rotor/ideal gas approximations, as 3RT +ZPE +Uvib -TS, where ZPE is the zero-point energy, Uvib is the vibrational component of the internal energy and S is the total entropy.Vibrational frequencies below 500 cm -1 were replaced with 500 cm -1 as previously recommended, 3 to avoid the artifactually high contribution of such low-frequency modes to the vibrational entropy.The use of analytical frequencies calculated on converged force-coupled geometries in this study is theoretically sound because the calculation is performed on the molecule plus its infinitely-compliant constraint (rather than just the molecule), which is a stationary point with all internal forces at 0.
Cartesian coordinates of conformational minima of the strain-free kinetically significant states of coumarin dimers 1-6, including with multiple functionals, and their force-dependent free energies are
1 H NMR and 13 C NMR spectra were recorded in CDCl3 or DMSO-d6 or D2O or DMF-d7 at 25 ℃ on a Bruker Avance III spectrometer (500 MHz 1H) or Bruker Avance II spectrometer (400 MHz 1H).The chemical shifts were given in ppm (δ) based on internal TMS or residual protonated solvent.The peak patterns are indicated as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; qui, quintet; sxt, sextet.The coupling constants, J are reported in Hertz (Hz).Mass spectroscopy data were collected on an autoflex maX MALDI-TOF MS instrument.The UV-Vis spectroscopy was performed on Thermo Scientific Multiskan SkyHigh and fluorescence spectra on F7000.Ultrasound experiments were carried out on a Sonics Vibra-Cell 505 liquid processor with a 12.8 mm (diameter) titanium solid probe.Gel permeation chromatography (GPC) was executed using Waters GPC system equipped with 2414 differential refractometer and four columns (7.8 × 300 mm).The eluent was N,Ndimethylformamide (DMF) (HPLC grade, ≥ 99.9%) containing 5.0 mM NH4BF4.HPLC was recorded by Aglilent LC-1200 equipped with an Eclipse XDB-C18 column and a UV-vis detector monitored at 280 nm (gradient :0 min, 50% CH3OH -50% H2O; 4 min, 60% CH3OH -40% H2O; 10 min, 100% CH3OH; flow 1.0 mL/min).Rheology test was performed on a TA discovery hybrid rheometer equipped with a 25 mm parallel plate and a ETC environmental control chamber at room temperature.

III. Method 1. Ultrasound sonication experiments
The sonication experiments were carried out on a Sonics Materials Vibra Cell 505 liquid processor equipped with a 12.8 mm (diameter) titanium solid probe using aqueous solutions of polymers at 6.0 mg/mL (0.033 mM) in a 3-neck glass cell immersed in an ice-water bath.Before sonication, the solution was purged with Ar for 20 min.Total experiments were carried out under Argon with a pulse sequence of 1.0 s on followed by 1.0 s off at a nominal power of 1.98 W/cm 2 (f = 20 kHz, 30% amplitude, E = 147361 J, t = 57600 s, A = 1.29 cm 2 ).The temperature of the system was kept at 0-5 o C. 0.6 mL or 1.0 mL solution were withdrawn periodically from the cell for UV-vis, fluorescence and GPC tests.

UV Irradiation experiments
The irradiation experiments were performed on a COUSZ LED UV light equipped with a standard focusing mirror UPUL008 (λ = 365 nm, 6500 mW/cm 2 ) at room temperature.300 μL solution containing P7 (6.0 mg/mL, 0.044 mM), P8 (6.0 mg/mL, 0.033 mM), sonicated P8 in H2O (6.0 mg/mL, 0.033 mM), 7 (0.05 mM), or 8 in DMF/H2O (2:1 by volume, 0.10 mM) was filled into a quartz cuvette (2 mm × 10 mm), which was then irradiated with UV light.The UV light was placed vertically at a distance of 1 cm from the top of container, and irradiation time range from 5 to 300 s with a step time in between 5 -30 s.After each irradiation step, the UVvis absorption spectra were recorded.
For NMR tests, the concentration of small molecule was 0.4 mM in DMF-d7/D2O (volume ratio of 2/1, 0.5 mL) and it was 200 mg/mL in D2O (0.5 mL) for polymer.The irradiation process was performed in an NMR tube covered by alumina foil.Due to the considerable overlap of NMR signals between reactant and product, the yield of photochemically released aniline was measured by HPLC.Generally, 1.8 mL solution containing P7 (6.0 mg/mL, 0.044 mM), P8 (6.0 mg/mL, 0.033 mM), sonicated P8 in H2O (6.0 mg/mL, 0.033 mM), 7 (10 μM) or 8 in 50% aqueous CH3OH (15 μM) was filled into a quartz cuvette (10 mm × 10 mm) and irradiated with UV light.After each irradiation step, the sample was analysed by HPLC equipped with a UV-vis detector.The signal at 280 nm was monitored, and the generated aniline was quantified using a calibration curve.The synthesis was adapted from the literature 6 .7-hydroxy-4-methyl-2H-chromen-2-one (4.4043 g, 25.0 mmol) and anhydrous potassium carbonate (6.9105 g, 50.0 mmol) were suspended in dry acetonitrile (50 mL) in a 100 mL round-bottom flask followed by dropwise addition of 2-iodoethanol through syringe.The reaction was refluxed and stirred under argon for 24 h.After completion, the mixture was filtered in hot and washed with dichloromethane to remove the starting material.The filtrate was evaporated under reduced pressure to give a white crude product S1 (4.7080 g, 85.5%).
The synthesis was adapted from the literature 8 .7 (151.2mg, 0.3 mmol) and benzophenone (328.0 mg, 5.0 mmol) were suspended in dry acetonitrile (30 mL) in a 50 mL Schlenk flask charged with argon.Under irradiation of ultraviolet light (45 W, wavelength = 365 nm), the reaction was stirred at 15 ℃ for 7 h.After completion, the mixture was concentrated and the residue was purified by column chromatography (dichloromethane : methanol = 100 : 2) and recrystallization to give a white solid 8 (120.1 mg, 78.9%). 1 H NMR   Under argon atmosphere, a Schlenk flask with a side arm was charged respectively with initiator 8 (25.2 mg, 0.025 mmol), copper wire (12.1 cm, diameter = 0.64 mm, polished by abrasive paper), Me6TREN (11.5 mg, 0.05 mmol), oligo(ethylene glycol) methyl ether methacrylate (OEGMA, Mn = 300, 26.3 mL, 90 mmol) and dry DMSO (26.3 mL).The mixture was subjected to three freeze-pump-thaw cycles and then stirred at 28 ℃ under vacuum for 3.5 h.Afterwards, the reaction was terminated upon exposure to air.The viscous solution was diluted with tetrahydrofuran (5.0 mL) and precipitated in cold ether three times.The obtained product was dried by vacuum pump to afford a colorless polymer P8.The step is similar to the synthesis of polymer P8.Under argon atmosphere, a Schlenk flask with a side arm was charged respectively with initiator 7 (25.2mg, 0.05 mmol), copper wire (12.1 cm, diameter = 0.64 mm, polished by abrasive paper), Me6TREN (11.5 mg, 0.05 mmol), oligo(ethylene glycol) methyl ether methacrylate (OEGMA, Mn = 300, 26.3 mL, 90 mmol) and dry DMSO (26.3 mL).The mixture was subjected to three freeze-pump-thaw cycles and then stirred at 28 ℃ under vacuum for 3.5 h.Afterwards, the reaction was terminated upon exposure to air.The viscous solution was diluted with tetrahydrofuran (5.0 mL) and precipitated in cold ether three times.The obtained product was dried by vacuum pump to afford a colorless polymer P7.

VI. Physical measurements, material characterization and data processing 1. Estimation of Mechanochemical Activation Yield
Fragmentation of P8 by retro [2 +2] cycloaddition of the coumarin dimer during ultrasound sonication can be represented as: where  is the fraction of the dimer that dissociated mechanochemically at sonication time t. is the slope of the linear dependence of fluorescence intensity of P8 solution on the molar concentration of the dimer. the linear dependence of fluorescence intensity of P7 solution on the molar concentration of the coumarin.
[]  is the concentration of the coumarin dimer at sonication time t.
[]  is the concentration of the coumarin at sonication time t. 0 is the fluorescent intensity of the P8 solution prior to sonication, I(t) is the fluorescent intensity of sonicated P8 solution at time t.The progress of photochemical reaction upon 365 nm irradiation of coumarin-terminated P7 and of the small-molecule analog 7 was also followed by UV-Vis spectroscopy (Figure S36, Figure S37, Figure S38 and Figure S39), demonstrating a gradual increase in the absorption intensity at 280 nm (aniline) and concomitant decrease in absorbance at 322 nm (coumarin/aniline adduct).The photochemistry of 7 was also investigated by UV-Vis and the results are shown in Figure S38 and Figure S39, and the phenomenon was similar to that of P7.We confirmed dimerization of coumarin under 365 nm UV irradiation by analyzing by HPLC a solution of 9 and by GPC a solution of P7 irradiated at 365 nm ((Figure S40-Figure S41).In HPLC samples the intensity of the peak at 7.5 min (9) gradually decreased and a new peak appeared at 5.8 min, which we attribute to dimer 92.GPCs of the irradiated solution of P7 developed a shoulder at higher molecular mass, with Mn of the sample increasing from 136.6 kDa to 161.3 kDa after 5 min of irradiation, suggesting photodimerization of coumarinterminated P7 to P8.        Wavelength, nm 0.0 mg/mL 2.0 mg/mL 4.0 mg/mL 6.0 mg/mL 8.0 mg/mL 10.0 mg/mL    .The number-average molar mass, Mn, of a sonicated solution of P7 (0.9 mg/mL in DMF containing 5 mM NH4BF4) at different sonication times.For each time, a 300 μL aliquote of this solution was freeze-dried and then dissolved in DMF containing 5 mM NH4BF4 for GPC analysis. .Absorption at 280 nm of HPLC of a solution of P7 (6.0 mg/mL, 0.044 mM in DI water) sonicated for 0 h, 0.5 h, 1.0 h, 1.5 h, 2.0 h, 4.0 h, 8.0 h, 16.0 h; sonicated for 16 h and then irradiated at 365 nm UV light for 300 s, and of a 5 μM solution of aniline in DI water. .Absorption at 280 nm of HPLC of aqueous solutions of P7 (6.0 mg/mL, 0.044 mM): before sonication or irradiation, after irradiation at 365 nm for 300 s; after sonication for 2 h; and after 2 h-sonication followed by irradiation at 365 nm for 300 s; of aqueous solutions of P8 (6.0 mg/mL, 0.033 mM): before sonication or irradiation, after sonication for 16 h; after sonication for 16 h and irradiation at 365 nm for 300 s; and of a 5 μM aqueous solution of aniline.

Photochemical release of aniline.
Absorption spectra of an aqueous solution of P8 sonicated for 16 h followed by irradiation at 365 nm displayed a gradually decreasing absorbance at 322 nm, corresponding to the coumarin/aniline construct (Figure S57).Hydroxy-terminated 4-arm star PEG (Mn = 10 kDa, ÐM = 1.02, 10.8376 g, 1.083 mmol) and triethylamine (0.6580 g, 6.5 mmol) were dissolved in dry dichloromethane (80.0 mL) in a 250 mL round-bottom flask charged with argon, followed by dropwise addition of a solution of 4-formly benzoyl chloride (1.0962 g, 6.5 mmol) in dry dichloromethane (20.0 mL) through syringe at 0 ℃.The resulting mixture was moved to room temperature and stirred for 16 h.Afterwards, the reaction was quenched by saturated sodium bicarbonate (40 mL) at room temperature for 0.5 h and extracted with dichloromethane five times (100 mL × 3).The organic solvent was combined and washed with saturated brine, dried over anhydrous sodium sulfate, filtered, and evaporated in vacuo.The residue was precipitated from dichloromethane with cold ether three times to afford 10 as a white powder (11.0311 g, 96.7% yield, 74.0% modification).The modification rate was determined by comparing the 1 H NMR peak of the benzaldehyde group on 10 with the CH2 peak of ethylene glycol (EG) repeat unit (3.5 ppm, Fig. S59).Each molecule of the precursor has 4x227=908 such protons, the measured 2.96:909 ratio of intensities of the two peaks suggests ~75% of the terminal OH groups in the precursor were terminated with benzaldehyde.

Figure S32 .Figure S33 .
Figure S32.The 280 nm UV-vis detector output of HPLC of a solution of aniline in deionized water at different concentrations.

Figure S34 . 7 Figure S35 .
FigureS34.The 280-nm output of the UV-vis detector of HPLC of a 6.0 mg/mL (0.044 mM) solution of P7 in deionized water at different irradiation (λ = 365 nm) times.

Figure
Figure S38.UV/vis spectra of 7 (as a 0.05 mM solution in DMF/H2O at 2:1 by volume) at different irradiation (λ = 365 nm) times; the spectrum of aniline at 0.05 mM is shown for reference.

Figure
Figure S40.RI detector output of HPLC of 10 μM of 9 in 50% aqueous CH3OH at different irradiation (λ = 365 nm) times.A peak at 5.8 min is the dimer.

Figure S45 .
Figure S45.Absorbances at 280 nm of HPLC of 8 (as a 15 μM in 50% aqueous CH3OH) at different irradiation (λ = 365 nm) times; for reference, the HPLC of a 5 μM solution of aniline in deionized water is shown at the top.

Figure S48 .
Figure S48.The fluorescent intensity at 389 nm of a solution of P8 (6.0 mg/mL, 0.033 mM in DI water) as a function of sonication times (λex= 322 nm).

Figure S50 .
Figure S50.The fluorescence emission intensity at 389 nm of an aqueous solution P8 as a function of its concentration.(λex = 322 nm).

Figure S51 .
Figure S51.The fluorescence emission spectra of an aqueous solution of P7 at different concentrations (λex = 322 nm).

Figure S52 .
Figure S52.The fluorescence emission intensity at 389 nm of aqueous solutions P7 as a function of its concentration (λex = 322 nm).
FigureS54.The number-average molar mass, Mn, of a sonicated solution of P7 (0.9 mg/mL in DMF containing 5 mM NH4BF4) at different sonication times.For each time, a 300 μL aliquote of this solution was freeze-dried and then dissolved in DMF containing 5 mM NH4BF4 for GPC analysis.
Figure S55.Absorption at 280 nm of HPLC of a solution of P7 (6.0 mg/mL, 0.044 mM in DI water) sonicated for 0 h, 0.5 h, 1.0 h, 1.5 h, 2.0 h, 4.0 h, 8.0 h, 16.0 h; sonicated for 16 h and then irradiated at 365 nm UV light for 300 s, and of a 5 μM solution of aniline in DI water.
Figure S56.Absorption at 280 nm of HPLC of aqueous solutions of P7 (6.0 mg/mL, 0.044 mM): before sonication or irradiation, after irradiation at 365 nm for 300 s; after sonication for 2 h; and after 2 h-sonication followed by irradiation at 365 nm for 300 s; of aqueous solutions of P8 (6.0 mg/mL, 0.033 mM): before sonication or irradiation, after sonication for 16 h; after sonication for 16 h and irradiation at 365 nm for 300 s; and of a 5 μM aqueous solution of aniline.

Figure 3 .
Figure S57.UV-Vis spectra of a solution of P8 (6.0 mg/mL, 0.033 mM in DI water) sonicated for

Table S1 .
The standard free energies (kcal/mol) of the kinetically significant stationary states of dissociation of the 4 isomeric unsubstituted coumarin dimers with 4 functionals and 6-31+G(d) basis set in vacuum.The rate-limiting TS is in bold.The UHF designation is omitted for brevity.

Table S2 .
Maximum force in nN at which the R/TS1 and Int/TS2 pairs exist.