Harnessing Proton-Coupled Electron Transfer for Hydrogenation of Aza-Arenes: Photochemistry of Quinoxaline Derivatives in Methanol

Three quinoxaline derivatives are investigated both experimentally and theoretically to assess their ability for the methanol oxidation and harvesting of hydrogen. In inert solvents, the nonplanar compounds exhibit very weak fluorescence from the lowest excited singlet state, whereas the planar and rigid chromophore emits non-Kasha fluorescence from the S2(ππ*) state despite the proximity of the S1(nπ*) state. In methanol, hydrogen-bonded complexes with solvent molecules are formed, and in all chromophores, the lowest singlet state is populated after excitation of the S2(ππ*) state. The switch from non-Kasha emission of the planar compound in inert solvents to regular emission in methanol is related to reduced symmetry of the hydrogen-bonded complex with methanol which results in effective mixing of ππ* and nπ* states and fast internal conversion to the lowest excited singlet state. The S1(nπ*) state of the hydrogen-bonded complex has charge-transfer character, and for all compounds in methanol, hydrogen transfer to the chromophore is observed. The chromophores retain the transferred hydrogen atoms, serving both as photocatalysts and as hydrogen storage materials. Undesired dark side reactions that occur are also discussed.


Synthesis of dibenzo[f
The reaction and purification was carried out without protection from oxygen and water from air.In a one necked-flask (250 ml) was put 9,10-phenanthrenedione (2.08 g, 10 mmol, 1 equivalent); 2,3diaminopyridyne (1.08 g; 10 mmol, 1 equivalent) and 2-propanol (any grade, 80 ml).The resulting brownorange suspension was vigorously refluxed with stirring under condenser for 24 hours.The reaction mixture was heterogeneous entire time.The monitoring of the reaction was done by TLC analysis using dichloromethane as eluent.After the condensation was practically completed, the flask was cooled below 50 o C. The crystalline powder was filtered off on a sinter funnel (G3, 1.5 inch wide) and washed with methanol (3 x 40 ml) and dried in air for 24 hours to yield brown-yellow crystalline powder (1.91 g).By cooling of the mother liquor (freezer), second crop was obtained and isolated similarly (0.64 g).The final very impure crop could be isolated from mother liquors (0.13 g).Total yield (95%).The first crop was crystallised from boiling 1-propanol (100 ml, p.a. grade, 2-propanol is too weak solvent to crystallise the compound) and putting in the fridge overnight.The light brown needles were filtered on a sinter funnel (G3, 1.5 inch wide) and washed with 2-propanol (2 x 30 ml) and dried in air for 24 hours.(1.79 g, 63 %) The product at this point had good NMR characteristics but it was unsuitable for the photophysical studies.220 mg of needles were sublimed under vacuum (220 o C, 10 -3 Torr, 4 hours) in a sublimation apparatus and yellow crystalline solid was obtained (200 mg).Everything was suspended in 2-propanol (40 ml, Uvasol) and ethyl acetate was added to dissolve the product (10 ml, Uvasol).The solution was boiled until the volume was reduced to 30 ml.The resulting pale yellow solution was put in the freezer (-25 o C) for 48 hours.
The crystallisation yielded pale yellow needles which were filtered on a sinter funnel (G3, 1 inch wide) and washed with 2-propanol (2 x 30 ml, Uvasol).The collected crystals were dried in air for 24 hours and in vacuum desiccator (6h, 75 o C, 2 mbar).Pale yellow needles (178 mg).The compound is known (I.The reaction and purification was carried out without protection from oxygen and water from air.In a one necked-flask (250 ml) was put 1,2-di(2-pyridyl)ethanedione (2,2'-Pyridil, 2.12 g, 10 mmol, 1 equivalent); 2,3-diaminopyridyne (1.08 g; 10 mmol, 1 equivalent) and 2-propanol (any grade, 80 ml).The resulting suspension was vigorously refluxed with stirring under condenser for 1 hour.The initial suspension promptly formed black solution.The monitoring of the reaction was done by TLC analysis using ethyl acetate as eluent.After the condensation was practically completed, the flask was cooled below 50 o C and filtered through a pad of cotton wool to remove insolubles.Next, the solution was put in the freezer (-25 o C) and gradually diluted with hexanes (3 x 50 ml) (the product crystallised capriciously at this step) to induce separation of the solid phase.After several days the beige solid was filtered off on a sinter funnel (G3, 1.5 inch wide) and washed with hexanes (3 x 30 ml) and dried in air for 24 hours.Beige powder (2.07 g, 72%).The mother liquor was kept.The obtained product was crystallised from boiling 2-propanol (20 ml, p.a. grade), quick diluting with n-heptane (3 x 20 ml) and leaving at ambient temperature.This time the crystallisation was quick and after 24 hours the light beige crystals were filtered on a sinter funnel (G3, 1.5 inch wide) and washed with hexanes (3 x 30 ml) and dried in air for 24 hours.(1.74 g, 61 %).
The product at this point had good NMR characteristics but it was unsuitable for the photophysical studies.250 mg of crystals were sublimed under vacuum (180 o C, 10 -3 Torr, 4 hours) in a sublimation apparatus and white crystalline solid was obtained (235 mg).Everything was dissolved in boiling 2-propanol (15 ml, Uvasol) and n-heptane was quickly added (2 x 25 ml, Uvasol) and the flask put in the freezer (-25 o C).Nothing happened and the crystallisation could only be induced by shock on the next day and after further 24 hours the colourless crystals were filtered off on a sinter funnel (G3, 1 inch wide) and washed with nheptane (3 x 20 ml, Uvasol).The collected crystals were dried in air for 24 hours and in vacuum desiccator (6h, 75 o C, 2 mbar).Very faint yellow crystals (130 mg).The compound is known (J.The reaction and purification was carried out without protection from oxygen and water from air.In a one necked-flask (250 ml) was put 1,2-di(2-pyridyl)ethanedione (2,2'-Pyridil, 2.12 g, 10 mmol, 1 equivalent); 1,2-phenylenediamine (1.08 g; 10 mmol, 1 equivalent) and 2-propanol (any grade, 80 ml).The resulting suspension was vigorously refluxed with stirring under condenser for 1 hour.The initial suspension slowly formed brown solution.The monitoring of the reaction was done by TLC analysis using ethyl acetate as eluent.After the condensation was practically completed, the flask was cooled below 50 o C and put in the freezer (-25 o C) for 24 hours.On the next day the obtained crystals were filtered off on a sinter funnel (G3, 1.5 inch wide) and washed with hexanes (3 x 30 ml) and dried in air for 24 hours to yield white crystals with brown tan.(2.54 g, 89%).The mother liquor was rejected.
The product at this point had good NMR characteristics but it was unsuitable for the photophysical studies.265 mg of crystals were sublimed under vacuum (180 o C, 10 -3 Torr, 4 hours) in a sublimation apparatus and white-yellowish crystalline solid was obtained (215 mg).Everything was dissolved in boiling 2-propanol (15 ml, Uvasol) and n-heptane was quickly added (25 ml, Uvasol) and the flask put in the freezer (-25 o C).After 24 hours the nearly colourless crystals were filtered off on a sinter funnel (G3, 1 inch wide) and washed with n-heptane (3 x 20 ml, Uvasol).The collected crystals were dried in air for 24 hours and in vacuum desiccator (6h, 75 o C, 2 mbar).White crystals (200 mg).The compound is known (J.

Experimental methods
Spectroscopic grade solvents from Sigma-Aldrich were used as received.
Optical measurements were performed at room temperature with dilute solutions in standard quartz cells (10 × 10 mm).Solutions were N2 purged and stoppers were wrapped with laboratory film to slow down the oxygen penetration.Absorption spectra at room temperature (21 o C) were recorded using a PerkinElmer Lambda 35 spectrophotometer.Emission spectra were obtained using a FLS 1000 of Edinburgh Instruments spectrofluorometer.Fluorescence kinetics studies were performed using the time correlated single photon counting technique.A mode-locked Coherent Mira-HP picosecond laser pumped by a Verdi 18 laser was used for excitation.The fundamental pulses of the Mira laser (tunable within 760 -800 nm) were upconverted to ~390 nm.The temporal width of the excitation pulses was ~280 fs and the instrument response function (IRF) about 40 ps.For shorter wavelength excitation Delta diode laser 336 nm was used.Fluorescence was dispersed with a 0.25 m Jarrell-Ash monochromator and detected with a HMP-100-07 hybrid detector coupled to an SPC-150 PC module, (Becker&Hickl GmbH).Fluorescence decays were analyzed with deconvolution software using a nonlinear least squares procedure with the Marquardt algorithm.[1] A standard 2 test and Durbin-Watson (DW) parameter along with residual and autocorrelation function plots were used to assess the quality of a fit.The estimated accuracy for the determination of decay time was below 10 ps.
For EPR spin measurements the samples were placed in suprasil EPR tubes.A EMXplus EPR Bruker CW X-band spectrometer equipped with an ER 4131VT nitrogen cryostat was used to measure spectra at 100 K.The samples were degassed with freezepumpthaw method and sealed to block the molecular oxygen.
1H FT-NMR spectra were recorded using Bruker AVANCE 500 MHz spectrometer.The chemical shifts are given with the residual solvent peak of methanol-d 4 .The sample was a saturated solution of HATN which was further transferred to NMR tube and efficiently purged with argon.
Mass spectrometry analyses were performed using Ultra-Performance Liquid Chromatograph ACQUITY UPLC I-Class (Waters) coupled with Synapt G2-S mass spectrometer (Waters) equipped with the electrospray ion source and quadrupole-Time-of-flight mass analyzer.The resolving power of the TOF analyzer was set to 20000 FWHM.The instrument was controlled and recorded data were processed using the MassLynx V4.1 software package (Waters).Both the positive and the negative ion modes were used in mass spectrometry.The measurements in positive mode were performed with capillary voltage set to 3.00 kV.The desolvation gas flow was 700 L/h and temperature 300 °C.The sampling cone voltage and source offset were set to 20 V and the source temperature was 120 °C.The measurements in negative ion mode were performed with capillary voltage set to 3.00 kV.The desolvation gas flow was 700 L/h and temperature 300 °C.The sampling cone voltage and source offset were set to 20 V, and the source temperature was 120 °C.Samples were dissolved in methanol, degassed and injected directly into the electrospray ion source.Methanol was used as a mobile phase with the flow rate 100 μl/min.The instrument worked with external calibration of sodium formate in the mass range of m/z = 50-1200.The Leucine-Enkephaline solution was used as the Lock-Spray reference material.The lock spray spectrum of the leucine-enkephalin was generated by the lock spray source and correction was done for every spectrum.The exact mass measurements for all peaks were performed within 3 mDa mass error.
Cyclic voltammetry measurements were performed using an SP-50 Potentiostat (BioLogic).All operations involving the electrolyte preparation and electrochemical measurements were carried out under dry argon atmosphere.Cyclic voltammetry studies were carried out for the 5x10-4 M solutions in 0.1 M [NBu 4 ][PF 6 ] (tetra(n-butyl)ammonium hexafluorophosphate) in appropriate solvent (dichloromethane, acetonitrile, methanol or their mixtures).The choice of the solvent was dictated by the solubility of the studied compound.The electrolyte was placed in a one-compartment, three electrode electrochemical cell with a glassy carbon disk working electrode of the surface area of 3 mm2, a platinum wire counter electrode and an silver wire pseudo-reference electrode reference electrode.To make the results comparable (using of the different solvents and pseudo-reference electrode) the potential of the Fc/Fc+ redox couple (internal reference)was checked at the end of each experiment and the obtained voltammograms were shifted to set zero at potential for the Fc/Fc + redox couple.
The rough evaluation of the polar solute -polar solvent interaction influence on the barrier of the PCET reaction can be straightforwardly carried out within the continuous dielectric medium and point dipole approximation.The polar molecule with dipole moment  and radius a induces in the continuous dielectric medium of dielectric constant the electric reaction field R which is proportional to the dipole moment  and solvent polarizability f (the polarizability of the solute is neglected) [2]: The f is a result of solvent's electronic polarizability f(n) = (n 2 -1)/(2n 2 +1) and dielectric polarizability f() = (-1)/(2+1).The reaction field is a sum of terms corresponding to the electronic and orientation polarizations [2]: where f(,n) = f() -f(n).The energy of the dipole moment  in reaction field R is: In a solvent the excited state energy E() is the calculated (for gas phase) value E v decreased by the E d : With this formula one may calculate the energy of the excited state in a polar solvent of a given dielectric constant at two OH distances and check how much the energy barrier between them changes.The radius of Onsager cavity for PQPhen was assumed a = 0.6 nm, dielectric constant of methanol =32.12 Å) = 13.21D at point of the energy minimum and at the charge transfer point in each state respectively were taken from calculations for the gas phase.With formula (3) one gets for the assumed parameters a bathochromic shift 0.0265 and 0.0281 eV for the S 1 state and 0.451 and 0.481 eV the S 2 state at minimum energy and PCET occurrence points respectively.The difference between them, 1.6 and 30 meV, gives the decrease of the barriers in the S 1 and S 2 respectively.This result is obtained for a nonpolarizable molecular dipole.It is known that usually the reaction field increases the dipole moment of the polarizable point dipole by 20 -50 % [3].However, even taking into account such correction, gas phase barriers are still substantially higher than the calculated change of the solvatochromic shift.

Theoretical methods
The ground-state equilibrium geometries of the compounds considered herein were determined with the second-order Møller-Plesset (MP2) method. 1 Vertical electronic excitation energies of singlet and triplet excited states were calculated with the algebraic-diagrammatic construction of second order (ADC(2)) method. 2 ADC(2) is a computationally efficient single-reference propagator method which generally yields similar results as the CC2 method 3 , and more accurate methods, such as CASPT2 or CC3. 4,5The correlationconsistent polarized valence-split double-ζ basis set (cc-pVDZ) 6 was employed in all calculations.The calculations were performed with the TURBOMOLE program package (V7.4) 7 making use of the resolution-of-the-identity (RI) approximation. 8), the vertical energy, E fl , and the oscillator strength, f, of the S 1 -S 0 fluorescence of the lowest-energy conformers of hydrogen bonded complexes of PQPhen, PQ2Py and Q2Py with methanol molecule determined with the ADC(2)/cc-pVDZ method at the MP2/cc-pVDZ equilibrium geometry of the ground state.Energies are in electronvolts.

System
PQPhen- Table S3.Vertical absorption energy (∆E), oscillator strength (f), dipole moment (µ), leading electronic configurations and relevant orbitals of the lowest excited states of the lowest-energy hydrogen-bonded complexes of PQPhen, PQ2Py and Q2Py with methanol molecule determined with the ADC(2)/cc-pVDZ method at the MP2/cc-pVDZ equilibrium geometry of the ground state.-0.946 eV Table S4.Wavelength of the i th absorption band maximum,  i , and molar absorption coefficient at that wavelength,  i , and Stokes shift (SS) for PQPhen, PQ2Py and Q2Py in solvents.

Figure S3 .
Figure S3.Normalized fluorescence excitation spectra (blue) and fluorescence spectra (green and red lines) of PQPhen (top), PQ2Py (middle) and Q2Py (bottom panel) in powder compared to absorption (black solid line) and fluorescence in n-hexane (black dot-dashed line).Legend specifies wavelength of observation for fluorescence excitation spectra and wavelength of excitation for fluorescence spectra in powder.

Figure S4 .
Figure S4.Effect of 375 nm photoirradiation on absorption spectra of PQ2Py in methanol.The legend specifies time of irradiation in minutes.

Figure S5 .
Figure S5.Effect of 375 nm photo-irradiation on fluorescence spectra of PQ2Py in methanol.Spectra obtained with 310 nm excitation.The legend specifies time of irradiation in minutes.Arrow visualizes the increase of fluorescence intensity with time of the irradiation.

Figure S6 .
Figure S6.Effect of 277 nm photo-irradiation on fluorescence spectra of Q2Py in methanol.Spectra obtained with 310 nm excitation.The legend specifies time of irradiation in minutes.Arrow visualizes the increase of fluorescence intensity with time of the irradiation.

Figure S7 .
Figure S7.Absorption and fluorescence excitation spectra (observed at 365 and 550 nm) of Q2Py in methanol recorded after 25 minutes of irradiation at 277 nm.

Figure S8 .
Figure S8.Fluorescence spectra (excitation at 290 and 370 nm) of Q2Py in methanol recorded after 25 minutes of irradiation at 277 nm.

Figure S17 .
Figure S17.The extended mass spectrum of the sample PQ2Py before laser irradiation and elemental prediction tables for the m/z = 286.11[PQ2Py+H] + and for the m/z = 308.09[PQ2Py+Na] + .

Figure S21 .
Figure S21.The mass spectrum of the sample PQ2Py after 375 nm laser irradiation plus additional 15 min of rest time and elemental prediction table for the m/z = 209.08[(PQ2Py-Pyr)+H] + .(The minus sign means the loss of the pyridine ring.)

Figure S23 .
Figure S23.The extended mass spectrum of the sample Q2Py in methanol before 277 nm LED irradiation and elemental prediction tables for the m/z = 285.12[Q2Py+H] + and for the m/z = 307.10[Q2Py+Na] + .

Figure S24 .
Figure S24.The extended mass spectrum of the sample Q2Py in methanol after 277 nm LED irradiation and elemental prediction tables for the m/z = 317.14[(Q2Py+CH 3 O)+H] + , m/z=338.34 and the m/z = 360.32(not assigned).

Figure S25 .
Figure S25.The extended mass spectrum of the sample Q2Py in methanol after 277 nm LED irradiation and elemental prediction table for the m/z = 219.08(not assigned).

Figure S26 .
Figure S26.The extended mass spectrum of the sample Q2Py in methanol after 277 nm LED irradiation and elemental prediction tables for the m/z = 256.09and the m/z = 258.10(not assigned).

Table S5 .
The proposed structure of some photoproducts found for PQPhen irradiated in methanol.

Table S6 .
The proposed structure of some photoproducts found for PQ2Py irradiated in methanol.