Supersonic Jet Spectroscopy and Density ... - ACS Publications

Sep 21, 2016 - Michał Kijak, Sebastian Peukert‡, Ephriem Mengesha†, Jerzy Sepioł, and Michał Gil. Institute of ... Tel: +48223433074. Cite this...
1 downloads 0 Views 965KB Size
Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Supersonic Jet Spectroscopy and DFT Study of Isomeric Diazines: 1,4- and 1,8-Diazatriphenylene. Why Do They Differ so Deeply? Micha# Kijak, Sebastian L. Peukert, Ephriem Tadesse Mengesha, Jerzy Sepiol, and Micha# Gil J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b06475 • Publication Date (Web): 21 Sep 2016 Downloaded from http://pubs.acs.org on September 27, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Supersonic Jet Spectroscopy and DFT Study of Isomeric Diazines: 1,4- and 1,8-Diazatriphenylene. Why Do They Differ so Deeply?

Michał Kijak, Sebastian Peukert, Ephriem Mengesha,# Jerzy Sepioł, and Michał Gil*

Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, PL-01-224 Warsaw, Poland.

*Corresponding Author, email: [email protected]; tel.: +48223433074 #

Present Address: CEA, CNRS, IRAMIS/LIDyL/Laboratoire Francis Perrin URA2453, CEA Saclay, 91191 Gif-sur-Yvette, France

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 41

ABSTRACT We report on laser induced fluorescence excitation and dispersed fluorescence spectra of two isomeric compounds: 1,4- and 1,8-diazatriphenylene (1,4- and 1,8-DAT) isolated in supersonic molecular jets, and their 1:1 complexes with protic solvents. We found that the ground and excited state vibronic patterns of bare 1,4-DAT differ significantly from those of 1,8-DAT, and those of the complexes of both isomers. A marked activity of several out-ofplane vibrations in 1,4-DAT and the symptoms of the distortion of the S1 excited molecule were diagnosed from the vibronic spectra, while planar structures were predicted for 1,8-DAT in S0 and S1 states. An anharmonic double-minimum potential along an out-of-plane coordinate has been derived and used to predict higher overtones of the S1 state vibration at 113 cm-1. Large enhancement of fluorescence was observed upon formation of 1:1 complexes of 1,4-DAT with water or methanol, which is explained in terms of an increased separation of interacting (n,π*) and (π,π*) electronic states in the H-bonded complexes, and/or a suppression of intersystem crossing process.

2 ACS Paragon Plus Environment

Page 3 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

1. INTRODUCTION Polycyclic aza-aromatic compounds are under continuous interest of researchers because their physical and chemical properties make them a promising choice for various applications. They are part of an aromatic backbone of larger (supra)molecular systems, where they play different roles, such as chelating ligands for lanthanides and metal ions,1,2 sensitizing moieties for intracellular luminescent probes,1,3 structural cores of discotic liquid crystals,4-6 or electron-accepting units in OLED materials.7 Among this relatively broad family of aza-aromatics, 1,4-diazines with nitrogen atoms in para position, forming a pyrazine ring, represent an interesting group of compounds. For example, 1,4diazanaphthalene (quinoxaline) is a parent structure for a range of molecules showing biological activity.8-14 In turn, a larger diazine, 1,4-diazatriphenylene (1,4-DAT) has been recently proposed as a core compound in new organic light emitting materials showing thermally activated delayed fluorescence, which can increase the quantum efficiency of fluorescent OLEDs without use of precious metals.7 Photophysical properties of nitrogen-heterocyclic molecules can be strongly affected by the presence of low-energy (n,π*) states located in close vicinity of (π,π*) ones, which results in efficient coupling of both states via vibronic interactions.15,16 This phenomenon (also called proximity effect) crucially depends on the energy gap between both interacting states. It can affect non-radiative decay rates of such molecules, and may be manifested by a geometric distortion of the lowest excited state (pseudo-Jahn-Teller, pJT, effect). As a result, luminescence (fluorescence or phosphorescence) properties of the molecule can be modified by the nature of surrounding environment, temperature, chemical substitution, or deuteration.15 Spectroscopy in supersonic jets proved to be especially useful in studying the photobehavior of such systems. For example, vibrational structures of isoquinoline and its methanol complexes have been investigated by high-resolution spectral and time-resolved 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 41

techniques in the gas phase showing an important role of out-of-plane (oop) vibrations in (n,π*) – (π,π*) interactions.17 Recently, a dramatic dependence of the fluorescence lifetime of 2-aminopurine on the site-selective hydration has been shown and correlated with an ordering and spacing of (n,π*) and (π,π*) singlet states.18 Detailed calculations of the excited-state energy surfaces using the ADC(2) method have demonstrated that the interaction of water with the amino group affects the energy barrier on the pathway from the S1 minimum to the conical intersection.19 1,4-diazines were reported in literature as systems showing unusual behavior in the excited electronic states, both in solution20-25 and when surrounded by rigid, apolar environments.16,26-28 The studies in frozen hydrocarbon matrices of Shpolskii-type have shown spectral signatures of strong coupling between adjacent excited states through an oop C-H bending vibrational mode. Other related di- and mono-aza compounds without a pyrazine subunit did not exhibit these anomalous features in high-resolution fluorescence spectra.29,30 Fluorescence and phosphorescence polarization studies of 1,4-DAT in 3methylpentene glass at 77 K have indicated a non-planar distortion of the lowest (π,π*) singlet state, in contrast to a planar geometry of the lowest triplet state.31 In solutions, it was shown that, contrary to isomeric structures, the basicity of both nitrogen atoms increases by several pK units in both lowest excited states, S1 and T1.21-23 This unusual behavior was interpreted as a result of the distortion of the protonated pyrazine ring upon excitation and the change of hybridization of nitrogen atoms from sp2 towards sp3 (partial pyramidalization), caused by increased electron density on the nitrogen atoms.23 The same group of molecules, mostly 1,4DAT, has been studied in ethanol solution from room temperature to the rigid solvent.25 The S1→S0 emission in conditions of fully relaxed environment revealed spectral anomalies: no “mirror image” symmetry in respect to absorption, a large Stokes shift, no vibrational structure and similarity to the cation emission spectrum. The anomalous emission band

4 ACS Paragon Plus Environment

Page 5 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

decreases upon lowering the temperature, and at ~120 K the fluorescence closely resembles that of 1,8-diazatriphenylene (1,8-DAT), an isomer without pyrazine ring. Due to a different stabilization of the (n,π*) and (π,π*) energy levels in the presence of a protic solvent, such photobehavior cannot be explained by pJT effect observed in non-polar environments. A cation-like structure has been proposed as a source of anomalous fluorescence.25 The main question to be answered is the following: Can the cold molecules isolated in molecular jets develop the complicated “chemistry” of distorted system? Thus, we studied the vibrational structure of 1,4-DAT by means of laser induced fluorescence excitation (LIF) and dispersed fluorescence (DF) spectroscopies, and compared it with that of the 1,8-DAT isomer used as a reference compound. We attempt to show how complexation with selected protic partners affects interactions between low-lying excited states. Our results show significant spectral differences observed for isolated molecules of 1,4- and 1,8-DAT, which diminish in the case of hydrogen-bonded complexes of both isomers. Analysis of the experimental spectra is supported by quantum chemical modeling, which helps in attribution of observed vibrations, suggests the type of distortion responsible for the “anomalous” behavior of 1,4DAT in the gas phase and provides insight into the energetics of bare diazines, as well as their complexes with protic solvents.

2. EXPERIMENTAL SECTION 1,4-diazatriphenylene (phenanthro[9,10-b]pyrazine) was synthesized according to the literature (for description, see Supporting Information).32 The jet experiments were performed with the home made setup described previously.33-35 The sample was heated up to 440 - 470 K, diluted by 3.5 atm of helium, and expanded to a vacuum chamber through a home-made pulsed valve (based on IOTA Series 9, General Valve) with a nozzle of 500 µm diameter. The pressure of the solvent vapor in 5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 41

helium, which produces complexes, was controlled by the temperature of the container attached to the supplying gas line. Excitation source was a narrow band (< 0.1 cm-1) laser based on an optical parametric oscillator (OPO, Sunlite Ex, Continuum) pumped by a seeded Nd:YAG laser (Powerlite 8000, repetition rate 10 Hz), or home-made dye laser (LDS698 dye, without etalon) having spectral half-width of ~0.6 cm−1.34 The tunable radiation from OPO or from the dye-laser was converted to the desired UV range by doubling it in KDP or BBO crystals, respectively. The final energy of excitation beam was in 0.1 – 1 mJ range. Fluorescence from the sample was collected by a toroidal mirror and directed either to a Hamamatsu R2949 photomultiplier connected to a Yokogawa DL9140 oscilloscope for measuring LIF spectra, or to a slit of SpectraPro 275 spectrograph (Acton Research) having a reciprocal linear dispersion of 3.0 nm/mm (with 1200 groove/mm grating) and equipped with a Princeton Research LN-cooled CCD to measure dispersed fluorescence (DF). Resolution of DF spectra was limited to ~50 cm-1 by the slit size of 0.2 mm used in the experiments. The molecular modeling of the ground and electronically excited states of studied chromophores and their complexes with water and methanol was performed using density functional theory (DFT) and its time-dependent (TD-DFT) formalism, respectively. The B3LYP and the BHLYP hybrid functionals, and the 6-31+G(d,p) basis set were employed. The basis set superposition error (BSSE) in the complexation energy was accounted for using the counterpoise method. Harmonic vibrational frequencies were calculated for all optimized structures. Vibrational frequency scaling factors 0.964 and 0.929 were applied for B3LYP and BHLYP functionals, respectively. The vibrationally-resolved absorption and emission spectra were simulated in the framework of the Franck-Condon (FC) principle including the Duschinsky rotation and within harmonic approximation.36 The half-widths of bands, 2 and 30 cm-1, for LIF and DF, respectively, have been arbitrarily chosen to match those of the experimental ones. The conventional linear-response Polarizable Continuum Model (PCM)

6 ACS Paragon Plus Environment

Page 7 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

was used to simulate the effect of a dielectric medium on the excited states separation and FC spectra. Partial atom charges were fitted to an electrostatic potential according to the MerzSingh-Kollman scheme.37 All computations were carried out within the Gaussian 09 suite of programs.38

3. RESULTS AND DISCUSSION

3.1 Electronic Singlet States and Geometry Optimizations The absorption spectrum of 1,4-DAT in n-hexane solution has its lowest energy maximum at 28250 cm-1 (see Figure S1 in Supporting Information). In more polar ethanol solution the lowest band has been reported to be at 28092 cm-1,20 while the spectral origin in cryogenic Shpolskii matrix is located at 27941 cm-1.26 In case of 1,8-DAT, the first absorption band has its intensity maximum at 29390, 29250, and 29429 cm-1, in n-hexane, ethanol,21 and Shpolskii matrix,29 respectively. Order and spacing of low-lying excited electronic states in azaaromatics have large impact on their mutual interactions.15 Table 1 gives energies and oscillator strengths of vertical excitations to the three lowest excited states of 1,4-DAT, while Table 2 presents data for two low-energy (π,π*) states (S1, S2) and the lowest (n,π*) state (S5) of 1,8-DAT, both calculated at the TD-DFT level. For both compounds, energy of the first (π,π*) transition calculated with the B3LYP functional seems to be overestimated by more than 2000 cm-1 when compared to the lowest energy maximum in the absorption spectra in n-hexane. However, due to the highly structured experimental absorption spectrum, its first maximum should be rather compared to the adiabatic excitation energy. For 1,8-DAT, the S1 state optimization gives us the energy of the 0-0 transition of 29520 cm-1, in perfect agreement with the experiment. It shows that the DFT method with B3LYP functional is a good choice for

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 41

description of our systems. Unfortunately, we found that the order of low-lying states calculated for 1,4-DAT with the B3LYP functional is affected by a well-known tendency of TD-DFT to underestimate energies of charge-transfer (CT) states.39 The presence of such a low lying state of partial charge transfer character was predicted for both compounds (see Figure S2 for orbital picture). This problem can be partly cured by employing the hybrid BHLYP functional which comprises a higher fraction of the non-local Hartree-Fock exchange potential. It has also been shown that the BHLYP functional gives a more balanced description of (n,π*) and (π,π*) states.40 Since the relative position and separation of the lowlying exited states are crucial in the discussion of the emissive properties, we had to use the BHLYP functional for all further excited state calculations, even if the exact values of the energies of vertical transitions are systematically overestimated (by ~3500 cm-1 in comparison to B3LYP values; Table 1 and 2, Figure S1). Figure 1 contains the graphical presentation of relative energies of vertically excited electronic singlet states for both studied isomers of DAT, as well as relative energies of their 1:1 complexes with water. For bare 1,4-DAT, the lowest electronic excited singlet state is of (π,π*) character, followed by a close-lying S2 state of (n,π*) character located 250 cm-1 above. The very small energy gap calculated for these two states is comparable with a previous estimation of ~160 cm-1, obtained using semi-empirical methods.26 The studies in frozen alkane matrices concluded that the (π,π*) state is the lowest one, and that a weak fluorescence of the isolated 1,4-DAT may be due to an efficient intersystem crossing involving a lowenergy (n,π*) triplet state.26 In contrast, in bare 1,8-DAT, the energy of the lowest excited singlet state of (n,π*) character is calculated at ~5600 cm-1 above the lowest (π,π*) state (Table 2). Previously, a significant mutual interaction between electronic states of different orbital origin in aza-aromatic compounds has been considered for energy gaps smaller than ~3000 cm-1.41 Therefore, a strong mixing of the lowest (π,π*) and (n,π*) states may occur for

8 ACS Paragon Plus Environment

Page 9 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

isolated 1,4-DAT, as it was suggested in the case of condensed phase studies. Such effects are not expected for 1,8-DAT, where a large energy separation is obtained regardless of the DFT functional used. Figure 2 shows geometries of bare 1,4- and 1,8-DAT and their 1:1 complexes with water calculated for the S1 state using the BHLYP functional (for the ground state geometries, see Figure S3). As a first step, we performed an excited-state geometry optimization of the lowest (π,π*) and (n,π*) states of 1,4-DAT with a constrained C2v symmetry of the ground state. Such obtained pure A1 (π,π*) and B1 (n,π*) states have adiabatic energies 32820 and 32590 cm-1, respectively. Both structures occurred to be saddle points on the potential energy surface with one imaginary frequency corresponding to out-of plane (oop) pyrazine ring deformation. Subsequent optimization of both states done upon lowering the symmetry constraint led to the same true minimum structure of Cs symmetry with adiabatic energy of 32300 cm-1, which is lower by 520 and 290 cm-1 than those of the C2v minima. The obtained S1 structure shows slight deviation from planarity mostly in the pyrazine ring. It is bent along the N-N axis (characterized by dihedral angle α = 176°, Fig. 2) and it deviates from the planar framework of the remaining molecule (angle β = 172°). The source of distortion lies in near degeneracy of its (π,π*) and (n,π*) states. Deformation from planarity, which allows configurational mixing, may lead to an energy lowering for some state (pseudo-Jahn–Teller effect). Indeed, analysis of the nature of the fully optimized S1 state of bare 1,4-DAT shows its almost half-and-half (π,π*) and (n,π*) orbital character. A similar phenomenon has been invoked to explain striking differences in the experimentally observed vibronic spectra of 9Hadenine and its 9-acetylo derivative.42 There exist two symmetry-equivalent nonplanar S1 minima whose structures are related by inversion. The consequences for the interpretation of the vibronic spectra will be discussed in section 3.2. Any attempts to find the non-planar minimum-energy geometry for the second of the interacting states failed. The same lowest-

9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 41

energy excited state structure was obtained, probably due to the proximity and strong mixing of states. In contrast, the geometry optimizations of 1,8-DAT show planar structures for both, S0 and S1 states (Fig. 2 and Fig. S3). Due to very large energy separation the mixing of (π,π*) and (n,π*) states is not efficient. 3.2 LIF and DF Spectroscopy of Bare Compounds Figure 3 shows the fluorescence excitation (LIF) spectra of jet-cooled bare 1,8-DAT and its 1:1 complex with methanol, and theoretically predicted spectrum of bare compound based on the calculated harmonic frequencies and FC approximation. The LIF spectrum of bare 1,8-DAT (Fig. 3A) shows the 0-0 band of the jet-isolated molecule at 29946 cm-1. The intensity of this band is substantially larger than those of other lines in the spectrum. The transitions in the low frequency region (100 – 350 cm-1) are very weak, while higher intensity bands are observed between 400 and 700 cm-1. A similar picture emerges from the absorption spectrum of 1,8-DAT frozen in n-hexane, where weak transitions at 130, 230, and 414 cm-1 were followed by more intense ones found at 660 cm-1 and above.29 Comparison of experimental LIF with the spectrum calculated in harmonic approximation (Fig. 3A top, see Table S1 for description of bands in the simulated spectrum) shows a qualitative agreement between both spectral patterns. The fully symmetrical vibrations of 1,8-DAT in the S1 state are calculated at 243, 397, 405, 585, and 649 cm-1 (see Figure S4 for shapes of vibrational modes) and are assigned to the experimentally observed bands at 228, 403, 412, 585, and 656 cm-1. However, the predicted activity of 243 cm-1 transition is substantially lower than that of the experimental one at 228 cm-1. The band at 119 cm-1 above the origin can be due to an overtone of an oop vibration of a2 symmetry calculated at 2 x 51 cm-1, with predicted relatively large intensity. For most of other experimental bands observed up to 500 cm-1, we can find counterparts in the simulated spectrum coming from overtones and combinations of the non-totally symmetric modes. However, above 500 cm-1 the experimental LIF spectrum

10 ACS Paragon Plus Environment

Page 11 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

contains quite intense bands (e.g. 571, 593, 651, 744, 754, and 786 cm-1), which are not reproduced by the FC simulation. Especially, the band at 754 cm-1 has large intensity in the experimental spectrum, while in the region of 650 – 1000 cm-1 the FC simulation predicts a presence of only very weak bands (Table S1). One has to notice, that the calculations predict the presence of a second optically allowed electronic B2 (π,π*) state, located close to the lowest one (~800 cm-1 above, Fig. 1), which may participate to the observed LIF spectrum or may induce activity of otherwise forbidden modes. Figure 4 shows the LIF spectra of bare 1,4-DAT and its complexes with water and methanol, as well as theoretically predicted spectrum of the monomer. For the isolated 1,4DAT molecule (Fig. 4A), the origin of the LIF spectrum is at 28736 cm-1. The spectrum shows very different pattern of active bands in comparison to that of 1,8-DAT. The 0-0 transition has lower intensity than the bands in the low-frequency region, but above ~500 cm-1 the intensity of the LIF spectrum dramatically drops down. The latter observation may be attributed to enhanced intramolecular vibrational-energy redistribution, or to an onset of an efficient non-radiative channel. Several bands observed in the LIF spectrum of 1,4-DAT in the 0 – 500 cm-1 region are not present in LIF of the parent molecule, triphenylene (TPH).43 Two very pronounced vibrations at 113 and 209 cm-1 are the most intense bands in the spectrum, and contrary to the situation in 1,8-DAT, several bands of medium intensity are present in the region between 250 and 400 cm-1. Harmonic frequency calculations predict five fundamental modes below 220 cm-1, and all of them are of oop type with respect to the molecular framework of 1,4-DAT. Both, for planar (C2v symmetry) and distorted (Cs symmetry) geometry in the excited state, the first in-plane (ip) mode is calculated at ~220 cm1

. A comparable mode is calculated in 1,8-DAT at 219 cm-1 but it has very weak intensity in

the simulated spectrum (it is of b2 symmetry). Thus, the two very strong bands observed in LIF must be due to oop vibrations. Their large intensity supports the non-planar geometry of

11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 41

the S1 state predicted by calculations (Fig. 2). An intensity borrowing mechanism within the planar structure can be excluded, as the next predicted excited state (Table 1) has much lower oscillator strength. The oop distortion implies existence of two symmetry-equivalent nonplanar S1 minima (C2v(M) molecular symmetry group). As a result, a symmetric doubleminimum potential should be present along at least one of the oop bending coordinates. Therefore, we attempted to analyze the LIF spectrum in terms of different choice of doubleminimum potential along possible oop modes. A mixed symmetric quartic-harmonic potential of the form V = z4 + Bz2 (in reduced coordinates) has been used.44 Such form of the potential has been successfully applied for ring puckering modes active in the vibronic spectrum of many systems.44-46 The existence of two symmetry equivalent distorted structures implies that only even levels (∆ν = 0, 2, 4,…) of such vibrational modes are active in the excitation from a planar ground state. In case of high inversion barrier, pairs of these levels are practically degenerate and spacing of low-energy levels is well approximated by harmonic frequency calculations within one of two minima. The “even level” selection rule applies in case of all non-totally symmetric modes in C2v(M) symmetry once we neglect Duschinsky rotation between modes. When the deformation conserves a symmetry plane perpendicular to the molecular framework this rule is strict for a2 and b2 modes (asymmetric in respect to that plane) in a harmonic approximation, even when Duschinsky rotation is present. However, for a1 and b1 modes it is only approximate since in general they can be of mixed ip and oop character. For example, molecular modeling shows that the two lowest oop a’ modes in the bent S1 state of 1,4-DAT can be reasonably described as b1 in C2v(M) symmetry. However, other a’ modes are, up to a different degree, of mixed ip (a1) and oop (b1) character, and they can be active without C2v(M) symmetry restrictions. Our harmonic calculations show three oop modes of a’ symmetry in the region below 220 cm-1. Two of them (calculated at 106 and 204 cm-1, Figure S4) correspond to the deformation of the S1 state presented in Figure 2. The

12 ACS Paragon Plus Environment

Page 13 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

vibration at 106 cm-1 corresponds to a butterfly type mode affecting to large extent a deviation of the pyrazine ring from the general plane of the framework (angle β in Fig. 2). The vibration at 204 cm-1 represents a pyrazine bending mode acting along the N-N axis (angle α, Fig. 2). Both modes are active in the calculated LIF spectrum (Fig. 4A), especially that at 106 cm-1. Thus, we chose the two strongest bands in the experimental LIF spectrum (at 113 and 209 cm1

) for simulation of the double-minimum potential, treating them as a progression of the same

mode. Two alternative potentials corresponding to two different values of the parameter B in the equation are possible (Figure 5). The predicted barrier heights for them are 320 and 205 cm-1, respectively, and are comparable to the inversion barrier calculated by TD-DFT method (290 cm-1). The potential with a larger barrier (Fig. 5A) predicts four additional even overtones at 276, 361, 476, and 608 cm-1. Their frequencies match nicely with the bands observed in LIF at 278, 361, 481, and 605 cm-1. In case of the potential with a lower barrier (Fig. 5B), two additional bands are predicted in analyzed spectral region (at 362 and 551 cm1

), which also agree well with the experimental bands at 361 and 561cm-1. Based on a

similarity of the calculated barriers (320 vs. 290 cm-1) and better matching of the predicted overtones with the bands of experimental LIF, we suggest the potential with higher barrier (Fig. 5A) as a preferred one. Still, the assignment of the remaining bands in the LIF spectrum is difficult. With a few exceptions, the cluster of experimental bands observed above 250 cm-1 has no clear correspondence in the harmonic FC simulation. We can assign a medium intensity band observed at 397 cm-1 to an ip mode calculated at 385 cm-1. Assuming a harmonic potential along other vibrational modes, we can only propose the following candidates for experimental bands at 268, 272, 291, 329, and 349 cm-1: ip a’ mode calculated at 231 cm-1, oop a’ + oop a’ at 292 (88+204) cm-1, 2 x oop a” at 269 (2 x 135) cm-1, and ip a’ at 395 cm-1. Additionally, there is the above mentioned oop a’ bending mode calculated at 204 cm-1. FC simulation predicts its activity in the spectrum, and this mode may be also

13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 41

strongly anharmonic. Thus, even allowing rather large tolerance for a frequency matching between the experiment and the harmonic frequency calculations, we are unable to assign the strong band at 329 cm-1. Moreover, a large intensity progression of the vibration at 113 cm-1 does not build up on any of those bands. We suggest that the excited-state deformation, analyzed along the 113 cm-1 vibrational mode, may be relevant for additional modes, for example, the pyrazine bending mode at 204 cm-1. However, in contrast to the mode at 113 cm1

, the experimental band at 203 cm-1 has weak intensity, and it is not clear which second band

should be chosen for the fitting of double minimum potential. These modes may be anharmonically coupled and their analysis would require two- or even multidimensional treatment. Such approach has been applied for analysis of ring-twisting and ring-bending modes of 1,4-benzodioxan in its S0 and S1 states.47 In our case it was not possible as the activity of vibronic bands in LIF diminishes above 500 cm-1. Figure 6 shows the experimental dispersed fluorescence (DF) spectra of bare 1,8-DAT and its 1:1 complex with methanol excited at the corresponding (0-0) transitions, as well as their theoretically predicted spectra using harmonic approximation. The spectrum of bare 1,8DAT excited at 29946 cm-1 (Fig. 6A) shows an intense resonance transition and a low number of active vibrations. Medium intensity bands at 622 and 698 cm-1 correspond reasonably well to ip modes observed in the LIF spectrum at 585 and 656 cm-1. In the simulated DF spectrum they are predicted at 608 and 678 cm-1, and they can be described as breathing vibrations of aromatic rings. A weak band observed at 1042 cm-1, as well as strong, overlapped vibrational bands with maxima around 1328 and 1411 cm-1 are in good agreement with a1 modes predicted at 1048, 1308, 1322, 1386, and 1434 cm-1, which engage both, the framework and hydrogen atoms. In the LIF spectrum of 1,8-DAT the medium intensity bands were observed at ~230 and ~400 cm-1, but corresponding transitions are absent in the dispersed fluorescence. In DF simulations, they are predicted to be very weak, and we suggest that the low signal-to-

14 ACS Paragon Plus Environment

Page 15 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

noise ratio of the experimental DF spectrum prevented their detection. These bands were observed in fluorescence of 1,8-DAT in Shpolskii matrix at 229 and 414 cm-1.29 Fluorescence bands observed in the DF spectrum of 1,8-DAT and assigned above to the fully symmetrical ip modes were also observed in the jet-cooled spectrum of parent triphenylene, indicating weak impact of heteroatoms on the spectrally active modes.43 Figure 7 shows the DF spectra of bare 1,4-DAT and its 1:1 complexes with water and methanol upon excitation at their (0-0) transitions. The spectrum of bare molecule (Fig. 7A) has much richer structure and very different intensity pattern, when compared to that observed for 1,8-isomer. Due to the existence of two symmetry equivalent S1 structures and the C2v(M)→C2v selection rules, the transitions to odd levels of b1 modes are forbidden, but their combinations and even levels are still allowed. In the region up to 500 cm-1, where 1,8-DAT shows no detectable fluorescence, two strong bands at 230 and 408 cm-1 are present. Based on the analysis of S1 vibrations observed in LIF, we assign them to the first overtone of the calculated 119 cm-1 oop mode, and to a combination of oop 119 cm-1 and oop 292 cm-1 modes. Very strong band at 984 cm-1 is assigned to a combination of two oop modes calculated at 119 and 854 cm-1. In analogy to 1,8-DAT, the bands at 572, 664, and 1404 cm-1 can be assigned to ip skeletal modes. The DF spectrum shows additional medium intensity bands at 124, 814, 1146, and 1260 cm-1. For the first of them, the only candidate is the first overtone of the lowest frequency mode of a2 symmetry, calculated at 67 cm-1. The FC simulation (Fig. 7A, bottom) shows a rich pattern of active vibrations. The agreement of predicted intensities and experimentally observed bands is poor in many cases. This is due to the above mentioned C2v(M)→C2v selection rules. They are not taken into account in the simulation, which vastly overestimates the intensity of odd levels of b1 modes. The calculated transitions are mostly due to progressions of two oop modes at 119 and 292 cm-1 (modes 31 and 81, see also Table S1-A). They correspond to 106 and 204 cm-1 modes calculated for the

15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 41

excited state. The simulation predicts also activity of oop C-H bending mode of pyrazine hydrogen atoms at 854 cm-1 (mode 301 in Fig. 7A), which is observed as strong 30131 combination band at 984 cm-1. A previously reported fluorescence spectrum of 1,4-DAT in Shpolskii matrix shows four fundamental vibrations at 231, 290, 574 and 858 cm-1.26 The latter is the most intense band with a clear progression. Also, the phosphorescence spectrum of 1,4-DAT in argon matrix at 10K (not published) confirms the existence of an intense transition at 862 cm-1. This suggests that in the rigid environment of frozen matrices (n-alkane or argon) the corresponding two symmetries are no longer equivalent and one minimum of double-well potentials along oop bending modes becomes energetically favored. In these conditions, 81 (292 cm-1) and 301 (854 cm-1) modes are allowed and they are observed at 290 and 858 cm-1 in Ref. 26.

3.3 Spectroscopy and Modeling of H-bonded Complexes Previous studies of various diazines in condensed phase showed that hydrogen-bond interactions and/or protonation of the nitrogen atoms may strongly influence their photophysical behavior.17-19,23-25 Complementary information can be obtained in the gas phase by studying effects of complexation by small clusters of protic solvents. The LIF spectra of 1,8- and 1,4-DAT in the presence of water and/or methanol are presented in Fig. 3B and 4B, respectively. Solvent concentration in the carrier gas has been adjusted for a preferential formation of 1:1 DAT/solvent complexes. Frequencies of the main bands due to these complexes are labeled. One can notice a significant contribution of the vibronic bands from the bare DAT molecules to the spectra. The spectra recorded at higher pressures of solvent (> 0.1 kPa) revealed additional bands, most probably due to the fluorescence of complexes having different stoichiometry. This observation can be supported by modeling. It predicts that binding of the second water molecule to the first one, already involved in the 1:1 complex

16 ACS Paragon Plus Environment

Page 17 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(Figure S5), is almost twice more energetically favorable (BSSE and zero point vibrational energy corrected complexation energies of -5.6 and -5.7 kcal/mol, for 1,4- and 1,8-DAT, respectively) than the first complexation step (-2.9 and -3.3 kcal/mol). Interestingly, it strengthens the hydrogen bond present in the 1:1 system. This kind of cooperativity can be explained by a cyclic form of that most stable 1:2 complex with a nonconventional C-H···O hydrogen bond linking the pyrazine ring and the second water molecule (H···O distance of 230 and 233 pm, for 1,4- and 1,8-DAT, respectively, Fig. S5). An attachment of the second water molecule to the second nitrogen atom (not involved in the 1:1 complex) is almost perfectly additive, as it is characterized by an energy gain similar to that of the first complexation step (-2.6 and -3.0 kcal/mol, for 1,4- and 1,8-DAT, respectively). The origin of the LIF spectrum of 1:1 complexes of 1,8-DAT with methanol (Fig. 3B) is at 29837 cm-1, while that with water (not shown) is at 29857 cm-1. This corresponds to a red shift of 109 and 89 cm-1, respectively, when comparing with the 0-0 band of the bare molecule. In a low frequency region of the spectrum, two weak bands are observed at 51 and 57 cm-1. Vibrations having such low frequencies are not present in the LIF spectrum of bare 1,8-DAT. Thus, they are due to the presence of intermolecular H-bond with methanol. Strong vibronic bands are observed at 586, 659, and 752 cm-1 above the 0-0 band of the complex. These frequencies match almost exactly with the vibronic modes of the intense bands found in the region of 550 – 750 cm-1 of the bare molecule spectrum (Fig. 3A), and by analogy they are assigned to the skeletal ip modes. The 0-0 bands of the 1:1 complexes of 1,4-DAT (Fig. 4B) are at 28661 and 28625 cm1

, for water and methanol. In respect to the origin of the parent structure they are shifted to the

red by 75 and 111 cm-1 (for the 1:1 ethanol complex it is 148 cm-1). The spectra of both complexes resemble each other. In the region of low frequencies new bands are observed at 48 and 95 cm-1 for the water, and 13, 52, 54, 90, and 131 cm-1 for the methanol complex.

17 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 41

Based on the correspondence of the experimental and calculated frequencies, we assign these bands to the H-bond wagging and twisting motions. Noticeably, the activity of vibronic bands in the LIF spectrum of the complexes decreases even faster than it is observed for LIF of bare molecule (Fig. 4A). In particular, the very intense bands at 113 and 209 cm-1, which in bare molecule are due to the progression of the oop deforming mode, are not present in H-bonded structures. Diminished activity of this mode agrees with a larger energy gap between the lowest (π,π*) and (n,π*) states predicted in the complex of 1,4-DAT by TD-DFT (Fig. 1). Figure 6B shows the DF spectrum of the 1:1 complex of 1,8-DAT with methanol excited at the (0-0) transition, and its theoretically predicted one. The spectrum clearly resembles that of the bare compound. Some broadening of the resonance transition of the complex is observed, which is due to a contribution of low frequency H-bond modes. Similarly as in the case of bare 1,8-DAT, the calculated spectrum of the complex gives very good agreement with the experimental one. Figure 7B shows the DF spectra of the 1:1 complex of 1,4-DAT with water and with methanol excited at their (0-0) transitions (28661 and 28625 cm-1, respectively). Both spectra are very different from that of bare 1,4-DAT. They are significantly simplified revealing “clusters” of bands centered approximately at 700 and 1400 cm-1, and they clearly resemble the DF spectrum of 1,8-DAT (see Fig. 6). All spectrally active vibrations observed for the complexes above 200 cm-1 have their correspondence in the spectrum of parent TPH.43 Calculated structures of the 1:1 complexes in S0 (Fig. S3) show a practically planar 1,4-DAT framework. One can notice an asymmetric broadening of the resonance DF band of the 1,4DAT complexes in comparison to that of bare molecule (Fig. S7). This is most probably due to the low frequency modes related to H-bond wagging and twisting motions, which were also active in LIF. According to our calculations, several vibrations are predicted in 0 – 100 cm-1 region.

18 ACS Paragon Plus Environment

Page 19 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 8 shows a comparison of the experimental DF spectrum of the 1:1 complex of 1,4-DAT and water with the DF spectra calculated in vacuum, n-hexane and acetonitrile. The spectral changes observed in the experimental DF upon complexation are qualitatively reproduced by the calculations in vacuum. However, much more modes of significant activity are predicted by this simulation. For example, the bands at 118 and 282 cm-1, due to oop deformations of the pyrazine ring, have a significant contribution to the calculated DF spectrum, but are practically absent in the experimental one. The low activity of these modes in the experiment may be related to the selection rules of C2v(M) molecular symmetry group, as discussed above for the DF spectrum of bare 1,4-DAT. However, in our opinion an asymmetric attachment of the molecule of a protic partner via a hydrogen bond distorts double-minimum potential to such extent, that it became essentially a single minimum (anharmonic) potential. Inspection of the calculated S1 geometry of the 1:1 complex (Figure 2) shows a distortion of the pyrazine ring similar to that in the bare compound. Despite the increased separation of the (π,π*) and (n,π*) states which is predicted for the complex (vertical energy gap of 250 cm-1 in the monomer increases to 1260 cm-1 in the complex, Fig. 1), the coupling between them seems to be still efficient. Indeed, an orbital character analysis shows a mixed nature of the optimized S1 state in the complex, although the (π,π*) contribution is larger (~80 %) than in case of the bare molecule (50 %). In order to verify how a further increase of separation of the interacting states affects the S1 geometry and the Franck-Condon analysis, we performed the calculations for the complex in the presence of dipolar environment by means of the PCM model. We used four solvents of increasing polarity: n-hexane (relative permittivity, ε = 1.88), o-xylene (2.55), fluorobenzene (5.42) and acetonitrile (35.7). In a dipolar surrounding, the more polar (π,π*) state (Fig. S2) should be relatively stabilized in comparison to the (n,π*) state. Indeed, the position of the (n,π*) state remains almost constant (within 100 cm-1) despite the increase of polarity, while the (π,π*)

19 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 41

state is substantially stabilized by environment. The vertical energy gap between them increases from 1260 cm-1 in vacuum, through 1650, 1810 and 2140 cm-1, up to 2460 cm-1 in acetonitrile). Significantly, deformation of the pyrazine ring in the S1-optimized complex gradually decreases with growing polarity of the surrounding. In the most polar acetonitrile, there is almost no deformation upon excitation (Fig. 2 and S3). Small deviations from a perfect planarity in both S0 and S1 states are similar, and are due to a steric hindrance within the complex. They are comparable to those obtained for 1:1 water complex of 1,8-DAT. The coupling between the (π,π*) and (n,π*) states, which induces pJT distortion, decreases with the increase of separation between them. It is reflected in the nature of the optimized S1 state, which shows growing contribution of the (π,π*) transition. The above mentioned effects have large influence on the simulated DF spectrum of the 1,4-DAT complex (Fig. 8). It simplifies gradually with increasing polarity of the surrounding. Large activity of oop deformation modes of the pyrazine ring obtained in vacuum, systematically diminishes. The shape of spectrum better resembles that of the experimental one. Even for the least polar n-hexane, the simulation gives a satisfactory result. However, due to the limited resolution of our DF spectrum, quantitative comparison of the calculated and experimental DF spectra was not possible. Thus, we are not able to say, which computational condition matches the experimental result the best. Clearly, the separation of the (π,π*) and (n,π*) electronic states in the H-bonded complexes predicted by modeling in vacuum (Fig. 1) is too small, and the actual value is in the range of 1500 – 2500 cm-1. In summary, a very similar vibrational pattern has been observed for the 1:1 complexes of 1,4- and 1,8-DAT with water and methanol, as well as for the bare molecule of 1,8-DAT. In contrast, the vibronic modes present in DF spectrum of the bare 1,4-DAT show very different spectral activity, which is due to the excited state deformation caused by pJT effect.

20 ACS Paragon Plus Environment

Page 21 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Another consequence of interaction of 1,4-DAT with protic molecules is the significant increase of fluorescence efficiency upon complex formation. It is illustrated in the upper panel of Figure 9 which shows change of intensity of the (0-0) bands of 1,4-DAT and its 1:1 complex with methanol upon increase of the solvent vapor pressure. The effect with water was comparable. Notice a very small uptake (short arrow) of free molecules and a large increase (long arrow) of fluorescence intensity of the complex. A very small population of Hbonded 1,4-DAT gives rise to the LIF band of a comparable intensity to that of the bare system. Fluorescence of the complex has to be much more efficient (~30-fold). In contrast, for a non-specific interaction with acetone (bottom panel) used here as a reference, much higher pressure is needed to give a comparable signal of the (0-0) band and it is accompanied with ~50% uptake of free molecules. These observations can be explained in terms of the abovementioned changes to energetics of the electronic excited states in hydrogen-bonded 1,4-DAT (see Fig. 1). Assuming that in a bare molecule the (n,π*) state is the lowest one, the large increase of emissive properties upon complexation may be attributed to an inversion of the (n,π*) and (π,π*) states. According to previous considerations,15,19 the phenomenon of fluorescence activation by protic solvents does not directly mean the change of the lowest excited singlet state from the (n,π*) type in non-polar solvents to the (π,π*) type in hydroxylic ones. Relative change of energies of those levels can affect the strength of vibronic coupling and as a result efficiency of non-radiative channels. This effect was observed in jet-cooled isoquinoline, manifested by a large (from ~350 ps to 4.6 ns) increase of fluorescence lifetime upon complexation with methanol.17 Additionally, the presence of a (n,π*) state in the triplet manifold located between the S1 and T1 states of (π,π*) character has been discussed as a reason for a very efficient intersystem crossing (ISC) process in 1,4-DAT, making a nonfluorescent deactivation channel.26 Significant changes of the energy levels of these states may suppress the ISC process. Stabilization of the singlet (π,π*) level in the complex was

21 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 41

already discussed, while a relative energy increase of the triplet (n,π*) state may be considered by analogy to the behavior of the singlet (n,π*) state in Fig. 1. In this context, it is worth to notice that the earlier attempts to detect the complexes of pyrazine with water in supersonic jets failed.48 This indicates that the specific order of excited electronic states, characteristic for larger aza-aromatic systems, plays a crucial role in the observed phenomena.

4. CONCLUSION By means of supersonic jet spectroscopy, we have demonstrated a very distinct photobehavior of 1,4-DAT, when compared with its 1,8- isomer, revealed both in the LIF and DF spectra. They show unusual activity of oop vibrational modes in bare 1,4-DAT, and molecular modeling predicts a distorted geometry of the excited molecule. We identified the main vibrational mode responsible for the deformation, and reproduced the characteristic members of its progression using a mixed symmetric quartic-harmonic potential with a barrier for inversion of 320 cm-1. Significant spectral changes occurring for 1,4-DAT upon complexation with a protic partner have been related to a theoretically predicted separation of the low-lying electronic singlet states. In bare molecule, nearby (π,π*) and (n,π*) singlet states are strongly coupled, which causes oop deformation of the pyrazine ring (pseudo-JahnTeller effect). This usually leads to enhanced internal conversion and/or intersystem crossing. Observed large increase of fluorescence from jet-isolated complexes of 1,4-DAT with water or methanol has been discussed in terms of a decreased vibrational coupling due to increased separation of participating states. Theoretical support by means of (TD-)DFT calculations proved to be a valuable help in the analysis of our experimental data. Being computationally inexpensive, the method gave insight into the character and energetics of the excited states, and revealed the nature of 22 ACS Paragon Plus Environment

Page 23 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

distortion within the pyrazine subunit. However, a mutual perturbation between electronic singlet states of (π,π*) and (n,π*) character, as well as possible anharmonicities in a multidimensional potential energy surface in a non-planar excited 1,4-DAT molecule make detailed analysis of the LIF and DF spectra a demanding task. For example, for thoroughly studied 9H-adenine, a similar effect of (π,π*) and (n,π*) states mixing was proposed based on a combined DFT/MRCI and CC2 theory.42 In this case, extended vibronic analysis including Herzberg-Teller contributions and a two-dimensional anharmonic potential based on CASPT2 modeling was necessary to propose the assignment of most prominent experimentally observed vibronic bands.49 It would be interesting to investigate further steps of complexation of 1,4-DAT by hydroxylic partners in the gas phase. How they affect photophysical parameters of our system, and whether it is possible to induce a cation-like emission, previously observed in bulk alcohols,25 by solvation in small clusters.

ACKNOWLEDGEMENTS This work was supported by the Polish National Science Centre grant 2014/15/B/ST4/05020, the PL-Grid infrastructure and computing grant G17-14 from the Interdisciplinary Centre for Mathematical and Computational Modeling of the Warsaw University. We wish to thank Professor Anna Grabowska for her interest and helpful discussions, and Dr Igor Czerski for the synthesis of the samples.

ASSOCIATED CONTENT Supporting Information. Absorption spectra, shape of orbitals, ground-state geometries of DAT compounds and their complexes, comparison of DF spectra of 1,4-DAT, and

23 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 41

visualization of selected vibrations. This material is available free of charge via the Internet at http://pubs.acs.org.

REFERENCES 1. Bakker, B. H.; Goes, M.; Hoebe, N.; Van Ramesdonk, H. J.; Verhoeven, J. W.; Werts, M. H. V.; Hofstaat, J.W., Luminescent Materials and Devices: Lanthanide Azatriphenylene Complexes and Electroluminescent Charge Transfer Systems, Coord. Chem. Rev. 2000, 208, 3-16. 2. Haftbaradaran, F.; Draper, N. D.; Leznoff, D. B.; Williams, V. E. Strained Silver(I) Coordination Polymers of 1,4-Diazatriphenylene. Dalton Trans., 2003, 2105-2106. 3. New, E. J.; Parker, D.; Peacock, R. D. Comparative Study of the Constitution and Chiroptical Properties of Emissive Terbium and Europium Complexes with a Common Tetraazatriphenylene Sensitiser; the Nature of the Sensitiser Determines Quenching Sensitivity and Cellular Uptake, DaltonTrans. 2009, 672–679. 4. Palma, M.; Levin, J.; Lemaur, V.; Liscio, A.; Palermo, V.; Cornil, J.; Geerts, Y.; Lehmann, M. Samori, P. Self-organization and Nanoscale Electronic Properties of Azatriphenylene-based Architectures: A Scanning Probe Microscopy Study, Adv. Mater. 2006, 18, 3313–3317. 5. Ichihara, M.; Suzuki, H.; Mohr, B.; Ohta, K. Different Disk Structures in the Hexagonal Columnar Mesophases of 2,3-Dicyano-6,7,10,11-tetraalkoxy-1,4-diazatriphenylenes and 2,3-Dicyano-6,7,10,11-tetraalkoxytriphenylenes, Liquid Crystals, 2007, 34, 401-410. 6. Boden, N.; Bushby, R. J.; Headdock, G.; Lozman, O. R.; Wood, A. Syntheses of New 'Large Core' Discogens Based on the Triphenylene, Azatriphenylene and Hexabenztrinaphthylene Nuclei, Liquid Crystals, 2001, 28, 139-144. 7. Takahashi, T.; Shizu, K.; Yasuda, T.; Togashi, K.; Adachi, C. Donor–acceptor-structured 1,4-Diazatriphenylene Derivatives Exhibiting Thermally Activated Delayed 24 ACS Paragon Plus Environment

Page 25 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Fluorescence: Design and Synthesis, Photophysical Properties and OLED Characteristics. Sci. Technol. Adv. Mater. 2014, 15, 034202. 8.

Pereira, J. A.; Pessoa, A. M.; Cordeiro, M. N.; Fernandes, R.; Prudêncio, C.; Noronha, J. P.; Vieira, M. Quinoxaline, its Derivatives and Applications: A State of the Art Review. Eur. J. Med. Chem. 2015, 97, 664-672.

9. Singh, D. P.; Deivedi, S. K.; Hashim, S. R.; Singhal, R. G. Synthesis and Antimicrobial Activity of Some New Quinoxaline Derivatives. Pharmaceuticals 2010, 3, 2416–2425. 10. Monge, A.; Martinez-Crespo, F. J.; Cerai, A. L.; Palop, J. A.; Narro, S.; Senador, V.; Marin, A.; Sainz, Y.; Gonzalez, M.; Hamilton, E. et al. Hypoxia-Selective Agents Derived from 2-Quinoxalinecarbonitrile 1,4-Di-N-oxides. 2. J. Med. Chem. 1995, 38, 4488–4494. 11. Ortega, M. A.; Morancho, M. J.; Martínez-Crespo, F. J.; Sainz, Y.; Montoya, M. E.; Ceráin, A. L.; Monge, A. New Quinoxalinecarbonitrile 1,4-Di-N-oxide Derivatives as Hypoxic-cytotoxic Agents. Eur. J. Med. Chem. 2000, 35, 21– 30. 12. Carta, A.; Paglietti, G.; Rahbar Nikookar, M. E.; Sanna, P.; Sechi, L.; Zanetti S. Novel Substituted Quinoxaline 1,4-Dioxides with in vitro Antimycobacterial and Anticandida Activity. Eur. J. Med. Chem. 2002, 37, 355-366. 13. Jampilek, J. Recent Advances in Design of Potential Quinoxaline Anti-Infectives. Curr. Med. Chem. 2014, 21, 4347-4373. 14. Gil, A.; Pabón, A.; Galiano, S.; Burguete, A.; Pérez-Silanes, S.; Deharo, E.; Monge, A.; Aldana, I. Synthesis, Biological Evaluation and Structure-activity Relationships of New Quinoxaline Derivatives as Anti-Plasmodium Falciparum Agents. Molecules 2014, 19, 2166-2180. 15. Lim, E. Proximity Effect in Molecular Photophysics: Dynamical Consequences of Pseudo-Jahn-Teller Interaction. J. Phys. Chem. 1986, 90, 6770-6777.

25 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 41

16. Kanamaru, N.; Lim, E. C. Radiationless Transitions in Aromatic Molecules with Nonbonding Electrons: Role of Vibronically Active Out-of-plane Modes. J. Chem. Phys. 1975, 62, 3252-3257. 17. Felker, P.; Zewail, A. H. Jet Spectroscopy of Isoquinoline. Chem. Phys. Lett. 1983, 5, 448-453. 18. Lobsiger, S.; Blaser, S.; Sinha, R. K.; Frey, H-M.; Leutwyler, S. Switching on the Fluorescence of 2-Aminopurine by Site-selective Microhydration. Nature Chem. 2014, 6, 989-993. 19. Barbatti, M.; Lischka, H. Why Water Makes 2-Aminopurine Fluorescent? Phys. Chem. Chem. Phys. 2015, 17, 15452-15459. 20. Badger, G. M.; Walker, I. S. Polynuclear Heterocyclic Systems. Part IX. n-π*-Transitions in the Spectra of Aromatic Aza-hydrocarbons. J. Chem. Soc. 1956, 122-126. 21. Bulska, H.; Chodkowska, A.; Grabowska, A.; Pakuła, B.; Slanina, Z. Excited States of Six-membered Aza-aromatic Rings. IX. 1,4- and 5,12-Diazatriphenylenes. J. Luminesc. 1975, 10, 39-57. 22. Grabowski, Z. R.; Grabowska, A. The Förster Cycle Reconsidered. Z. Phys. Chem. 1976, 101, 197-208. 23. Herbich, J.; Grabowska, A. Unusual Structure and Kinetics of Singly Protonated Quinoxaline in Fluorescent and Phosphorescent States. Chem. Phys. Lett. 1977, 46, 372376. 24. Waluk, J.; Grabowska, A.; Lipinski, J. Charge Density Flow as a Driving Force of Distortion in Excited Protonated Azaaromatics. Chem. Phys. Lett. 1980, 70, 175-179. 25. Grabowska, A.; Pakuła, B.; Sepioł, J. Double Fluorescence and Hydrogen Bonding of Anomalous 1,4-Diazines: 1,4-Diazaphenanthrene and 1,4-Diazatriphenylene. Nouv. J. Chim. 1979, 3, 287-292.

26 ACS Paragon Plus Environment

Page 27 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

26. Chojnacki, H.; Olszowski, A.; Ruziewicz, Z. High-resolution Electronic Spectra and Electronic States of 1,4-Diazatriphenylene. J. Mol. Struct. 1970, 5, 297-307. 27. Brenner, K.; Ruziewicz, Z. The Vibrational Structure of the Electronic Spectra of Quinoxaline in Shpolskii-type Hydrocarbon Matrices. J. Luminesc. 1977, 15, 235-254. 28. Brenner, K.; Lipiński, J; Ruziewicz, Z. Highly Resolved Electronic Spectra and Electronic Structure of Benzo[f]quinoxaline. J. Luminesc. 1983, 28, 13-26. 29. Lewanowicz, A.; Lipiński, J.; Ruziewicz, Z. Low-temperature Electronic Spectra, Luminescence Properties and Electronic Structure of 1,8-Diazatriphenylene. A Comparison with 1,4-Diazatriphenylene. J. Luminesc. 1981, 26, 159-175. 30. Ruziewicz, Z.; Olszowski, A.; Chojnacki, H. A Study of Quasilinear Electronic Spectra and Electronic States of 1-Azatriphenylene. Acta Phys. Polon. 1968, 34, 801-819. 31. Lim, E. C.; Li, Y. H.; Non-planar Distortions of, and Polarization of Luminescence from, π,π* States of Some Nitrogen Heterocyclic Compounds. Chem. Phys. Lett. 1969, 4, 6870. 32. Mason, A. T. Beiträge zur Kenntniss der Alkylendiamine. Ber. Dtsch. Chem. Ges. 1886, 19, 112-113. 33. Vdovin, A.; Sepioł, J.; Jasny, J.; Kauffman, J. M.; Mordziński, A. Excited State Proton Transfer in Jet-cooled 2,5-Di-(2-benzoxazolyl)phenol. Chem. Phys. Lett. 1998, 296, 557565. 34. Nosenko, Y.; Jasny, J.; Pietraszkiewicz, M.; Mordziński, A. Laser Spectroscopy of Porphycene Derivatives: A Search for Proton Tunneling in 2,7,12,17-Tetra-tertbutylporphycene. Chem. Phys. Lett. 2004, 399, 331-336 35. Sepioł, J.; Stepanenko, Y.; Vdovin, A.; Mordziński, A.; Vogel, E.; Waluk, J. Proton Tunnelling in Porphycene Seeded in a Supersonic Jet. Chem. Phys. Lett. 1998, 296, 549.

27 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 41

36. Barone, V.; Bloino, J.; Biczysko, M.; Santro, F. Fully Integrated Approach to Compute Vibrationally Resolve Optical Spectra: From Small Molecules to Macrosystems. J. Chem. Theory Comput. 2009, 5, 540-554. 37. Singh, U. C.; Kollman, P. A. An Approach to Computing Electrostatic Charges for Molecules. J. Comp. Chem. 1984, 5, 129-145. 38. Gaussian 09, Revision B.01, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A. et al. Gaussian, Inc., Wallingford CT, 2009. 39. Dreuw, A.; Weisman, J. L.; Head-Gordon, M. Long-range Charge-transfer Excited States in Time-dependent Density Functional Theory Require Non-local Exchange. J. Chem. Phys. 2003, 119, 2943-2946. 40. Silva, M. R.; Schreiber, M.; Sauer, S. P. A.; Thiel, W. Benchmarks for Electronically Excited States: Time-dependent Density Functional Theory and Density Functional Theory Based Multireference Configuration Interaction. J. Chem. Phys. 2008, 129, 104103. 41. Hochstrasser, R. M.; Marzzacco, C. Perturbations between Electronic States in Aromatic and Heteroaromatic Molecules. J. Chem. Phys. 1968, 49, 971-984. 42. Engler, G.; Seefeld, K.; Schmitt, M.; Tatchen, J.; Grotkopp, O.; Müller, T. J. J.; Kleinermanns, K. Acetylation Makes the Difference: a Joint Experimental and Theoretical Study on Low-lying Electronically Excited States of 9H-adenine and 9Acetyladenine. Phys. Chem. Chem. Phys. 2013, 15, 1025-1031. 43. Kunishige, S.; Kawabata, M.; Baba, M.; Yamanaka, T.; Morita, Y.; Higashibayashi, S.; Sakurai, H. Jet Spectroscopy of Buckybowl: Electronic and Vibrational Structures in the S0 and S1 States of Triphenylene and Sumanene. J. Chem. Phys. 2013, 139, 044313.

28 ACS Paragon Plus Environment

Page 29 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

44. Laane, J. Eigenvalues of the Potential Function V=z4±Bz2 and the Effect of Sixth Power Term. Appl. Spectrosc. 1970, 24, 73-80. 45. Frey, J. A.; Leist, R.; Tanner, C.; Frey, H.-M.; Leutwyler, S. 2-Pyridone: The Role of Out-of-plane Vibrations on the S1↔S0 Spectra and S1 State Reactivity. J. Chem. Phys. 2006, 125, 114308. 46. Trachsel, M. A.; Lobsiger, S.; Schar, T.; Leutwyler, S. Low-lying Excited States and Nonradiative Processes of 9-Methyl-2-aminopurine. J. Chem. Phys. 2014, 140, 044331. 47. Yang, J.; Wagner, M.; Laane J. Ring-Twisting and Ring-Bending Vibrations of 1,4Benzodioxan in Its S0 and S1(π,π*) States. J. Phys. Chem. A 2006, 110, 9805-9815. 48. Wanna, J.; Menapace, J. A.; Bernstein, E. R. Hydrogen Bonded and Non-hydrogen Bonded van der Waals Clusters: Comparison between Clusters of Pyrazine, Pyrimidine, and Benzene with Various Solvents. J. Chem. Phys. 1986, 85 1795-1805. 49. Conti, I.; Di Donato, E.; Negri, F.; Orlandi, G. Revealing Excited State Interactions by Quantum-Chemical Modeling of Vibronic Activities: The R2PI Spectrum of Adenine. J. Phys. Chem. A 2009, 113, 15265–15275.

29 ACS Paragon Plus Environment

The Journal of Physical Chemistry

+ H2O

1,4-DAT

BHLYP

+ CH3OH

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 41

B3LYP

Symmetry

E /cm-1

∆E /cm-1

f

E /cm-1

∆E /cm-1

f

A1 (π,π*)

34020

(0)

0.147

30530

(0)

0.093

B1 (n,π*)

34270

(+250)

0.004

29870

(-660)

0.003

B2 (π,π*)CT

34880

(+860)

0.007

29720

(-810)

0.013

(π,π*)

33540

(0)

0.157

30000

(0)

0.098

(n,π*)

34800

(+1260)

0.004

30400

(+400)

0.004

(π,π*)CT

34150

(+610)

0.012

28710

(-1290)

0.014

(π,π*)

33530

(0)

0.155

29990

(0)

0.097

(n,π*)

34740

(+1210)

0.004

30290

(+300)

0.004

(π,π*)CT

34150

(+620)

0.012

28690

(-1300)

0.014

Table 1. Energies (E) and oscillator strengths (f) of vertical electronic transitions from S0 to the two lowest (π,π*) and the lowest (n,π*) excited singlet states in 1,4-DAT and its 1:1 complexes with water and methanol calculated by TD-DFT with specified functionals at the DFT/B3LYP optimized ground state geometries. ∆E is the relative energy of a state in respect to that of the A1(π,π*) state.

30 ACS Paragon Plus Environment

Page 31 of 41

+ H2O

1,8-DAT

BHLYP

+ CH3OH

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

a

B3LYP

Symmetry

E /cm-1

∆E /cm-1

f

E /cm-1

∆E /cm-1

f

A1 (π,π*)

35290

(0)

0.060

31800

(0)

0.030

B2 (π,π*)CT

36090

(+800)

0.016

32050

(+250)

0.040

B1 (n,π*)

40900

(+5610)

0.006

35350a

(+3550)a

0.000a

(π,π*)

35120

(0)

0.068

31810

(0)

0.027

(π,π*)CT

35980

(+860)

0.011

31530

(-280)

0.041

(n,π*)

40850

(+5730)

0.003

35290

(+3480)

0.002

(π,π*)

35110

(0)

0.066

31780

(0)

0.027

(π,π*)CT

35960

(+850)

0.012

31510

(-270)

0.040

(n,π*)

40840

(+5730)

0.003

35280

(+3500)

0.002

(n,π*) state with A2 symmetry

Table 2. Energies (E) and oscillator strengths (f) of vertical electronic transitions from S0 to the two lowest (π,π*) and the lowest (n,π*) excited singlet states in 1,8-DAT and its 1:1 complexes with water and methanol calculated by TD-DFT with specified functionals at the DFT/B3LYP optimized ground state geometries. ∆E is the relative energy of a state in respect to that of the A1 (π,π*) state.

31 ACS Paragon Plus Environment

The Journal of Physical Chemistry

(n,π*)

(n,π*)

500

(n,π*) 0

(π,π*)

1000

(n,π*)

(π,π*)CT 250 cm-1

(π,π*)CT

(π,π*)CT

500

0

(π,π*)

(π,π*)

(π,π*) -500

-500

1,4-DAT bare

1,4-DAT / H2O 1:1 complex

1,8-DAT bare

1,8-DAT / H2O 1:1 complex

Figure 1. Relative energies of the two lowest (π,π*) and the lowest (n,π*) vertically excited electronic singlet states of 1,4-DAT and 1,8-DAT, and their 1:1 complexes with water calculated using the TD-DFT/BHLYP method for the DFT/B3LYP optimized ground state geometries.

32 ACS Paragon Plus Environment

5730 cm-1

(π,π*)CT

1260 cm-1

1000

5610 cm-1

5500

Relative energy ∆ν (cm-1)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 41

Page 33 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 2. The excited-state geometries of 1,4- and 1,8-DAT and their 1:1 complex with water calculated using TD-DFT/BHLYP approach. Geometry of 1,4-DAT complex has been additionally optimized in a polar medium (acetonitrile, relative permittivity ε = 35.7). Values of dihedral angles α and β (see text for description) are listed.

33 ACS Paragon Plus Environment

29900

30300

649 (221)

585 (201)

30500

659

1,8-DAT + methanol

30700

752

30100

571 585 593 651 656 744 754 786

397 (121) 405 (141) 464 (62) 477 (72)

232 (42) 289 (7111) 310 (6121)

167 (4111)

Experiment

586

0-0 51,57

B)

Calculations

119 172 226 228 248 294 303 403 412 458 468

( x 0.1 )

29946 cm-1 29900

102 (12)

1,8-DAT

A)

Fluorescence Intensity (a.u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 41

( x 0.5 )

The Journal of Physical Chemistry

30100 30300 30500 Wavenumber (cm-1)

30700

Figure 3. Laser induced fluorescence excitation (LIF) spectra of (A) bare 1,8-DAT, theoretically predicted using the DFT/BHLYP approach (top) and experimental one (bottom) and, (B) 1:1 complex of 1,8-DAT with methanol. Positions (in cm-1) and/or descriptions of bands are given. Calculated vibrational frequencies were scaled by 0.929.

34 ACS Paragon Plus Environment

Page 35 of 41

1,4-DAT

Calculations 596 (11132)

545 (181)

491 (11131)

539 561 605

Experiment 446 481

385 (111)

310 (5131) 318 (33)

268 272 291 278 329 349 361 397

194 (3121) 204 (51) 212 (32)

209

88 (21) 106 (31)

113

28736 cm-1

0-0

Fluorescence Intensity (a.u.)

A)

28700 28800 28900 29000 29100 29200 29300

237 247

1,4-DAT + water

237

52,54 90 131

0-0

95

48

0-0

75

B)

13

Fluorescence Intensity (a.u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

1,4-DAT + methanol

28700 28800 28900 29000 29100 29200 29300 Wavenumber (cm-1)

Relative wavenumbers ∆ν (cm

Figure 4. Laser induced fluorescence excitation spectra of (A) bare 1,4-DAT, theoretically predicted using the DFT/BHLYP approach (top) and experimental one (bottom) and, (B) 1:1 complex of 1,4-DAT with water (top) and with methanol (bottom). Positions (in cm-1) and/or descriptions of bands are given. Calculated vibrational frequencies were scaled by 0.929. In the upper panel a theoretically modeled anharmonic progression of the oop deforming mode is marked (see Figure 5).

35 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 41

Figure 5. Two possible mixed quartic-harmonic potentials fitted to the first two vibronic bands at 113 and 209 cm-1 with Cs(M)←Cs allowed (solid lines and numerical values) and forbidden (dashed line) vibrational levels. Barrier heights are indicated.

36 ACS Paragon Plus Environment

1,8-DAT + methanol 675

(29837 cm-1)

1411

1328

1308 (541) 1322 (561) 1386 (571) 1434 (591)

-1000

Experiment

Calculations

-1500 1317 1373 1411 1433

-500

1029 (411) 1048 (421)

622 698 608 (201) 678 (221)

0

-2000 Experiment

0

-500

-1000

1387 1432

1315

Calculations 41

B)

1042

1,8-DAT

(29946 cm-1)

Fluorescence intensity (a.u.)

A)

Fluorescence intensity (a.u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

679

Page 37 of 41

-1500

-2000

-1

Relative wavenumber (cm

)

Figure 6. Dispersed fluorescence (DF) spectra of (A) bare 1,8-DAT and, (B) 1:1 complex of 1,8-DAT with methanol, excited at their corresponding (0-0) transitions. Bottom panels show theoretically predicted spectra, calculated using the DFT/BHLYP approach (vibrational frequency scaling factor 0.929). Positions (in cm-1) and/or descriptions of bands are given.

37 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1404

30131 30132 1 1 30 8 3018131

83 8132 82 1 3 83 2 1 83 8232 301

814

1146 1260

408

984

Experiment

Calculations

33

32 81

31

1 1

124

572 664

230

(28736 cm -1)

0

-500

1045

240

627 697

1,4-DAT + methanol

-1000

-2000

1275 1350 1398 1453

-1500

1287 1401 1454

235

1046

- 500 -1000 1,4-DAT + water 624 699

0 (28661 cm -1)

B)

1,4-DAT

(28625 cm -1)

Fluorescence intensity (a.u.)

A)

Fluorescence intensity (a.u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 41

-1500 -1

Relative wavenumber (cm

-2000

)

Figure 7. Dispersed fluorescence spectra of (A) bare 1,4-DAT, theoretically predicted using the DFT/BHLYP approach (top) and experimental one (bottom) and, (B) 1:1 complex of 1,4DAT with water (top) and methanol (bottom), excited at their corresponding (0-0) transitions. Positions (in cm-1) and/or descriptions of bands are given. Calculated vibrational frequencies were scaled by 0.929.

38 ACS Paragon Plus Environment

- 500

- 1000

- 1500

Experiment

1320 1429

1002

Calculations ( vacuum )

1002

609 671

119 239 286 400

1239 1317 1370 1429

Calculations ( ε = 1.88 )

49

118

607 679

38 282

- 2000

1275 1350 1398 1453

624 699

1,4-DAT + water

235

28661 cm -1

0

Fluorescence intensity (a.u.)

0

1234 1312 1370 1429

1002

609 676 760

400

Calculations ( ε = 35.7 ) 133 239

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

1046

Page 39 of 41

- 500 - 1000 - 1500 - 2000 -1 Relative wavenumber (cm )

Figure 8. (From top to bottom) Experimental dispersed fluorescence spectrum of the 1:1 complex of 1,4-DAT with water excited at the (0-0) transition, and the spectra of this complex calculated in vacuum and in polar surroundings (n-hexane, relative permittivity ε = 1.88, and acetonitrile, ε = 35.7 ). The DFT/BHLYP approach with a vibrational frequency scaling factor of 0.929 was used.

39 ACS Paragon Plus Environment

The Journal of Physical Chemistry

Complex Fluorescence intensity (a.u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 41

Bare molecule

Methanol: p = 0.06 kPa p < 0.006 kPa

28620

28635

28725

28740

Acetone: p = 0.6 kPa p < 0.08 kPa

28530

28545

28725

Wavenumbers

28740

(cm-1)

Figure 9. Intensity changes of the (0-0) transition bands in LIF of 1,4-DAT and its 1:1 complexes with methanol (up) and acetone (bottom) at the indicated vapor pressure of the solvent.

40 ACS Paragon Plus Environment

Page 41 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

TOC graphic

41 ACS Paragon Plus Environment