Synthesis, electronic structure and spectral fluorescent properties of vinylogous merocyanines derived from 1,3-dialkyl-benzimidazole and malononitrile

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa

Synthesis, electronic structure and spectral fluorescent properties of vinylogous merocyanines derived from 1,3-dialkyl-benzimidazole and malononitrile Andrii V. Kulinich ⁎, Elena K. Mikitenko, Alexander A. Ishchenko Institute of Organic Chemistry, National Academy of Sciences of Ukraine, Murmans'ka str. 5, 02660 Kyiv, Ukraine

a r t i c l e

i n f o

Article history: Received 6 May 2016 Received in revised form 10 August 2016 Accepted 11 August 2016 Available online 13 August 2016 Keywords: Merocyanine Electronic structure Solvatochromism Quantum chemical calculation TDDFT

a b s t r a c t A vinylogous series of merocyanines were synthesized with 1,3-dibutyl-benzimidazole and malononitrile residues as the donor and acceptor terminal groups. These dyes do not comprise carbonyl groups, which are prone to the strong specific solvation by polar solvents up to hydrogen bond formation, and nevertheless they possess distinct reversed solvatochromism, i.e. their molecules have very high dipolarity. At that, they are soluble in a wide range of solvents from n-hexane to ethanol and do not aggregate readily. They were studied thoroughly by UV/Vis, fluorescence, IR, and NMR spectroscopy methods. Their structure and spectral properties in the ground and excited fluorescent states were modelled at the DFT level both in vacuum and in solvents of various polarities by using the PCM solvent field simulation. The calculations were performed using several hybrid functionals (B3LYP, CAM-B3LYP, and wB97XD) and the split-valence 6-31G (d,p) basis set. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The advent and development of a great deal of modern technologies, such as ink-jet printing, holography, liquid crystal and OLED displays, optical data storage, bio-medical imaging, entailed the ever increasing demand for novel functional dyes to meet the new and challenging criteria [1–4]. Obviously, the tailored design of functional compounds requires revealing the regularities connecting their properties with both their chemical structure and environment. In this respect, the donor-acceptor substituted polymethines, i.e. merocyanines, are among the most suitable objects to study, since their spectral-luminescent and other physical-chemical properties change in a very wide range [5–8]. Depending on the donor-acceptor properties of their terminal groups (D and A), the polymethine chain length, and the solvent polarity, their electronic structure can vary from the neutral polyene (А1), via the ideal polymethine (А2), to the dipolar polyene (А3) – three virtual limiting states suggested by S. Dähne to simplify description of the polymethine dyes; [5] the unconventional limiting structure A2 being commended because of a series of unique characteristics – the minimum excitation energy to the first excited singlet state, narrowing and growing of the peak intensity of the long-wavelength spectral bands due to the least changes of the chromophore π-bond

⁎ Corresponding author. E-mail address: [email protected] (A.V. Kulinich).

http://dx.doi.org/10.1016/j.saa.2016.08.015 1386-1425/© 2016 Elsevier B.V. All rights reserved.

orders and dipole moment upon electronic excitation, and by the highest fluorescence quantum yield (Φf) [6].

Consequently, the deviation from the ideal polymethine state A2 to either side is followed by a hypsochromic shift and broadening of the absorption and fluorescence bands of merocyanines. These changes can be traced most easily by investigation of their solvatochromism and solvatofluorochromism. The chemical structure of dyes is kept constant in this case. Therefore, all the spectral changes observed can be attributed to the electronic structure variations. The reversed solvatochromism is the most interesting situation in this respect, since there can be relatively broad ranges of positive (A1-to-A2) or negative (A2-to-A3) solvatochromism but only a single point in which the solvatochromism reversal takes place. In addition to physical-chemical methods, the quantum chemistry provides us with ever increasing possibilities to study, predict, and hence to design the properties of novel functional molecules. However, there exist some unsettled issues in application of quantum chemistry for merocyanines. For example, the common ab initio and DFT methods overestimate considerably the energy of the long-wavelength electronic transition in their molecules [9,10], while more sophisticated approaches are too resource consuming. Also, the bulk solvent modelling approaches can be unequal to the task when strong solute-solvent interactions (SSIs) are involved, which is often the case with carbonyl-

318

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

containing dyes in protic solvents [11,12]. One more issue is the quantum chemical modelling of the excited states of polymethine dyes, which is both resource-intensive and challenging task. Consequently, there is still a demand for some reper dyes to verify the results of such calculations. Again, merocyanines possessing reversed solvatochromism are probably the most convenient objects in this respect. Therefore, in the present paper, we designed and studied by using the physical-chemical and quantum-chemical methods the vinylogous series of malononitrile-based merocyanine dyes. The strong electrondonating properties of the 1,3-dialkyl-benzimidazole residue was conjectured to provide high dipolarity of the studied molecules and consequently their reversed solvatochromism. At that, they do not comprise carbonyl groups in their molecules, so the solvent effects can be treated more reliably via the standard PCM approach. The dipolarity of donor-acceptor molecules is known to decrease gradually with the πconjugated system lengthening [6,12,13]. So, in this case we had at least two criteria to verify the results of the TDDFT calculations – the point of solvatochromism reversal and the change of dyes electronic structure upon the polymethine chain elongation.

2. Materials and methods Merocyanines 1–3 were synthesized by heating equal amounts of 1,3-dibutyl-2-methyl-benzimidazolium bromide [14] and the corresponding hemicyanine [15] in pyridine for 30–90 min in the presence of DBU (1,8-diazabicycloundec-7-ene).

IR spectra were recorded in KBr pellets using a Bruker Vertex 70/80 FTIR spectrometer. 1H NMR spectra were registered on a Varian VXR300 spectrometer (299.943 MHz for H-atoms) in CDCl3 or (CD3)2SO, 13 C NMR spectra with power-gated decoupling, 13C NMR APT spectra with J-compensation, HMBC, and HMQC spectra — on a Bruker Avance III spectrometer (100.61 MHz for C-atoms) in CDCl3; TMS was used as an internal standard in both cases. The atom labelling used in the discussion is shown below.

The long-wavelength absorptions and fluorescence bands of merocyanines 1–3 were analysed by the method of moments, which allows the quantitative characterization of their position and shape [20]. Three parameters obtained by this way are discussed alongside with the band maxima (λmax) and the molar extinctions (ε): M−1 is the centre of gravity of a band in the scale of wavenumbers (M− 1 = 107/ν, where ν is the wavenumber); the value of σ characterizes the deviation of a spectral band from its gravity centre and is similar to the widely used full width at half maximum (FWHM); f is the oscillator strength of the long-wavelength absorption transition (f = 4.317 × 10−9 ∫ενdν). Indices ‘a’ and ‘f’ denote the parameters relative to absorption and fluorescence correspondingly. The deviations by maxima (Daλ) of the absorption bands of merocyanines 1–3 were calculated using the spectral data of the corresponding cationic and anionic dyes. The Stokes shifts were calculated both by the band maxima (SSλ) and by the band centres (SSM). 3. Results and discussion

The reaction mixture was diluted with water-ethanol (1:1), left overnight, and the resulting precipitate was filtered off, dried and refined by column chromatography on neutral alumina-80 using chloroform as an eluent. The resulting product was then crystallized from ethanol. Melting points were measured in an open capillary and were not corrected. n-Hexane, toluene, chloroform, dichloromethane (DCM), DMF, ethanol (96%), and other solvents used were refined according to the methods given in ref. [16] The UV/Vis spectra were recorded using a Shimadzu UV-3100 spectrophotometer. Solutions of merocyanines 1–3 were confirmed to obey the Lambert–Beer law in the concentration range of 1 × 10−6–5 × 10−5 M in toluene, chloroform, DCM, DMF, and ethanol. In n-hexane the studied dyes are not soluble enough to carry out the same test. However, no sign of aggregation was traced by their absorption spectra in this solvent. The fluorescence spectra were recorded using a CM2203 spectrofluorometer (“Solar”, Belarus, wavelength scale 220–900 nm). Solutions of merocyanines were not degassed since the fluorescent characteristics for degassed and non-degassed solutions were identical. The fluorescence quantum yield of merocyanine 1 was measured in relation to Coumarin 1 in ethanol (Φf = 73% [17]), of dye 2 – in relation to Rhodamine 6G in ethanol (Φf = 95% [18]), of dye 3 — in relation to Nile blue in ethanol (Φf = 27% [19]). The optical densities of solutions for fluorescent measurements were kept below 0.1 to avoid inner filter effects. The values of Φf were corrected taking into account refractive indices of the solvents. The fluorescence excitation spectra have been found to coincide nearly with the absorption spectra of merocyanines 1–3 and to be independent of the emission wavelength, which confirms both the fluorescence purity of the studied dyes and the absence of aggregation in their solutions at the given concentrations.

Both 1H and 13C NMR spectra of compounds 1–3 can be used to evaluate roughly their electronic structure. The most obvious characteristics here are the vicinal spin-spin coupling constants (SSCCs) between the polymethine chain H-atoms, which are known to correlate with the corresponding C–C-bond orders [21–23]. Even in medium-polarity CDCl3, the 3J(H,H) SSCCs alternation for dyes 2 and 3 corresponds to the structure from the A2–A3 range. Going from CDCl3 to (CD3)3SO as the solvent should naturally enhance dipolarity of dyes 1–3, which is shown up specifically in the further increase of the polymethine chain 3J(H,H) SSCCs alternation and in the downfield shifts of the NCH2, Har, and H(1) signals of vinylogue 3. For comparison, in CDCl3 the corresponding 3J(H,H) SSCCs values for the relative dyes with 1,3-diphenyl-benzimidazole as the donor terminal group are nearly equalized and only in high-polarity (CD3)2SO they alternate distinctly [22]. The Cα and Cω atoms chemical shifts can also be used to evaluate the merocyanine electronic structure [22]. The Cω signal for merocyanine 3 is shifted upfield by 1.4 ppm relative to that in its N,N-diphenyl-analogue, once more confirming the greater dipolarity of N,N-dibutyl-benzimidazole based 3. In the series 1 - 2 - 3, the Cα and Cω signals go upfield and downfield correspondingly with the polymethine chain lengthening, thus demonstrating a minor decrease of dipolarity for the higher vinylogues, just as expected. The characteristic C`N frequencies of the malononitrile residue in the IR spectrum should decrease upon growth the electronic density on this group. Malononitrile itself (νC`N = 2272 cm−1) and malononitrile anion (νC`N = 2158 cm−1) are the two obvious extreme cases. To find the state A2 characteristic frequency of malononitrile-based polymethines, the IR spectra of the three corresponding symmetric anionic dyes (see ESI) were recorded. For all these dyes, an intensive peak of the νC`N symmetric stretch of the malononitrile residue (the attribution is based on the DFT quantum chemical calculations) lies at 2189– 2192 cm−1, with the νC`N asymmetric stretch shoulder at 2170–

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

319

2180 cm−1. The corresponding peaks of merocyanine 1 (νC`N(s) = 2193 cm−1, νC`N(as) = 2177 cm−1) well-nigh coincide with those of the anionic dyes. Hence, its electronic structure is very close to the ideal polymethine state. For vinylogues 2 and 3 these peaks are slightly shifted to the low-frequency region (νC`N(s) = 2180 cm−1, νC`N(as) = 2159 cm−1), which implies that in a solid state these dyes are even more dipolar than 1, falling in the range A2-to-A3. This result seemingly disagrees with the conclusions based on the NMR spectral data, which can be attributed to the crystal packing effect, viz. to π-stacking interactions, which should be greater for the higher vinylogues due to the greater polarizability of a longer chromophore. The UV/Vis and fluorescence spectral characteristics of merocyanines 1–3 are collected in Table 1. 3.1. UV/Vis spectra

Fig. 1. UV/Vis absorption spectra of dye 1 in n-hexane (black), toluene (blue), and ethanol (red).

A bathochromic shift of the long-wavelength absorption band of merocyanine 1 by both λamax and M−1 is observed when going from a less-polar n-hexane to toluene (Table 1). Certainly, it can be explained by the greater refractive index of toluene, but there are other spectral effects indicative of the dye electronic structure change upon this solvent replacement: an increase in the absorption band intensity (ε, f) and the absorption bandwidth (σa) decrease, the latter in spite of the stronger solute-solvent interactions (SSIs) in more polar medium. A further increase in solvent polarity proceeding to chloroform, DCM, DMF, and ethanol results in the hypsochromic shifts, abatement of intensity, and broadening of the long-wavelength absorption band of dye 1 (Table 1, Fig. 1). Hence, merocyanine 1 has a reversed solvatochromism and its ground state electronic structure verges on the ideal polymethine state A2 in toluene, biasing to the A1–A2 range in less-polar n-hexane and to the A2–A3 range in medium- and high-polarity solvents. It is noteworthy, that the absorption band of dye 1 is shifted hypsochromically while going from polar aprotic DMF to ethanol, it also becomes wider and less intensive in the latter solvent (Table 1). Hence, the electronic structure of merocyanine 1 diverges more from the structure A2 towards the A3 in ethanol than in DMF. For reference, the same solvent change causes much smaller hypsochromic shift of the absorption band of the analogous merocyanine derived from 1,3-diphenyl-benzimidazole, at that the bandwidth retains the value of 1100 cm−1 and the molar extinction increases [24]. Given these data, one can suggest that the increased electron density on the malononitrile residue of dye

1, induced by the greater electron donating ability of the donor terminus, makes the former more susceptible to electrophilic solvation. To the best of our knowledge, this phenomenon has not been known for malononitrile-based merocyanines so far [25]. Only faint signs of this effect could be traced by juxtaposition of solvatochromism of 1,3,3trimethylindole and 1,3-diphenylbenzimidazole based merocyanines [6]. In all the solvents chosen, the long-wavelength absorption band of vinylogue 2 is characterized by greater molar extinctions and oscillator strengths than that of dye 1 (Table 1) – the usual outcome of the polymethine lengthening for cyanine dyes whose electronic structure does not deviate considerably from the state A2. The pattern of spectral changes under solvent polarity variations is very similar for dyes 1 and 2 and will not therefore be discussed in detail. The most noticeable effect of this polymethine chain lengthening is narrowing of the absorption band in toluene, chloroform, DCM and DMF and its broadening in n-hexane and ethanol, i.e. in the solvents in which the electronic structure of dye 2 deviates farthermost from the state A2. Such behaviour can be construed in the terms of two principal reasons of band broadening in electronic spectra – the inhomogeneous broadening (SSIs) and the vibronic interactions (VIs), the latter being determined mostly by the change of the chromophore bond orders upon electronic transition. It has been shown by many examples that the VIs abate with the polymethine chain lengthening for polymethine dyes whose electronic structure does

Table 1 Characteristics of the UV/Vis absorption and steady state fluorescence bands of dyes 1–3. Dye

Solvent

λamax (nm)

ε × 10−4 (L × mol−1 × cm−1)

λfmax (nm)

Φf (%)

f

M−1 a (nm)

σa (cm−1)

M−1 f (nm)

σf (cm−1)

SSλ (cm−1)

SSM (cm−1)

1

n-Hexane Toluene CHCl3 CH2Cl2 DMF EtOH n-HEXANE toluene CHCl3 CH2Cl2 DMF EtOH n-Hexane Toluene CH3CCl3 CHCl3 CH2Cl2 DMF MeCN EtOH MeOH

428 431 427 425 422 417 519 529 526 522 515 511 610 632 630 632 625 610 603 608 599

8.95 9.57 9.29 8.97 7.95 7.71 –a 15.90 15.63 15.12 11.67 11.05 –a,b 19.40 22.24 23.02 21.08 11.59 10.97 10.80 8.80

450 457 452 450 450 447 542 553 552 549 545 541 633 660 659 660 657 655 643 650 643

0.07 0.18 0.27 0.16 0.21 0.05 0.10 1.2 2.4 1.8 1.3 0.59 0.42 11 31 38 30 14 13 12 11

0.830 0.875 0.919 0.920 0.907 0.910 – 1.104 1.135 1.176 1.166 1.138 – 1.322 1.387 1.402 1.455 1.368 1.349 1.356 1.355

414.0 418.2 415.0 412.4 408.1 404.1 501.4 516.2 513.2 508.0 496.9 491.3 579.7 613.2 614.3 617.1 606.9 574.8 567.5 568.1 553.9

1100 1080 1090 1130 1200 1230 1110 930 940 970 1170 1260 1260 980 920 860 980 1500 1540 1650 1840

494.6 500.5 494.9 493.4 498.8 495.4 580.3 579.7 572.6 569.9 565.6 561.2 679.7 683.1 679.0 676.5 672.7 669.4 662.5 665.1 661.2

2250 2160 2340 2360 2490 2580 1390 1190 1070 1070 1090 1090 1170 830 780 740 770 820 830 830 840

1140 1320 1300 1310 1470 1610 820 820 900 940 1070 1090 600 670 700 670 780 1130 1130 1060 1140

3940 3930 3890 3980 4460 4530 2710 2120 2020 2140 2440 2540 2540 1670 1550 1420 1610 2460 2530 2570 2930

2

3

a b

Low-solubility. For 1,3-didodecyl-benzimidazole-analogue (3a) ε = 9.01 × 104 L × mol−1 × cm−1, Φf = 14% in EtOH and 0.83% in n-hexane.

320

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

not deviate considerably from the state A2 [6,8,20]. On the contrary, the SSIs should increase for higher vinylogues due to the greater polarizability of the longer chromophore. Certainly, in n-hexane the SSIs are too weak to be considered as a cause of the band broadening at issue. It has been mentioned above that with the polymethine chain lengthening the contribution of the non-polar structure A1 usually increases for merocyanines in low-polarity solvents [6]. In less-polar n-hexane this tendency should be most pronounced, what can be traced in the series 1 - 2 - 3 by the solvatochromic shift rise in the pair n-hexane - toluene: 160 cm−1– 360 cm−1–570 cm−1. Consequently, the broadening of the spectral bands in n-hexane while going to dye 2 and then to 3 should be explained by the growth of the VIs due to the more polyenic nature of the higher vinylogues. In more polar solvents, in which the studied dyes are close to the state A2, reducing of VIs with the polymethine chain lengthening results in the narrowing of the spectral bands, but the increased SSIs cause the band broadening in ethanol and for the pair 2–3 even in DMF (Table 1). Merocyanine 3 retains the reversal solvatochromism of its predecessors. At that, the maxima of its long-wavelength absorption band coincide in toluene and chloroform. Considering the lesser refractive index of chloroform, it should be deduced that the electronic structure of dye 3 is closer to the state A2 in this solvent, which becomes more evia 1 dent from comparison of the corresponding values of ε, M− a , and σ (Table 1). UV/Vis spectra of dye 3 was also recorded in 1,1,1trichloroethane, which falls in between toluene and chloroform by the relative permittivity. In this solvent, the long-wavelength absorption band of dye 3 is blue-shifted, wider, and less intensive (by the value of ε) relative to that in chloroform. The polymethine chain lengthening in the pair 2–3 is accompanied by growth of the integral intensity (f) of the long-wavelength absorption band (Table 1). To the contrary, the peak intensity (ε) increases in toluene, chloroform, and DCM but decreases in high-polarity DMF and ethanol, in which the considerable broadening of the absorption band takes place, evidently, due to the greater SSIs for the highest vinylogue [6,10]. The absorption band narrowing in going from 2 to 3 occurs only in chloroform, thus confirming that for vinylogue 3 it is “the ideal” solvent. Dyes 1–3 demonstrate the moderate range of solvatochromism, which, unlike to that of their thiobarbituric analogues [13], does not rise with the polymethine chain lengthening. For example, the solvatochromic shifts in the pair toluene - ethanol amount to 780 cm−1, 670 cm−1, and 620 cm−1 in the series 1 - 2 - 3. Hence, the site-specific solvation of molecules 1–3 is much weaker in comparison with “common” carbonyl comprising merocyanines. Vinylene shifts and deviations of the absorption band maxima are important characteristics of polymethine dyes used to study their electronic structure [6]. In the series 1 - 2 - 3 the first vinylene shifts fall within an interval of 91–99 nm by maxima and 87.2–98.2 nm by band centres. The second vinylene shifts vary within a wider range of 91– 106 nm by λamax and 76.8–103.9 nm by M−1 a . Both shifts are maximal in toluene, chloroform, and DCM. They are very close in these solvents to the value of 100 nm typical for symmetric polymethines, and decrease both in less-polar n-hexane and in high-polarity DMF and ethanol. The declension of the electronic structure of molecules 1–3 from the state A2 in the solvents of extreme polarities are better traced by the band centres (M−1 a ), since this parameter includes the change of the band shape also. The values of deviations well confirm the above conclusions on the dyes 1–3 electronic structure: they are very small in medium-polarity DCM but increase in going to high-polarity ethanol (see ESI).

comparison the parameters of their absorption and fluorescence spectra. The solvatofluorochromic shifts are smaller than solvatochromic for all three vinylogues, e.g. in the solvent pair toluene - ethanol they are equal to 490 cm−1, 400 cm−1, and 230 cm−1. This implies weakening of the specific SSIs in their fluorescent excited state. The only exception from the above regularity takes place while going from n-hexane to toluene, i.e. within the solvent polarity range in which the studied merocyanines demonstrate positive solvatochromism and consequently their dipole moments increase upon excitation. Weaker solvation of dyes 1–3 in the fluorescent state in polar solvents is confirmed also by the greater vinylene shifts of their fluorescent bands in DMF and ethanol. The violation of the mirror symmetry of absorption and fluorescence bands of merocyanines 1–3 (Fig. 2) is an important issue. The discrepancy is most conspicuous for vinylogue 1 – its fluorescence bands are considerably wider than the absorption bands in all the solvents (cf. the corresponding values of σa and σf in Table 1). Moreover, the fluorescence bands of dye 1 are 300–800 cm−1 (in various solvents) wider than those of its N,N-diphenyl analogue [24], while their absorption bands have almost the same bandwidths. This effect cannot be explained by an instrumental function, i.e. by the white noise, since the fluorescence quantum yields of these merocyanines are sufficient to obtain the reliable spectral curves. Formation of trans-cis stereoisomers as an explanation of fluorescence band broadening must be dismissed in the case of dye 1 due to the axial symmetry of both its termini. In addition, the malononitrile residue is not a bulky one to suppose any steric hindrances in molecule 1. Hence, there should be some intrinsic source(s) of the fluorescence band broadening. The absorption and fluorescence bands of higher vinylogues 2 and 3 are seemingly very close to the mirror symmetry. Nevertheless, juxtaposition of their σa and σf values demonstrates that the fluorescence bands of dye 2 are wider than absorption in low- and medium-polarity solvents and narrower in DMF and ethanol. The latter narrowing can be explained by weakening of the SSIs in the fluorescent state in polar media. The probable reason of the fluorescence band broadening of merocyanine 1 in all the solvents used and of dye 2 in the low-polarity media is an increase of the VIs in the fluorescence transition. The fluorescence bands of vinylogues 3 are narrower than its absorption bands, the difference increasing with the solvent polarity growth. The fluorescence quantum yields of merocyanines 1–3 correlate well with their solvatofluorochromism. They are maximal in medium-polarity chloroform and decrease in low- and high-polarity solvents in which the dyes electronic structure deviate more from the ideal polymethine state A2. At that, the Φf values of vinylogue 1 are very low, but they soar up with the polymethine chain lengthening (Table 1). There are many paths of fluorescence quenching in polymethines: vibrational deactivation of the excited state which is enhanced by strong

3.2. Fluorescence spectra Reversible solvatofluorochromism of dyes 1–3 replicates their solvatochromism for all practical purposes (Table 1). Nonetheless, there are several essential distinctions which can be traced by

Fig. 2. Normalized absorption and fluorescent spectra of dye 3 in n-hexane (black), chloroform (blue), and glycerol (red).

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

VIs and SSIs, trans-cis isomerization around the polymethine chain C\\C bonds or TICT state formation [26], presence of low-lying non-fluorescent nπ⁎-states [3], intersystem crossing [27]. Evidently, and this has been confirmed by the quantum chemical calculations, there are no low-lying nπ⁎-states in the molecules of dyes 1–3. They are also devoid of any factors favouring intersystem crossing. Strong SSIs as an effective source of excited state deactivation can be ruled out taking into account the above conclusion about weakening of the SSIs in the fluorescent state of merocyanines 1–3 in polar solvents. Hence, only two non-radiative excited state deactivation pathways remain to be discussed: the vibronic interactions and rotations around the polymethine chain bonds, the latter including both isomerization and TICT. A weakening of the VIs with the polymethine chain lengthening in the series 1 - 2 - 3 is proved by the spectral data, first of all by the bandwidths of their fluorescent bands (σf). So, the fluorescence rise of the higher vinylogues could be explained by this only factor. Nevertheless, the rotational pathway of the excited state deactivation cannot be ignored. Indeed, it has been shown earlier [28] that the principal site of trans-cis photoisomerization in simple merocyanines is the C\\C-bond connecting the donor terminal group and the polymethine chain (the Cα\\C(1) bond in the current notation). The volume of a rotor in this case increases steadily with the polymethine chain lengthening. Hence, the maximum effect of trans-cis photoisomerization onto the fluorescence quenching should take place for dimethine-merocyanine 1. This suggestion is in good agreement with the actual data. The lengthening of the polymethine chain in the series 1 - 2 - 3 by one vinylene group results in the fluorescence quantum yield increase by nearly an order of magnitude (Table 1), with a tendency that this effect enhances with the solvent polarity growth. The only exception to the rule is DMF in which the effect is much diminished. Moreover, in this solvent the Φf value of dye 1 is greater than that in medium-polarity DCM. Both these irregularities can be rationalized by the greater viscosity of DMF. The higher medium viscosity suppresses the solute internal rotations and this effect should be maximum in the case of vinylogue 1. To obtain the more reliable data, the UV/Vis and fluorescence spectra of dyes 1– 3 were recorded in highly viscous glycerol (see ESI). It was found that in this solvent the Φf values of merocyanines 1–3 are 16, 9 and 2.25 times greater than in ethanol. Hence, the rotational deactivation of the excited state actually plays an important role for all the studied molecules, the impact of this factor decreasing with the polymethine chain lengthening. Spectra of dyes 1–3 in glycerol once more testified the fact that the electrophilicity of a solvent indeed has a great effect on the solvatochromism of malononitrile-based merocyanines in the case of high basicity of the donor terminal group. For all the vinylogues, their absorption maxima in more electrophilic glycerol are shifted hypsochromically relative to those in ethanol (although the refractive index of glycerol is considerably greater) and the absorption bands become wider. At that, the position and shape of the fluorescence bands in these two solvent are very close. The same effect can be observed also in going from ethanol to methanol as the solvent (cf. the corresponding σa and σf values in Table 1). 3.3. Quantum chemical simulation Quantum chemical simulation was performed using Gaussian-09 software package [29]. The geometry optimization of molecules 1–3 was carried out at the DFT level using B3LYP, CAM-B3LYP, and wB97XD functionals with the split-valence 6-31G (d,p) basis set. The CAM-B3LYP functional, which comprises the greater share of HF exchange over a long range than B3LYP, was shown to provide better agreement with experimental data in the case of long highly polarizable molecules [30,10]. The wB97XD includes both long-range correction and Grimme's empirical dispersion [31]. The convergence criterion on the residual forces has been set to 1 × 10−5 Hartree Bohr−1 or Hartree Rad−1. The calculation of force constants and vibrational frequencies

321

was performed to verify that the geometry located was a minimum. The vertical excitation energies as well as the first singlet excited state relaxed geometry have been calculated with the adiabatic TDDFT approximation. The polarisable continuum model (PCM) was applied to model the solvent effects; [32] the inclusion of nuclear relaxations in the case of fast absorption and fluorescence processes (Franck–Condon principle) was avoided by employing a state-specific solvation calculation in terms of fast electronic cloud reorganizations and slow solvent and solute nuclear motions [33]. As it was explained above, merocyanine 1 cannot have any isomers. The all-trans conformation of vinylogues 2 and 3 follows from the values of the polymethine chain 3J(H,H) SSCCs. The geometry optimizations (B3LYP, CAM-B3LYP, and wB97XD) predict the chromophores of dyes 1–3 are non-planar in the ground state. The torsion angle (N(1)\\Cα_C(1)\\C(2)) is equal to ≈ 165° in vacuo and decreases gradually with the solvent polarity growth, going to 157–158° in ethanol. There exist also the out-of-plane distortion of the N(2)-atom that decreases in the polar media as well. Probably, this non-planarity originates from the repulsion of the (N2)-CH2–group and the H(2)-atom of the polymethine chain. In more polar media the Cα_C(1)-bond elongates (Table 2), consequently its bond order decreases and the rotational distortion becomes easier. To estimate the position of a merocyanine on the A1–A2–A3 scale, we used the bond length alternation (BLA) parameter, which is specified as the difference between the average bond lengths of the formal single and double bonds in the polymethine chain [7]. Irrespective of a functional used, the DFT calculations predict inversion of the BLA sign with the solvent polarity growth (Table 2) – the result which correlates well with the reversal solvatochromism of the studied dyes. With the polymethine chain lengthening, the minimum absolute BLA value is obtained for more polar solvents, again in good consistency with the experimental data. The TDDFT calculations indicate the only allowed long-wavelength transitions for the studied molecules is the polymethine-type ππ⁎-transition which involves mainly the HOMO and LUMO (Fig. 3). The higher electronic transitions have considerably greater energies and cannot therefore influence the Vis-absorption and fluorescence optical properties of dyes 1–3. At that, the n-orbitals of the malononitrile residue contribute only to the very deep transitions, S11 ← S0 and higher. Another noteworthy feature of the molecules 1–3 is very small changes of the molecular dipole moment upon excitation (Table 2), which implies that the light induced CT gives only a minor contribution to their S1 ← S0 transition. The long-wavelength transition energies in merocyanines 1–3 are overestimated considerably in all cases, the B3LYP functional giving the comparatively better result (cf. the experimental values of λamax in Table 1 and the calculated λamaxc values in Table 2). On the other hand, the solvatochromism and solvatofluorochromism of the studied dyes are better reproduced by the CAM-B3LYP and wB97XD simulations. Indeed, the B3LYP predicts negative solvatochromism for molecules 1 and 2 and reversed solvatochromism for the longest vinylogue 3. Two other functionals give negative solvatochromism for 1 and reversed for 2 and 3, i.e. they overestimate the dipolarity of merocyanines 1–3 to a smaller degree. There were some problems during the first excited state optimization of molecules 1 and 2. In some cases the optimization led to the TICT-like geometry of the molecules with the benzimidazole terminal group twisted strongly relative to the polymethine chain and then run into a cyclic path (hence the empty cells in Table 2). However, in most cases the excited state geometry was located successfully. The chromophore of merocyanines 1–3 becomes more planar in the S1 state, the (N(1)\\Cα_C(1)\\C(2)) torsion angle increases to 165–169° in all media, most considerably in high-polarity ethanol. Relaxation to the first excited state affects most strongly the polymethine chain atoms. As a result the polymethine bonds alternation in the excited state (BLA⁎) differs considerably from that in the ground state. The

322

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

Table 2 Some results of quantum chemical calculations for dyes 1–3. Medium

μ (S0) (D)

μ (SFC 1 ) (D)

μ (S1) (D)

μ (SFC 0 ) (D)

BLA (Å)

Cα–C(1) (Å)

BLA⁎ Å

λamaxc (nm)

fa

λfmaxc (nm)

ff

Vacuo n-Hexane Toluene CHCl3 DCM EtOH Vacuo n-Hexane Toluene CHCl3 DCM EtOH Vacuo n-Hexane Toluene CHCl3 DCM EtOH

11.8 13.8 14.4 15.9 16.7 17.4 15.3 18.3 19.2 21.4 22.8 23.9 18.2 22.2 23.5 26.7 28.7 30.4

10.1 11.7 12.2 13.6 14.4 15.1 15.2 17.6 18.3 20.1 21.1 22.0 19.4 22.6 23.6 26.0 27.4 28.6

– – – 14.6 15.6 16.3 – 18.0 18.9 20.7 21.7 22.5 19.4 23.1 24.1 26.5 27.8 28.8

– – – 15.8 16.3 16.8 – 19.2 19.9 21.5 22.4 23.2 18.7 23.1 24.4 27.1 28.4 29.5

0.006 −0.006 −0.010 −0.018 −0.024 −0.028 0.019 0.006 0.001 −0.009 −0.016 −0.021 0.026 0.012 0.008 −0.004 −0.012 −0.018

1.397 1.403 1.405 1.410 1.413 1.415 1.394 1.401 1.403 1.409 1.412 1.414 1.393 1.400 1.402 1.407 1.411 1.414

– – – −0.030 −0.023 −0.019 – −0.027 −0.025 −0.022 −0.019 −0.017 0.008 −0.011 −0.014 −0.016 −0.016 −0.014

362.0 356.8 355.3 350.4 347.2 344.1 412.2 411.2 410.4 406.2 402.3 397.9 462.6 465.3 465.2 461.3 456.6 450.9

1.006 1.010 1.009 1.007 1.004 1.000 1.598 1.606 1.605 1.595 1.584 1.572 2.108 2.141 2.146 2.139 2.121 2.097

– – – 372.0 371.0 370.6 – 428.5 427.2 424.5 422.9 421.6 478.0 480.0 479.8 477.4 475.2 473.2

– – – 1.001 1.016 1.020 – 1.564 1.575 1.585 1.589 1.591 2.127 2.141 2.137 2.128 2.126 2.125

CAM-B3LYP 1 Vacuo n-Hexane Toluene CHCl3 DCM EtOH 2 Vacuo n-Hexane Toluene CHCl3 DCM EtOH 3 Vacuo n-Hexane Toluene CHCl3 DCM EtOH

11.8 13.9 14.6 16.1 17.0 17.7 15.0 18.1 19.2 21.7 23.3 24.6 17.2 21.3 22.8 26.8 29.6 32.0

11.5 13.0 13.5 14.6 15.2 15.7 16.4 18.6 19.2 20.6 21.4 22.1 20.8 23.5 24.3 26.3 27.6 28.7

11.5 13.4 14.0 15.1 15.8 16.3 16.5 18.8 19.4 20.8 21.6 22.1 20.7 23.8 24.7 26.5 27.4 28.1

12.6 14.0 14.5 15.5 16.1 16.6 15.9 18.7 19.5 21.2 22.0 22.7 18.7 23.0 24.2 26.7 27.9 28.7

0.012 −0.003 −0.007 −0.018 −0.025 −0.030 0.030 0.012 0.006 −0.009 −0.019 −0.026 0.041 0.023 0.016 −0.003 −0.016 −0.027

1.392 1.400 1.402 1.408 1.411 1.414 1.388 1.397 1.400 1.407 1.412 1.416 1.385 1.393 1.397 1.406 1.412 1.418

−0.027 −0.017 −0.015 −0.011 −0.008 −0.006 −0.006 −0.011 −0.011 −0.009 −0.007 −0.005 0.005 −0.006 −0.008 −0.007 −0.005 −0.003

340.2 337.0 335.7 330.8 327.0 323.2 391.6 395.4 394.9 388.8 381.7 373.9 438.5 452.9 454.1 447.3 435.6 420.8

1.078 1.074 1.071 1.061 1.054 1.047 1.643 1.651 1.650 1.638 1.623 1.607 2.120 2.159 2.168 2.169 2.148 2.115

370.7 361.9 360.6 358.3 357.2 356.5 416.2 416.2 415.7 413.2 411.2 409.9 465.8 472.8 472.8 469.8 467.1 464.8

0.964 1.014 1.022 1.033 1.036 1.037 1.621 1.624 1.624 1.622 1.621 1.621 2.163 2.171 2.170 2.164 2.161 2.160

11.9 14.1 14.8 16.3 17.2 18.0 15.2 18.5 19.6 22.2 23.9 25.2 17.4 21.6 22.2 25.6 30.6 33.1

11.6 13.1 13.6 14.6 15.2 15.8 16.6 18.7 19.3 20.7 21.5 22.2 21.1 23.8 24.6 26.5 27.8 28.9

– 13.5 14.0 15.2 15.8 16.4 16.6 18.9 19.6 20.9 21.6 22.2 21.1 24.1 24.9 26.6 27.5 28.1

– 14.0 14.5 15.6 16.2 16.7 16.1 18.9 19.7 21.3 22.1 22.8 19.2 23.3 24.5 26.9 28.1 28.9

0.011 −0.004 −0.009 −0.021 −0.027 −0.033 0.029 0.010 0.004 −0.012 −0.022 −0.030 0.041 0.022 0.015 −0.006 −0.021 −0.033

1.394 1.402 1.405 1.410 1.414 1.417 1.391 1.400 1.403 1.411 1.416 1.420 1.387 1.396 1.400 1.410 1.417 1.423

– −0.014 −0.012 −0.010 −0.006 −0.001 −0.006 −0.010 −0.010 −0.007 −0.005 −0.002 0.002 −0.006 −0.007 −0.006 −0.003 −0.001

340.3 336.6 335.0 329.5 325.1 320.8 392.0 395.3 394.2 386.2 377.5 368.5 438.0 453.7 454.6 445.0 429.7 411.0

1.065 1.062 1.059 1.047 1.039 1.031 1.633 1.636 1.634 1.619 1.602 1.584 2.118 2.149 2.155 2.149 2.122 2.085

– 364.3 362.8 359.3 358.2 357.3 417.8 417.2 416.4 413.7 412.1 410.4 468.5 475.0 474.7 471.4 468.9 466.4

– 0.984 0.992 1.013 1.017 1.018 1.604 1.609 1.609 1.609 1.611 1.610 2.155 2.157 2.155 2.150 2.150 2.148

Dye B3LYP 1

2

3

wB97XD 1

2

3

Vacuo n-Hexane Toluene CHCl3 DCM EtOH Vacuo n-Hexane Toluene CHCl3 DCM EtOH Vacuo n-Hexane Toluene CHCl3 DCM EtOH

FC FC μ (S0), μ (SFC 1 ), μ (S1), μ (S0 ) – molecular dipole moments in the ground (S0), excited Frank-Condon (S1 ), excited (S1), and ground Frank-Condon states correspondingly.

BLA⁎ values change lesser than the BLA with solvent polarity alternation and are negative in all cases but for dye 3 in vacuo (Table 2). The absolute values of BLA⁎ decreases with the polymethine chain lengthening, in agreement with the conclusion about abatement of the VIs for higher vinylogues. The Mulliken atomic charges on the polymethine chain are redistributed considerably upon excitation, although they do not actually change their sign (Fig. 4). Their values in the fluorescent state S1 remain very near to those in the Frank–Condon state SFC 1 ; the same is true for the states S0 and SFC 0 . One can see also from Fig. 4 that the charge

alternation decreases essentially in the excited state of molecule 3. This result correlates with the abatement of the solvent effects in the fluorescent spectra of the studied dyes. The excited state optimization also gave an alternative explanation of the excited state deactivation of molecules 1–3. The Ci –Cω bond connecting polymethine chain and the malononitrile residue elongates considerably in the S 1 state of dye 1, but this effect decreases for vinylogue 2 and becomes very small for 3. It is possible therefore, that just this bond is the most active in the excited state rotational deactivation of dyes 1–3. Still the Cα –C (1) bond cannot

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

323

compare spectroscopic data and the results of theoretical calculations. Having clear reversed solvatochromism and solvatofluorochromism and at that not comprising carbonyl groups in their acceptor terminus, they allowed to trace more precisely how well the solvatochromic behaviour of donor-acceptor compounds can be modelled by the methods of quantum chemistry. As a vinylogous series, these molecules also were used to see the effects of the polymethine chain lengthening on the physical-chemical properties of merocyanines and compare the gained results with the theoretical models. The quantum chemical simulations of molecules 1–3 in the ground and excited fluorescent states were carried out at the (TD)DFT level using various hybrid functionals. Their results were then juxtaposed with the actual experimental data. Probably due to absence in the studied molecules of functional groups prone to strong site-specific solvation, the solvatochromic effects in their absorption and fluorescence spectra were represented fairly well in these calculations. The better agreement between the experimental results and the quantum chemical calculations was observed when the long-range corrected functionals, CAM-B3LYP and wB97XD, were applied. It has been revealed also that the malononitrile residue in the studied dyes are after all susceptible to electrophilic solvation. This feature has not been observed for malononitrile-based merocyanines so far and is explained by the high dipolarity of the studied probes. It has been shown that the fluorescence quantum yield of the dyes 1–3 depends regularly on the polymethine chain length as well as the medium polarity and viscosity. The analysis of the probable effecting forces to these changes allowed a conclusion that both vibronic interactions and rotations around the polymethine chain bonds are crucial pathways of non-radiative excited state deactivation in these molecules. Acknowledgements This work was performed using computational facilities of the joint computational cluster of State Scientific Institution “Institute for Single Crystals” and Institute for Scintillation Materials of the National Academy of Science of Ukraine incorporated into the Ukrainian National Grid. Fig. 3. The frontier MOs of dye 3 (contour value is 0.03) and dependence of their energy on the solvent polarity (CAM-B3LYP/6-31G (d,p), PCM).

be ruled out completely in this respect, since there exists some initial distortion in this place and the steric factors also could favour its rotation in the excited state.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.saa.2016.08.015. Notes

4. Conclusion The presented here highly-dipolar malononitrile based merocyanines was found to be the convenient reper compounds to

Fig. 4. Alternation of the polymethine chain atoms Mulliken charges (with hydrogens summed into heavy atoms) of dye 3 in the ground (black squares) and fluorescent (red dots) states; CAM-B3LYP/6-31G(d,p), chloroform as a solvent for PCM.

‡ Characterization of dyes 1–3. Dye 1 – 2-[2-(1,3-dibutyl-1,3-dihydro-2H-benzimidazol-2ylidene)ethylidene]malononitrile. Yield 47%. Saffron plates. M.p. = 157–158 °C. 1H NMR (300 MHz; (CD3)2SO): δH 0.91 (6H, t, J = 7.3, 2 × CH3), 1.35 (4H, sex, J = 7.4, 2 × CH3CH2), 1.73 (4H, quin, J = 7.4, 2 × NCH2CH2), 4.27 (4H, t, J = 7.3, 2 × NCH2), 5.44 (1H, d, J = 14.7, H1), 7.35–7.42 (2H, m, 2 × Har), 7.45 (1H, d, J = 14.7, H2), 7.66–7.74 (2H, m, 2 × Har). 13C NMR (100 MHz; CDCl3): δC 13.63, 20.07, 30.45, 45.14, 50.78 (Cω), 80.27 (C1), 109.90, 118.17, 120.76, 124.37, 132.07, 149.20 (Cα), 149.45 (C2). (Visual schemes with detailed attribution of all signals in 13C NMR spectra of dyes 1–3 are given in ESI). Calcd for C20H24N4 (320.43): C, 74.97; H, 7.55; N, 17.48. Found: C, 75.08; H, 7.59; N, 17.37. Dye 2 – 2-[(E)-4-(1,3-dibutyl-1,3-dihydro-2H-benzimidazol-2ylidene)but-2-enylidene]malononitrile. Yield 60%. Dark-wine needles. M.p. = 161–162 °C. 1H NMR (300 MHz; CDCl3): δH 1.06 (6H, t, J = 7.3, 2 × CH3), 1.52 (4H, sex, J = 7.5, 2 × CH3CH2), 1.87 (4H, quin, J = 7.7, 2×NCH2CH2), 4.14 (4H, t, J = 7.7, 2×NCH2), 5.46 (1H, d, J = 14.2, H1), 6.21 (1H, dd, J1 = 13.1, J2 = 12.1, H3), 7.05 (1H, d, J = 13.1, H4), 7.22–7.34 (3H, m, H2 + 2 × Har), 7.34–7.41 (2H, m, 2 × Har). 13C NMR (100 MHz; CDCl3): δC 13.77, 20.18, 30.65, 45.23, 52.11 (Cω), 85.88 (C1), 109.89, 111.06 (C3), 118.77, 121.04, 124.55, 132.26, 147.59 (C2),

324

A.V. Kulinich et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 171 (2017) 317–324

148.93 (Cα), 154.17 (C4). Calcd for C22H26N4 (346.47): C, 76.27; H, 7.56; N, 16.17. Found: C, 76.12; H, 7.55; N, 16.09. Dye 3 – 2-[(2E, 4E)-6-(1,3-dibutyl-1,3-dihydro-2H-benzimidazol-2ylidene)hexa-2,4-dienylidene]malononitrile. Yield 38%. Cobalt blue metalescent needles. M.p. = 168–169 °C. 1H NMR (300 MHz; CDCl3): δH 1.05 (6H, t, J = 7.3, 2 × CH3), 1.51 (4H, sex, J = 7.5, 2 × CH3CH2), 1.88 (4H, quin, J = 7.6, 2 × NCH2CH2), 4.16 (4H, t, J = 7.7, 2 × NCH2), 5.51 (1H, d, J = 14.1, H1), 6.05 (1H, dd, J1 = 13.1, J2 = 11.6, H5), 6.16 (1H, dd, J1 = 13.7, J2 = 12.1, H3), 6.89 (1H, dd, J1 = 13.7, J2 = 11.6, H4), 7.01 (1H, d, J = 13.1, H6), 7.18–7.32 (3H, m, H2 + 2 × Har), 7.32– 7.39 (2H, m, 2 × Har). 1H NMR (300 MHz; (CD3)2SO): δH 0.91 (6H, t, J = 7.3, 2 × CH3), 1.36 (4H, sex, J = 7.5, 2 × CH3CH2), 1.75 (4H, quin, J = 7.5, 2 × NCH2CH2), 4.39 (4H, t, J = 7.4, 2 × NCH2), 5.76 (1H, dd, J1 = 13.6, J2 = 11.9, H5), 6.14 (1H, dd, J1 = 13.4, J2 = 11.4, H3), 6.20 (1H, d, J = 14.9, H1), 6.79 (1H, d, J = 13.6, H6), 6.89 (1H, dd, J1 = 13.4, J2 = 11.9, H4), 7.41 (1H, dd, J1 = 14.9, J2 = 11.4, H2), 7.44–7.51 (2H, m, 2 × Har), 7.76–7.83 (2H, m, 2 × Har). 13C NMR (100 MHz; CDCl3): δC 13.73, 20.15, 30.68, 45.26, 52.99 (Cω), 88.67 (C1), 109.96, 112.65 (C5), 116.60 (C3), 118.84, 121.19, 124.69, 132.34, 147.46 (C2), 148.21 (Cα), 151.89 (C4), 152.97 (C6). Calcd for C24H28N4 (372.51): C, 77.38; H, 7.58; N, 15.04. Found: C, 77.30; H, 7.63; N, 14.98. Dye 3a – 2-[(2E, 4E)-6-(1,3-didodecyl-1,3-dihydro-2Hbenzimidazol-2-ylidene)hexa-2,4-dienylidene]malononitrile. The analogue of dye 3 with improved solubility in low-polarity media was synthesized using the general synthetic procedure. Its UV/Vis and fluorescence spectra practically coincide with those of dye 3. Yield 22%. Greenish-black powder. M.p. = 89–94 °C. 1H NMR (300 MHz; (CD3)2SO): δH 0.84 (6H, t, J = 7.1, 2 × CH3), 1.15–1.35 (36H, m), 1.77 (4H, quin, J = 7.3, 2 × NCH2CH2), 4.16 (4H, t, J = 7.2, 2 × NCH2), 5.77 (1H, dd, J1 = 13.6, J2 = 11.8, H5), 6.06–6.23 (2H, m, H1 + H3), 6.76 (1H, d, J = 13.4, H6), 6.98 (1H, dd, J1 = 13.2, J2 = 11.8, H4), 7.42 (1H, dd, J1 = 14.8, J2 = 11.4, H2), 7.44–7.52 (2H, m, 2×Har), 7.76–7.83 (2H, m, 2 × Har). Calcd for C40H60N4 (596.95): C, 80.48; H, 10.13; N, 9.39. Found: C, 80.25; H, 10.16; N, 9.30. References [1] P. Gregory, in: K. Hunger (Ed.), Industrial Dyes: Chemistry, Properties, Applications, Wiley-VCH, Weinheim 2003, pp. 543–584.

[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]

S.-H. Kim (Ed.), Functional Dyes, Elsevier, Amsterdam, 2006 (650 pp.). D. Gao, A. Li, L. Guan, X. Zhang, L.Y. Wang, Dyes Pigments 129 (2016) 163. Liess, M. Stolte, T. He, F. Würthner, Mater. Horiz. 3 (2016) 72. S. Dähne, Chimia 45 (1991) 288. A.V. Kulinich, A.A. Ishchenko, Russ. Chem. Rev. 78 (2009) 141. G.U. Bublitz, R. Ortiz, S.R. Marder, S.G. Boxer, J. Am. Chem. Soc. 119 (1997) 3365. G. Bach, S. Dähne, in: M. Sainsbury (Ed.), Rodd's Chemistry of Carbon Compounds, vol. IVB, Elsevier, Amsterdam 1997, pp. 383–481. D. Jacquemin, Y. Zhao, R. Valero, C. Adamo, I. Ciofini, D.G. Truhlar, J. Chem. Theory Comput. 8 (2012) 1255. A.E. Masunov, D. Anderson, A.Y. Freidzon, A.A. Bagaturyants, J. Phys. Chem. A 119 (2015) 6807. C. Reichardt, T. Welton, Solvents and Solvent Effects in Organic Chemistry, 4th ed. Wiley-VCH, Weinheim, 2010 (708 pp.). M.C. Rezende, M. Domínguez, J. Mol. Model. 21 (2015) 212. A.V. Kulinich, E.K. Mikitenko, A.A. Ishchenko, Phys. Chem. Chem. Phys. 18 (2016) 3444. D.R. Chapman, C.A. Reed, Tetrahedron Lett. 29 (1988) 3033. M. Strell, W.B. Braunbruck, W.F. Fühler, O. Huber, Liebigs Ann. Chem. 587 (1954) 177. W.L.F. Armarego, C.L.L. Chai, Purification of Laboratory Chemicals, 6th ed. Elsevier, Oxford, 2009 (743 pp.). G. Jones II, W.R. Jackson, C. Choi, W.R. Bergmark, J. Phys. Chem. 89 (1985) 294. R.F. Kubin, A.N. Fletcher, J. Lumin. 27 (1982) 455. R. Sens, K.H. Drexhage, J. Lumin. 24 (1981) 709. A.A. Ishchenko, V.A. Svidro, N.A. Derevyanko, Dyes Pigments 10 (1989) 85. R. Radeglia, G. Engelhardt, E. Lippmaa, T. Pehk, K.-D. Nolte, S. Dähne, Org. Magn. Reson. 4 (1972) 571. A.V. Kulinich, A.A. Ishchenko, U.M. Groth, Spectrochim. Acta A 68 (2007) 6. N.A. Murugan, K. Aidas, J. Kongsted, Z. Rinkevicius, H. Ågren, Chem. Eur. J. 18 (2012) 11677. A.A. Ishchenko, A.V. Kulinich, S.L. Bondarev, V.N. Knyukshto, J. Phys. Chem. A 111 (2007) 13629. K.C. Kreß, T. Fischer, J. Stumpe, W. Frey, M. Raith, O. Beiraghi, S.H. Eichhorn, S. Tussetschläger, S. Laschat, ChemPlusChem 79 (2014) 223. Z.R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 103 (2003) 3899. A.C. Benniston, K.S. Gulliya, A. Harriman, J. Chem. Soc. Faraday Trans. 93 (1997) 2491. C. Benniston, A. Harriman, J. Chem. Soc. Faraday Trans. 94 (1998) 1841. M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, et al., Gaussian 09, Rev. D.01, Gaussian, Inc., Wallingford, CT, 2009. M. Dierksen, S. Grimme, J. Phys. Chem. A 108 (2004) 10225. J.-D. Chai, M. Head-Gordon, Phys. Chem. Chem. Phys. 10 (2008) 6615. J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 105 (2005) 2999. R. Improta, G. Scalmani, M.J. Frisch, V. Barone, J. Chem. Phys. (2007) 127 (074504(19)).