Solvent and substituent effect on intramolecular charge transfer in 5-arylidene-3-substituted-2,4-thiazolidinediones: Experimental and theoretical study

2.1 Materials

Details on materials are given in Supplementary material.

2.2 Synthesis of 2,4-thiazolidinedione, 3-methyl-2,4-thiazolidinedione, diphenylthiourea and 3-phenyl-2,4-thiazolidinedione

Details related to syntheses of 2,4-thiazolidinedione (Jain et al., 2009; Brown, 1961), 3-methyl-2,4-thiazolidinedione (Macháček et al., 1981), diphenylthiourea (Yamin and Yusof, 2003) and 3-phenyl-2,4-thiazolidinedione (Singh et al., 1981) are given in Supplementary material.

2.3 Synthesis of 5-arylidene-3-substituted-2,4-thiazolidinedione

Two series of 5-arylidene-3-methyl-2,4-thiazolidinediones (series 1) and 5-arylidene-3-phenyl-2,4-thiazolidinediones (series 2) were synthesized by the Knoevenagel condensation of corresponding 3-methyl- or 3-phenyl-2,4-thiazolidinediones with the appropriate aryl aldehydes (Rančić et al., 2012).

2.4 Characterization methods

Full details of characterization methods are given in Supplementary material.

2.5 Theoretical calculations

Methods used for theoretical calculation are given in Supplementary material.

2.6 Results of the characterization of series 1 and series 2

¹H and ¹³C NMR, FTIR and elemental analysis data of known compounds 1a–c, 1h, 2a–c, 2h and 2i are given in Supplementary material, while data of newly synthesized compounds are given in detail:

5-benzylidene-3-methyl-2,4-thiazolidinedione (1a). White solid; Yield: 79%, Melting point (m.p.): 130–131 °C (lit. m.p. 133 °C (Yang et al., 2005));

5-(4-methylbenzylidene)-3-methyl-2,4-thiazolidinedione (1b). White solid; Yield: 73%, m.p.: 148–150 °C (lit. m.p. 125–127 °C (Chandrappa et al., 2008));

5-(4-methoxybenzylidene)-3-methyl-2,4-thiazolidinedione (1c). Yellow solid; Yield: 88%, m.p.: 148–150 °C (lit. m.p. 145–147 °C (Yang et al., 2005));

5-(4-bromobenzylidene)-3-methyl-2,4-thiazolidinedione (1d). Yellow solid; Yield: 82%, m.p.: 162–166 °C; ¹H NMR: 3.1 (3H, s, CH₃); 7.5–7.7 (4H, dd, Ph); 7.9 (1H, s,⚌CH—); ¹³C NMR: 28.1 (CH_3, N-CH₃); 115.8 (C5, TZD); 123.4 (C4′, Ph); 128.1; 128.2 (C2′, C6′, Ph); 132.5; 133.0 (C3′, C5′, Ph); 130.4 (C6, ⚌CH—); 134.5 (C1′, Ph); 165.8 (C4, 4-C⚌O); 166.4 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3023 (C—H stretching of phenyl group); 1736 (C⚌O stretching); 1682 (amide band I – C2⚌O and C4⚌O stretching); Elemental Analysis:

Calculated: %C, 44.31; %H, 2.70; %Br, 26.80; %N, 4.70; %O, 10.73; %S, 10.75.

Found: %C, 44.27; %H, 2.73; %Br, 26.75; %N, 4.74; %O, 10.70; %S, 10.81.

5-(4-chlorobenzylidene)-3-methyl-2,4-thiazolidinedione (1e). Yellow solid; Yield: 89%, Melting point: 171–172 °C; ¹H NMR: 3.2 (3H, s, N-CH₃); 7.3 (2H, d, Ph); 7.8 (1H, s, ⚌CH—); 7.4 (2H, d, Ph); ¹³C NMR: 28.0 (CH_3, N-CH₃); 129.5 (C3′, C5′, Ph); 131.3 (C2′, C6′, Ph); 122.1 (C5, TZD); 136.6 (C6, ⚌CH—); 131.6 (C1′, Ph); 132.2 (C4′, Ph); 166.2 (C4, 4-C⚌O); 167.4 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3021 (C—H stretching of phenyl group); 1737 (C⚌O stretching); 1685 (amide band I – C2⚌O and C4⚌O stretching); Elemental Analysis:

Calculated: %C, 52.08; %H, 3.18; %Cl, 13.97; %N, 5.52; %O, 12.61; %S, 12.64.

Found: %C, 51.73; %H, 3.21; %Cl, 14.04; %N, 5.60; %O, 12.70; %S, 12.72.

5-(4-fluorobenzylidene)-3-methyl-2,4-thiazolidinedione (1f). Yellow solid; Yield: 85%; m.p.: 170–171 °C; ¹H NMR: 3.1 (3H, s, CH₃); 7.3–7.4 (2H, m, Ph); 7.6–7.7 (2H, m, Ph); 7.9 (1H, s, ⚌CH—); ¹³C NMR: 28.1 (CH_3, N-CH₃); 115.6 (C5, TZD); 116.6; 117.0 (C3′, C5′, Ph); 131.4 (C6, ⚌CH—); 131.7 (C1′, Ph); 132.7; 132.9 (C2′, C6′, Ph); 160.7 (C4′, Ph); 165.7 (C4, 4-C⚌O); 166.0 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3024 (C—H stretching of phenyl group); 1741 (C⚌O stretching); 1690 (amide band I – C2⚌O and C4⚌O stretching); Elemental Analysis:

Calculated: %C, 55.69; %H, 3.40; %F, 8.01; %N, 5.90; %O, 13.49; %S, 13.52.

Found: %C, 55.70; %H, 3.37; %F, 7.97; %N, 5.93; %O, 13.51; %S, 13.52.

5-(4-dimethylaminobenzylidene)-3-methyl-2,4-thiazolidinedione (1g). Orange solid; Yield: 92%, m.p.: 170–173 °C; ¹H NMR: 3.0 (3H, s, N-CH₃); 3.1 (6H, s, CH₃); 7.3–7.4 (2H, m, Ph); 6.8–6.9 (2H, m, Ph); 7.8 (1H, s, ⚌CH—); ¹³C NMR: 27.8 (CH_3, N-CH₃); 39.8 (CH₃, N(CH₃)₂); 112.3; 117.0 (C3′, C5′, Ph); 113.7 (C5, TZD); 120.0 (C1′, Ph), 132.5 (C2′, C6′, Ph); 134.0 (C6, ⚌CH—); 151.8 (C4′, Ph); 166.3 (C4, 4-C⚌O); 172.6 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3007 (C—H stretching of phenyl group); 1732 (C⚌O stretching); 1672 (amide band I – C2⚌O and C4⚌O stretching); Elemental Analysis:

Calculated: %C, 59.52; %H, 5.38; %N, 10.68; %O, 12.20; %S, 12.22.

Found: %C, 59.49; %H, 5.36; %N, 10.73; %O, 12.19; %S, 12.23.

5-(4-nitrobenzylidene)-3-methyl-2,4-thiazolidinedione (1h). Yellow solid; Yield: 75%, m.p.: 220–222 °C (lit m.p. 222–224 °C (Romagnoli et al., 2013));

5-(3,4-dimetoxybenzylidene)-3-methyl-2,4-thiazolidinedione (1i). Yellow solid; Yield: 91%, m.p.: 172–175 °C; ¹H NMR: 3.3 (3H, s, N-CH₃); 3.5 (3H, s, OCH₃); 3.6 (3H, s, OCH₃) 7.2 (H, d, CH (C₂′)); 7.4 (H, d, CH(C3′)); 7.4 (H, d, CH (C6′)); 8.1 (1H, s, ⚌CH—); ¹³C NMR: 27.8 (CH_3, N-CH₃); 55.9, 56.1 (CH₃, OCH₃); 111.3 (C3′, C5′, Ph); 122.6 (C2′, C6′, Ph); 112.3 (C5, TZD); 148.6 (C6,=CH-); 126.0 (C1′, Ph); 149.2 (C4′, Ph); 166.4 (C4, 4-C⚌O); 177.6 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3019 (C—H stretching of phenyl group); 1735 (C⚌O stretching); 1683 (amide band I – C2⚌O i C4⚌O stretching); Elemental Analysis:

Calculated: %C, 55.90; %H, 4.69; %N, 5.01; %O, 22.91; %S, 11.48.

Found: %C, 56.00; %H, 4.64; %N, 5.00; %O, 22.89; %S, 11.47.

5-benzylidene-3-phenyl-2,4-thiazolidinedione (2a). White solid; Yield: 89%, m.p.: 209–209.8 °C (lit m.p. 208–209 °C (Yang et al., 2008));

5-(4-methylbenzylidene)-3-phenyl-2,4-thiazolidinedione (2b). Yellow solid; Yield: 80%, m.p.: 190–191 °C (lit m.p. 190–191 °C (Yang et al., 2008));

5-(4-methoxybenzylidene)-3-phenyl-2,4-thiazolidinedione (2c). White solid; Yield: 79%, m.p.: 199–201 °C (lit m.p. 200–202 °C (Yang et al., 2008));

5-(4-bromobenzylidene)-3-phenyl-2,4-thiazolidinedione (2d). Yellow solid; Yield: 78%, m.p.: 211–215.5 °C; ¹H NMR: 7.0 (H, s, N-Ph); 7.3–7.4 (4H, m, Ph), 7.5–7.6 (4H, m, Ph), 7.8 (1H, s, ⚌CH—); ¹³C NMR: 121.0 (C5, TZD); 122.1 (C4′, Ph), 128 (C2″, C4″, C6″, N-Ph); 129.2 (C2′, C6′, Ph, C3″, C5″, N-Ph); 130.1 (C2′, C6′, Ph); 131.3 (C1″, N-Ph); 131.3 (C3′, C5′, Ph); 132.6 (C1′, Ph); 134.6 (C6, ⚌CH—); 163.4 (C4, 4-C⚌O); 166.2 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3033 (C—H stretching of phenyl group); 1716 (C⚌O stretching); 1644 (amide band I – C2⚌O i C4⚌O stretching); Elemental Analysis:

Calculated: %C, 53.35; %H, 2.80; %Br, 22.18; %N, 3.89; %O, 8.88; %S, 8.90.

Found: %C, 53.34; %H, 2.78; %Br, 22.20; %N, 3.87; %O, 8.90; %S, 8.91.

5-(4-chlorobenzylidene)-3-phenyl-2,4-thiazolidinedione (2e). Yellow solid; Yield: 81%, m.p.: 203–204 °C; ¹H NMR: 7.3 (H, s, N-Ph); 7.5 (4H, m, N-Ph); 7.5 (2H, m, Ph); 7.6 (2H, m, Ph); 7.9 (1H, s, ⚌CH—); ¹³C NMR: 121.7 (C5, TZD); 127.2 (C2″, C4″, C6″, N-Ph); 129.4 (C3′, C5′, Ph); 129.6 (C3″, C5″, N-Ph); 130.1 (C4′, Ph); 131.4 (C2′, C6′, Ph); 131.7 (C1″, N-Ph); 133.0 (C1′, Ph), 136.8 (C6, ⚌CH—); 165.5 (C4, 4-C⚌O); 166.8 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3020 (C—H stretching of phenyl group); 1738 (C⚌O stretching); 1696 (amide band I – C2⚌O i C4⚌O stretching); Elemental Analysis:

Calculated: %C, 60.86; %H, 3.19; %Cl, 11.23; %N, 4.44; %O, 10.13; %S, 10.15.

Found: %C, 60.88; %H, 3.20; %Cl, 11.21; %N, 4.47; %O, 10.09; %S, 10.15.

5-(4-fluorobenzylidene)-3-phenyl-2,4-thiazolidinedione (2f). Yellow solid; Yield: 82%, m.p.: 165–166 °C; ¹H NMR: 7.0 (H, s, N-Ph); 7.1 (4H, m, N-Ph); 7.5 (4H, m, Ph); 7.8 (H, s, ⚌CH—); ¹³C NMR: 120.82 (C3′_, C5′, Ph); 120.98 (C5, TZD); 124.60 (C2″, C4″, C6″, N-Ph); 127.9 (C3″, C5″, N-Ph); 129.0 (C2′, C6′, Ph); 120.6 (C1′, Ph); 129.3 (C1″, N-Ph); 136.5 (C6, ⚌CH—); 154.9 (C4′, Ph); 169.7 (C4, 2-C⚌O); 171.4 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3052 (C—H stretching of phenyl group); 1723 (C⚌O stretching); 1637 (amide band I – C2⚌O i C4⚌O stretching); Elemental Analysis:

Calculated: %C, 64.20; %H, 3.37; %F, 6.35; %N, 4.68; %O, 10.69; %S, 10.71.

Found: %C, 64.13; %H, 3.40; %F, 6.34; %N, 4.71; %O, 10.70; %S, 10.72.

5-(4-dimethylaminobenzylidene)-3-phenyl-2,4-thiazolidinedione (2g). Yellow solid; Yield: 85%, m.p.: 239–240 °C; ¹H NMR: 3.1 (6H, s, N(CH₃)₂); 6.7 (2H, d, Ph); 7.3 (1H, s, N-Ph); 7.3 (2H, d, Ph); 7.5 (4H, m, N-Ph); 7.9 (1H, s, ⚌CH—); ¹³C NMR: 39.9 (CH_3, N(CH₃)₂); 111.9 (C3′, C5′, Ph); 120.6, 113.7 (C5, TZD); 127.4 (C1′, Ph); 129.3 (C2″, C4″, C6″, N-Ph); 132.6 (C3″, C5″, N-Ph); 135.5 (C2′, C6′, Ph); 136.5 (C6, ⚌CH—); 185.4 (C1″, N-Ph); 147.8 (C4′, Ph); 151.6 (C4′, Ph); 165.6 (C4, 4-C⚌O); 166.2 (C2, 2-C⚌O); IR (KBr, cm⁻¹): 3066 (C—H stretching of phenyl group); 1735 (C⚌O stretching); 1686 (amide band I – C2⚌O i C4⚌O stretching); Elemental Analysis:

Calculated: %C, 66.64; %H, 4.97; %N, 8.64; %O, 9.86; %S, 9.88.

Found: %C, 66.61; %H, 5.00; %N, 8.65; %O, 9.89; %S, 9.85.

5-(4-nitrobenzylidene)-3-phenyl-2,4-thiazolidinedione (2h). Orange solid; Yield: 80%, m.p.: 249–250 °C (lit m.p. 238–240 °C (Yang et al., 2008));

5-(3,4-dimetoxybenzylidene)-3-phenyl-2,4-thiazolidinedione (2i). Yellow solid; Yield: 85%, m.p.: 208–209 °C (lit m.p. 209–210 °C (Yang et al., 2008));

3.1 Spectral properties of 5-arylidene-3-substituted-2,4-thiazolidinediones

It is well known that spectral behavior of molecule, i.e. position, intensity and shape of absorption band, depends on the properties of solvent used. Indeed, a change of solvent is accompanied by a variation of dipolarity/polarizability and hydrogen bonding intermolecular solvent-solute interactions, which means that the study of the solvent effects on the absorption spectra allows assessment of fundamental molecular properties (Reichardt, 2003). A generally accepted concept, concerning the relationship between molecular structure and shift in absorption spectrum, explains that more planar structure supports higher contribution of extended π-electron delocalization which is reflected in bathochromic shift. Based on this, two series of TZDs coupled with substituted benzaldehydes were synthesized, supposing that the obtained π-conjugated systems could show different sensitivities toward solvent and substituent effects. The exocyclic double bond takes Z-configuration, as it was evidenced by higher stability of these isomers (Section 3.3). Two factors could have favorable effect: lower crowdedness and appropriate contribution of intramolecular hydrogen bonding which could exist between C6-H hydrogen and C4⚌O oxygen (Fig. S1) (Luo et al., 2010).

The absorption spectra of the investigated compounds were recorded in nineteen solvents of different properties, and the characteristic spectra are shown in Fig. 1. Two absorbance bands at 250–280 nm and at 320–380 nm, corresponding to π–π^∗ and n–π^∗ (Rančić et al., 2012) transitions, respectively, were distinctly observed. From the presented UV–vis spectra (Fig. 1), a diversity of the spectra could be noticed, and broad band at higher wavelength originates from n–π^∗ transitions. The data from Tables 1 and 2 indicate that values of absorption frequencies, ν_max, depend on substituent and solvent effect, though the absorption bands of both electron-donor and electron-acceptor substituted derivatives appear at similar wavelengths (in the range 10–20 nm). The absorption maxima of 5-arylidene-3-methyl-2,4-thiazolidinediones showed bathochromic shift compared to parent compound in most solvents (except 1f in 2-PrOH, 2ME, 2CE, Et₂O and Pen, and 1g in 2CE; abbreviations are given in Table 1). The introduction of electron-donor substituents produces larger bathochromic shifts suggesting a more pronounced ICT. Low absorption band at ∼400 nm, observed for 1f in AcN, could originate from photoisomeration of more stable Z- to E-isomer (Uchiyama et al., 2003). On the other hand, absorption maxima of 5-arylidene-3-phenyl-2,4-thiazolidinedione showed red shift for all substances, and the largest one was found for dimethylamino derivatives, i.e. 1g and 2g.

Close

Figure 1 Absorption spectra of compounds 1a–1i in acetonitrile (a) and DMSO (b), and compounds 2a–2i in acetonitrile (c) and DMSO (b).

Export to PPT

3.2 Solvent effects on the UV–Vis absorption spectra: Correlation with multi-parameter solvent polarity scales (LSER analysis)

The effect of various types of solvent–solute interactions on the absorption maxima shifts was interpreted by means of Kamlet-Taft and Catalán LSER models. The correlation analysis was carried out using Microsoft Excel software at confidence level of 95%. The goodness of fit is discussed using the correlation coefficient (R), the standard error of the estimate (sd), and Fisher’s significance test (F). The regression values h, s, a, and b (Kamlet-Taft), and c, d, a, and b (Catalán) fit are given in Tables 3–6, respectively.

Presented results indicate that the solvent effects on UV–Vis spectra are complex due to diversity of contribution of both solvent and substituent effects. It can be noticed, from the results given in Tables 3 and 4, that in most cases non-specific solvent effect is a factor of the highest contribution to UV–vis spectral shifts. The negative sign of coefficient s for most compounds, except coefficient s for 2a and 2b (Table 4), and a coefficient for all studied compounds, indicates bathochromic (red) shift with increasing solvent dipolarity/polarizability and hydrogen-bond donating capability. This suggests better stabilization of the electronic excited state relative to ground state with increasing solvent polarity and solvent acidity. The highest absolute values of coefficient s were found for compounds 1g and 2d, and for compounds 1g and 2b regarding coefficient a. The positive sign of coefficient b for most compounds (Tables 3 and 4), except for 1g and 2i, indicates a hypsochromic (blue) shift relative to increasing solvent hydrogen-bond accepting capability. This suggests better stabilization of the ground state relative to the excited state. The highest absolute values for coefficient b were found for compounds 1e and 2b.

Results of the quantitative separation of the non-specific solvent effect into polarizability and dipolarity term (coefficients c and d, respectively), performed by using Catalán equation, are given in Tables 5 and 6.

The results obtained by the use of Catalán equation provide better understanding of attractive/repulsive solvent/solute interactions, and enable an estimation of their appropriate contribution to ν_max shift. The correlation results, given in Tables 5 and 6, imply that the solvent polarizability is the principal factor influencing the shift of ν_max, whereas solvent dipolarity, acidity and basicity have moderate to low contribution. These results are in accordance with previous study related to 5-arylidene-2,4-thiazolidinediones (Rančić et al., 2012). Negative values of the coefficient c, except for compound 1i, indicate higher contribution of the polarizability effect to the stabilization of the excited state. The noticeable polarizability indicates that the overall π-electronic density shifts take place by two cooperative mechanisms: overall π-network and polarization of localized π-unit (Fig. S1). Interaction of 3-substituted-2,4-thiazolidindione moiety (TZD), which act as an electron-acceptor, with the substituents of different electronic properties causes variation in the mobility of the π-electrons, and thus, a wide range of coefficient c values were found. The sensitive balance of both electronic substituent effect and π-electronic interaction in TZD moiety contributes to highest polarizability in compound 1h (nitro substituted derivate) in series 1, and compound 2i in series 2.

A greater extent of resonance interaction is operative within π₁- and π₂-resonance units contributing to higher polarizability of π-conjugated system (Fig. S1). In that context, dimethylamino substituent exerts larger π-electron delocalization supported by electron-accepting TZD moiety (similar c values were found for both 1g and 2g). The phenyl ring present at N3 position in series 2 participates in electron delocalization over TZD which causes electron-density shift to electron-accepting nitro substituent present contributing to lower polarizability, i.e. coefficient c of −2.00 was obtained for 2h (Table 6).

Effect of solvent dipolarity, assigned with d term, is of lower significance and showed complex behavior, with the highest absolute values found for compounds 1d and 2i. Dipolar solvent–solute interactions lead to blue shift when a moderate and strong electron-acceptor is present in the arylidene part of the molecule in series 1. Higher contribution of the solvent dipolar effect, in compounds with substituent displaying low/moderate effect, could be due to balanced contribution of two opposite effects: electronic effect of the substituent and TZD moiety that cause separation of charges by creation of dipolar structure differently oriented in space. The better stabilization of the excited state was found mostly for electron-donor substituted compounds, including 2h, which indicates that directional shift of electronic density, causes more effective charge separation.

Specific solvent–solute interactions realized through hydrogen bonding, i.e. HBD effect can be attributed mainly to the carbonyl groups of the TZD moiety, while solvent basicity (HBA) could be established with generated positive charge at molecular structure. Negative values of the coefficient a found in both series 1 and 2, except 2f, 2h and 2i, indicate moderate to low contribution of the solvent acidity to the stabilization of the excited state. Long-range transmission of substituent effects, which supports the larger polarization of both carbonyl groups, enhances hydrogen bond accepting capabilities of carbonyl oxygens. These observations indicate that solvent achieves the higher extent of hydrogen bonding at nucleophilic sites in TZD moiety.

3.3 LFER analysis of UV–vis data

In an attempt to assess substituent effects on ICT of the investigated compounds, the principles of liner free energy relationships (LFER) were applied to the UV–vis spectral data by using Hammett Eq. (3). The correlation results are given in Tables 7 and 8.

The correlation results (Tables 7 and 8), obtained separately for electron-donating and electron-accepting substituents, reflect different transmission modes of electronic substituent effects through differently oriented π-electronic units (Fig. S1). Obviously, the UV–vis absorption of studied compounds is more substituent–dependent than solvent–dependent (Tables 7 and 8). The introduction of an electron-donor substituent enhances the negative solvatochromism of both series 1 and 2 (Tables 7 and 8, respectively), and also results in a higher susceptibility of the absorption maxima shifts to electronic substituent effects. Somewhat higher sensitivity of ν_max to substituent effect was found in solvent with higher proton-donating ability. Significantly lower sensitivity was found for electron-acceptor substituted derivatives in both series 1 and 2. Such behavior was explained by destabilization of the TZD moiety by electron-accepting groups (Rančić et al., 2012, 2013), whereas the introduction of either methyl or phenyl group at TZD contributes to low sensitivity to solvent effects. The lowest sensitivities of ν_max to substituent effects in nonpolar solvents (pentane and heptane) can be explained by the fact that, at low relative permittivity of surrounding medium, the energy necessary to bring about charge separation in the excited state is relatively high, which gives rise to a lower susceptibility to substituent effects.

3.4 LFER analysis of NMR data

Furthermore, an analysis of the experimental and calculated ¹³C NMR chemical shifts has been performed to get a better insight into the influence of the substituent effects on SCS shifts. The calculated ¹³C NMR chemical shifts, obtained by the use of GIAO/WP04/aug-cc-pVDZ basis set, are given in Tables S4 and S5. Results of the intercorrelation of experimental and theoretical NMR data are good which indicates that adequate methodology was applied for calculation of theoretical NMR data (Figs. S4–S7). The general conclusion derived from the ¹³C NMR chemical shifts is that all substituents influence chemical shifts of the carbon atoms of interest (C5, C6, C1′) via their electronic effects. LFER principles were applied in the analysis of the substituent effect, defined by σ_p or ${\sigma }_p^{+}$ substituent constants, on the SCS of the carbon atoms of interest. The best correlation results obtained for C5, C6 and C1′ carbons are presented in Table 9.

The observed ρ values for these carbons indicate different susceptibilities of the SCS to substituent effects, and it is apparent that chemical shifts of C5 and C1′ showed normal substituent, while reverse substituent effect was observed at C6 for both series 1 and 2. The negative sign of reaction constant, ρ, for C6 atom means reverse behavior, i.e. the value of SCS decreases although the electron-withdrawing ability of the substituents, measured by σ_p, increases. Similar results were obtained in a previous study related to 5-arylidene-2,4-thiazolidinediones (Rančić et al., 2013) (Tables S6 and S7). The reverse substituent effect operative at C6 can be attributed to localized π-polarization, which predominates over the extended π-polarization (Fig. S8). The π-polarization concept was introduced by Reynolds (Craik and Brownlee, 1983; Reynolds et al., 1983) to explain substituent field effect operative in the side chain of para- and meta-substituted benzenes. The field effect, induced by substituent dipole, causes secondary polarization of π-electrons in the subsequent independent π-electronic system without net π-electron transfer. According to Reynolds (Craik and Brownlee, 1983; Reynolds et al., 1983), the polar effect mainly arises as a result of the substituent dipole induced field effect, and this effect alters the electron density at C5 by two mechanisms: (i) field-induced polarization of the side chain vinyl group (localized or direct π-polarization) (Fig. S8a), and (ii) field-induced π-electron transfer (extended π-polarization) (Fig. S8b) (Rančić et al., 2013). The second term is the major effect mainly operative in planar systems. Similar influence of substituent from the meta- and para-positions on C5 supported Reynolds' conclusion that the polar effect is rather of field than of inductive origin (Craik and Brownlee, 1983; Reynolds et al., 1983). The resonance interaction in the extended conjugated system of the substituted styrene molecules in the presence of electron-acceptor substituent has complementary effect to the polarization mechanism (Fig. S8c), and opposite is true for electron-donor substituted compounds.

It can be noticed, from Table 9, that correlations with electrophilic substituent constants, σ_p⁺, are of better quality for C6 and C1′ atoms in series 1, and C5 and C1′ atoms in series 2. It indicates higher contribution of extended resonance interaction for these carbons. Introduction of the methyl or phenyl group in series 1 and 2, respectively, decreases susceptibility of the C5 and C1′ atoms to substituent effects in comparison with previous result (lower ρ) (Rančić et al., 2013), while increased susceptibility of the C6, for both series 1 (−7.03), and series 2 (−3.32; Table 9), was found. Sensitivity of SCS to substituent effects is mainly determined by three factors: conformational arrangement, substituent effects and introduction of methyl or phenyl group at N3-position of TZD which contributes to specific shift of electronic density at the carbon atom of interest.

To evaluate separate contributions of the polar (inductive/field) and resonance effects, the regression analysis according to DSP Eq. (4), i.e. Reynold’s equation, with σ_F and σ_R^o substituent constants has been performed and the results are given in Table 10.

The success of the fitting is not improved by using DSP equation, but it helps in better understanding of the transmission mode of substituent effect. The positive ρ_F and ρ_R values were obtained for the C5 and C1′ atoms, while negative values are obtained for C6 carbon. The positive values of ρ_F and ρ_R for the C5 atom and negative values for the C6 atom show that substituents in a different way influence the π electron density at carbon atoms in a side chain. The negative ρ_F value for the C6 atom of both series is an indication of the opposite substituent effect, the so-called reverse substituent effect (RSE), which is due to π-polarization. The π-polarization concept can also explain the larger magnitude of polar effects at C6 in series 1 (ρ_F = –8.69), which is related to the strength of the resonance interaction between the para-carbon (C1′) and para-substituent present. All λ values are higher than 1, except for C6 atoms in series 1 and 2, which means that resonance effect predominates over field effect, and highest effect was found for C1′ carbon. The extent of resonance interaction depends on spatial arrangement of the molecules, i.e. the values of torsional angle θ, and thus, the resonance substituent effect is most effectively transmitted to C1′ carbon.

Value of the term ρ_R, i.e. the effect of π-electron delocalization, and the term ρ_F, attributed to the effect of substituent induced π-polarization, significantly depends on the electronic system present at C5 carbon. The ρ_F value for C5 and C6 atoms reflects mainly field effect, which is the most pronounced in series 1 when methyl group is present at N3 atom. Introduction of phenyl ring at N3-position compensates for electron-accepting power of TZD by increasing π-electron density, and causing lower extent of extended π-electron delocalization (ρ_R = −2.92 and −5.20 for series 2 and 1, respectively; Table 10).

The electron-accepting carbonyl group present in TZD moiety tends to decrease electron density at the vinyl side chain resulting in strong resonance interaction with para-electron–donating substituents. Electron-donors cause the increase of electron density in the arylidene ring which, therefore, acts as π-donor and causes increase in resonance interaction with the vinyl side chain. Synergism of these effects causes pronounced electron density changes at the C5 atom. Oppositely, in the case of electron-accepting substituent, balanced and oppositely directed interactions of two electron-accepting moieties determine such behavior where both parts exert appropriate response to electronic demand of the rest of molecule, and the consequence is the higher extent of localized π-polarization at C5 atom. The net result is that the electron-accepting substituents increase the electron density on the C6 atom due to π-polarization.

3.5 Comparative LFER analysis of SCS for C5 and C6 in styrene series 1–10

The study of the mechanism of transmission of electronic substituent effects in an aromatic framework with a side chain, i.e. β-substituted styrene derivatives (Fig. S9) is very useful in order to distinguish the contribution of polar (field/inductive) and polarization effects (Craik and Brownlee, 1983) at carbon of interest. The substituent dipole polarizes the vinyl π-electron resulting in separation of opposite charges at neighboring atoms which depends on heterocyclic system attached at C5 carbon. In order to get deeper insights into factors determining such phenomena, a comprehensive analysis of the electronic and structural effect was performed for series 1 and 2 and compared with results obtained for eight styrene series 3–10 (Fig. S9) (Rančić et al., 2013; Reynolds et al., 1983; Hamer et al., 1973a, 1973b; Tan et al., 1987; Krabbenhoft, 1978; Cornelis et al., 1981). Saleh et al. (2008) have studied 19 styrene series and concluded that different groups (Fig. S9) of the side chain have significant effects on SCS of C5 carbon. Due to this selection of a wide range of structurally different groups present at C5 atom in series 1–10 offered a good base of data used in a systematic LFER analysis.

Results of LFER analysis for the C5 and C6 atoms of series 3–10, in the form of SSP equation with σ_p or σ_p⁺ constants, are presented in Tables S8 and S9, respectively (Rančić et al., 2013). Normal and reverse substituent effects were observed at C5 and C6 atoms, respectively, for all studied compounds. Two groups of correlation results could be recognized for the C5 atom: series 3–6, including series 1 (9.94; Table 10), and series 7–10, including series 2 (6.63; Table 10). Values of the coefficient ρ for C6 are similar for series 2–10 with the highest value of −3.34 found for series 9 (Table S9). Exceptionally higher ρ values of −7.03 for series 1 were found (Table 9). The correlation results for the C5 atom are significantly improved when electrophilic substituent constant σ_p⁺ is used in correlation for series 1 and 4–9, while this does not apply for the series 2, 3 and 10. The terminal C5 atom in vinyl group suffers from electron deficiency, and thus SCS in series 1 and 4–9 better correlates with σ_p⁺. The correlation results are of lower quality for the correlation for C6 SCS with σ_p⁺ indicating lower contribution of the extended resonance interaction.

Previous results of LFER analysis for series 3–10, obtained by using DSP model are given in Tables S10 and S11 (Rančić et al., 2013). The magnitude of resonance coefficient, ρ_R, for C5 is twice (series 5) or more than twice larger (series 3 and 7–9) than the field effect emphasizing the significance of resonance interaction at C5 atom. It is worth of noting a significantly lower contribution of ρ_R, i.e. lower λ values for C6 of series 4–8 and 10. The number of systematic studies also revealed higher sensitivity of SCS at C5 due to the greater contribution of resonance effects (Craik and Brownlee, 1983; Reynolds et al., 1983; Hamer et al., 1973a, 1973b; Tan et al., 1987; Krabbenhoft, 1978; Cornelis et al., 1981; Saleh et al., 2008). The Reynolds ρ_F values for the C5 atom reflect the pure field effect which is most pronounced in series 3. Delocalization of the lone electron pair in amide group (structure I; Fig. S10) contributes to an excess of electron density and, hence, the sensitivity of the C5 atom to the field effect is increased. Competitive resonance interaction (structures I and II; Fig. S10) participates in a different extent in SCS change at C5 and C6 depending on substituent present. Electron-donors support increased contribution of structure II (Fig. S10), oppositely, electron-acceptors suppress to some extent extended π-delocalization favoring amide resonance (structure I; Fig. S10). Introduction of methyl group, present in series 1, exerts negligible effect by supporting amide group resonance. In the other hand, the presence of phenyl ring in series 2 contributes to aniline type n,π-resonance interaction supporting resonance presented by structure II (Fig. S10).

The structural similarity of series 3 and 9 offers a possibility to compare correlation results in relation to structural difference of heterocyclic structure attached at C5 atom. The correlation results for the C5 and C6 atoms (Tables S10 and S11, respectively) indicate that, not only electronic properties of heteroatoms, but also geometrical characteristics, i.e. planar hydantoines (Tan et al., 1990), and non-planar 5-arylidene-2,4-thiazolidindiones (Table S12) influence their specific behavior. Moreover, similar λ values for the C5 (2.09 and 2.82) and C6 (1.46 and 1.14) atom for series 3 and 9, respectively, were found, while larger differences were observed with respect to other series. In hydantoines there is a competitive contribution of two resonance structures (Fig. S11) depending on substituent present. According to correlation results, N1 lone pair delocalization contributes to lower ρ_R at C5 (7.52 versus 17.03 for series 9 and 3, respectively; Table S10). Anyway, resonance interaction at C5 is of high contribution (λ = 2.82; Table S10), indicating significant π,π-delocalization within heterocyclic moieties. Comparing the series 6–8 to series 4 and 5, it can be noted that strong electron-accepting substituents, chloro and cyano group, enhance sensitivity of C5 carbon to resonance effect (Table S10), while two carbonyl groups adjacent to the C5 atom in series 6–8 exert appreciable extent of electron-accepting properties. Correlation results for series 6–8 show a decrease in field and resonance effect for C5 (Table S10), while the λ values are high with decreasing order from series 4 to series 6. This order indeed reflects ability of electronic density at the C5 carbon to interact with two neighboring carbonyl groups which in turn might competitively resonate with lone electron pair at oxygen (series 4) or nitrogen (series 5) and phenyl π-electrons (series 6).

A more complex situation was found for C6 because this carbon senses the electronic effect of the substituted phenyl ring and heterocyclic moiety; hence, no regularity of correlation results was found for this carbon. There are some similarity regarding ρ_F values (Table S11), with the highest ones found for series 1, 2, 7 and 8. However, low λ values, obtained for all series (λ < 1, except for series 3 and 9), indicate a lower contribution of resonance effect which is likely due to balanced/opposed effect of the substituent and heterocyclic moiety resulting in the higher contribution of the π-polarization at this carbon.

Generally, it might be concluded that, according to results of LFER studies (Tables 9 and 10 and S8–S11), transmission of substituent electronic effects through defined π-resonance units indicates that these units, behave either as isolated or as conjugated fragments depending on substituents present in the corresponding molecule. The significance of extended π,π-delocalization through defined π-resonance units of series 1 (Scheme 1) is mainly operative within π-resonance units with high contribution of π-polarization at C5 carbon. The opposite is true for C6 atom for series 2–8 and 10 where significant π,π-delocalization is observed within heterocyclic moieties. Also, Brownlee et al. (Craik et al., 1982) concluded that the differences in geometry affect the ratios of the ρ_R values for different styrene series, and not the ratio of the ρ_R values for the C5 and C6 carbons in the same series.

3.6 MP2 and TD-DFT calculations. Evaluation of electronic transition and ICT

Theoretical calculations can complement experimental investigations, i.e. to validate experimental conclusions, or even upgrade with some confidence into experimental unexplored area. In that context, as first step, it was necessary to perform optimization of the geometries of the investigated molecules by the use of ab initio MP2 method. The significant stability of the Z isomer was obtained, and energies related optimized geometries are presented in Table S13, and optimized structures of series 1 and 2, in acetonitrile, are given in Figs. S11 and S12, respectively. Similar results were obtained for gas phase. Elements of the optimized geometries of investigated compounds, (Z)-5-arylidene-3-methyl-2,4-thiazolidinediones (series 1) and (Z)-5-arylidene-3-phenyl-2,4-thiazolidinediones (series 2), obtained by the use of MP2/6-31G(d,p) method, are given in Tables S14 and S15.

It can be seen that geometric features of the studied TZDs of both series follow the similar trend related to geometry parameter change versus substituent effect. Consideration of the torsional angles θ (Fig. S1), defined by out-of-plane rotation of the phenyl ring with respect to the exocyclic double bond, and their changes (Tables S14 and S15) indicates that introduction of substituent induces decreasing the θ values. Similar trend was found for 5-arylidene-2,4-thiazolidinediones (Table S8) (Rančić et al., 2013). The introduction of electron-donor substituent causes increase of N3—C4 and both carbonyl groups’ bond lengths, while, oppositely, decrease of C4—C5, N3—CH₃, as well as N3—Ph, C6—C1′ and C2—N3 bond lengths was found. The most pronounced effect was found for dimethylamino group which supports an electron density shift from the π₁-unit to the electron-accepting TZD moiety contributing to higher planarization of molecule and greater extent of the π,π-delocalization. This result is an additional support for the extended conjugation operative in the π₁- and π₂-unit, i.e., the π-electron density shifts toward C4⚌O carbonyl group causing increase in both carbonyl groups bond lengths.

The opposite is true for electron-acceptor substituted compounds. Higher electron-acceptor ability of the substituent causes increase in θ values and C4—C5 bond length which indicates that contribution of two opposite effects: electron-accepting arylidene group and TZD carbonyl groups cause geometry adjustment together with low sulfur positive resonance participation as a response to electronic demand of the electron deficient environment. The normal carbonyl groups’ polarization is suppressed and causes a slight bond length decrease.

Furthermore, it was shown that interpretation of the UV–Vis spectra was achieved by the use of TD-DFT quantum chemical calculations given on 1-(1-naphthoyl)-3-(halo-phenyl)-thioureas (Saeed et al., 2015). Mechanism of electronic excitations and changes in the overall charge distribution in both ground and excited states of the investigated molecules were studied by calculation of HOMO/LUMO energies (E_HOMO/E_LUMO) and E_gap values in solvent acetonitrile. Obtained results are presented in Figs. 2 and 3, and Table S16. Generally, E_gap values were lower than for unsubstituted compound in both series, and also lower values were obtained for series 2 with the lowest ones found for compounds 1g and 2g. No significant ICT could be noticed from the structure of frontier orbitals, but it could be observed similarity of orbital independently on the presence of methyl or phenyl group in N3-position of studied compounds. Unsubstituted compounds have the largest energy gaps of 6.582 eV (1a) and 6.542 eV (2a), while the lowest energy gaps were observed for compounds 1g and 2g with 5.776 and 5.727 eV, respectively. This is consistent with the largest bathochromic shift of their absorption maxima. Obviously, the ICT process is more feasible in more planar compounds 1g and 2g due to higher polarizability of excited state which can be, under influences of surrounding solvent environment, preferentially stabilized with respect to the ground state (Table 11). Due to this, the excited and ground states are closer and more rapid internal conversion is permitted. Variation of substituent patterns clearly indicates that contributions of both structural and substituent/compound donor-acceptor character are involved in the ICT mechanism of the investigated molecules. Additional results of TD-DFT calculations, oscillator strengths, vertical excitation energies and electronic transitions are provided in Table 11, since the TD-DFT results have indicated a large contribution of single HOMO to LUMO excitations in ground to first excited state transition (greater than 90% for all compounds; Table 11). Only in the case of nitro substituted compounds (1h and 2h), there is small participation of HOMO to LUMO + 1 excitation (<5%). In conclusion, all compounds showed bathochromic shift compared to unsubstituted compounds in solvent acetonitrile for both series, and the lowest E_gap values were observed for compounds 1g and 2g form. These results are consistent with experimental ones.

Close

Figure 2 The HOMO/LUMO orbitals and E_gap of compounds 1a–1i in acetonitrile.

Export to PPT

Close

Figure 3 The HOMO/LUMO orbitals and E_gap of compounds 2a–2i in acetonitrile.

Export to PPT

ICT occurs when photon absorption promotes a significant deformation of the electronic cloud due to appropriate (partial/total) electronic shift from one moiety of a molecule to another. In order to quantify this phenomenon, a simple and efficient model was used for definition/evaluation of ICT on the basis of the overall charge distribution computed and changes in the both ground and excited states (Le Bahers et al., 2011; Jacquemin et al., 2012; Ciofini et al., 2012). TD-DFT method was used for quantification of the ICT by calculations of the charge-transfer distance (D_CT), amount of transferred charge (Q_CT), and the value of variation of dipole moment between the ground and excited states (μ_CT). Excitation of a molecule by a photon may cause a transfer of charges from one part to another part of a molecule leading to conformational changes in the excited state, which result in a change of dipole moment in the excited state compared to ground state.

Practically, the knowledge of the excited state dipole moment is quite useful in designing nonlinear materials or allows one to estimate the site of electrophilic or nucleophilic attacks in some photochemical reactions. Results of calculated distance between two barycenters of the density depletion and the density increment zones (D_CT), Q_CT and μ_CT change during electron excitation are presented in Table 12. In general most of ICT take place by electron density shift from TZD to arylidene part of molecule during electronic transition. It is noticeable that the strongest ICT occurs in 1g and 2g, i.e. dimethylamino substituted compound, showing transfer of 0.567 e⁻ over 2.939 Å and 0.566 e⁻ over 2.980 Å, respectively. On the other hand, in molecules 1h and 2h nitro substituted derivatives, intramolecular charge transfer of 0.737 e⁻ over 0.989 Å and 0.736 e⁻ over 0.988 Å, respectively, is an evidently local process which occurs within nitro substituent (Figs. 4 and 5).

Close

Figure 4 ICT processes in compounds 1a–1i; Left images – difference between densities in excited and ground state (red and blue – density increase and decrease upon transition, respectively); Right images – positions of barycenters describing charge loss (cyan circle) and charge gain (violet circle) upon transition.

Export to PPT

Close

Figure 5 ICT processes in compounds 2a–2i; Left images – difference between densities in excited and ground state (red and blue – density increase and decrease upon transition, respectively); Right images – positions of barycenters describing charge loss (cyan circle) and charge gain (violet circle) upon transition.

Export to PPT

The variation of substituent patterns indicates that contributions of both conformational arrangement and donor–accepting character are involved in the ICT mechanism of the investigated molecules (Figs. 4 and 5).

MEP (molecular electrostatic potential) analysis was used to evaluate and visualize charge distribution over investigated compounds, and illustrate the three dimensional charge distributions overall investigated molecules. MEP potential at a point in space around a molecule gives information about the net electrostatic effect produced at that point by total charge distribution (electron + proton) of the molecule and correlates with dipole moments, electro-negativity, partial charges and chemical reactivity of the molecules. It provides a visual method to understand the relative polarity of the molecule (Reynolds et al., 1983). An electron density isosurface mapped with electrostatic potential surface depicts the size, shape, charge density and site of chemical reactivity of the molecules. MEP shown in Figs. S13 and S14 illustrates the three dimensional charge distributions overall investigated molecules. The values of the electrostatic potential at the surface are represented by different colors; red represents regions of most electronegative electrostatic potential, it indicates the region of high electron density, i.e. sites favorable for electrophilic attack (favorable site for HBD solvent to solute interactions); blue represents regions of the most positive electrostatic potential, i.e. region of low electron density favorable for nucleophilic attack (preferential solvent/solute HBA interactions), and green represents regions of zero potential. Potential increases in the order red < orange < yellow < green < blue. The blue color indicates the strong attractive potential, region favorable for HBA solvent interactions, while red color indicates the repulsive potential, and includes sites favorable for HBD solvent interactions. As can be seen from the MEP map of the molecules, negative regions are mainly localized over the carbonyl groups from the TZD moiety and over the nitro and metoxy substituents in the compounds 1h, 1i, 2g and 2i (Figs. S13 and S14). The positive region is localized on the methyl and phenyl groups attached on nitrogen atom in TZD moiety. This effect is the least pronounced in molecules 1g and 2g with strong electron-donating dimethylamino substituent is present at phenyl ring causing shifting electronic density to electro-accepting TZD ring. Most of HBA capabilities of TZD molecules could be assumed to be from 2,4-dioxo groups as strong electron-acceptor groups that cause increase in electron density at these two sites creating favorable interaction with proton-donating solvents. As can be seen from the MEP map of the compounds, the regions having the negative potential are over the electronegative atoms, the regions having the positive potential are over the methyl and phenyl groups attached to nitrogen atom and the remaining species are surrounded by zero potential.

The experimental and theoretical study implies that the solvatochromic properties are the consequence of the overall effect of the molecule geometry influenced by the electronic substituent effects transmitted through π-conjugated systems and results obtained in this study help in assessing the potential application of the investigated compounds in different homogeneous media.