The conformational analysis of push–pull enaminoketones using Fourier transform IR and NMR spectroscopy, and quantum chemical calculations: II. β-Dimethylaminoacrolein

Spectrochimica Acta Part A 74 (2009) 1010–1015

Contents lists available at ScienceDirect

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

The conformational analysis of push–pull enaminoketones using Fourier transform IR and NMR spectroscopy, and quantum chemical calculations: II. ␤-Dimethylaminoacrolein Sergey I. Vdovenko ∗ , Igor I. Gerus, Elena A. Fedorenko Institute of Bioorganic Chemistry and Petrochemistry, National Academy of Sciences of Ukraine, Str. Murmanska 1, 02094 Kiev, Ukraine

a r t i c l e

i n f o

Article history: Received 31 March 2009 Received in revised form 4 June 2009 Accepted 11 August 2009 Keywords: NMR and IR-spectra Conformational analysis ␤-Dimethylaminoacrolein Quantum chemical calculations

a b s t r a c t IR Fourier and 1 H NMR spectra of ␤-dimethylaminoacrolein (DMAA) were investigated in various pure solvents. Quantum chemical calculations by the method AM1 also was carried out to evaluate relative energy and dipole moment of each conformer. On the basis of NMR and IR-spectra we showed that the (DMAA) presented in solutions as equilibrium of two conformers, (E-s-Z)  (E-s-E). Constant of this equilibrium, Keq = C(E-s-E)/C(E-s-Z), depended strongly on the total (DMAA) concentration: ln Keq = ln Keq0 + a(1 − e−bCtotal ). Besides, (E-s-Z) conformer of the (DMAA) was more polar and more stable than the (E-s-E) conformer. Correlation of the out-of-phase ˜ (C O) and in-phase ˜ (C C) vibrations with solvatochromic parameters of Kamlet, Abbot, and Taft (KAT) revealed that the main contribution to the shift of the out-of-phase ˜ (C O) vibrations of the both conformers made solvent’s hydrogen bond acceptor (HBA) (ˇ) term, whereas hydrogen bond donor (HBD) acidity (˛) term influenced predominantly on the shift of the in-phase ˜ (C C) vibrations of the conformers. Moreover, influence of these dominated terms was more pronounced for the (E-s-Z) conformer in comparison with the (E-s-E) conformer, hence the first conformer was more polarized than the last. Investigations of the enthalpies of the (E-s-Z)  (E-s-E) equilibrium in carbon tetrachloride, 1,4-dioxane and their mixtures showed that these enthalpies depended predominantly on the solvent’s atomic and electronic polarization and dispersive interactions. © 2009 Published by Elsevier B.V.

1. Introduction It is well known that enaminoketones (R1 CO–CR2 CR3 –NR4 R5 ) are typical push–pull ethylenes, with low energy barriers around the formal carbon–carbon double bond, and increased energy barriers around the O C C and the C N single bonds which allow the existence of several stereoisomeric forms. Literature concerned to this problem is voluminous (see, e.g. [1–12]), whereas the data available for the simplest representative of this class of compounds, namely ␤-dimethylaminoacrolein (R1 H) are restrained and conflicting [3,5,14]. Particularly, authors [14] showed that ␤-dimethylaminoacrolein exists, presumably, as the (E-s-E) conformer (in CDCl3 at −100 ◦ C). Moreover, authors [5] evaluated percentage of this conformer (86%) in CF3 COOH at very low temperature. All these data were obtained with NMR method in polar solvents at various low temperatures. At the same time FT-IR spectroscopy enables to register reliably various stereoisomeric forms of the push–pull enaminones [2,4,10,11,15,16]. Enaminoketones of general formula R1 CO–CR2 CR3 –N(CH3 )2 are known to exist in the

∗ Corresponding author. Tel.: +380 44 573 25 93; fax: +380 44 573 25 52. E-mail address: [email protected] (S.I. Vdovenko). 1386-1425/$ – see front matter © 2009 Published by Elsevier B.V. doi:10.1016/j.saa.2009.08.034

following stereoisomeric forms: (Z-s-Z), (Z-s-E), (E-s-Z), and (E-sE) [3,13]. The first letter in this labelling denotes isomer of the studied compound, in other words it shows mutual orientation of dialkylamino group and carbonyl group relative to C C double bond. The conformational possibilities for carbonyl containing compounds are given by the rotation of the carbonyl group around the C–C single bond with the carbonyl group oriented away or towards the double C C bond what is denoted by the second capital letter E or Z, respectively. As dimethylamino group is symmetrical relative to C N bond, so the rotation of this group does not influence on molecular conformation. All possible stereoisomeric forms are presented in Fig. 1. Recently [15] we investigated structure of ␣,␤-unsaturated enaminoketones (R1 CH3 , CF3 ) and showed that the conformational composition of these enaminoketones is strongly dependent on R1 substituent. Earlier [16] we have shown that the reactivity of similar push–pull systems, namely, ␤alkoxyvinyl methyl ketones is closely sequestered with geometry of enone. Hence, the stereochemistry of enaminone can be controlled to some extent by equilibrium of stereoisomeric forms. From this point of view it was reasonable to investigate the stereochemistry of ␤-dimethylaminoacrolein and solvent influence on it. Thus, in this work we studied the stereoisomeric forms of ␤dimethylaminoacrolein (DMAA) in various solvents at ambient

S.I. Vdovenko et al. / Spectrochimica Acta Part A 74 (2009) 1010–1015

Fig. 1. Possible stereoisomeric R1 CO–CR2 CR3 –NR4 R5 .

forms

of

the

enaminoketones

temperature and evaluated equilibrium constants and thermodynamic parameters of equilibrium of these forms. To our knowledge no such study has been carried out with FT-IR method until this time and therefore we have decided to present basic experimental and theoretical information about isomers and conformers of the mentioned compound and the factors influencing on the equilibrium of the (DMAA) stereoisomers. 2. Experimental

1011

0.0022, 0.0063 0.01028, 0.0212, 0.0521, and 0.102 cm (for dilution measurements). Temperature measurements were carried out in thermostated (±0.05 ◦ C) NaCl cell with pathlength 0.00875 cm. The solutions were scanned at the same conditions as a background. The Bruker Opus software Version 6.0 was used for all data manipulation. In IR-spectra of the (DMAA) in nonpolar solvents (n-hexane, chexane, and methylcyclohexane) two bands ˜ (C O) with maxima, respectively, at 1677 and 1671 cm−1 were observed, whereas in polar solvents the bands ˜ (C O) were strongly overlapped, but the analysis of 4th derivative revealed existence of two bands ˜ (C O). Temperature and concentration investigations enabled us to attribute the band with higher wavenumber to the ˜ (C O) band of the (E-s-Z) and that with lower wavenumber to the ˜ (C O) band of (E-s-E) conformer, respectively (Table 1). The ˜ (C C) of the enaminoketone (DMBN) was broad, single band at about 1578 cm−1 , but in IR-spectra of the (DMAA) in alcohols there were two distinct bands ˜ (C C). Hence the fitting also required two components at about, 1630 and 1624 cm−1 (in cyclohexane). This result was confirmed by the analysis of 4th derivative of this band. After band fitting the integrated intensity of each ˜ (C O) and ˜ (C C) band was calculated according to Eq. (1).

A = (Cs × d)

−1

 ×

log band

I  0

I

d

(1)

2.1. General Solvents (obtained from Aldrich) were purified using standard techniques and were dried over the appropriate drying agent before use. ␤-Dimethylaminoacrolein (DMAA) was purchased (Aldrich), distilled in vacuum and stored under dry nitrogen at +4 ◦ C. 2.2.

1H

and 13 C NMR spectra

1 H and spectra were recorded on Bruker DRX-500 instrument at

25 ◦ C

using standard conditions. As it follows from Table 1 of Supplementary Part the studied compound was exclusively (E)-isomer. The comparison of 1 H and 13 C NMR spectra data of the (DMAA) in deutarated chloroform and DMSO was very useful to understand the character of electron density distribution in the molecule. Thus at 25 ◦ C N,N-dimethylamino group appears as two good resolved signals in 1 H and 13 C NMR spectra in both solvents that can be explained by hindered rotation around C–N bond in the (DMAA) as a result of stronger positive charge at the nitrogen atom (higher level of C N+ double bond). It is well known the 13 C chemical shift of a carbon incorporated in a conjugated system depends on the electron density at the carbon, the 13 C chemical shift difference between the two olefinic carbons reflects the C C double bond polarization of the N–C C– system and the degree of the n,␲conjugation. The larger the ı (13 C) value, the more extended is the bond polarization and the more efficient is the n,␲-conjugation. 13 C NMR investigations showed (Table 1 in the Supplementary Part) the ı (13 C) value for the (DMAA) in DMSO-d6 (60.5 ppm) to be larger than that in CDCl3 (58.7 ppm). So the polarization of C C double bond was greater in DMSO-d6 than in CDCl3 , that confirming a preference of structure B in the (DMAA) in more polar solvents. 2.3. Infrared spectra Infrared spectra were recorded on a Bruker Vertex 70 FTIR spectrometer with KBr beamsplitter and RT-DLaTGS detector at the room temperature (20 ± 1 ◦ C). For all spectra 32 scans recorded at 2 cm−1 resolution were averaged. Solution spectra were measured in carbon tetrachloride using standard NaCl cells with pathlengths

where A (cm/mol) is integrated intensity of the ˜ (C O) band, Cs (mol/l) is the concentration of the given stereoisomer in solution, and d (cm) stands for the cell pathlength. The main problem was to evaluate the concentration of each stereoisomeric form, Cs . Since the studied (DMAA) was presented as equilibrium (E-s-Z)  (E-s-E), which depended on total concentration of the (DMAA) and temperature, we used the method, proposed by authors [17] for an analysis of conformational equilibria assuming that the absorption coefficients of corresponding bands were temperature independent. We plotted integrals of ˜ (C O) (E-s-Z) vs. integrals of appropriate ˜ (C O) (E-s-E) [15]. Intercepts of this plot with axes were integrals of ˜ (C O) bands of (E-s-Z) and (E-s-E) conformers, respectively, but for all that the concentration of the conformer equaled the total concentration of enaminoketone, Ctotal . Knowing the cell pathlength and Ctotal , it was easy to calculate integrated intensities of the ˜ (C O) band of each conformer, and, hence, to evaluate the concentration of each stereoisomeric form, Cs , at given temperature in accord with Eq. (1). Integrated intensities of the ˜ (C O) bands of the two stereoisomeric forms of studied acrolein in carbon tetrachloride and 1,4-dioxane are listed in Table 2. Solvent effects on ˜ (C O) and ˜ (C C) bands were investigated in 20 various solvents and wavenumbers obtained are listed in Table 1. We correlated these data with solvatochromic parameters of Kamlet, About and Taft in accord with Eq. (2): ˜ = ˜ 0 + (s␲∗ + dı) + a˛ + bˇ + hı2H

(2)

where ˜ is the vibration wavenumber of solute [such as ˜ (C O) and ˜ (C C) vibration] in a pure solvent and ˜ 0 is the regression value of the ˜ (C O) and ˜ (C C) vibration in cyclohexane as a reference solvent, ␲* is an index of solvent dipolarity/polarizability, ı is a discontinuous polarizability correction term for aromatic and polychlorosubstituted aliphatic solvents, ˛ is a measure of the solvent hydrogen bond donor acidity, ˇ is a measure of the solvent hydrogen bond acceptor basicity, and ıH is the Hildebrand’s solubility parameter. In Table 3 we presented the results of KAT correlations for investigated enaminal.

1012

S.I. Vdovenko et al. / Spectrochimica Acta Part A 74 (2009) 1010–1015

Table 1 Wavenumbers of ˜ (C O) and ˜ (C C) bands of (E-s-Z) and (E-s-E) stereoisomeric forms of the (DMAA) and solvent parameters (␲*), (˛), (ˇ), (ı) and (ı2H /100) of Eq. (2). n/n

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Solvent

E-s-Z

Cyclohexane n-Hexane Tetrachloromethane Trichloromethane Dichloromethane 1,2-Dichloroethane 1,1,1-Trichloroethane Tetrachloroethylene 1,4-Dioxane Triethylamine Di-n-butyl ether Tetrahydrofuran Acetonitrile n-Butanol tert-Butanol i-Propanol n-Propanol Ethanol Methanol 2,2,2-Trifluoroethanol

E-s-E

Solvent parameters

˜ (C O)

˜ (C C)

˜ (C O)

˜ (C C)

␲*

˛

ˇ

ı

ı2H /100

1677.58 1679.17 1671.75 1668.16 1681.44 1681.44 1671.79 1672.60 1672.18 1674.88 1673.81 1673.81 1682.13 1645.16 1642.12 1645.21 1645.04 1645.33 1645.86 1647.89

1624.88 1626.20 1622.08 1611.16 1618.30 1618.3 1622.96 1622.26 1623.32 1621.99 1623.14 1623.14 1617.52 1602.28 1605.96 1603.88 1602.16 1602.30 1601.12 1572.49

1671.38 1674.08 1659.48 1656.57 1654.34 1654.34 1659.52 1661.7 1657.69 1666.64 1664.15 1664.15 1652.44 1640.71 1641.45 1640.89 1640.77 1640.97 1641.57 1643.83

1630.72 1631.65 1627.13 1617.41 1621.48 1621.48 1626.86 1627.86 1626.83 1630.00 1629.78 1629.78 1622.52 1617.98 1619.78 1618.82 1618.06 1617.91 1617.03 1593.63

0 −0.08 0.28 0.58 0.82 0.81 0.49 0.28 0.55 0.14 0.24 0.58 0.75 0.47 0.41 0.48 0.48 0.54 0.6 0.73

0 0 0 0.44 0.185 0 0 0 0 0 0 0 0.19 0.79 0.68 0.76 0.76 0.83 0.93 1.51

0 0 0 0 0 0 0 0 0.37 0.71 0.46 0.55 0.31 0.88 1.01 0.95 0.95 0.77 0.62 0

0 0 0.5 0.5 0.5 0.5 0.5 0.5 0 0 0 0 0 0 0 0 0 0 0 0.5

0.67 0.55 0.74 0.87 0.98 0.96 0.49 0.86 1.00 0.78 0.97 0.83 1.42 1.3 1.1 1.32 1.333 1.69 2.1 1.371

2.4. Quantum chemical calculations As we have demonstrated previously [15] the calculated energy differences of stereoisomeric forms of enaminones obtained by both AM1 and ab initio methods were very close and for this reason now we confined in only AM1 calculations. Quantum chemical calculations of energies of formation for conformers of the (DMAA) were performed with a program HyperChem (Hypercube Inc., release 8.0.3) by semiempirical AM1 method with the algorithm Polack–Ribiere in vacuum at RMS gradient 0.048 kJ mol−1 . Results of the calculations are listed in Table 4. 3. Results and discussion As it was mentioned above ␤-dialkylamino-␣,␤-unsaturated carbonyl compounds are able to generate various spatial forms due to rotational isomerism around C C vinylic double bond and C–C single bond between vinyl and carbonyl group. In general, four isomeric forms are to be anticipated for the enaminoketones (when R4 R5 ) (Fig. 1), hence four pairs of (C O) and (C C) bands should be expected in IR spectrum if all stereoisomeric forms exist as an equilibrium. However, for the investigated (DMAA) the number of

bands was less (see, e.g. Supplementary Part, Fig. 1a and b). It can be explained either by the absence of some of these isomers or by coincidence of some of their bands. It is worth to note that ˜ (C O) and ˜ (C C) for enaminoketones are strongly coupled, possessing, in large measure, the character of out-of-phase and in-phase vibrational modes, respectively [15,18]. Nevertheless, as previously, we shall continue to use the ˜ (C O) and ˜ (C C) as a convenient form of shorthand. Earlier [5,14] the ␤-dimethylaminoacrolein was investigated by the NMR 1 H and 13 C method at low temperatures and it was shown the (DMAA) to be exclusively in (E)-configuration. Our 1 H NMR investigations confirmed that the (DMAA) is in (E)-configuration even at ambient temperatures. Moreover, it was elucidated that in solutions of polar solvents this compound formed the equilibrium of (E-s-Z) and (E-s-E) conformers [3,5,14]. Hence, in IR spectrum of this compound it should be expected two pairs of the ˜ (C O) and ˜ (C C) band. In fact, there were three bands in the region of double bond vibrations, viz. 1678, 1671 and 1625 cm−1 (in cyclohexane, Table 1, Fig. 1b of Supplementary part). The general profile of the ˜ (C O) bands of the (DMAA) depended on total concentration of the acrolein (Fig. 2). The same was true for profile of the ˜ (C C) band and precise analysis of 4th derivative revealed two bands at 1632

Table 2 Integrated intensities of stretching vibrations ˜ (C O) of existing conformers of the (DMAA) and parameters ln Keq (0), a and b and correlation coefficient R of Eq. (3). Solvent

Stereoisomer

A(C O), M−1 cm−1

Carbon tetrachloride

E-s-E E-s-Z

6237 9262

1,4-Dioxane

E-s-E E-s-Z

8996 8851

˛

b

ln Keq (0)

R

2.1806

3.0294

0.4397

0.9949

0.9534

7.2218

0.5334

0.9801

Table 3 The coefficients (s, d, a, b, and h), error of estimate (S.D.), Fisher index of reliability (F), and correlation coefficients (R2 ) of the equation Eq. (4–7) for ˜ (C O) and ˜ (C C) bands of the (DMAA). Stereoisomeric form

Independent variable

Intercepta

s

E-s-Z

˜ (C O) ˜ (C C)

1684.48 1624.35

14.730 −1.897

−1.433b 2.197

−19.463 −25.656

−23.043 −2.212

E-s-E

˜ (C O) ˜ (C C)

1677.51 1625.34

−10.695 −9.971

1.289b −0.185b

−8.526 −20.844

−16.212 5.088

a b

Wavenumber, cm−1 . Not significant.

d

a

b

R2

h

N

S.D.

F

−6.601 1.009

20 20

4.4143 0.9984

38.530 0.9323 526.24 0.9947

4 5

20 20

2.8164 1.6941

53.883 0.9506 91.887 0.9704

6 7

1.289b 6.266

Eq. (No)

S.I. Vdovenko et al. / Spectrochimica Acta Part A 74 (2009) 1010–1015

1013

Table 4

.

Calculated total energy (kJ/mol), dipole moment, , and torsion angle (degree) by AM1 method. Stereoisomeric form

Energy, kJ/mol

Relative energy, kJ/mol

, D

Tetrahedral angle C C–C O

E-s-E E-s-Z

−124624.29 −124628.06

3.77 0.00

5.30 4.33

179.1◦ 0.2◦

and 1626 cm−1 , respectively (see Table 1). We presented the results of the profile fitting of IR-spectra of the (DMAA) at different concentrations in Figs. 3 and 4, from which it was clear that under the increase of total dimethylaminoacrolein concentration the intensities of the bands at 1672 and 1622 cm−1 decreased, at the same time intensities of the bands at 1659 and 1627 cm−1 increased. Temperature investigations (not shown) also showed that the intensities of the bands at 1672 and 1622 cm−1 decreased and those of the bands at 1659 and 1627 cm−1 increased when temperature rose. For correct attribution of these pairs of bands we calculated energies and dipole moments of the (DMAA). As it was clear from results of calculations (Table 4) the (E-s-Z) conformer was more stable, but less polar in comparison with the (E-s-E) conformer. Dipole moments of both conformers were relatively high therefore the increase of the (DMAA) concentration raised the solvent polarity significantly thus stabilizing the more polar (E-s-E) conformer. Hence, we attributed the pair of bands at 1672 and 1622 cm−1 to the ˜ (C O) and ˜ (C C) bands of the (E-s-Z) conformer, whereas the bands at 1659 and 1627 cm−1 to the ˜ (C O) and ˜ (C C) bands of the (E-s-E) conformer, respectively. In addition, in more polar sol-

vents the intensity of the ˜ (C O) and ˜ (C C) bands, ascribed by us to the more polar (E-s-E) conformer, increased with the simultaneous intensity decrease of the ˜ (C O) and ˜ (C C) bands of less polar (E-s-Z) conformer. This fact was additional vindication of accuracy of our attribution. Moreover, from Table 1 it was followed that our attribution corresponded to the criterion of Smith and Taylor [10], C=O (E-s-Z) > C=O (E-s-E). according to which C=C C=C Integral intensities of the ˜ (C O) bands of each conformer (Table 1) enabled us to calculate constant Keq of the equilibrium (E-s-Z)  (E-s-E). As long as this equilibrium depended on total concentration of the (DMAA), we estimated Keq for various concentrations of the studied enaminal (see Table 2 of Supplementary Part) and plotted ln Keq vs. Ctotal (Fig. 5). Graphs obtained (see, e.g. Fig. 5) we approximated by Eq. (3): ln Keq = ln Keq (0) + a(1 − e−bCtotal );

(3)

As it easily follows from Table 2 equilibrium constant at infinite dilution was greater in 1,4-dioxane [Keq (0) = 1.7047] than in car-

Fig. 4. FT-IR spectrum of the (DMAA) in the region of the ˜ (C O) and ˜ (C C) bands in carbon tetrachloride, C = 0.18 M, pathlength = 0.01028. Fig. 2. FT-IR-spectra of the (DMAA) in the region of the ˜ (C O) bands. Dilution in carbon tetrachloride, C × pathlength = 1.82 × 10−3 M cm.

Fig. 3. FT-IR spectrum of the (DMAA) in the region of the ˜ (C O) and ˜ (C C) bands in carbon tetrachloride, C = 0.018 M, pathlength = 0.1024 cm.

Fig. 5. Plot of ln Keq vs. Ctotal for the (DMAA) in carbon tetrachloride approximated by Eq. (3).

1014

S.I. Vdovenko et al. / Spectrochimica Acta Part A 74 (2009) 1010–1015

bon tetrachloride [Keq (0) = 1.5522] as a result of additional solvent stabilization of more polar (E-s-E) conformer. The same was true for enthalpies of the (E-s-Z)  (E-s-E) equilibrium (see below). Solvent influence on IR-spectra depends on electron distribution in solute molecule. The model equation of the linear solvation energy relationship (LSER) applied in infrared spectroscopy is shown as Eq. (2). Besides the dielectric effects, measured by solvent polarity/polarizability term (␲*) of Eq. (2), solvent molecules interact directly with the C O and C C group by two kinds of attractive forces. The first of these forces is that the solvent molecule forms a hydrogen bond with oxygen of the carbonyl group (or with its ␲ system). The capability of solvent to donate proton in solvent-to-solute (DMAA) hydrogen bond is described by parameter ˛, solvent HBD acidity. The second is that the solvent molecule which has electronegative atoms O, N and S can form hydrogen bonds with a solute as proton donor. Parameter ˇ describes the ability of a solvent to accept proton in solute-to-solvent hydrogen bond, i.e. its hydrogen bond acceptor capability with respect to a reference hydrogen bond donor. We correlated the wavenumbers the ˜ (C O) and ˜ (C C) bands with solvatochromic parameters of twenty solvents (Table 1) according to Eq. (2) and presented the results in Table 3 as regression coefficients of parameters ␲*, ı, ˛, ˇ, and ı2H of Eqs. (4)–(7) for stereoisomeric forms of studied enaminal. Only terms at the 0.95 significance level or higher were retained. The quality of multiple linear regressions is indicated by the standard error of the estimate (S.D., the smaller the better), the Fisher index of reliability (F, the larger the better), and the correlation coefficient (R, variance = R2 , the closer to 1 the better). A correlation equation was considered as acceptable if the correlation coefficient, R, indicates that the equation accounts for more than 80% of the variance (R2 ≥ 0.80). From Eqs. (4)–(7) it followed that the conformers of studied compound differed notably in sensitivity to solvent influence, therefore it was reasonable to examine the multiple correlations separately and then to compare the results obtained. Analysis of the data (Table 3) revealed some general features: all equations obtained were statistically significant (R2 ≥ 0.80). Another similarity was that almost in all equations coefficients d of parameter ı (a discontinuous polarizability correction term) was insignificant, i.e. less than standard error of estimate (S.D.). Negligible contribution of the discontinuous polarizability correction term ı in Eq. (2) we noted earlier for correlations of the ˜ (C O) and ˜ (C C) bands of ␤-dimethylaminovinylmethyl ketone and ␤-dimethylaminovinyltrifluoromethyl ketone [15] and (3Z,E)4-ethoxy-1,1,1-trifluoro-5,5-dimethylhex-3-en-2-one [18]. 3.1. (E-s-Z) conformer of the (DMAA) The most influence on the ˜ (C O) vibrations of the (E-s-Z) conformer exerted the solvent’s hydrogen bond acceptor basicity (␤) term [Eq. (4)]: its regression coefficient b was the largest in absolute value. Influence of solvent’s hydrogen bond donor (HBD) acidity (˛) and solvent’s dipolarity/polarizability term (␲*) were also considerable. Coefficients a and b were highly negative, whereas the coefficient s was highly positive. In other words, increase of solvent’s hydrogen bond donor (HBD) acidity and solvent’s hydrogen bond acceptor (HBA) basicity decreased wavenumber of the ˜ (C O) mode while increase of solvent’s dipolarity/polarizability increased the wavenumber of this mode. From our point of view the only explanation of the ˜ (C O) shift to higher wavenumbers with an increase of the solvent dipolarity/polarizability (␲*) is the disruption of the conjugation in C C–C O system as a result of easier rotation of the carbonyl group out of the C C double bond plane due to the carbonyl specific solvation with solvent molecules [18]. In Eq. (5) for the ˜ (C C) vibrations the main contribution was made by solvent’s hydrogen bond donor (HBD) acidity (˛) term (coefficient a was most negative among all correlation coefficients, Table 3), so

Table 5 Enthalpies of equilibrium, mol.% of carbon tetrachloride and [g(εr ) − f(n)] function of mixtures of carbon tetrachloride with 1,4-dioxane. H, kJ/mol

Mol% of carbon tetrachloride

g(εr )a − f(n)b

f(n)b

17.63 13.89 11.49 10.43 7.13

0.00 35 60 70 100

0.0729 0.0597 0.0468 0.0408 0.0229

0.3371 0.3451 0.3508 0.3531 0.3598

a b

Function g(εr ) = (εr − 1)/(εr + 1) [19]. Function f(n) = (n2 − 1)/(n2 + 1) [19].

the shift of the ˜ (C C) mode to lower wavenumbers was mostly due to hydrogen bond formation between vinyl moiety of the (DMAA) and solvent molecules. Influence of other terms in Eq. (5) was much less than solvent’s hydrogen bond donor (HBD) acidity (˛) term, although all of them were negative. 3.2. (E-s-E) conformer of the (DMAA) In contrast to Eq. (4) the coefficient s in Eq. (6) was highly negative, hence interaction of carbonyl group of the (E-s-E) conformer with solvent did not exert distortion of the O C–C C system. Here again the main contribution made solvent’s hydrogen bond acceptor (HBA) basicity (ˇ) term, hence similarly to (E-s-Z) conformer [Eq. (4)] the shift of the ˜ (C O) vibrations to lower wavenumbers was determined, predominantly, by solvent’s hydrogen bond acceptor (HBA) basicity. Similarly to Eq. (5) coefficient b was the most negative, hence hydrogen bond between –CH CH– fragment and solvent molecules exerted maximal shift of the ˜ (C C) mode to lower wavenumbers. It should be noted that coefficient b in Eq. (7) was slightly positive indicating that solvent’s hydrogen bond acceptor (HBA) basicity shifted negligibly the ˜ (C C) mode to higher wavenumbers. The comparison of Eqs. (4) and (5) with Eqs. (6) and (7) revealed that the influence of solvent’s hydrogen bond aceptor (HBA) basicity (ˇ) on the ˜ (C O) vibrations was greater for the (E-s-Z) conformer. At the same time for the ˜ (C C) vibrations solvent’s hydrogen bond donor (HBD) acidity (␣) term in Eq. (5) for the (E-s-Z) conformer was greater than that in Eq. (7) for the (E-s-E) conformer, so the C O and C C bonds were more polarizable in the (E-s-Z) conformer than in the (E-s-E) form. The value of a also was more negative for the ˜ (C O) vibrations of the (E-s-Z) conformer than that of the (E-s-E) [cf. Eqs. (4) and (6)] indicating that weight of structure B was greater for the (E-s-Z) conformer comparing with the (E-s-E) stereoisomeric form. Hence, conjugation in the (E-s-Z) conformer was greater than that in the (E-s-E) form:

3.3. Solvent influence on enthalpy of the (E-s-Z)  (E-s-E) equilibrium Temperature investigations of the ˜ (C O) vibrations enabled us to estimate enthalpies of the (E-s-Z)  (E-s-E) equilibrium of the (DMAA) in pure carbon tetrachloride, 1,4-dioxane and their mixtures (Table 5). It was reasonable to suggest that the study of medium effects on the enthalpy of the (E-s-Z)  (E-s-E) equilibrium be carried out first using f(n) and [g(εr ) − f(n)] separately [19] and then as linear combination of both. As it can be easily seen from Figs. 6 and 7 there were good linear relationships between enthalpies of the (E-s-Z)  (E-s-E) equilibrium and function of the relative permittivity, εr , and refractive index, n. The use of these

S.I. Vdovenko et al. / Spectrochimica Acta Part A 74 (2009) 1010–1015

1015

the atomic and electronic polarization part, together with dispersive interactions made the main contribution, increasing the enthalpy of the (E-s-Z)  (E-s-E) equilibrium while pass aging from carbon tetrachloride to 1,4-dioxane (see Table 5) due to additional solvent stabilization of the (E-s-E) conformer. It is worth to note that in the correlation of solvent ␲*-scale and the functions f(n) and [g(εr ) − f(n)] the contribution of the first was fourfold greater, than the last [19]. 4. Conclusion

Fig. 6. Plot of H vs. [g(εr ) − f(n)] in mixture of carbon tetrachloride and 1,4-dioxane H = 2.1044 + 205.895 [g(εr ) − f(n)]; R2 = 0.9891.

Studied ␤-dimethylaminoacrolein was presented as an equilibrium of two conformers, viz. (E-s-Z) and (E-s-E), which ratio strongly depended on total acrolein concentration. The higher this concentration the greater fraction of more polar (E-s-E) conformer was. The investigation of solvent influence on stretching vibrations of C O and C C bonds showed that influence of solvent’s hydrogen bond acceptor (HBA) basicity (ˇ) on the ˜ (C O) vibrations was greater for the (E-s-Z) conformer. At the same time for the ˜ (C C) vibrations solvent’s hydrogen bond donor (HBD) acidity (˛) term for the (E-sZ) conformer was greater than that for the (E-s-E) conformer, so the C O and C C bonds were more polarized in the (E-s-Z) conformer than in the (E-s-E) form. Coefficient a also was more negative for the ˜ (C O) vibrations of the (E-s-Z) conformer than that of the (E-s-E) indicating that a contribution of canonical structure B was greater for the more polarized (E-s-Z) conformer comparing with (E-s-E) stereoisomeric form. So, conjugation in the (E-s-Z) conformer was greater than that in the (E-s-E) form. In solvents and in their mixtures where substantial ‘specific’ interactions (hydrogen bonding and/or donor–acceptor effects) were absent, enthalpies of the (E-s-Z)  (E-s-E) equilibrium depended predominantly on the atomic, electronic polarization and dispersive interactions. Appendix A. Supplementary data

Fig. 7. Plot of H vs. f(n) in mixture of carbon tetrachloride and 1,4-dioxane H = 2.1044 + 205.895 [g(εr ) − f(n)]; R2 = 0.9891.

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.saa.2009.08.034.

functions is based on fundamental concepts of reaction field theory as applied to electrostatic and dispersive interactions. f(n) and g(εr ) are defined through Eqs. (8) and (9):

References

f (n) =

n2 − 1 n2 + 1

g(εr ) =

εr − 1 εr + 1

(8) (9)

Moreover, as it was shown in [19] the orientation component of the solvent polarization is appropriately described by the function [g(εr ) − f(n)] while the atomic and electronic polarization part, together with dispersive interactions, can be described by f(n). Thus, for solvents and their mixtures examined here we considered that ‘polarity-polarizability’ scale should be amenable to a treatment as linear combination of these two functions. Hence, we expected for the enthalpy of the (E-s-Z)  (E-s-E) equilibrium in a given medium Eq. (10) to be hold: H = a f (n) + b[g(εr ) − f (n)] + c

(10)

Actually, multiple regression according to Eq. (10) (a = −409.287, b = 22.252, c = 153.947, R2 = 0.9993) was excellent and revealed that

[1] G.O. Dudek, E.P. Dudek, J. Am. Chem. Soc. 88 (1966) 2407. [2] J. Dabrowski, U. Dabrowska, Chem. Beer. 101 (1968) 2365; J. Dabrowski, U. Dabrowska, Chem. Beer. 101 (1968) 3392.

[3] A. Gómez-Sa nchez, R. Paredes-León, J. Campora, Man. Reason. Chem. 36 (1998) 154. ´ [4] J. Dabrowski, K. Kamienska-Trela, Spectrochim. Acta 22 (1966) 211. [5] J. Dabrowski, L.J. Kozerski, Org. Man. Reason. 4 (1972) 137. [6] L. Kozerski, J. Dabrowski, Org. Man. Reason. 5 (1973) 459. ´ [7] J. Dabrowski, K. Kamienska-Trela, L. Kozerski, Org. Man. Reason. 6 (1974) 499. ´ [8] E. Czerwinska, L. Kozerski, J. Boksa, Org. Man. Reason. 8 (1976) 345. ´ [9] L. Kozerski, K. Kamienska-Trela, L. Kania, Org. Man. Reason. 12 (1979) 365. [10] D. Smith, P.J. Taylor, Spectrochim. Acta A 32 (1976) 1477. [11] P.J. Taylor, Spectrochim. Acta A 32 (1976) 1477; P.J. Taylor, Spectrochim. Acta A 32 (1976) 1503. ´ [12] M. Bugaj, K. Kamienska-Trela, Pol. J. Chem. 80 (2006) 1217. [13] S.F. Tayyari, M. Fazli, F. Milani-nejad, J. Mol. Struct. (Theochem.) 541 (2001) 11. [14] M.-L. Filleux-Blanchard, H. Durand, G.J. Martin, Org. Man. Reason. 2 (1970) 539. [15] S.I. Vdovenko, I.I. Gerus, V.P. Kukhar, Vib. Spectrosc., in press. [16] S.I. Vdovenko, I.I. Gerus, V.P. Kukhar, J. Phys. Org. Chem. 20 (2007) 190. [17] K.O. Hartmann, G.L. Carlson, R.E. Witkowski, W.G. Fateley, Spectrochim. Acta A 24 (1968) 157. [18] S.I. Vdovenko, I.I. Gerus, V.P. Kukhar, Spectrochim. Acta A 72 (2009) 229. [19] J.-L. Abbot, R. Notario, Pure Appl. Chem. 71 (1999) 645.