SPECTROCHIMICA ACTA

PART A

E L S E V I E R Spectrochimica Acta Part A 51 (1995) 1891-1901

Iden t i f i ca t ion o f a p r e d o m i n a n t ke to c o n f o r m e r o f m a l o n y l dichlor ide: A n inf ra red and theore t ica l s tudy

N. Pietri, J.P. Aycard, A. Allouche, P. Verlaque, T. Chiavassa *

Physique des Interactions loniques et Moleculaires, URA CNRS 773, ~quipe de Spectrom~tries et Dynamique Mol~culaire, Boite 542, Universit~ de Provence, F-13397 Marseille Cedex 20, France

Received 21 November 1994; accepted 2 February 1995

Abstract

Fourier transform infrared spectra of malonyl dichloride (propanedioyl dichloride), both as a neat liquid and isolated in an argon matrix at l0 K, have been analysed. These spectra reveal the existence of a predominant keto conformer ( ~ 90%). The semiempirical PM3 method was used to evaluate the relative stability of the keto conformers. The calculations predict the existence of three stable conformations. The lowest energy structure belongs to the C~ symmetry group with one C(O)C1 group rotated by ¢ I (CCCO)= 90 ° and the other C(O)C1 group lying in the CCC plane (¢2(CCCO) = 0°). The second conformer has C 2 symmetry with both C(O)C1 groups rotated by 80 °. The last conformer has C2v symmetry with the C(O)CI in the same plane (¢ 1 = ¢2 = 0°). By comparison of theoretical and experimental vibrational frequencies, it is determined that the C~ conformer is the predominant conformer.

I. Introduction

For many years, fl-dicarbonyl compounds have been of considerable interest to study the keto-enol equilibrium. For example, malonaldehyde has been used as an important model for experimental [1-4] as well as theoretical [5-8] studies.

In solution or in the gas phase the enol tautomer (Cs symmetry) generally predomi- nates because it is stabilized by a strong intramolecular hydrogen bond. Exceptions favoring the keto form have been reported, such as in the case of bulky ~-substituents [9] or for the malonyl difluoride [10] which leads only to the keto form.

The purpose of this work is two-fold: firstly to study the vibrational spectra of malonyl dichloride as a neat liquid and isolated in an argon matrix and secondly to identify the main existing species. Thus, a theoretical study of the relative stabilities of differents enol and keto conformers of malonyl dichloride was undertaken. The main conformations, the five diketo stereoisomers with three planar and two gauche forms and two enol stereoisomers, with an intramolecular hydrogen bond, are displayed in Fig. 1. In order to evaluate the molecular structures of all the possible conformations we used the more recent semiempirical method, MNDO-PM3 (PM3) [l 1]. This method proposed new

* Corresponding author.

0584-8539/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0584-8539(95)01420-9

1892 N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891 1901

parameter sets leading principally to better values for the heats of formation and geometries than semiempirical MNDO or AM1 methods. More, recently, Seeger et al. [12] have shown that the PM3 method predicts accurate force fields and vibrational spectra for small organic molecules. In order to confirm the validity of these calculations, a comparison with an ab initio calculation was performed for the main conformer of malonyi dichloride.

2. Experimental

Malonyl dichloride supplied from Aldrich (purity ~, 97%) was used after purification by vacuum distillation Teb = 40 °C (12 mmHg). Matrix was obtained by mixing evapo- rated malonyl dichloride with the rare gas (argon) at the ratio 1:500 and depositing the gas onto a CsBr window cooled to 20 K. The deposition time was about 60 min and the matrix thickness was about 250 ~tm. The spectrum was measured on the Nicolet 7000 FTIR Spectrometer with a resolution of 0.12 cm -1.

Neat, liquid infrared spectra were measured on the Mattson 1000 FTIR spectrometer with a resolution of 2 cm - t . All of the infrared spectra were recorded in the range 4000-400 cm ~.

Keto form

,i: H ~ '"H 9

U-cis, C2v

(-87.23)

O

S-gauche, C1

(-88.05)

O C1 CI CI

S-trans, Cs W-trans, C2v

(-85.54) (.82.08)

S-gauche, C2

(-87.20)

H o '/ .... o o" ~'" cI

L MA1, Cs MA2, Cs

(-83.27) (-81.56)

Fig. 1. Main conformations of keto and enol tautomers of malonyl dichloride. The adopting numbering system is referred to in the Tables. The calculated enthalpies of formation in kcal tool - ~ are given in parentheses for each form,

N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901 1893

4-

J

3.

2"

1-

0. .., _ A

Wavenumbers ' 1060 500

Fig. 2. Infrared spectrum of pure malonyl dichloride in the 4000-400 cm t region. The weak line observed next to 2300 c m - ~ denoted by an asterisk is attributed to CO,. The very weak shoulder at 1334 c m - t results probably from an impurity due to a malonyl dichloride decomposition.

3. Vibrat ional spectra

The infrared spectrum of malonyl dichloride as neat liquid (Fig. 2) is dominated by the existence of many broad, intense vibration bands. The most interesting spectral features are observed in the range > 1500 cm- ~. Here, there is not indication of the enol form in the liquid at room temperature. Typical vibrations for this form would be expected in the range 1560-1700cm -t for the C--O stretching vibration v(C-O), 1630-1500cm -~ for the C--C stretching vibration v(C--C) and 3000-2500cm ~ for the OH stretching vibration (v(O-H)) reflecting the bond softening due to the electron delocalization [4,8]. The only absorptions observed in this range are those assigned to the keto form: 1796 cm -j (vs, v(C--O)), 3580 cm -1 ((vw, 2v(C---O)) and 2938 and 2986 cm -t which are expected to be the C-H vibrational bands. In the region 1500-1200cm-~, the three vibrational lines at 1389, 1306 and 1200 cm-~ are probably caused by CH2 deformations (CH: bend, wag and twist) [13]. The region < 1200cm -~ reveals a broad doublet near 1000cm -~ in which the C-C stretching mode could be expected and an intense absorption band at 629 cm- ~ typical of a C-C1 stretching vibration. An exhaustive listing of the frequencies and relative intensities of the vibrational bands with their correspond- ing assignments are summarized in Table 1. A definite assignment will be given after comparison with calculations.

Moreover, it is interesting to note that the NMR spectra of malonyl dichloride in the liquid phase shows that this fl-dicarbonyl compound exists only as the keto tautomer. In the tH spectrum there is only one proton signal at 4.4 ppm corresponding to the two methylene hydrogens, and no signal is observed in the range g = 10-18 ppm, where the enol protons are expected.

The infrared spectrum of malonyl dichloride isolated in an argon matrix (Fig. 3) shows some interesting differences compared with the liquid. Firstly, the presence of new, weak, absorption bands belonging to hydrogen chloride and ketenecarbonyl chloride induced by thermal degradation of malonyl dichloride before deposition. Secondly, a higher resolution of the different lines is observed because ofthe absence of rotational structure when the molecule is embedded in a crystalline environment. Lastly, multiplet structures which are the result of site effects in the matrix are observed in most infrared bands. The frequencies of the main peaks indicated in bold type (Table 1) will designate the bands

1894 N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901

in the following. All of the vibrational frequencies with their relative intensities are reported in Table 1. The band positions are very close to those obtained for the liquid and suggest the existence of the same keto conformers in the two cases. The higher resolution enables one to identify six well separated features, two of strong intensity (1805 and 1799cm -~) and four of much weaker intensity (1817, 1811, 1794, 1784cm 1), in the spectral region, 1825-1690 cm -1 (Fig. 3a). These spectral features suggest the existence of two modes relative to three steroisomers. One of these is predominant (about 90%, from integrated absorbance measure). In the range below 1500 cm-l (Figs. 3b and 3c), we observe also principal bands occurring at 1395, 1302, 1033 and 974cm -l and very weak vibrational bands at 1406, 1291, 1057, 1043, 1023 and 998 cm -l. These bands are assigned to the vibrational modes of the major and minor conformers, respectively. We note that a 30 K annealing, followed by cooling to 10 K, involves no significant change in the spectrum; the temperature is indeed too low to allow a conversion between keto forms.

To identify these species and particularly the predominant species, we perform a conformational study in order to evaluate the respective stability of the conformers and determine for each of them, the relative rotational orientation of the two carbonyl groups.

Table 1 Observed infrared frequencies (cm - I ) and vibrational assignment for malonyl dichloride ~'b. Frequencies separated by a hyphen are attributed to a multiplet structure due to site effects in the matrix. Main peak frequencies are printed in bold type and, for the sake of simplicity, the values designate the band. Braces indicate overlapping bands

Approximate description Liquid Rel. int. Argon matrix Rel. int

v3 + v4 Combinat ion 3580 vw 3625 vw v l CH 2 antisymmetric stretch 2986 vw 2995 vw v2 CH2 symmetric stretch 2938 vw 2945 vw v' C--O stretch 1817-1816-1815 w v" C--O stretch 1811 w v 3 C=O stretch 1807-1805 vs

{1796 bd,vs v4 C---O stretch 1801 - 1799 vs v' C=O stretch 1794-1793-1791-1790 w v" C=O stretch 1784 w v" CH2 deformation 1406-1401 vw v5 CH 2 deformation 1389 w 1395-1391 w v6 CH2 wag 1306 m 1302-1299-1298-1296 m v' or v" CH 2 wag 1291-1289 vw 2 v 12 Combinat ion 1206-1203 vw v7 CH 2 twist 1200 vw 1198-1196-1190 w v' or v" 1057 vw v' or v" 1043 vw v8 C - C stretch 1034 m 1037-1033-1031 m v' or v" 1023 vw v' C - C stretch 998 sh 1001-999-998-995 vw v9 C42 stretch 978 m 978-974-967 m vl0 CH 2 rock 923 sh, vw 920 w v I 1 CCC deformation 805 vw 800 w v12 C-CI stretch 629 s 631-624 s v 13 C(O)CI wag 546 w 543 w v14 C(O)CI wag 490 sh, vw 486 vw v 15 C-CI stretch 460 w

{461 w vl6 C(O)C1 deformation 457 vw

Abbreviations used: s, strong; m, moderate; w, weak; v, very; sh, shoulder; bd, broad. b On the basis of the calculations vi, v', v" refer respectively to the assignments made for the main conformer

(S-gauche, Ct symmetry) and minor conformers (U-cis, C2v symmetry; S-gauche, C 2 symmetry). The population of the major and minor conformers in malonyl dichloride is proportional to the ratio of the intensities of v(C--O) vibrational bands due to major and minor conformers (about l0 to l).

N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901 1895

(a)

8

WAVIm',iUMBHR

1430 l~lJ0

(b)

12/0 1190 1110

WA~I'dl~IBHR

1070 930 790 ~ $10

WAVHI~D~BER

Fig. 3. Infrared spectrum of malonyl dichloride isolated in an argon cryogenic matrix. (a) v(C--O) region, 1825-1690 cm - ~; (b) 1430-1070 cm - 1; (c) 1070-440 cm- ~. The weak vibrational bands denoted by an asterisk are attributed to ketenecarbonyl chloride. The vibrational band next to 660 cm-~ is attributed to CO2. The channel spectra, clearly visible, denote the good optical quality of the matrix and allow one to calculate the matrix thickness (~ 250 lam).

4. Conformational analysis

4. I. Calculations

The semiempirical PM3 method implemented through the Ampac program package [14] was used. The calculat ions were performed on an IBM-3090 computer .

4.2. Enol forms

Calculat ions for the two most stable enol tautomers displaying an in t ramolecular hydrogen bond, led to the MA1 form (Fig. 1) with a O - H . . . O interact ion lower in energy than MA2 with a O - H . . - C I interaction. Nevertheless, the former is higher in

1896 N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901

0 C1 Cl C1

s- . ~ ~ L_.& w-~,~ " , ~

Fig. 4. Potential energy surface and contour plot for the keto form of malonyl dichloride as a function of the 4,1 and 4,2 dihedral angles ( - 180 ° < 4,1 < 180 ° and - 180 ° < 4, 2 < 180°). Same units as in Fig. 1.

ene rgy t h a n the s tab les t k e t o t a u t o m e r by 4.78 kca l m o l -~ (see be low) , so this resul t is

cons i s t en t w i th o u r e x p e r i m e n t a l resul t a n d we will no t c o n s i d e r f u r t he r the eno l fo rms .

4.3. Ke to f o r m s

In the first step relative energies for various dihedral angles (~1(01C2C3C4) and (~2(O2C3C4C6) in respective steps of 30 ° were calculated leading to the potential energy surface V(~ 1, 4,2) (Fig. 4), in which the S-trans and W-trans conformations correspond- ing to transition states are easily identified. The repulsion and the steric hindrance existing between the oxygen and chlorine atoms (S-trans) or between the two chlorine atoms (W-trans) explain the high energy of these conformations and also make easier a loss of planarity. The detail of the potential energy surface for 4,1 and ~ 2 values between - 9 0 and 90 ° (Fig. 5) shows three local minima, corresponding to three stable conforma-

0 0 0

~ . . . .CI ~.. C1 C1 C1

C I ~ O ~ , uH .Hcv H H H H £K/AJ--f-~I , 2 S-g~u,n~.c~

Kc.~

-87.5

-88

o .. -50 O1 50

Fig. 5. Potential energy surface and contour plot for the keto form of malonyl dichloride as a funct ion o f the 4,1 and 4,2 dihedral angles ( - 9 0 ° < 4, l < 90 ° and - 9 0 ° < 4,2 < 90°). Same units as in Fig. l .

N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901 1897

Table 2 PM3 optimized geometries and relative energies for the three stable keto conformations of malonyl dichloride. Bond lengths are in ~mgstroms, and bond angles are in degrees. The relative energies are in kcal mol-

S-gauche (C 1 symmetry) S-gauche (C 2 symmetry) U-cis (C:v symmetry)

C-H 8 1.110 1.110 1.111 C-H9 1.109 1.110 1.111 C2-C 3 1.503 1.509 1.501 C~-C4 1.508 1.509 1.501 C2=OI 1.201 1.201 I. 197 C4=O 5 1.201 1.201 1.197 C 2 C17 1.770 1.765 1.783 C4-C16 1.761 1. 765 1. 778 C2C3C 4 113.6 111.4 117.7 C2H 106.2 106.2 105.6 OIC2C3 128.1 126.5 129.8 CsC40 ~ 126.0 CI7C2C 3 111.3 112.9 110.2 C3C4C16 113.0 ~b 1 (Oi C2C3C4) -0 .5 80.0 0 ~b 2(C2C3C405) 91.2 - 80.0 AE/kcal mol- 1 0 + 0.87 + 0.82

tions. The variation in energy between these different conformations is weak (AE < 1 kcal mol-1) and one must be cautious when comparing the relative stability of the conformers. It appears that the conformer with a C1 symmetry, which can be described as S-gauche and possessing a dihedral angle of 90 ° between the two C(O)C1 groups, is lower in energy than the two other conformers that can be described as S-gauche with a C2 symmetry (dihedral angle ~bl= -~b2 = 80 °) and U-cis with a C2v symmetry (~bl = ~b2 = 0 °) (see Fig. 1). The higher stability of the first form can be explained by a minimal dipole-dipole interaction between heteroatoms of each C(O)C1 group. A similar result was found for malonyl difluoride where the most stable con- former was described by ab initio calculation (6-31G** basis) as S-gauche with a C1 symmetry but with a dihedral angle of 115 ° between the two C(O)F groups [10]. This increase of the torsional dihedral angle indicates a larger C16 • ' ' 0 1 steric repulsion in the malonyl dichloride than the corresponding steric repulsion in malonyl difluoride.

Further geometry optimizations and corresponding relative energies are given (Table 2). For every minimum determining a stable conformer, very slight deviations around the dihedral angles are observed because there is no constraint for q~l and ~b2~

Now we wish to verify the presence of these conformers in the liquid or in an argon matrix; therefore, we calculated the vibrational frequencies in order to compare them with the experiment.

5. Vibrational frequency calculations and discussion

The cartesian harmonic force fields of the three keto conformers of malonyl dichloride have been calculated using the same semiempirical method. The PM3 optimized geome- try of each conformer, listed in Table 2, was the starting point for the calculation of the second derivatives of the total molecular energy. The internal coordinate calculation was performed with the help of the REDONG program [15] which transforms the cartesian force field into the complete set of 24 internal coordinates, eight bond lengths (C2--O1, C4---O5, C2-C3, C3-C4, C2-C17, C4-C16, C3-H9) , 12 in-plane angles (OIC2C3, C3C2C17, C17C201, C2C3H8, C2C3H9, HsC3H9, C2C3C4, H8C3C4, HqCsC4, C3C405, O5C4C16, C16C4C3) , four dihedral coordinates (the two torsions around the C-C bonds and the two out-of-plane bending angles of the C-CI bonds). Unscaled frequencies (v b) and an approximate description of the vibrational modes for each keto form are indicated in

1898 N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891-1901

Table 3. The most striking feature seen in the comparison between the calculated and experimental frequencies is that the PM3 method predicts values which are the correct order of magnitude. Only the frequencies correponding to C=O stretching modes (v3 and v4) and the CH2 twisting mode (v7) are badly estimated and need both scale factors to be compared with the experimental frequencies. According to Seeger et al. [12] the scale factors computed for C--O stretching and CH2 twisting coordinates are taken as 0.8 and 1.35, respectively. For the C-H and C-C stretches, and angular deformation coordinates scale factors will be assumed equal to unity [12]. Lastly, we note that the unscaled frequencies including the C-C1 stretching coordinate (v12), which values are calculated for the three conformers between 562 and 574 cm-l, are underestimated compared with the strong absorption band observed at 624 cm -1. A good fitting could be obtained by applying a scale factor greater than unity (1.2). Thus, for each conformer these same scaling factors have been introduced and the scaled frequencies (v9 are reported in Table 3.

As it can be seen in Table 3, below 1000 cm -1, the scaled frequencies relative to the S-gauche conformer with C~ symmetry (C~ conformer) and U-cis conformer with C2v symmetry (C2v conformer) are very similar. Some are also clearly distinguishable (v9, v 10, v 1 l, v 14) from scaled frequencies of the S-gauche conformer with C2 symmetry ((72 conformer). Moreover, in this same range, we note that the calculated frequencies of the Cj and C2v conformers are in good agreement with the more intense absorption bands in the spectrum. As a result the C2 conformer is probably not the major conformer. In the range above 1000 cm 1, important differences between the experimental and calculated frequency shifts of the two C--O stretching modes, and CH2 twisting modes frequencies are key ingredients to the spectral identification of the main conformer. Thus, the measured shift between the two very intense C--O stretching vibrational bands (6 cm-1) is closer to the calculated shift for the C l conformer (8 cm -~) than the two other conformers, C 2 (28 cm - j ) and C2v (22 cm-l). The CH2 twisting calculated frequency of

Table 3 PM3 vibrational frequencies of the three keto conformers of values

malonyl dichloride: bunscaled values; ~scaled

Fundamental S-gauche, C1 S-gauche, C2 U-cis, C2~

vt ' (cm 1) v ~ (cm- ' ) vt ' (cm-1) v ~(cm 1) v b ( c m - t ) v ~ (cm- ' )

vl CH 2 antisymmetric stretch 3019 3019 3022 3022 3012 3012 v2 CH_, symmetric stretch 2947 2946 2954 2953 2932 2931 v3 C--O stretc/f 1987 1806 1994 1813 2011 1828 v4 C--O stretch 1977 1798 1966 1785 1986 1806 v5 CH2 deformation 1363 1398 1376 1399 1342 1380 v6 CH 2 wag/C-C stretch 1314 1322 1300 1341 1338 1337 v7 CH 2 twist 1092 1160 1072 1097 1042 1097 v8 C-C stretch/CH 2 wag 1035 1032 1062 1070 1023 1022 v9 C-C stretch 1007 1007 1044 1059 998 994 vl0 CH2 rock 925 932 955 966 893 909 vl 1 CCC deformation/C(O)Cl wag 755 772 741 747 780 796 v12 C-CI stretch/C(O)Cl deformation 570 610 574 606 562 604 v 13 C(O)C1 wag 536 552 539 544 538 538 v14 C(O)CI wag 492 494 508 541 451 480 v 15 C-C1 stretch 442 470 452 470 448 456 v16 C-CI stretch/C(O)Cl deformation 434 454 435 455 425 443 v17 CCCI deformation 338 339 337 338 296 296 v18 CCCI deformation 233 237 266 267 244 250 v 19 CCC deformation 150 150 164 164 131 131 v20 Torsion 25 25 20 20 25 25 v21 Torsion 19 19 2 2 1 1

Scaled PM3 calculations with scaling factors of 1.0 for C-H and C-C stretches, 0.8 for C--O stretch, 1.2 for C-CI stretch, 1.35 for CsC4H8 bend and 1.0 for all the other angular deformations.

N. Pietri et al . /Spectrochimica Acta Part A 51 (1995) 1891-1901 1899

Table 4 Calculated vibrational frequencies of the Ct conformer in the 3000--400cm-~ region: comparison of PM3 method and ab initio (DZ) method results

Experiment

Argon PM3

vexp 1 (a.u.) v ' Av/v (%) Assignment

Calculation

Ab initio

v b v ~ 1 (a.u.) Av/v (%) Assignment

vl 2995 3019 1 vCHs/vCH 9 3377 3012 2 1 vCHs/vCH 9 v2 2945 2946 0 vCH9/vCH 8 3295 2938 2 0 vCHg/vCH 8 v3 1805 76 1806 0 vC-O1/vC--O 5 1960 18(13 67 0 vC-OI / vC- -O 5

1,4 1799 100 1798 0 vC Os/vC=Oi 1953 1797 100 0 vC-Os/vC=O I v5 1395 8 1398 0 3CH 2 1615 t403 10 1 6 C H 2

v6 1302 16 1322 2 c~CHs/6CH9/vC2-C 3 1504 1318 13 0 6CHs/6CH 9 1,7 1190 6 1160 - 3 OCHs/vC 3 C 4 1364 1191 6 0 6 C H s / d C H , ~

1'8 1033 20 1032 0 (~CH9/I ,C 3 C a 1166 1054 46 2 t'C~ C 4

v9 974 20 1007 3 vC 2 C 3 11078 971 18 0 ¢5CH9/;'C 2 C 3 vl0 920 10 932 1 6 C H 2 / 6 C H 8 1023 933 11 1 vC 2 C3/vC3-C4 v I 1 800 8 772 - 3 6 CCC/TCCI 6 880 800 13 0 ~ CCC/TCCI 6 v12 624 38 610 - 2 6C(O)C16/vCCI 6 693 650 50 3 ~C(O)C16/vCCI 6

v13 543 7 552 2 7CC16 567 538 9 - 1 7CC16 v14 486 494 2 ;'CCI 7 497 455 7 - 7 7CC17 v15 460 6 470 2 vCCI 6 473 439 3 - 5 vCC16 v16 457 454 - 1 vCCI7/~C(O)C] 7 462 422 2 - 8 vCC17/6~C(O)C17

v designates a stretching coordinate and 6, 7, an in-plane and out-of-plane bending coordinate, repectively. Relative intensities rounded off and normalized to 100 for the largest band are interested in each case. Columns 6 and 11 display only internal coordinates whose contribution is larger than 20% in the potential energy distribution, b Unscaled values, c Scaled values. Av/v indicates the relative discrepancy between experimental and calculated values: A v / v = (vc~ I - vexp)/vc~ I. Scaled ab initio calculation with scaling factors of 0.8 for C-H stretch, 0.85 for C-O and C-C stretches, 0.95 for C-C1 stretch and C(O)C1 bend and 0.75 for all the other angular deformations.

the C~ conformer ( l l 60cm -~) is also in best agreement with the principal absorption band observed in the matrix at 1190 cm-] than the frequency of the C2 (1097 cm -~) and C2v (1097 cm -~) conformers. Consequently, it is highly probable that the major con- former is the C] conformer. Moreover, its identification is supported by the very good correlation existing between all the calculated and the most intense experimental vibra- tion frequencies. These frequencies, with the corresponding assignment for the Cj conformer, have been listed in Table 4, where only the internal coordinates that contribute by more than 20% to the potential energy distribution are indicated. The relative differences Av/v between observed and calculated frequencies are less than Y'/o in the 4000-400 cm ~ range and show the accuracy of the PM3 approximation in predict- ing vibrational spectrum.

In order to confirm the validity of this result we have compared, only for the predominant C] conformer, the values obtained by this semiempirical calculation with an ab initio calculation, performed with the help of the HONDO-8 program [16] using the Duning double-zeta (DZ) basis set. The ab initio optimized geometry reported in Table 5 shows no major difference in bond angles and lengths with PM3 values except for the C-C1 bond lengths which are longer for the ab initio calculation. The unscaled (v b) and scaled (v c) frequencies with the corresponding calculated intensities are listed in Table 4 along the PM3 and experimental results. The scaling factors are standard scaling factors [17], i.e. 0.85 for C---O and C-C stretching, 0.95 for C-C1 stretching and C(O)C1 wag, and 0.75 for all the other angular deformations. The scaling factor relative to the C-H stretching vibrations has been chosen to fit the experimental frequencies. The calculated ab initio values are in close agreement with the experimental values except for the low-frequency modes (v13 to v16) identified with the C-CI stretch and C(O)CI wag. These are underestimated, with relative differences Av/v reaching 8%. These frequency

1900 N. Pietri et aL/Spectrochimica Acta Part A 51 (1995) 1891-1901

Table 5 Ab initio (DZ) optimized geometries for the C~ conformer of malonyl dichloride. Bond lengths are in ~.ngstr6ms, and bond angles are in degrees

S-gauche (Ci symmetry) S-gauche (C~ symmetry)

C-H 8 1.081 C2C3C 4 112.2 C-H 9 1.077 HC_~ H 108.6 C2-C s 1.513 O I C2C 3 128.1 C3-C 4 1.513 CSC405 126.1 C2--O I 1.188 C17C2C 3 110.9 C4-q35 1.189 C3C4C16 113.2 C2-C17 1.834 q~ 1(O I C2C3C4) - 0.5 C4-C16 1.833 ~b 2(C2C sC405) 91.8

shifts are consistent with our previous analysis of the C-C1 bond lengths. The compari- son between the semiempirical and ab initio assignment reveals some discrepancies. For instance, for the modes v7, v8, v9 and vl0, the two calculations show that the main internal coordinates involved in these modes are C2-C3, C3-C4, C-C-H8 and C-C-H 9 and that they may be distributed differently amongst these four modes. Therefore, at this level it seems difficult to give a correct description of these modes. Lastly, we note that the intensities calculated for each mode have roughly the same order of magnitude as the experimental intensities.

Only a few weak bands are difficult to assign and remain subject to some uncertainties (see Table 1). These are probably related to the presence of a small amount of the two other conformers, C2v (v' modes) and 6"2 (v" modes). For instance, the v(C--O) region exhibits two sets of vibrational bands (1817 and 1794 cm-1) and (1811-1784 cm-1) with different multiplet structures which indicate the existence to two v(C--O) modes for each conformer. The measured shifts between the two v(C--O) modes (Av = 23 cm -1 and Av = 27 cm -I, respectively) are very close to the calculated shifts for the C:v and C2 conformers (Av = 22 cm -~ and Av = 28 cm-t). This result can be compared with previ- ous studies of malonyl difluoride by gas electron diffraction (GED) [10] where a mixture of two conformers was observed: a primary conformer (90%) described as S-gauche with C~ symmetry (~bl = 112 °) and a minor conformer described as W-gauche with C2 symmetry (q~l=-~b2 = 120°). However, the inclusion of a small amount of a U-cis conformer in the GED analysis is to be taken into account.

6. Conclusions

The semiempirical PM3 method predicts, for malonyl dichloride, the existence of three conformers: S-gauche with C1 symmetry, S-gauche with 6"2 symmetry and U-cis with C2v symmetry. The C1 conformer characterized by one C(O)CI group rotated by ~b I(CCCO)= 90 ° and the other COC1 group lying in the CCC plane (~b2(CCCO)= 0°), i.e. C--O cis to C-C, is the most stable. The argon cryogenic matrix infrared spectrum exhibits the predominance of this conformer and allows one to verify the reliability of the calculation, displaying experimental frequencies in good agreement with the calculated frequencies. The most striking feature is that the PM3 method leads to frequencies of comparable accuracy to the ab initio method. The existence of the two other conformers, in small amounts is also revealed by the presence of weak absorption bands in the infrared spectrum. This gives also evidence of the enol form absence. The instability of the enol tautomer relative to dihalogen fl-dicarbonyl compounds could be explained by a more electron withdrawing than releasing effect induced by halogen atoms, which tends to weaken the hydrogen bond.

N. Pietri et al./Spectrochimica Acta Part A 51 (1995) 1891 1901 1901

Acknowledgements

W e are very gra tefu l to Professor J. P o u r c i n a n d D o c t o r P. R o u b i n for f ru i t fu l d iscuss ions . T h e C N R S ( I D R I S ) a n d the reg ion P r o v e n c e - A l p e s - C 6 t e d ' a z u r (Cen t re r6g iona l de calcul scient if ique) are gra tefu l ly a c k n o w l e d g e d for their pa r t i a l f inanc ia l suppor t .

References

[1] S.L. Baughcum, R.W. Duerst, W.F. Rowe, Z. Smith and E.B. Wilson, J. Am. Chem. Soc., 103 (1981) 6296.

[2] R.S. Brown, A.T. Nakashima and R.C. Haddon, J. Am. Chem. Soc., 101 (1979) 3157. [3] S.H. Bertz and G. Dabbaagh, J. Org. Chem., 55 (1990) 5161. [4] T. Chiavassa, P. Verlaque, L. Pizzala and P. Roubin, Spectrochim. Acta Part A, 50 (1994) 343. [5] G. Buemi, C. Gandolfo, J. Chem. Soc. Faraday Trans. 2, 85 (1989) 215. [6] N. Shida, P.F. Barbara, J.E. Alml6f, J. Chem. Phys., 91 (1989) 4061. [7] M.J. Frisch, A.C. Scheiner, H.F. Schaefer II1, J. Chem. Phys., 82 (1985) 4194. [8] T. Chiavassa, P. Roubin, L. Pizzala, P. Verlaque, A. Allouche and F. Marinelli, J. Phys. Chem., 96 (1992)

10659. [9] J. Emsley, N.J. Freeman, M.B. Hursthouse, P.A. Bates, J. Mol. Struct., 161 (1986) 1479.

[10] A. Jin, H.G. Mack, A. Waterfeld and H. Oberhammer, J. Am. Chem. Soc., 113 (1991) 7847. [11] J.J.P. Stewart, J. Comput. Chem., 10 (1989) 209. [12] D.M. Seeger, C. Korzeniewski and W. Kowalchyk, J. Phys. Chem., 95 (1991) 6871. [13] (a) N.B. Colthup, Appl. Spectrosc., 34 (1980) 1.

(b) J.R. Durig, W. Zhao, D. Lewis and T.S. Little, J. Chem. Phys., 89 (1988) 1285. [14] (a) M.J.S. Dewar, E.G. Zoebish, E.F. Healy and J.J.P. Stewart, J. Am. Chem. Soc., 107 (1985) 3902.

(b) M.J.S. Dewar, Q.C.P.E. Bulletin, program no. 506 (1986). [15] (a) A. Allouche, J. Pourcin, Spectrochim. Acta Part A, 49 (1993) 4, 571.

(b) A. Allouche, Vibrational Analysis and Coordinate Transformation from a GAVSSIAN 88 Calculation, Q.C.P.E. Bulletin, 13 (1993) 3.

[16] M. Dupus, A. Farazdel, S.P. Karna, S.A. Maluendes, in E. Clementi (Ed.), Modern Techniques in Computational Chemistry: Mottec-90, ESCOM, Leiden, 1990.

[17] G. Fogarasi, P. Pulay, in J.R. Durig (Ed.), Vibrational Spectra and Structure, Elsevier, Amsterdam, 1985.