Conformational behaviour of trimethyl phosphate studied by infrared spectroscopy

Journal of

MOLECULAR STRUCTURE ELSEVIER

Conformational Roman

Journal of Molecular Structure 376 (1996) 277-287

behaviour of trimethyl phosphate infrared spectroscopy’

Strecka, Austin J. BarnesbT*, Wouter

A. Herreboutc,

Benjamin

studied by J. van der Vekenc

Abstract The conformational behaviour of trimethyl phosphate was studied by infrared spectroscopy in the liquid phase and as 1% solutions in a wide range of solvents. Variable temperature studies carried out in carbon dlsulphide and liquid xenon solutions, and in the pure liquid phase, showed a reversal in the relative stability of the two conformers between liquid xenon and carbon disulphide solutions. In liquid xenon solution the less polar conformer 11 is ~1.8 k.1 mol-’ lower in energy, whereas in CS2 solution the more polar conformer 1 is rr 1.4 kJ mol-’ more stable. In the pure liquid, the energy difference between the two conformers is similar to that in carbon disulphide solution. Ab initio calculations predict that conformer 11 has c‘? symmetry with all the methoxy groups in the gozrche configuration, whereas conformer 1 has C, symmetry with one trnns and two gauclze methoxy groups.

1. Introduction There have been numerous studies of the conformational behaviour of trimethyl phosphate using vibrational spectroscopy [lo 121. dipole moment and molar Kerr constant measurements [68,13,14], electron diffraction in the vapour phase [15], NMR spectroscopy [16,17], and ab initio cal[ 12,18]. The interaction of trimethyl culations phosphate with hydrogen chloride and water in the gas phase has been studied by infrared spectroscopy [ 191 and the enthalpy of mixing of trimethyl phosphate with water has been investigated by calorimetry [20]. * Corresponding author. ’Dcdicatcd to Professor James E. Boggs on the occasion of his 75th birthday.

Doublets observed [l-12] in the liquid phase or solution vibrational spectra of trimethyl phosphate in the P=O stretching region (zzl282/1269 cm ‘), the PO3 symmetric stretching region (- 753/737 cm-‘), and the PO3 symmetric deformation region (~4551440 cm-‘) have been attributed to the existence of two conformers, I and II. It has also been suggested [5,1 I] that splittings in other regions of the spectrum (C-O stretching near 1050 cm-‘, PO3 antisymmetric stretching near 850 cm-’ and PO3 antisymmetric deformation near 500 cm-l) have a conformational origin. One component of each conformational doublet (1269, 753, 440 cm-‘) disappears in the solid phase spectra and increases in relative intensity with increasing temperature in the liquid phase; these bands have accordingly been assigned to the higher cncrgy conformer II. The

0022-2860/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0022-2860(95)09106-g

278

R. Streck et al.lJournal

of Molecular

enthalpy difference in the liquid phase is small, Marsault-H&rail [5] reporting a value of 0.3 kcal mol-’ (1.2 kJ mol-‘). The relative intensities of the doublets are strongly influenced by the dielectric constant of the solvent, several authors [1,6,10,11] reporting that the bands at ~1269 and 753 cm-’ decrease in relative intensity in polar solvents and thus belong to the less polar conformer (II); hence the conformer I which persists in the solid phase is the more polar one. Structural studies using a variety of techniques have led to contradictory conclusions. Maiants et al. [2] proposed, on the basis of vibrational calculations, that conformer I1 has C,,, symmetry with the methoxy groups in the cis orientation (torsional angle 0”) and that conformer I has C3 symmetry with a torsional angle of 90”. On the basis of dipole moment measurements combined with infrared spectral data, Raevskii [7] suggested that conformer I contains one methoxy group in the trans orientation (torsional angle ISO’), and conformer II has the methoxy groups arranged cis and gauche (torsional angle 60”). A vapour phase study by electron diffraction [15] gave the best fit to the experimental data by assuming two conformers of C, symmetry, with torsional angles of 153” (near truns) and 68” (guuche), in a ratio of -3: 1. NMR investigation [ 171 of partially oriented trimethyl phosphate in liquid crystals, combined with molecular mechanics calculations, assumed at least C3 symmetry and also suggested gauche and near tram (with a torsional angle of 155”) conformers, but with the gauche form predominant. Ab initio calculations [12,18] predicted that the lower energy conformer has C, symmetry with torsional angles of w 35”, 50” and 178” and a dipole moment of4 D, whereas the higher energy conformer has C3 symmetry with torsional angles of N 45’ and a dipole moment of 1 D (using the 6-31G’ basis set). The energy difference was computed to be 1 kcal mol-’ (4.2 kJ molP’). Matrix isolation trapping experimknts using two different nozzle temperatures [ 121 were claimed to give qualitative support to the conclusion that the C3 conformer is the higher energy conformer in the vapour phase (however, the spectra illustrated actually show the higher wavenumber component, i.e. the more polar conformer I, increasing in relative intensity with increasing nozzle temperature).

Structure

376 (1996)

277-287

The present investigation was undertaken to clarify the conformational behaviour of trimethyl phosphate by determining the relative stabilities of the two conformers under different conditions. Ab initio calculations were carried out using GAUSSIAN 92 [21] at the RHF/6-31G* level, the effect of different solvents on the conformational equilibrium was re-examined and variable temperature studies were undertaken in carbon disulphide and liquid xenon solutions and in the pure liquid phase.

2. Experimental Solutions of trimethyl phosphate (Aldrich, 99% + , dried over molecular sieve) were prepared as 1% by weight in the various solvents (analytical or spectroscopic grade, dried over molecular sieve). Infrared spectra were recorded on a Nicolet Magna 750 FTIR spectrometer at 2 cm-’ resolution (1 cm-’ for the variable temperature studies). The vapour spectrum was recorded using a 10 m pathlength gas cell. Solution spectra were measured using 0.05 mm pathlength KBr cells, except for alcohols, acetic acid and water, for which a 0.015 mm pathlength CaF2 cell was used. Low temperature experiments were carried out using a VLT-2 cell; the pure liquid was measured as a capillary film between KBr plates. and for the CS2 solution spectra a 0.05 mm pathlength AgCl cell was used. Fourier self-deconvolution was used to determine the positions and intensities of overlapping bands. The liquid xenon solution experiments were carried out using a copper cell with a pathlength of 7 cm, equipped with CaF2 windows and suspended in a vacuum shroud. The cell was connected to a stainless steel filling manifold, allowing the cell to be pressurised to 15 bar. The vapour pressure of trimethyl phosphate proved to be too low to obtain spectra by premixing the solute and xenon in the Table 1 Results of ab imrio calculations for trimethyl phosphate AEjkJ m&’

Conformer

Symmetry

Angles

g, g, g 1, K. x I, -g, g

C3 C, C,

459,450, 45” 0 184”, 35”. 49” 4.2 ISO”, -33^, 33” 8.7

p/D 1.0 4.0 3.1

R. &reck

et al.lJoumal

of Molecular

I. Schematic

representations

of some possible

Table 2 Observed and calculated (scaled by a constant conformers of trimethyl phosphate Mode

376 (1996)

conformations

factor of 0.9) vibrational

Obs. (liquid)

Calc. Is. g, g)

1282 1269 118X 116ba 1076

1262 1196 116X 1110

104-I

1070

of trimethyl

wavenumbers

X42 7.53 737 529 521 499 455 440

843

1285

1283

119s 1168 1103 1082

1194 1169 1105 1082

840 X26

x32 828

741 723

126 523

512 507

504 495 440

435 418 361

372a

a Observed

351

243a

228

205a

191

in Raman

spectra (refs. [5] and [9]).

344 231 208

350 346 230 206

171

164

(g = gauche, t = tram).

of bands between

Calc. 0, -g, g)

1053 850

phosphate

Calc. (I, g, 8)

I063

1033

279

277-287

6 1, t

t, t, g

1, g, -g Fig.

Structure

1300 and 200 cm-’ for the

R. Srreck ct al./Journal of’Molecular Srrucrure 376 ( 1996) 277-287

280

phosphate are shown schematically in Fig. 1. Energy minima were found only for the three conformers

gas phase; accordingly, a small quantity of trimethyl phosphate dissolved in carbon disulphide was injected directly into the cell and the solvent was pumped off prior to condensing in the xenon. Infrared spectra were recorded on a Bruker 66v or 113~ spectrometer, using a Ge/KBr beamsplitter and a LN2 cooled broadband MCT detector. For all spectra. 100 scans recorded at 1 cm-’ resolution were averaged, Happ Genzel apodized and Fourier transformed using a zero filling factor of 4. Integrated intensities were obtained by use of the curve fitting procedure in the OPUS software.

g, g,g

C,), t,g! .Y (symmetry CI) and C,) (Table 1). The first two correspond to the conformers previously identified by ab initio calculations [12,18], but with their relative stabilities reversed compared with the results of the previous calculations. The torsional angles and dipole moments agree well with previous predictions. The third conformer is sufficiently close in calculated energy that it may be significant in the more polar solvents. Calculated wavenumbers for these three conl’ormers (scaled by a constant factor of 0.9) are compared with the bands observed in liquid trimethyl phosphate in Table 2. The calculated band positions fit very closely with the conformational splittings reported for the P=O stretch, PO?

3. Results 3.1. Ab initio calculations A number Tahlc 3 Conformational Phase/solution

of possible conformations

behawour

of trunethyl

Acceptor

of trimethyl

phosphate

under different

number

(symmetry

t, -6, g (symmetry

c

1/ (P=O)

“s (PO,)

i25”c’)

vapour Xe (-60°C) n-c,H~ c-C6H,j (CIH, )20 C-C,,HRO Cn& cc14 CS2 (CH&C=O &H&N liquid CbHSNOz CH,CN (CH,),S=O CH,NO: solid (-5o’C) CHCI, (CH,),COH (CH,),CHOH C,H,OH CH,OH CCl,CH,OH CH,COOH H:Ob

I.rlrl

I xx

14.8 18.9 19.3 20.5

1.X8 2.02 4.20 7.58 2.27 2.23 2.64 20.6 25.2 21.3 34.8 35.9 46.5 35.9

23.1 29.1 33.5 37.1 41.3

4.81 12.5 1Y Y 24 5 32 7

52.9 54 8

6.17 78.5

3.9 8.8 8.2 8.6 12.5 15.5

a Int(I)/int(ll)_ ’ v, IPO~) data from ref. [lo] (Kaman)

I

11

rel.INa

II

1317 I304 I101 1300 1295 1291 1290 1290 1289 1255 1282 1282

1292 1282 12x0 1280 1277 1274 1273 1271 1271 1272 1271 1269 1269 1266 1264 1265

mn.5 0.7 0.9 1 1 10 1.7 1.5 1.0 1.4 2.0 1.9 1.5 2.2 3.3 a3 x3

159

1261 1259 1258 1255 I257 1230

=4 1.4 13 I2 13

12x2 1279 1276 1277 1276 1275 1276 I272 1270 I267 I262 1232 1238

I

rel int.a

754 154 754 753 753 753 753 753 752 753 752 754 752 759

738 738 738 737 737 739 737 737 737 737 737 736 736 739 740

0.3 0.4 0.4 0.5 0.5

749

737

27

759

743

9

0.5 0.6 0.4 0.7 0.6 0.6 0.6 2.7

R. Streck et al.:Journal

of Molecular

--r----r

123Oh'

0

5

10

15

20

281

Structure 376 (1996) 277-287

25

30

35

40

45

Acceptor Number

Fig. 2. Variation of P=O symmetric

It is

also

splittings

stretch and

PO3

clear

the calculated

from

symmetric

stretching band positions with solvent acceptor number.

3.2. Solution studies

deformation. data

that

the

Infrared spectra of trimethyl phosphate at 1% concentration in a wide range of solvents were recorded and the relative intensities of the doublets in the P=O and PO3 stretching regions were measured (after spectral subtraction of the pure solvent). In many cases, spectral interference by the solvent was too great to obtain meaningful data in the PO, symmetric stretching region. Results for

observed

for the C-O antisymmetric PO3 antisymmetric stretching and PO3 antisymmetric deformation modes originate in part from the presence of two conformers. However,

stretching,

in

these

regions

the

is

lifted

in

lower

the

and thus two bands tions in which

degeneracy

of

symmetry

are observed

the C, symmetry

the

modes

conformer(s)

even under condiconformer

is absent.

3.5 I

31

I.

2.5

0.2

03

0.4

0.5 (epsilon

Fig. 3. Relative

intensity

of the conformer

bands in the P=O

0.6 -

0.7

0.8

0.9

1

l)/(epsilon t 2)

stretching

region

as a function

of the dielectric

constant

of the solvent.

R. Streck et al.~JournalqfMolecular

282

126C

1_. 100

~__~.~ 120

140

~_~~ 16C

180

Structure 376 (1996)277-287

,

_7~~~

2OC

220

,~____~ 240

rp 280

260

300

Temperature/K Fig. 4

vertical

Temperature van&ion of P=O strctchmg line indicates the normal freezing point).

band positions

in pure trimcthyl

the solutions studied are listed in Table 3, with those for the vapour, pure liquid and solid phases of trimethyl phosphate included for comparison. In the P=O stretching region, both components of the doublet move to lower wavenumber in more polar solvents. The band positions show no systematic variation with the dielectric constant of the solvent, but correlate well with the acceptor number of the solvent (Fig. 2). The band due to

phosphate:

?? , hquid

phase;

A, solid phase (the

conformer I exhibits larger shifts than that due to conformer II, so that the separation of the two components falls steadily from 25 cm-’ in the vapour phase to 11 cm -’ in acetonitrile solution. The intensity of the band due to conformer I increases relative to that of conformer II as the dielectric constant of the solvent increases (Fig. 3). In the strongly hydrogen bonding solvents, the separation of the bands due to conformers I and II

756 754 y 752 f

750

: m 748 b <746 r 2," 7LL ;

7L2

d

740 738 736l 130

120

140

160

180

200

220

240

260

280

300

Temperature/K

Fig. 5. Temperature variation of PO, symmetric stretching (the vertical line indicates the normal freezing point).

band positions

in pure trimethyl

phosphate:

?? , liquid

phase, A. solid phase

R. Srreck et al./Journal

points

were cxcludcd

in the Icast squares

of Molecular Structure376 11996) 277-287

fit).

increases again to about 15 cm-’ in the alcohols; in ethanoic acid or water solution only one band was observed. The behaviour in the PO3 symmetric stretching region shows little evidence of change in conformational stability with solvent polarity, there being no systematic shift in the positions of the bands in different solvents, and the change in the relative intensity of the two components with the dielectric constant of the solvent is smaller and less clear than that in the P=O stretching region. The only

3.0

283

3.5

4.0

solvents in which there is a marked increase in the relative intensity of conformer I are nitromethane (where the conformer II band is not observed in the P=O stretching region) and tertbutanol (which shows inconsistent behaviour in the two regions). 3.3. Variable temperature

studies

The spectrum of pure trimethyl phosphate was investigated as a function of temperature from

4.5

SO

55

6.0

phosphate:

W. cooling,

1DOOK,‘T Fig. 7. Van’t Hoff

plot for

PO? symmetric

stretching

bands of liquid trimethyl

A, heating

284

R. Streck et al.jJournai of Molecular Structure 376 (1996) 277 287

c

-0 44

-0 46 t 3.3

Fig. 8. Van? Hoff plot for P=O experiments).

stretching bands of trimethyl phosphate in carbon disulphide solution (W and A represent two separate

room temperature to 103 K. The bands in the P=O stretching region (Fig. 4) shifted slowly to lower wavenumbers as the temperature was lowered: from 1282/1269 cm-’ at room temperature to 1279/1266 cm-l at the freezing point (233 K). The bands continued to shift to lower wavenumbers in the supercooled liquid, reaching 1275/1263 cm ’ at 168 K. In the solid phase, the lower wavenumber component disappeared and the remaining band shifted from 1276 cm-’ just below the

freezing point to 1273 cm-’ at 103 K. In the PO3 symmetric stretching region (Fig. 5). the two components of the doublet shifted -1 cnC1 to higher wavenumbers as the temperature was lowered from room temperature to the freezing point. In the supercooled liquid, the higher wavenumber component continued to shift as the temperature was lowered whereas the lower wavenumber component showed little variation. On freezing, the lower wavenumber component shifted -2 cm-’

Fig. 9. Van? Hoff plot for PO, symmetric stretching bands of trimethyl phosphate in carbon disulphide solution (W and A represent two separate experiments).

R. Srrcck EI ai./Journal

of .44olecular Srructure

I

320

1300

1280

1075

stretching

and C-O

stretchmg

regions

285

277-287

I

1

1025

1050 G/cm-l

V/cm-l

Fig. 10. P=O

376 (1996)

of trimethyl

to higher wavenumbers and continued to shift slowly to higher wavenumbers as the temperature was lowered, reaching 742 cm-’ at 103 K. The higher wavenumber component of the doubiet persisted as a weak band below the freezing point until ~213 K. Measurements of the relative intensities of the components of each of the doublets in the liquid phase, over the temperature range 25X-168 K (supercooled liquid), gave the van? Hoff plots shown in Figs. 6 and 7. From these, enthalpy differences (conformer II- conformer I) of I .O f 0.1 kJ mol-’ and 1.8 * 0.2 kJ mol-‘, respectively, were obtained. The former value agrees well with that reported by Marsault-H&ail [5]. A variable temperature study was carried out in carbon disulphide solution from room temperature to 183 K. Over this temperature range. the P=O stretching doublet shifted down from 128911271 ’ to 128211267 cm ’ and the PO3 stretching irublet shifted up from 753/737 cm-’ to 7551739 cm-‘. Use of these data to determine the enthalpy difference between the conformers had to be limited to the range from room temperature to significantly above 223 K; at lower temperatures the spectra showed anomalies suggesting that the trimethyl phosphate was freezing out of the solution. Measurement of the relative intensities of the components of the doublet in the P=O stretching

phosphate

in liquid xenon solution:

A, -6O’C;

U, -100°C.

region from two separate experiments gave the van? Hoff plot shown in Fig. 8. from which an enthalpy difference (conformer II- conformer I) of i .4 f 0. i kJ moi-! was obtained. However, similar measurements for the doublet in the PO3 symmetric stretching region were difficult to carry out because of the low band intensities, and gave anomalous results. The data are shown in Fig. 9 in the form of a van? Hoff plot, which apparently shows conformer 11 to be the more stable with an enthalpy difference (conformer I - conformer II) of 1.1 f 0.4 kJ mol-‘. It can be seen that there is a poor correlation between the data from the two separate experiments carried out, and quite possibly the result should be disregarded as arising from errors in the intensity measurements. In liquid xenon solution, at high dilution, the lower wavenumber band (conformer II) in the P=O stretching region is the more intense (Fig. 10). The higher wavenumber component of the C0 antisymmetric stretching absorption is relatively intense, compared with the spectrum of the pure liquid, and increased in intensity as the temperature was lowered: thus the assignment (Table 2) of this component as primarily due to the C3 conformer is confirmed. The relative intensities of the components of the doublet in the P=O stretching region were measured in two separate experiments over the

286

R. Strcck et ai.!Journal

of Molecular

Structure

376 11996) 277-287

4

-0.30

x 2 -0.35

-0.55 4.7

Fig. Il. Van? Hoff plot for P=O experiments).

4.8

4.9

5.0

5.1

5.2 5.3 1OOOK,'T

5.4

5.5

5.6

5.1

5

stretching bands of trimethyl phosphate in liquid xenon solution (M and A represent two separate

temperature range from 213 to 173 K (the doublet in the PO3 symmetric stretching region could not be studied because the cell had calcium fluoride windows). The data are shown in Fig. 11 as a van’t Hoff plot, from which an enthalpy difference (conformer I - conformer II) of 1.8 & 0.3 kJ mol-’ was obtained.

4. Discussion The doublet observed in the P=O stretching region of organophosphorus esters is generally accepted to result from the presence of two conformers. It is apparent from the variation of the band positions with acceptor number (Fig. 2) that the P=O stretching mode of trimethyl phosphate is strongly affected by intermolecular interaction with the solvent. The higher wavenumber component (conformer I) shows the greater shift with increasing solvent acceptor number, but it is evident from the monotonic variation that there is no reversal in the order of the bands due to the two conformers (at least in the solvents with acceptor numbers in the range O-20). The lower wavenumber component (conformer II), which is the more intense in the vapour phase and in non-polar solvents, decreases in relative intensity as the dielectric

constant of the solvent increases (Fig. 3) and thus belongs to the less polar conformer. In liquid xenon solution this conformer is the more stable by ~1.8 kJ mall’ In more polar solvents, the higher wavenumber component, due to conformer I, is the more intense. As measured from the P=O stretching doublet, this conformer is the more stable by ~1.0 kJ mol-’ in the pure liquid phase and by -1.4 kJ mol ’ in carbon disulphide solution. It is clear from Fig. 4 that the single band remaining in the solid phase is due to conformer I. Comparison of these results with the ab initio predictions suggests that conformer I is the t, g, g conformer (with one tram methoxy group), and conformer II IS the g, g,g conformer (with all three methoxy groups gnuche). The relative stability of the two conformers in a non-polar medium is in agreement with the predictions of the present ab initio calculations (but the reverse of that predicted by previous ab initio studies [12:18]). The manifestations of conformational isomerism in the PO1 symmetric stretching region are much less clear-cut than those in the P=O stretching region. As for the P=O stretch, the doublet is generally accepted to result from two conformers. The higher wavenumber component clearly disappears in the solid phase (Fig. 5), but there are

R. Swrck et al./Journal

ofMolecular

inconsistencies in the solvent and temperature dependence of the relative intensity of the two components of the doublet. The band positions show little variation between the vapour phase and solution, or between different solvents, and there is no clear change in the relative intensity of the two bands with increasing dielectric constant of the solvent. In the liquid phase, the bands move slowly to higher wavenumbers with decreasing temperature, and the band due to conformer I also shifts to higher wavenumber on freezing. The effects of intermolecular interactions on these bands are difficult to interpret, as the mode will certainly contain significant contributions from vibrations other than the PO3 symmetric stretch, and the precise nature of the mode will be different for the two conformers. The variation of band intensities with temperature gave a higher value for the enthalpy difference between conformer II and conformer I in the pure liquid phase (1.8 kJ mol-’ compared with 1.0 kJ mall’) and an inconsistent value (conformer II apparently more stable by 1.1 kJ mol-‘, compared with conformer 1 more stable by 1.4 kJ molI’) in carbon disulphide solution. These inconsistencies may be partly explained by the difficulties in accurately measuring the intensities of these relatively weak bands. but it seems likely that other factors arc contributing to the intensity variations in this region. One possible perturbation is Fermi resonance with the overtone of the POC antisymmetric deformation mode, which occurs at 372 cn--’ in the liquid phase [9]. Such an interaction could explain the persistence of a weak band at the position of the high wavenumber component in the solid phase near the freezing point. Consequently, enthalpy differences derived from this region should be treated with caution.

5. Conclusions The experimental data clearly demonstrate that the more stable conformer in the vapour phase or non-polar solvents is the less polar form (the C, symmetry g, g, g structure) whereas the more stable conformer in the liquid phase or in polar solvents is the more polar form (the Ci symmetry t,g, g structure).

Structure

376 119961 277-287

287

Acknowledgements NFWO (Belgium) is thanked for financial support towards some of the spectroscopic equipment used in this study. One of us (WAH) thanks NFWO (Belgium) for a grant.

References [l] F.S. Mortimer. Spectrochim. Acta. 9 (1957) 270 [?I L.S. Malants. E.M. Popov and M.I. Kabachnik, Opt. Spektrosk., 7 (1959) 170. [3] F. H&ail and V. V&sat, C. R. Acad. Sci ,259(1964) 4629. [4] F. H&ail, C. R. Acad. Sci., 261 (1965) 3375. [5] F. Marsault-H&rail, J. Chim. Phys. Phys.-Chim. Biol.. 68 (1971) 274. [6] O.A. Raevskii, A.N. Vereshchagin and F.G. Khalitov, Izv. Akad. Nauk SSSR, Ser. Khim., (1972) 353. [7] O.A. Raevskii. J. Mol. Struct.. 19 (1973) 275. [8] O.A. Raevskii, A.N. Vereshchagin, Yu.A. Donskaya, A.G. Abul’khanov and Ya.A. Levin. Izv. i\kad. Nauk SSSR. Ser. Khim., (1976) 2013. [9] C. Garrigou-Lagrange and 0. Bouloussa, C. R. Acad. Sci., Ser. C, 282 (1976) 15. [IO] A.V. Yarknv and O.A. Raevskii, Zh. Ohshch Khim., 55 (1985) 50. [1 I] R.A. Nyquist and C.W. Puehl, Appl. Spectrosc.. 46 (1992) 1564. [12] V. Sablinskas, A. Hornand P. Klaeboe, J. Mol. Struct,349 (1995) 157. 1131 M.J. Aroney, L.H.L Chia, R.J.W. Le Fivre and J.D. Saxby, J. Chem. Sot., (1964) 2948. [I41 W. Waclawek and W. Guzowski, Rocz. Chem., 42 (1968) 2183. [IS] H. Oberhammer, Z. Naturforsch.. Teil A, 28a (1973) 1140. [I61 B. Donaldson and L.D. Hall. Can. J. Chem.. 50 (19i2) 2111. 1171 C.L. Khetrapal, G. Govil and H.J.C. Yeh. J. Mol. Struct., II6 (1984) 303. [I81 J.R. Van Wazer and C.S. Ewig, J. Am. Chem. Sot., 108 ( 1986) 4354. [I91 A. Bertoluzla, S. Bonara, G. Fini and M.A Morclli, Can. J. Spectrosc., 32 (1987) 107. [2U] A.S. Kertes and L. Tsmering, J Phys Chem., 81 (1977) 120 [21] M.J. Frisch. G.W. Trucks, M. Head-Gordon, P.M.W. Gill, M.W. Wang, J.B. Foresman, B.G. Johnson, H.B. Schlegel, M.A. Robb, E.S. ReplogIc. R. Gomperts, J.L. Andres, K. Raghauchari, J.S. Binkley, C. Gonzahz, R.L. Martin, D.J. Fox, D.J. Defrees, J. Baker, J.J.P. Stewart and J.A. Pople, C~,~SSIANYZ.Gaussian Inc.. Pittsburgh, PA, 1992.