View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/B104786H (Paper) New J. Chem., 2002, 26, 145-152

On the structure, carbonyl-stretching frequencies and relative stability of trans- and cis-[W(CO)4(η2-alkene)2]0/+: a theoretical and IR spectroelectrochemical study†

Jarosław Handzlik a, Frantisek Hartl *b and Teresa Szymańska-Buzar *c
aInstitute of Organic Chemistry and Technology, Cracow University of Technology, , 31-155 Kraków, Poland
bInstitute of Molecular Chemistry, Universiteit van Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands. E-mail: hartl@science.uva.nl; Fax: +31 20 525 6456; Tel: +31 20 525 6450
cFaculty of Chemistry, University of Wrocław, 14 F. Joliot-Curie, 50-383 Wrocław, Poland. E-mail: tsz@wchuwr.chem.uni.wroc.pl; Fax: +48 71 3282348; Tel: +48 71 3204221

Received (in Montpellier, France) 31st May 2001 , Accepted 17th September 2001

First published on 7th January 2002

Abstract

The geometries, ν(CO) frequencies, relative enthalpies and Gibbs energies of the alkene carbonyl complexes [W(CO)4(η2-C2H4)2]0/+ and [W(CO)5(η2-C2H4)] were calculated by means of the GAUSSIAN 98 program using the hybrid B3-LYP density functional. The predicted geometries and ν(CO) vibrational frequencies agree with the experimental data. The calculated relative energies (ΔG298) show that trans-[W(CO)4(η2-C2H4)2] is more stable by 10 or 12 kJ mol−1 (depending on the basis set applied) than cis-[W(CO)4(η2-C2H4)2]. In contrast to this, the stability of their one-electron oxidation products, the corresponding 17-electron cationic complexes, is reversed, the cationic form of the cis isomer being preferred by 14 or 10 kJ mol−1. Comparison of the calculated and experimental vibrational spectra has elucidated the electrochemical oxidation path of trans-[W(CO)4(η2-alkene)2] compounds. The electrochemical oxidation of trans-[W(CO)4(η2-1-butene)2] produces the corresponding 17-electron cation, which undergoes spontaneous isomerisation to cis-[W(CO)4(η2-1-butene)2]+. The identity of the latter species has been established by cyclic voltammetric and IR spectroelectrochemical experiments at low temperature.

Introduction

Zero-valent tungsten bis(alkene) complexes of the type trans-[W(CO)4(η2-alkene)2], first characterised by IR spectroscopy by Stolz et al. in 1963,1 remain the focus of much attention.1–20 The structure of several trans-bis(alkene) complexes has been elucidated by X-ray crystallographic studies.2,11,18,20 The two alkene ligands are mutually orthogonal and eclipsed with respect to the OC–W–CO axes of the square-planar W(CO)4 unit. The neutral cis-bis(alkene) complexes were identified as transients at low temperatures3,4,6,16–19 or in the gas phase.13

Studies of this class of complexes have focused on their participation as important intermediates in reactions of alkenes catalysed by transition metal carbonyls.3,9,10,12,14–19 Of particular interest have been the catalytic cycles of [W(CO)6] regarding isomerisation and metathesis reactions of alkenes.

The widely studied photochemical isomerisation of trans-[W(CO)4(η2-alkene)2] complexes to their thermally unstable cis isomers can be monitored conveniently at low temperatures with IR and NMR spectroscopies.16–19 Preliminary studies have revealed that the trans–cis isomerisation of trans-[W(CO)4(η2-alkene)2] is also induced by one-electron electrochemical oxidation.12 However, the cationic d5 tungsten(I) product has only been identified in frozen electrolysed solution by EPR spectroscopy. In addition, the published voltammetric data have been ambiguous with regard to the more positively lying W0/WI redox couple for the cis isomers, hence pointing to the electrocatalytic nature of the trans–cis isomerisation. In this paper we rectify the assignment of the cyclic voltammetric response with support from an IR spectroelectrochemical study of the oxidation at variable temperatures and from theoretical calculations of structural parameters and IR ν(CO) frequencies. In the case of transition metal carbonyls IR spectroscopy is often used to identify reactive intermediates. Comparison of calculated and experimental vibrational spectra has become one of the principal means of identifying unusual molecules. The low-spin d5 chromium(I) cationic complex [Cr(CO)4(3,4,7,8-tetramethyl-1,10-phenanthroline)]+, thoroughly studied by combined low-temperature IR, UV-VIS and EPR spectroelectrochemical and DFT computational methods,21 provides a suitable example.

Several previous theoretical studies dealt with trans-[Mo(CO)4(η2-C2H4)2]22–24 and [W(CO)5(η2-C2H4)],25,26 however, little is known about the structure of cis-[W(CO)4(η2-C2H4)2] and of the corresponding cation. Understanding the factors controlling their conformational preferences is important, as various catalytic cycles for the transformation of olefins involve the W(CO)4 unit with two organic ligands in mutually cis positions. In metathesis reactions of olefins, the formation of such catalytically active intermediates is expected, being induced by the oxidation of zero-valent tungsten compounds by Lewis acids such as SiCl4, GeCl4 or SnCl4.27,28 Therefore, it has been our aim to explain and predict the geometry and relative stability of trans- and, in particular, cis-[W(CO)4(η2-alkene)2] complexes and their 17-electron cationic forms. In the present paper, the theoretical part precedes the spectroelectrochemical study of the oxidation-induced trans–cis isomerisation.

Experimental

Materials

Dichloromethane (Acros; analytical quality) was dried over anhydrous CaCl2 and freshly distilled under nitrogen. The supporting electrolyte Bu4NPF6 (Aldrich) was recrystallised twice from absolute ethanol and dried in vacuo at 80[thin space (1/6-em)]°C for 10 h. The complexes trans-[W(CO)4(η2-alkene)2], alkene[thin space (1/6-em)]=[thin space (1/6-em)]ethene (1) and 1-butene (1A), were synthesized by photochemical reaction of [W(CO)6] with the alkene in n-hexane, according to published procedures.17–19

Computational details

Geometries. The geometries and ν(CO) frequencies of the alkene carbonyl complexes trans-[W(CO)4(η2-C2H4)2]0/4 (1/1+), cis-[W(CO)4(η2-C2H4)2]0/+ (2/2+), and [W(CO)5(η2-C2H4)] (3), were calculated with the GAUSSIAN 98 program29 by means of density functional theory (DFT). The hybrid B3-LYP density functional was applied in all calculations.30 Two basis sets were employed in parallel. The first one, denoted A hereafter, includes a pseudopotential LANL2DZ basis set31 for the W atom and a Dunning–Huzinaga valence double-ζ basis with a set of d-polarization functions D95V(d)32 for the C, O and H atoms. In the second basis set denoted B, the W atom is also described by a LANL2DZ basis set, while other atoms are described by a 6-31G(d) basis. The calculations of the neutral complexes were carried out at the restricted B3-LYP level. The unrestricted B3-LYP approach was used in the case of the cationic forms. In all the cases, the stability of the wavefunction was tested.33

All the structures were fully optimised with the Berny algorithm using redundant internal coordinates.34,35 The calculations were initially performed without any symmetry constraints. Then, the structures were optimised with the molecular symmetries obtained during the earlier calculations. In general, the B basis set yields slightly longer W–C bonds and shorter C[double bond, length half m-dash]C, C–H and C[triple bond, length half m-dash]O bonds than the A basis.

To estimate the accuracy of the theoretical method applied, the geometry calculations for the well-known structure of tungsten hexacarbonyl has been performed. The W–C bond lengths of 2.062 and 2.068 Å obtained, with the B3-LYP/A and B3-LYP/B methods, respectively, are consistent with the reported experimental value of 2.058 Å.36 A similar accuracy for DFT calculations of the W–C bond lengths in [W(CO)6] was obtained by Ziegler et al.37,38

Harmonic vibrational frequencies were calculated for each structure to confirm the potential energy minimum and to obtain ν(CO) vibrational frequencies, enthalpies and Gibbs energies.

Vibrational spectra. The ν(CO) frequencies were calculated employing the B3-LYP/B basis set. The scaling factor of 0.9614,39 recommended for systems containing first and the second row elements at the B3-LYP/6-31G(d) level, was used after its re-examination to calculate ν(CO) vibrations for [W(CO)6] (2109 a1g, 2019 eg and 1996 t1a cm−1). Despite the fact that within the B basis the tungsten atom is not described by the 6-31G(d) basis set, the scaled values of the ν(CO) vibrational frequencies are in satisfactory agreement with the recorded experimental data.

Additionally, we obtained a new scaling factor of 0.9648 for the W–CO interactions by a least-squares fit of the calculated data to the experimental vibrational frequencies of [W(CO)6].40 Comparison of the values of the ν(CO) vibration frequencies obtained using both scaling factors indicates that the scaling factor obtained by Scott and Radom39 provides a better fit to experiment.

Electrochemistry

All (spectro)electrochemical experiments were conducted under an inert atmosphere of dry nitrogen or argon. Cyclic voltammetry was carried out with an EG&G PAR Model 283 potentiostat, using a gas-tight three single-compartment cell equipped with a carefully polished Pt disk (0.42 mm2 apparent surface area) working, Pt gauze auxiliary and Ag wire pseudo-reference electrodes. The reported electrode potentials are given against the ferrocene/ferrocenium (Fc/Fc+) redox couple used as an internal standard.

IR spectroelectrochemical experiments at variable temperature were performed with a previously described OTTLE cell on a Pt minigrid working electrode positioned between CaF2 optical windows (ca. 0.1 mm optical path).41 The course of the redox reactions was monitored with a Bio-Rad FTS 7 FT-IR instrument and thin-layer cyclic voltammetry. Potential control was achieved with a PA4 potentiostat (EKOM, Czech Republic). In a typical experiment the cell was loaded with a 10−3 M solution of the bis(alkene) complex 1A in dichloromethane containing 0.3 M Bu4NPF6.

Results and discussion

Geometries and relative stability of the alkene carbonyl complexes of tungsten

trans-[W(CO)4(η2-alkene)2] (1). The coordination sphere of tungsten in trans-[W(CO)4(η2-alkene)2] compounds represents a distorted octahedron with the two alkene molecules mutually trans, orthogonal and each C[double bond, length half m-dash]C bond eclipsing one of the two OC–W–CO axes in the equatorial W(CO)4 plane. This geometry has been documented by X-ray crystallographic studies for compounds containing ethene,18cis- and trans-cyclooctene11,20 and methyl acrylate2 ligands. Selected optimised geometrical parameters are compared with experimental values in Table 1. The theoretically predicted structure has D2d symmetry. The complete calculated geometry for 1 is given in the ESI. There is good agreement between the previous X-ray structure and the data calculated in this work (Fig. 1). Note that the theory has predicted strictly the same W–C(ethene) bond lengths for the two carbons atoms of the ethene ligand, even though the two experimental crystallographic values differ by 0.016 Å. The calculated W–C(ethene) bond lengths of 2.349 (A basis) and 2.362 Å (B basis) slightly exceed the experimental values reported for 1 [2.299(9) and 2.315(9) Å].18 This difference is to be expected due to the presence of packing forces in the crystal lattice.
Optimised geometry of trans-[W(CO)4(η2-C2H4)2] (1).
Fig. 1 Optimised geometry of trans-[W(CO)4(η2-C2H4)2] (1).
Table 1 Selected experimental and calculated structural parameters for the trans-[W(CO)4(η2-alkene)2]0/+ complexesa
Alkene Ref. W–C(alkene) W–CO C[double bond, length half m-dash]C OC–W–CO W–C–O
a Bond lengths are given in angstroms, bond angles in degrees. b This work.
18-Electron complexes
C2H4 B3-LYP/Ab 2.349 2.050 1.418 174.24 179.54
  B3-LYP/Bb 2.362 2.055 1.411 172.76 179.54
  X-Ray18 2.299(9), 2.315(9) 2.033(10), 2.045(9) 1.413(13) 171.8(7) 175.1(13), 176.2(14)
cis-C8H14 X-Ray11 2.36(1), 2.38(1) 2.01(1), 2.03(1) 1.39(1) 173.9(3), 174.8(3) 176(1), 177(1), 178(1)
trans-C8H14 X-Ray20 2.327(3), 2.328(3) 2.026(5), 2.049(4) 1.412(6), 1.424(5) 176.2(1), 176.6(1) 176.8(3), 177.9(3)
CH2[double bond, length half m-dash]CHCO2CH3 X-Ray2 2.292(8), 2.310(7) 2.062(9), 2.043(10) 1.416(11), 1.396(13) 172.5(3), 172.3(3) 177.8(8), 175.7(8)
    2.306(9), 2.292(9) 2.039(8), 2.034(8)     179.2(7), 173.9(8)
17-Electron complex
C2H4 B3-LYP/Ab 2.466 2.081 1.388 175.10 178.71
  B3-LYP/Bb 2.483 2.086 1.381 176.85 178.77

cis-[W(CO)4(η2-C2H4)2] (2). cis-Bis(alkene) complexes have been proposed as key intermediates in catalytic cycles for [W(CO)6] catalysed reactions of alkenes. However, an exact structure for this kind of compound has not been determined by X-ray crystallographic studies up to now. In both condensed and gas phases, cis-[M(CO)4(η2-C2H4)2] (M[thin space (1/6-em)]=[thin space (1/6-em)]Cr, Mo, W) is thought to have C2ν symmetry with the two parallel ethene ligands.13,23 In contrast to this, no theoretical paper reporting the full geometry optimisation of a cis-bis(alkene) metal carbonyl complex has appeared so far. In the present work, the calculations on the B3-LYP/A and B3-LYP/B level were carried out for several initial orientations of the two ethene ligands in 2. We have localised four critical points on the potential energy surfaces.

Fig. 2 shows the four optimised structures of different symmetry: 2a (C2) 2a′ (C2ν) 2b (Cs) with mutually perpendicular C[double bond, length half m-dash]C bonds) and 2c (C2ν, with both olefinic bonds in the equatorial plane). The 2a′ structure possesses the postulated C2ν symmetry with two parallel ethene ligands. However, on the B3-LYP/B level, this structure is calculated to be a transition state with an imaginary frequency corresponding to the rotation of ethene ligands about the tungsten–ethene axes. This motion leads to the 2a structure of C2 symmetry. On the other hand, the 2a′ conformation is predicted to be a ground state at the B3-LYP/A level of calculations. The relative enthalpies and Gibbs energies of the four calculated conformers of 2 are given in Table 2. The results obtained for the 2a′ geometry on the B3-LYP/B level have been omitted. The conformation 2a of C2 symmetry with two slightly staggered C[double bond, length half m-dash]C bonds has the lowest enthalpy, but the enthalpies predicted for the 2a′ and 2b conformers is only about 1 kJ mol−1 higher. On the other hand, the predicted Gibbs energy difference between the four conformers indicates that 2a′ is thermodynamically favoured in comparison to 2b and 2a. However, the presented differences are very low (in the range of 1–3 kJ mol−1) and their order could possibly reversed if different theoretical methods are applied. Therefore, on the basis of the current calculations, it seems that both 2a (or 2a′) and 2b structures cannot be excluded as the most stable conformation. The third structure (2c) has clearly both higher enthalpy and Gibbs energy compared to the 2a, 2a′ and 2b conformers. Complete results of the thermochemical analysis of the calculated possible conformers of 2 are given in the ESI.


Theoretically predicted geometries for cis-[W(CO)4(η2-C2H4)2] (2).
Fig. 2 Theoretically predicted geometries for cis-[W(CO)4(η2-C2H4)2] (2).
Table 2 Enthalpies and Gibbs energies (kJ mol−1) calculated for the four idealised conformations of cis-[W(CO)4(η2-C2H4)2]  2 relative to 2a
  Method 2a (C2) 2a′ (C2ν) 2b (Cs) 2c (C2ν)
ΔH298 B3-LYP/A 0 0.6 1.4 10.5
  B3-LYP/B 0 — 1.2 11.4
ΔG298 B3-LYP/A 0 −3.1 −0.8 8.8
  B3-LYP/B 0 — −1.4 8.2

Selected structural parameters for 2a are presented in Table 3. The complete data for all four possible conformers of 2 are given in the ESI. The two carbons in each ethene ligand in 2a are bound at slightly different distances to tungsten (2.479 and 2.466 Å for B3-LYP/A, and 2.490 and 2.474 Å for B3-LYP/B). The two mutually trans carbonyls bind at a longer distance from tungsten (2.050 Å for B3-LYP/A and 2.056 Å for B3-LYP/B) than those in trans positions to the ethene ligands (1.993 and 2.000 Å with using A and B basis sets, respectively).

Table 3 Selected calculated structural parameters for the cis-[W(CO)4(η2-C2H4)2]0/+ complexes
Bond lengths and anglesa cis-[W(CO)4(η2-C2H4)2], 2a cis-[W(CO)4(η2-C2H4)2]+, 2+
B3-LYP/A B3-LYP/B B3-LYP/A B3-LYP/B
a Bond lengths are given in angstroms, bond angles in degrees.        
W–C1; W–C2 1.993 2.000 2.074 2.079
W–C3; W–C4 2.050 2.056 2.064 2.070
W–C5; W–C8 2.479 2.490 2.460 2.477
W–C6; W–C7 2.466 2.474 2.460 2.477
C5–C6; C7–C8 1.387 1.380 1.387 1.379
C1–W–C2 90.91 91.98 86.95 88.30
C3–W–C4 172.31 172.72 176.69 176.78
W–C1–O1; W–C1–O1 179.74 179.55 179.79 179.27
W–C3–O3; W–C4–O4 178.24 178.55 179.47 179.62

The DFT calculations indicate a thermodynamic instability of 2a (C2) relative to its trans-isomer 1 by about 10–12 kJ mol−1 (ΔG298, Table 4). The calculated enthalpy difference (ΔH) is in a good agreement with ab initio calculations at the SCF level22,23 that gave a value of 16.7 kJ mol−1. Complete results of the thermochemical analysis of the trans and cis isomers of the neutral molecules and their cationic forms are given in the ESI.

Table 4 The enthalpy and Gibbs energies differences (in kJ mol−1) between the redox pairs trans-[W(CO)4(η2-C2H4)2]0/+ (1/1+) and cis-[W(CO)4(η2-C2H4)2]0/+ (2/2+)
  1/2 1 +/2+
B3-LYP/A B3-LYP/B B3-LYP/A B3-LYP/B
ΔH298 14.2 15.3 −16.0 −13.0
ΔH213 14.0 15.1 −15.9 −13.0
ΔG298 10.0 11.9 −13.6 −9.8
ΔG213 11.2 12.8 −14.3 −10.7

[W(CO)5(η2-C2H4)] (3). The X-ray structure analysis of 3 (Fig. 3) shows the C[double bond, length half m-dash]C bond eclipsed to one of the two equatorial OC–W–CO axes of the square-pyramidal W(CO)5 unit, and perpendicular to the trans-located W–CO bond. This is in accordance with the results of a previous theoretical study of 325,26 and our DFT data. Selected structural parameters for 3 are presented in Table 5. Complete data set for the calculated geometry is given in the ESI. The three different W–C(carbonyl) bond lengths have been calculated for 3, two longer (in the equatorial plane) and one shorter (in an axial position). The two carbonyls staggered to the C[double bond, length half m-dash]C bond lie slightly closer to tungsten (2.056 Å) than those eclipsed to the olefinic bond (2.062 Å), using the B3-LYP/B level of calculations. The structural parameters obtained here are in good agreement with the results of previous MP2 calculations (see Table 5).26
Optimised geometry of [W(CO)5(η2-C2H4)] (3).
Fig. 3 Optimised geometry of [W(CO)5(η2-C2H4)] (3).
Table 5 Selected experimental and calculated structural parameters for the [W(CO)5(η2-alkene)] complexesa
Alkene Ref. W–C(alkene) W–CO C[double bond, length half m-dash]C OC–W–CO W–C–O
Eclp.b Stag.c trans d Eclp b Stag c Eclp b Stag c trans d
a Bond lengths are given in angstroms, bond angles in degrees. b W–CO bond eclipsed to C[double bond, length half m-dash]C bond. c W–CO bond staggered to C[double bond, length half m-dash]C bond. d W–CO bond trans to C[double bond, length half m-dash]C bond. e This work.
C2H4 B3-LYP/Ae 2.473 2.056 2.051 2.009 1.389 174.54 178.17 178.81 179.54 180.00
  B3-LYP/Be 2.486 2.062 2.056 2.017 1.382 174.95 176.90 179.10 179.68 180.00
  MP226 2.372 2.054   2.026 1.402          
cis-C8H14 X-Ray11 2.49(1), 2.51(1) 2.02(1) 1.97(1) 1.96(1) 1.38(2) 173.4(4) 178.3(5) 178(1) 178(1) 179(1)
      2.01(1) 2.00(1)         177(1) 179(1)  
trans-C8H14 X-Ray20 2.425(4), 2.451(5) 2.031(4) 2.047(5) 2.011(5) 1.384(6) 175.4(2) 177.0(2) 178.4(4) 179.3(4) 176.5(4)
      2.050(5) 2.069(5)         178.1(4) 177.6(4)  

It is interesting to compare the W–C(ethene) and the C[double bond, length half m-dash]C bond lengths in 1, 2 and 3 calculated with the B3-LYP/B method. The W–C(ethene) distances to both carbon atoms in 1 are distinctly shorter (2.362 Å) than those in 2 (2.490 and 2.474 Å) and 3 (2.486 Å), while the C[double bond, length half m-dash]C bond lengths show the opposite trend: 1 (1.411); 2 (1.380); and 3 (1.382 Å).

The mutually trans carbonyl ligands bind at nearly the same distance to the tungsten atom in alkene carbonyl compounds: 1 2.055; 2 2.056; and 3 2.062 and 2.056 Å, eclipsed and staggered to the C[double bond, length half m-dash]C bond, respectively.

Another significant diagnostic feature of these alkene carbonyl complexes is the W–CO bond length. For mutually trans carbonyl ligands, the W–CO distances are larger (2.055 Å for 1) than those for more firmly bound carbonyl ligands trans to ethene in 2 (2.000) and 3 (2.017 Å), consistent with the trans influence of the σ-donor and weaker π-acceptor alkene. As expected, analogous results were obtained using the B3-LYP/A method (Tables 1, 3 and 5).

trans- and cis-[W(CO)4(η2-C2H4)2]+ (1+ and 2+). Fig. 4 shows the optimised structures of the 17-electron cationic form of the trans and cis bis(ethene) compounds. Selected structural parameters are presented in Tables 1 and 3. The complete data sets are given in the ESI.
Optimised geometries of formally 17-electron d5 tungsten(I) complexes trans-[W(CO)4(η2-C2H4)2]+ (1+) and cis-[W(CO)4(η2-C2H4)2]+ (2+).
Fig. 4 Optimised geometries of formally 17-electron d5 tungsten(I) complexes trans-[W(CO)4(η2-C2H4)2]+ (1+) and cis-[W(CO)4(η2-C2H4)2]+ (2+).

The calculations (B3-LYP/B) have shown that one-electron oxidation of compound 1 causes an increase in the W–C(ethene) distances by 0.121 Å (Table 1). The elongation of the W–C(carbonyl) bond is smaller (0.031 Å). Conversely, there is a slight shortening of the C[double bond, length half m-dash]C bond length from 1.411 Å in 1 to 1.381 Å in 1+ (Table 1). Apparently, the single electron causes diminished dπ(W) π-back-donation to both CO and ethene ligands. In contrast to this, the one-electron oxidation of the cis-isomer 2 affects mainly the distance between tungsten and the carbonyl ligand in the trans position to the ethene ligand: 2.000 Å in 2 and 2.079 Å in 2+ (Table 3). The elongation of the W–C(carbonyl) distances for mutually trans carbonyl ligands in 2+ is smaller (0.014 Å). Also in this case, this phenomenon arises from diminished dπ(W) to π*(CO) π-back-donation upon oxidation. However, the π-back-donation, diminished upon oxidation of 2, is probably partly compensated for by increased σ-donation from the ethene ligands in 2+. This explanation nicely agrees with the shorter W–C(ethene) distances in 2+ compared to those in parent compound 2 (Table 3) and with the higher stability of 2+ compared to 1+. Indeed, there is an opposite energetic difference between the 1/2 and 1+/2+ redox pairs of isomers. The calculated potential energies indicate that 17-electron cis isomer 2+ is more stable than trans1+ by about 14 or 10 kJ mol−1 (ΔG at B3-LYP/A or B3-LYP/B levels, respectively, Table 4).

IR characterisation of the alkene carbonyl complexes of tungsten

The calculated (B3-LYP/B, scaling factor of 0.961439) and experimental ν(CO) vibration frequencies for 1, 2 and 3 and the corresponding 17-electron cations 1+ and 2+ are presented in Table 6. The comparison of the calculated and experimental ν(CO) wavenumbers and relative intensities show good agreement, in particular for the intense e mode of compound 1.
Table 6 Calculateda and experimental freqencies (cm−1) and intensities (in parentheses) for CO stretching bands of the studied alkene carbonyl complexes of tungsten
[W(CO)6] (Oh)
 
a1g eg t1u Ref. Medium
 
2109 (0) b 2019 (0) b 1996 This work Gas phase
2126 (w) 2021 (vw) 1998 (vvs) 40 Gas phase
 
[W(CO)5 (η2-C2H4)] (C2ν)
a1 a1 B2 a1 b1 Ref. Medium
2077 (0.07) 2002 (0.001) 1981 (0.92) 1976 (0.43) 1967 (1) This work Gas phase
2083 (w) —c —c 1973 (s) 1955 (vs) 18 n-Hexane, 293 K
 
trans-[W(CO)4(η2-C2H4)2] (D2d)
a1 b2 e 13CO modes Ref. Medium
2055 (0) c 1996 (0.04) 1968 (1)   This work Gas phase
2058 (vw) 1990 (w) 1966 (vs) 1928 (vw), 1917 (vw) 18 n-Hexane, 293 K
2053.9 (vw) 1999 (w) 1967.6 (vs) 1936.8 (vw) 18 Ar, 14 K
2058.6 (vw) 1989.8 (w) 1953.6 (vs) 1928 (vw), 1917 (vw) This work d n-Hexane, 233 K
2056 (w) 1986 (w) 1944 (vs)   This work d CH2Cl2, 223 K
 
trans-[W(CO)4(η2-C2H4)2]+ (D2)
a b1 b3 b2 Ref. Medium
2114 (0) c 2051 (0.001) 2041 (0.88) 2040 (1) This work Gas phase
 
cis-[W(CO)4(η2-C2H4)2] (C2)
a a b b Ref. Medium
2042 (0.21) 1971 (0.27) 1967 (1) 1938 (0.66) This work Gas phase
2057.3 (w) 1957.5 (s) 1954.1 (s) 1897.2 (m) 18 Ar, 14 K
2050 (w) —c —c 1913 (m) 18 n-Hexane, 293 K
2050 1957 —c 1910 4 Xe, 195 K
 
cis-[W(CO)4(η2-C2H4)2]+ (C2ν)
a1 a1 b1 b2 Ref. Medium
a The scaling factor 0.961439 at the B3-LYP/B level was used. b IR inactive. c Not observed. d For [W(CO)4(η2-C4H8)2].
2108 (0.25) 2055 (0.21) 2052 (0.48) 2037 (1) This work Gas phase
2127 (0.06) 2058 (0.41) 2035 (0.97) 2005 (1) This work d CH2Cl2, 223 K

The IR spectrum of compound 1 in n-hexane solution shows a characteristic ν(CO) pattern, with a very intense band at 1966 cm−1, a weak one at 1990 cm−1 and a very weak one at 2058 cm−1. These spectral features are fully consistent with the pseudo-octahedral geometry and D2d symmetry18 of 1 (Table 6). Two additional very weak ν(CO) bands at 1928 and 1917 cm−1 are observed in solution. In an argon matrix at ca. 14 K only one band was detected in this region at 1937 cm−1. It has been assigned to the isotopomer trans-[W(12CO)3(13CO)(η2-C2H4)2].18

The cis isomer of [W(CO)4(η2-alkene)2] was first spotted by IR spectroscopy as a thermally unstable product formed during the photolysis of [W(CO)6] in the presence of C3H6 in a rigid alkane glass at 77 K.3 Compound 2 was identified as a photoproduct of [W(CO)6] in ethene-doped LXe solution at 195 K.4 Furthermore, it has been proven that cis-bis(alkene) complexes also result from the photochemical isomerisation of trans-[W(CO)4(η2-alkene)2] compounds at low temperatures.16 Assignment of the cis geometry was based on their IR spectra, which exhibit four CO stretching bands, as expected from the C2 symmetry of 2 (Table 6). For cis-[W(CO)4(η2-C2H4)2](2) in LXe at 195 K,4 three ν(CO) bands were detected at 2050, 1957 (probably two unresolved bands) and 1910 cm−1, while only two of the four ν(CO) bands were observed18 at 2050 and 1913 cm−1 in n-hexane at room temperature (the other two ν(CO) bands are obscured by absorptions associated with the trans isomer 1). A comparison of the calculated and observed ν(CO) frequencies for 2 reveals the largest difference (ca. 25 cm−1) for the lowest energy b mode (vibration of the two CO groups trans to the ethene ligands). The data in Table 6 clearly shows the influence of solvent dipoles if we compare the position of the ν(CO) bands in the gas phase, n-hexane and CH2Cl2 solution.

It is generally accepted that predominantly metal-centred one-electron oxidation of metal carbonyls leads to an increased CO stretching frequency (Δν) by ca. 100 cm−1.21,42–45 This shift results from significantly reduced π-back-donation to the π*(CO) orbitals and strengthening of the C[triple bond, length half m-dash]O bonds. A comparison of the predicted ν(CO) frequencies for the e mode of 1 and b2 mode of 1+ gives a Δν value of 72 cm−1 (Table 6).

Upon one-electron oxidation of compound 2 the four ν(CO) bands shift to larger wavenumbers by 66, 84, 70 and 114 cm−1, respectively. The order of the two b modes of 2 (1967 and 1938 cm−1) is reversed in 2+. The higher energy b mode for 2 (1967 cm−1) and the b2 mode for 2+ (2037 cm−1) correspond to the vibration of the two mutually trans CO ligands. The lower energy b mode for 2 (1938 cm−1) and the b1 mode for 2+ (2052 cm−1) are assigned to vibrations of the two CO ligands in trans positions to the alkene. The different Δν of 70 cm−1 for the two mutually trans CO ligands and Δν of 114 cm−1 for the CO ligands trans to the alkenes correlate with the distinct effects of the one-electron oxidation of the tungsten centre on the W–CO bond distances. Calculated elongation of the W–CO bonds in 2 is indeed larger for carbonyls in positions trans to alkene ligands (0.079 Å) compared to that for the mutually trans carbonyl ligands (0.014 Å).

Oxidation-induced trans–cis isomerisation of [W(CO)4(η2-alkene)2]

Isomerisation of Group 6 metal(0) carbonyls induced by electrochemical oxidation of the metal centre is a known phenomenon since the 1980's.42–45 In the early 1990's oxidation of trans-[W(CO)4(η2-alkene)2] (alkene[thin space (1/6-em)]=[thin space (1/6-em)]1-C5H10, 1-C6H12, C5H10, C6H10, C7H12 and C8H14) was also shown to produce the corresponding 17-electron cations in the cis configuration.12 However, uncertainty has remained with regard to the mutual position of the WI/W0 redox couples for the trans and cis isomers. Herein, trans-[W(CO)4(η2-1-butene)2] (1A) has been chosen for a more detailed spectroelectrochemical study.
Cyclic voltammetry. The cyclic voltammogram of 1A in dichloromethane shows a well-defined anodic wave at E1/2[thin space (1/6-em)]=[thin space (1/6-em)]+0.62 V vs. Fc/Fc+ (ΔEp[thin space (1/6-em)]=[thin space (1/6-em)]90 mV, comparable with 80 mV for Fc/Fc+ internal standard). At room temperature, the oxidation becomes a completely chemically reversible one-electron process at scan rates ν higher than 1 V s−1. For scan rates between 50–200 mV s−1 a small cathodic peak is observed on reverse cathodic scanning at Ep[thin space (1/6-em)]=[thin space (1/6-em)]−0.51 V vs. Fc/Fc+ [Fig. 5(a)]. Its assignment to back-reduction of cis-[W(CO)4(η2-1-butene)2]+ (2A+) was facilitated by a thin-layer cyclic voltammogram recorded during the IR spectroelectrochemical experiment at 223 K [Fig. 5(b)].
(a) Cyclic voltammogram of ca. 10−3 M trans-[W(CO)4(η2-1-butene)2] (1A) in CH2Cl2–0.1 M Bu4NPF6 at 293 K, ν = 200 mV s−1. (b) Thin-layer cyclic voltammogram of ca. 10−2 M complex 1A in CH2Cl2–0.3 M Bu4NPF6 at 223 K, ν = 2 mV s−1, recorded during the IR spectroelectrochemical experiment depicted in Fig. 6. The dashed vertical lines indicate decreased current during electrolyses at constant potential. Anodic peak A: oxidation of 1A producing 2A+; cathodic peak B: reverse reduction of 2A+ regenerating 1A.
Fig. 5 (a) Cyclic voltammogram of ca. 10−3 M trans-[W(CO)4(η2-1-butene)2] (1A) in CH2Cl2–0.1 M Bu4NPF6 at 293 K, ν[thin space (1/6-em)]=[thin space (1/6-em)]200 mV s−1. (b) Thin-layer cyclic voltammogram of ca. 10−2 M complex 1A in CH2Cl2–0.3 M Bu4NPF6 at 223 K, ν[thin space (1/6-em)]=[thin space (1/6-em)]2 mV s−1, recorded during the IR spectroelectrochemical experiment depicted in Fig. 6. The dashed vertical lines indicate decreased current during electrolyses at constant potential. Anodic peak A: oxidation of 1A producing 2A+; cathodic peak B: reverse reduction of 2A+ regenerating 1A.

IR spectral changes in the CO stretching region resulting from reversible one-electron oxidation of ca. 10−2 M trans-[W(CO)4(η2-1-butene)2] (1A) in CH2Cl2 at 223 K within an OTTLE cell.41
Fig. 6 IR spectral changes in the CO stretching region resulting from reversible one-electron oxidation of ca. 10−2 M trans-[W(CO)4(η2-1-butene)2] (1A) in CH2Cl2 at 223 K within an OTTLE cell.41

There is no indication of electrocatalytic conversion of trans-[W(CO)4(η2-1-butene)2] (1A) upon oxidation to neutral cis-[W(CO)4(η2-1-butene)2] (2A ). The latter cis isomer is thus oxidised less positively than the parent trans isomer 1A.

IR spectroelectrochemistry. IR ν(CO) changes associated with the one-electron oxidation of 1A were first studied in CH2Cl2 at 223 K within a low-temperature IR OTTLE cell [Fig. 5(b) and 6]. The ν(CO) bands due to 1A [2056 (w), 1986 (w-m) and 1944 (vs, br) cm−1] were gradually replaced with a set of four new bands of a cationic carbonyl product at 2127 (w), 2058 (m), 2035 (s) and 2005 (s) cm−1 (Fig. 6).

The intensity pattern of the ν(CO) bands is in agreement with C2ν symmetry the W(CO)4 skeleton (Table 6). For the neutral cis isomer 2A, only two out of the predicted four ν(CO) bands (2041 and 1901 cm−1) were observed17 in n-hexane at 263 K. The IR spectrum of the electrochemically generated cation 2A+ with cis configuration shows the expected high frequency ν(CO) shift. According to data in Table 6, an intense ν(CO) band around 2030 cm−1 should be observed for the trans isomer 1A+ (2040 cm−1 predicted for the ethene derivative 1+). At this frequency, however, a strong band due to 2A+ also arose, thus a small quantity of the primary oxidation product 1A+ is hardly detectable by IR spectroscopy in a mixture with 2A+. However, no 1A+ was detected by thin-layer cyclic voltammetry in CH2Cl2 at 223 K upon reverse cathodic scanning beyond the anodic peak of parent 1 [Fig. 5(b)]. The only back-reduction occurred at an electrode potential of ca. −0.5 V vs. Fc/Fc+; a potential difference larger than 1 V was observed between the oxidation of 1 and the reverse cathodic step. This observation proves exclusive formation of 2A+, which is sufficiently stable at 223 K. This result points to rapid oxidation-induced trans–cis isomerisation of 1A+ to 2A+, although 1A+ is still observable by cyclic voltammetry even at room temperature (see above). It is noteworthy that the back-reduction of 2A+ [Fig. 5(b)] resulted in direct complete recovery of the starting neutral trans isomer 1. The reverse reduction induced cis–trans isomerisation of 2A+ to 1 is therefore comparably fast.

The anodic IR spectroelectrochemical experiment performed at room temperature resulted in rapid decarbonylation, indicated by disappearance of the ν(CO) absorption and evolution of CO gas at the anodic surface. Only a potential-step experiment and rapid oxidation of 1 within a few seconds led to a detectable weak ν(CO) absorption due to the cis isomer 2A+. According to the cyclic voltammetric response, the primary trans isomer 1A+ is thermally unstable at room temperature on a subsecond time scale and could therefore hardly be detected in the course of the IR spectroelectrochemical oxidative experiment. Also this result supports assignment of the ν(CO) bands at 2127 (w), 2058 (m), 2035 (s) and 2005 (s) cm−1 (Fig. 6) to 2A+. Thus, direct spectroscopic evidence for the existence of cis-[W(CO)4(η2-alkene)2]+ has now been obtained.

Conclusions

The results presented on the cyclic voltammetric and IR spectroelectrochemical studies at variable temperatures have unequivocally confirmed the identity of the formally 17-electron cationic species cis-[W(CO)4(η2-alkene)2]+ as the ultimate carbonyl product of the one-electron oxidation of trans-[W(CO)4(η2-alkene)2]. Our theoretical results support the experimental data, showing that trans-[W(CO)4(η2-C2H4)2] is more stable by about 10 kJ mol−1 (ΔG298, B3-LYP/A) or 12 kJ mol−1 (ΔG298, B3-LYP/B) than cis-[W(CO)4(η2-C2H4)2], but stability of the cationic forms of these compounds is reversed; the cationic cis isomer is preferred by about 14 or 10 kJ mol−1 (ΔG298, B3-LYP/A and B3-LYP/B, respectively).

Acknowledgements

Computing resources from the Academic Computer Centre CYFRONET UMM (SGI Origin2000 computer, grant no. KBN/SGI_ORIGIN_2000/PK/109/1999) and grant no. KBN/3 T09A09317 are gratefully acknowledged. We also owe our thanks to K. Kern for synthesising the studied alkene complexes.

References

  1. I. W. Stolz, G. R. Dobson and R. K. Sheline, Inorg. Chem., 1963, 2, 1264 CrossRef CAS.
  2. F.-W. Grevels, M. Lindemann, R. Benn, R. Goddard and C. Kröger, Z. Naturforsch., B: Anorg. Chem. Org. Chem., 1980, 35, 1298 Search PubMed.
  3. K. R. Pope and M. S. Wrighton, Inorg. Chem., 24, 2792 Search PubMed.
  4. M. F. Gregory, S. A. Jackson, M. Poliakoff and J. J. Turner, J. Chem. Soc., Chem. Commun., 1986, 1175 RSC.
  5. S. A. Jackson, R. K. Upmacis, M. Poliakoff, J. J. Turner, J. K. Burdett and F.-W. Grevels, J. Chem. Soc., Chem. Commun., 1987, 678 RSC.
  6. F.-W. Grevels, J. Jacke and S. Özkar, J. Am. Chem. Soc., 1987, 109, 7536 CrossRef CAS.
  7. F.-W. Grevels, J. Jacke, P. Betz, C. Krüger and Y.-H. Tsay, Organometallics, 1989, 8, 293 CrossRef CAS.
  8. G. Davidson and C. L. Davies, Inorg. Chim. Acta, 1989, 165, 231 CrossRef CAS.
  9. S. A. Jackson, P. M. Hodges, M. Poliakoff, J. J. Turner and F.-W. Grevels, J. Am. Chem. Soc., 1990, 112, 1221 CrossRef CAS.
  10. P. M. Hodges, S. A. Jackson, J. Jacke, M. Poliakoff, J. J. Turner and F.-W. Grevels, J. Am. Chem. Soc., 1990, 112, 1234 CrossRef CAS.
  11. J. M. Dalla Riva Toma, P. H. Toma, P. E. Fanwick, D. E. Bergstrom and S. R. Byrn, J. Crystallogr. Spectrosc. Res., 1993, 23, 41 Search PubMed.
  12. M. Wilgocki, T. Szymańska-Buzar, M. Jaroszewski and J. J. Ziółkowski, in Molecular Electrochemistry of Inorganic, Bioinorganic and Organometallic Compounds, ed. A. J. L. Pombeiro and J. A. McCleverty, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1993, p. 573. Search PubMed.
  13. H. Takeda, M. Jyo-o, Y. Ishikawa and S. Arai, J. Phys. Chem., 1995, 99, 4558 CrossRef CAS.
  14. M. Jaroszewski, T. Szymańska-Buzar, M. Wilgocki and J. J. Ziółkowski, J. Organomet. Chem., 1996, 509, 19 CrossRef CAS.
  15. T. Szymańska-Buzar, M. Jaroszewski, W. Wilgocki and J. J. Ziółkowski, J. Mol. Catal. A: Chem., 1996, 112, 203 CrossRef CAS.
  16. T. Szymańska-Buzar, M. Jaroszewski, A. J. Downs, T. M. Greene and L. J. Morris, J. Organomet. Chem., 1997, 531, 207 CrossRef CAS.
  17. T. Szymańska-Buzar, K. Kern and D. J. Stufkens, New J. Chem., 1998, 22, 1539 RSC.
  18. T. Szymańska-Buzar, K. Kern, A. J. Downs, T. M. Greene, L. J. Morris and S. Parsons, New J. Chem., 1999, 23, 407 RSC.
  19. T. Szymańska-Buzar and K. Kern, J. Organomet. Chem., 1999, 592, 212 CrossRef CAS.
  20. F.-W. Grevels, J. Jacke, W. E. Klotzbücher, F. Mark, V. Skibbe, K. Schaffner, K. Angermund, C. Krüger, C. W. Lehmann and S. Özkar, Organometallics, 1999, 18, 3278 CrossRef CAS.
  21. I. R. Farrell, F. Hartl, S. Záliš, M. Wanner, W. Kaim and A. Vlček, Jr., Inorg. Chim. Acta, 2001, 318, 143 CrossRef CAS.
  22. C. Bachmann, J. Demuynck and A. Veillard, J. Am. Chem. Soc., 1978, 100, 2366 CrossRef CAS.
  23. C. Daniel and A. Veillard, Nouv. J. Chim., 1986, 10, 83 Search PubMed.
  24. C. Daniel and A. Veillard, Inorg. Chem., 1989, 28, 1170 CrossRef CAS.
  25. T. A. Albright, R. Hoffmann, J. C. Thibeault and D. L. Thorn, J. Am. Chem. Soc., 1979, 101, 3801 CrossRef CAS.
  26. U. Pidun and G. Frenking, Organometallics, 1995, 14, 5325 CrossRef CAS.
  27. T. Szymańska-Buzar, J. Mol. Catal., 1988, 48, 43 CrossRef CAS.
  28. T. Szymańska-Buzar, J. Mol. Catal. A: Chem., 2000, 160, 133 CrossRef CAS.
  29. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al- Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, and J. A. Pople, GAUSSIAN 98, rev. A.9, Gaussian, Inc., Pittsburgh PA, 1998..
  30. A. D. Becke, J. Chem. Phys., 1993, 98, 5648 CrossRef CAS.
  31. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299 CrossRef CAS.
  32. M. J. Frisch, E. Frisch and J. B. Foresman, GAUSSIAN 94 User’s Reference, Gaussian, Inc., Pittsburgh PA, 1996. Search PubMed.
  33. R. Bauernschmitt and R. Ahlrichs, J. Chem. Phys., 1996, 104, 9047 CrossRef CAS.
  34. H. B. Schlegel, J. Chem. Phys., 1982, 77, 3676 CrossRef CAS.
  35. C. Peng, P. Y. Ayala, H. B. Schlegel and M. J. Frisch, J. Comput. Chem., 1996, 17, 49 CrossRef CAS.
  36. B. Jost and A. Rees, Acta Crystallogr., Sect. B, 1975, 31, 2649 CrossRef.
  37. J. Li, G. Schreckenbach and T. Ziegler, J. Am. Chem. Soc., 1995, 117, 486 CrossRef CAS.
  38. A. W. Ehlers, Y. Ruiz-Morales, E. J. Baerends and T. Ziegler, Inorg. Chem., 1997, 36, 5031 CrossRef CAS.
  39. A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16[thin space (1/6-em)]502 CAS.
  40. L. H. Jones, R. S. McDowell and M. Goldblatt, Inorg. Chem., 1969, 8, 2349 CrossRef CAS.
  41. F. Hartl, H. Luyten, H. A. Nieuwenhuis and G. C. Schoemaker, Appl. Spectrosc., 1994, 48, 1522 Search PubMed.
  42. A. M. Bond and R. Colton, Coord. Chem. Rev., 1997, 166, 181 CrossRef CAS and references therein.
  43. S. Sun and D. A. Sweigart, Adv. Organomet. Chem., 1997, 40, 171 Search PubMed and references therein.
  44. K. A. Conner and R. A. Walton, Organometallics, 1983, 2, 169 CrossRef CAS.
  45. R. D. Rieke, H. Kojima, T. Saji, P. Rechberger and K. Öfele, Organometallics, 1988, 7, 749 CrossRef CAS.

Footnote

† Electronic supplementary information (ESI) available: Tables S1–S18 Cartesian coordinates for fully optimised geometries of trans-[W(CO)4(η2-C2H4)2], trans-[W(CO)4(η2-C2H4)2]+, cis-[W(CO)4(η2-C2H4)2], cis-[W(CO)4(η2-C2H4)2]+, [W(CO)5(η2-C2H4)], and [W(CO)6], and their corresponding energies (E0, H and G). See http://www.rsc.org/suppdata/nj/b1/b104786h/
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002
Click here to see how this site uses Cookies. View our privacy policy here.