Tertiary phosphine induced migratory carbonyl insertion in cyclopentadienyl complexes of iron(II)

Article abstract of DOI:10.1016/j.jorganchem.2005.06.020

Cyclopentadienyldicarbonylmethyliron, [CpFe(CO)₂Me] (1), undergoes migratory carbonyl insertion under the influence of isosteric phosphine ligands P(4-FC₆H₄)₃ and P(4-MeC ₆H₄)₃. The products of the reaction, [CpFe(CO)(COMe)P(4-FC₆H₄)₃] (2a) and [CpFe(CO)(COMe)P(4-MeC₆H₄)₃] (2b), were characterised by X-ray crystallography. In both structures, the iron atom adopts a pseudo octahedral coordination geometry. Fe-P bond distances are the same at 2.1932(8) ? in 2a and 2b, respectively. Thus, contrary to what was expected, X-ray data could not be used to quantitatively differentiate between the two phosphine ligands in 2a and 2b. Therefore, additional spectroscopic techniques such as IR and NMR were employed. Similarly, the Fe-C bond lengths of the carbonyl (Fe-CO) and acetyl (Fe-COMe) are 1.748(3) and 1.955(3) in 2a, and 1.744(3) and 1.951(3) ? in 2b, respectively. The migratory carbonyl insertion was studied by NMR, IR, and UV-vis spectroscopies to determine the mechanism and the rate law. Results from NMR spectroscopy show that the formation of the product is accompanied by oxidation of the corresponding phosphine ligand. An increase in the reactivity of migratory carbonyl insertion for P(4-MeC₆H₄)₃ was observed when the solvent was changed from CH₂Cl₂ to MeCN. The kinetic data showed that P(4-MeC₆H₄)₃ reacts faster than P(4-FC₆H₄)₃.

Full text of DOI:10.1016/j.jorganchem.2005.06.020

Journal of Organometallic Chemistry 690 (2005) 4159–4167
www.elsevier.com/locate/jorganchem
Tertiary phosphine induced migratory carbonyl insertion
in cyclopentadienyl complexes of iron(II)
b,
b
b,
*
Ntaoleng M. Makunya a,1, Reinout Meijboom *, Alfred Muller , Andreas Roodt
a
Department of Chemistry, University of Johannesburg, P.O. Box 524, Auckland Park, 2006 Johannesburg, South Africa
b
Department of Chemistry, University of the Free State, P.O. Box 339, 9300 Bloemfontein, South Africa
Received 12 May 2005; received in revised form 7 June 2005; accepted 20 June 2005
Available online 26 July 2005
Abstract
Cyclopentadienyldicarbonylmethyliron, [CpFe(CO)2Me] (1), undergoes migratory carbonyl insertion under the influence of isos-
teric phosphine ligands P(4-FC6H4)3 and P(4-MeC6H4)3. The products of the reaction, [CpFe(CO)(COMe)P(4-FC6H4)3] (2a) and
[CpFe(CO)(COMe)P(4-MeC6H4)3] (2b), were characterised by X-ray crystallography. In both structures, the iron atom adopts a
˚
pseudo octahedral coordination geometry. Fe–P bond distances are the same at 2.1932(8) A in 2a and 2b, respectively. Thus, con-
trary to what was expected, X-ray data could not be used to quantitatively differentiate between the two phosphine ligands in 2a and
2b. Therefore, additional spectroscopic techniques such as IR and NMR were employed. Similarly, the Fe–C bond lengths of the
˚
carbonyl (Fe–CO) and acetyl (Fe–COMe) are 1.748(3) and 1.955(3) in 2a, and 1.744(3) and 1.951(3) A in 2b, respectively.
The migratory carbonyl insertion was studied by NMR, IR, and UV–vis spectroscopies to determine the mechanism and the rate
law. Results from NMR spectroscopy show that the formation of the product is accompanied by oxidation of the corresponding
phosphine ligand. An increase in the reactivity of migratory carbonyl insertion for P(4-MeC6H4)3 was observed when the solvent
was changed from CH2Cl2 to MeCN. The kinetic data showed that P(4-MeC6H4)3 reacts faster than P(4-FC6H4)3.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Iron; Migratory carbonyl insertion; X-ray crystallography; Kinetics
1. Introduction
Factors known to govern the rate of migratory CO
insertion reaction include the entering ligand [3], solvent
[4,5] and alkyl group [6]. Ligand steric and electronic ef-
fects play a key role in determining organometallic reac-
tivity trends and catalytic behavior. Phosphine based
ligands have found widespread application both in orga-
nometallic chemistry and in industrial application of
homogeneous catalysis, because their steric demands
as well as their r-donor and p-acceptor properties are
subject to deliberate control [7]. For example, in the
hydroformylation process, phosphine modified systems
offer great benefits over unmodified systems, including
increased catalyst stability [8], improved reaction rates
and selectivity, and enhanced partitioning in two-phase
systems [9].
Migratory carbonyl insertion serves as a key C–C
bond formation pathway in catalytic carbonylations in
the chemical industry, such as in the acetic acid synthesis
from methanol with CO [1]. Transition metal alkyl and
acyl compounds involved can be used as models for al-
kyl intermediates in the Fischer–Tropsch reactions. In
general, Groups 8–10 metals are known to be active cat-
alysts for the Fischer–Tropsch reaction, but to date only
iron based catalysts are used industrially [2].
*
Corresponding authors. Tel.: +27 51 401 2547; fax: +27 51 430
7805.
E-mail addresses: meijboomr.sci@mail.uovs.ac.za (R. Meijboom),
roodta.sci@mail.uovs.ac.za (A. Roodt).
1
Two main aspects are involved in the coordination of
tertiary phosphine ligands to transition metal atoms.
Current address: Department of Chemistry, University of the Free
State, P.O. Box 339, 9300 Bloemfontein, South Africa.
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.06.020

4160
N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
K1, k1
k2
Intermediate
[CpFe(COMe)(CO)PR3]
[CpFe(CO)2Me] + PR3
k-1
1
2a/2b
Scheme 1. General Scheme for the two-step reaction with a fast pre-equilibrium as reported previously [16]. kobs = (k2 K1[PR3])/(1 + K1[PR3]); kobs
represents the pseudo-first order rate constant ([P] ꢁ [Fe]).
Firstly, the electronic character of the M-PR3 bond
which is a combined effect of the r-bond formed by
donation of lone pair electrons from the phosphine li-
gand to the metal atom, and the ability of the phosphine
ligand to accept electron density from the metal back
into the appropriate phosphine orbital [10]. The r-basi-
city and p-acidity of the phosphorous ligand has been
studied by Tolman by comparing the carbonyl stretch-
ing absorption bands of the coordinated CO ligands in
[Ni(CO)3(PR3)] complexes [11]. Strong r-donor ligands
give a high electron density on the metal centre and
therefore a substantial back-donation to the CO ligands
and lowered IR frequencies. Strong p-acceptor ligands
on the other hand, compete with CO for the electron
back-donation and the CO stretch frequencies remain
high [7]. The r-donor ability of the tertiary phosphine
ligands is expected to correlate with the pKa values of
HPRþ3 . Substituents with better electron donating capa-
bilities promote Lewis basicity of the phosphine, while
p-acidity is promoted by electron withdrawing groups
[12].
the phosphine ligand is followed by simultaneous co-
ordination of the phosphine ligand and migration of
the alkyl group. Our long term interest into elucidation
of the electronic and steric effects of Group 15 ligands
[17] led us to the present study, which was conducted
to structurally and kinetically investigate the electronic
effects of two isosteric ligands, P(4-FC6H4)3 and P(4-
MeC6H4)3. The kinetics of the migratory CO insertion
reaction in [CpFe(CO)2Me] (1), was studied using UV–
vis, IR, 19F and 31P NMR spectroscopies, which con-
firmed the general two-step reaction represented in
Scheme 1. The resulting products were fully character-
ised and their molecular structures were determined by
single crystal X-ray spectroscopy.
2. Experimental
2.1. General
All syntheses of air and moisture sensitive com-
pounds were performed using standard Schlenk tech-
niques under prepurified N2 [18]. All solvents were
pre-dried by passage over alumina (neutral, Brockmann
grade I). MeCN was distilled from P2O5, tetrahydrofu-
ran from Na/benzophenone and CH2Cl2 from CaH2
prior to use [19]. [CpFe(CO)2]2 and [CpFe(CO)2Me]
(1), were prepared according to literature procedures
[20]. Fe(CO)5, dicyclopentadiene and the phosphine li-
gands were obtained from Sigma–Aldrich and used as
received. Melting points of 2a and 2b were determined
on a Kofler hot-stage microscope (Reichert-Therm-
ovar). Infrared spectra were recorded on a Bruker Equi-
nox 55 FT-IR spectrometer and analysed with the
Bruker OPUS-NT software (32 scans, 4 cmꢀ1 resolu-
tion, Blackman-Harris 3-Term apodisation). Infrared
data for solution spectra of compounds 2a and 2b were
collected using NaCl windows (optical path 0.1 mm).
NMR spectra were recorded on a Varian Gemini 2000
spectrometer (1H: 300 MHz, 19F: 282.3 MHz, 31P:
121.5 MHz, 13C: 75.5 MHz). NMR spectra were refer-
enced relative to TMS (1H, 13C), SOCF3 (19F) or 85%
H3PO4 (31P) using either the residual protonated impu-
rities in the solvent (1H NMR), the solvent resonance
(13C NMR), or an external reference (19F, 31P). UV–
vis spectra were recorded on a Varian Cary 50 spectro-
meter equipped with a thermostated water bath accurate
to 0.1ꢁ in a double-sided 1 cm tandem quartz cells with
Teflon caps.
According to the Quantitative Analysis of Ligand Ef-
fects (QALE) hypothesis [13], the nature of the metal
phosphorous bond has a strong influence on the electron
density around the metal center. Strong r-donors en-
hance the electron density, whereas p-acceptors decrease
the electron density. This model leads to the conclusion
that the Fe–P bond distance can be utilised to diagnose
the r-donor or a p-acceptor character of the phosphine
ligand. It predicts that pure r-donor ligands exhibit long
˚
Fe–P bond distances of ca. 2.20 A whereas r donor/p-
acceptor ligands should have shorter distances of ca.
˚
2.10 A [14]. The here reported acyl complexes [CpFe-
(CO)(COMe)PR3] (R = 4-FC6H4 in 2a and 4-MeC6H4
in 2b) are crystalline materials and consequently, charac-
terised by X-ray crystallography to evaluate both the
electronic and steric properties of the ligands.
Previous kinetic studies indicated a mechanism in
which alkyl migration in [CpFe(CO)2Me] from metal
to CO occurs in the first step and an unsaturated 16-
electron, or solvent saturated acyl intermediate, complex
is generated [15]. Attack of a nucleophile on the interme-
diate produces the final product. These studies showed
that nucleophilic and polar solvents increase the inser-
tion rate, and the authors attributed this effect to the sta-
bilisation of the unsaturated intermediate by the solvent.
Bassetti and co-workers [16] proposed a mechanism
in which a rapid formation of an outer-sphere complex
between the alkyl complex, [g5-C9H7Fe(CO)2Me], and

N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
4161
2.2. Structure determination
2.3. Preparation of [CpFe(CO)(COMe)P(4-FC6H4)3]
(2a)
Crystals of [CpFe(CO)(COMe)P(4-FC6H4)3] (2a),
and [CpFe(CO)(COMe)P(4-FC6H4)3] (2b), were grown
from pentane as described below. X-ray diffraction data
for 2a and 2b were collected on the Bruker SMART 1K
CCD area detector diffractometer using a graphite
A solution of P(4-FC6H4)3 (160 mg, 0.5 mmol) in
MeCN (2.0 ml) was added to a solution of 1 (100 mg,
0.5 mmol) in MeCN (2.0 ml). The mixture was heated
under reflux for ca. 48 h. The solvent was evaporated
in vacuo and the residue dissolved in acetone and pen-
tane (1:20). The product was purified by column chro-
matography over silica, eluting with hexane. The first
yellow band (Rf = 0.34) was collected. Evaporation of
the solvent gave crude 2a as a crystalline yellow mate-
rial. Recrystallisation from pentane afforded crystals
suitable for X-ray studies. (0.14 g, 50%) m.p. 151 ꢁC.
˚
monochromated Mo Ka (k = 0.71073 A) collecting x-
scans at 293(2) K. After completion, the first 50 frames
were recollected to check for decay, which was not ob-
served. All the reflections were merged and integrated
using SAINT+ [21]. The structures were solved by the hea-
vy atom method and refined through full-matrix least
squares cycles using the SHELXL-97 [22] software package
P
2
with
ðjF o j ꢀ jF c jÞ being minimised. All non-H
m
max/cmꢀ1 (CHCl3) 1918 (CO); mmax/cmꢀ1 (KBr) 1904
atoms were refined with anisotropic displacement
parameters, while the H atoms were constrained to par-
ent sites using a riding model. The Diamond Visual
Crystal Structure information system software was used
for the graphics [23]. Crystal data and details of data
collection and refinement are presented in Table 1.
CCDC reference numbers 269347 and 269348. See
http://www.ccdc.cam.ac.uk.
(CO) 1607 (COMe); d {H} (CDCl3, 300 MHz,) 2.32
(3H, s, COCH3), 4.39 (5H, s, C5H5), 7.05–7.68 (12H,
m, C6H4); dF{H} (CDCl3, 282.3 MHz) ꢀ112.30; dP{H}
(CDCl3, 121.5 MHz) 75.40; dC{H} (CDCl3, 75.5 MHz)
52.5 (COCH3), 85.6 (C5H5), 115.9 (dd, J = 10 Hz, 3,5-
(C6H4), 128.4 (d, J = 9 Hz, (C6H4), 133.6 (t, J =
10 Hz, 4-C6H4), 135.7 (t, J = 10 Hz, 2,6-(C6H4), CO
and COMe were not observed; kmax/nm (CH2Cl2)
Table 1
Crystal data and structural refinement for [CpFe(CO)(COMe)P(4-FC6H4)3] (2a) and [CpFe(CO)(COMe)P(4-MeC6H4)3] (2b)
2a
2b
Emprical formula
Formula weight
T (K)
Wavelength A
Crystal system
C
26H20F3FeO2P
C26H29FeO2P
496.34
293(2)
0.71073
Triclinic
508.24
293(2)
0.71073
Triclinic
˚
ꢀ
ꢀ
Space group
P1
P1
˚
a (A)
8.2497(16)
8.7164(17)
16.655(3)
8.256(3)
9.552(3)
17.245(6)
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
80.074(4)
78.618(4)
86.783(3)
1156.2(4)
106.063(6)
91.285(9)
98.721(6)
1288.9(7)
3
˚
V (A )
Z
2
2
Calculated density (mg mꢀ3
)
1.460
0.766
520
1.279
0.670
520
Absorption coefficient (mmꢀ1
F (000)
)
Crystal size (mm)
h Range (ꢁ)
0.38 · 0.20 · 0.18
1.26–26.50
ꢀ10 6 h 6 10
ꢀ10 6 k 6 10
ꢀ20 6 l 6 14
7236
4759
3646
0.0404
0.46 · 0.22 · 0.16
1.23–26.00
ꢀ9 6 h 6 10
ꢀ11 6 k 6 10
ꢀ21 6 l 6 20
7735
5019
3758
0.0224
Index ranges
Collected reflections
Unique reflections
Observed reflections
Rint
Completeness to 2h
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R1 indices [I > 2r(I)]
R1 (all data)
99.2%
99.00%
Full-matrix least-squares on F2
4759/0/299
Full-matrix least-squares on F2
5019/0/302
1.031
1.078
R1 = 0.0392, wR2 = 0.0997
R1 = 0.0566, wR2 = 0.1080
0.411 and ꢀ0.256
R1 = 0.0357, wR2 = 0.0881
R1 = 0.0537, wR2 = 0.1056
0.291 and ꢀ0.214
ꢀ3
˚
Largest different peak and hole (e A
)

4162
N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
337.7 (e 1168 Mꢀ1 cmꢀ1); kmax/nm (CH3CN) 336.55
(e 2417 Mꢀ1 cmꢀ1).
immediately placed in the spectrometer and the spectral
change was recorded in the wavelength range of 200–
600 nm. Least-squares data analyses for the individual
runs, using first order kinetics and the limiting rate
law as reported in the literature [16] were conducted
by means of MicroMath Scientist [24].
2.4. Preparation of [CpFe(CO)(COMe)P(4-FC6H4)3]
(2b)
A solution of P(4-MeC6H4)3 (150 mg, 0.5 mmol) in
MeCN (2.0 ml) was added to a solution of 1 (100 mg,
0.5 mmol) in MeCN (2.0 ml). The mixture was heated
under reflux for ca. 72 h after which the reaction mixture
was cooled and filtered. The product was purified by col-
umn chromatography over silica eluting with hex-
ane:acetone = 9:1. The yellow band (Rf = 0.28) was
collected. Evaporation of the solvent gave the title com-
pound as an orange crystalline material. Recrystallisa-
tion from pentane afforded crystals suitable for X-ray
studies. (0.23 g, 80%) m.p. 156 ꢁC. mmax/cmꢀ1 (CHCl3)
1915 (CO); mmax/cmꢀ1 (KBr) 1898 (CO), 1597 (COCH3);
3. Results and discussion
3.1. Synthesis
Scheme 2 outlines the overall procedure for the syn-
thesis of complexes 1, 2a and 2b. Complex 1 was pre-
pared by reduction of the iron carbonyl dimer
[CpFe(CO)2]2 with excess sodium amalgam, followed
by reaction with methyl iodide [20]. The product was
isolated as a moderately oxygen sensitive yellow crystal-
line material.
d
{H} (CDCl3, 300 MHz) 2.30 (3H, s, COCH3) 2.33 (9H,
s, CH3C6H4), 4.38 (5H, s, C5H5), 7.30–7.38 (12H, m,
C6H4); dP{H} (CDCl3, 121.5 MHz) 73.68; dC{H} (CDCl3,
75.5 MHz) 21.7 (C3), 52.3 (COCH3), 88.6 (C5H5), 128.5
(4J = 9 Hz, 4-C6H4), 129.2 (3J = 10 Hz, 3,5-(C6H4),
133.6 (2J = 10 Hz, 2,6-(C6H4), 136.9 (1J = 43 Hz, 1-
(C6H4), CO and COMe were not observed; kmax/nm
(CH2Cl2) 334.7 (e 2514 Mꢀ1 cmꢀ1); kmax/nm (CH3CN)
326.4 (e 4588 Mꢀ1 cmꢀ1).
Treatment of one equivalent of the alkyl complex 1
with one equivalent of the phosphine ligand P(4-
FC6H4)3 or P(4-MeC6H4)3 in acetonitrile gives the
corresponding acyl complexes [CpFe(CO)(COMe)P(4-
FC6H4)3] (2a) and [CpFe(CO)(COMe)P(4-MeC6H4)3]
(2b), in moderate to good yields. The products were iso-
lated as air stable, orange crystalline materials which
slowly decompose in solution.
IR spectroscopy of the two complexes showed the ex-
pected mCO absorption bands at 1904 and 1898 cmꢀ1 for
2a and 2b, respectively. The acyl mCO absorptions bands
are observed at 1607 and 1597 cmꢀ1 for 2a and 2b,
respectively. The complexes 2a [13,25] and 2b
[13,25,26] were reported previously, but only mCO
absorption bands in cyclohexane were reported
(mCO = 1924 for 2a and 1920 for 2b).
Upon methyl migration, one of the CO groups is re-
placed by a phosphine ligand inducing the formation
of the acetyl group. Consequently the electron density
on the metal center increases, leading to the lowering of
the mCO absorption band due to increased Fe–CO bond
order. In 2a compared to 2b the mCO absorption band is
shifted to a lower value which is a good indication that
P(4-MeC6H4)3 in complex 2b is a better r-donor than
2.5. Kinetics studies
2.5.1. NMR and IR spectroscopies
A mixture of alkyl complex 1 (31 mM) and P(4-FC6-
H4)3 or P(4-MeC6H4)3 (120/140 mM) in MeCN was
thermostated at 58.0 ꢁC. Aliquots (0.10 ml) were taken
at timed intervals of ca. 3 h for P(4-MeC6H4)3, or ca.
10 h for P(4-FC6H4)3, and thermally quenched by stor-
age under inert atmosphere below 10 ꢁC. All samples
were analysed at the end of the kinetic run and reconsti-
tuted with CDCl3.
2.5.2. UV–vis studies
Kinetic measurements were carried out by UV–vis
spectroscopy on a Varian Cary-50 spectrometer at
50.0 ꢁC under pseudo-first order conditions, using an ex-
cess of phosphine ligand. Blank experiments on solu-
tions of alkyl complex 1 in the absence of ligand
showed no significant decomposition during the time re-
quired for the kinetic runs.
Complex 1 (0.52 mM) and a set of P(4-FC6H4)3 or
P(4-MeC6H4)3 solutions, ranging in [P(4-FC6H4)3] from
5.0150 mM (or [P(4-FC6H4)3] from 3.3–26.3 mM), in
MeCN were added to different compartments of tandem
quartz cells. The cells were stoppered and placed into a
thermostated cell holder. Time was allowed for thermal
equilibration before the cells were shaken to ensure mix-
ing of the two solutions. The resulting mixture was
∆
2 [Fe(CO)5] + C10H12
6 CO + H2 + [CpFe(CO)2]2
1) Na/Hg
THF
2) MeI
PR3
CH3CN, ∆
2 [CpFe(CO)2Me]
[CpFe(COMe)(CO)PR3]
2a/2b
1
Scheme 2. The sequence for the synthesis of 2a and 2b. PR3 = P(4-
FC6H4)3 (2a), P(4-MeC6H4)3 (2b).

N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
4163
P(4-FC6H4) in 2a. An upfield shift of the Cp protons in
the 1H NMR, from d 4.73 ppm in 1 to d 4.39 ppm in 2a
and d 4.38 ppm in 2b was also observed.
C41
C42
In the 31P NMR spectrum, complex 2a is observed at
d 75.40 ppm and 2b at 73.68 ppm, while the respective
free ligands resonate at d 8.5 ppm (P(4-FC6H4)3) and d
7.6 ppm (P(4-MeC6H4)3). Extra peaks were observed
in the 31P spectrum during kinetic runs at d 27.50 and
d 29.67 ppm and these were found to correspond to
the phosphine oxides.
Fe
O2
C2
P
C22
C1
O1
3.2. X-ray crystallography results
C3
Orange crystals of 2a and 2b were characterised by X-
ray crystallography and the numbering scheme of both
2a and 2b are presented in Figs. 1 and 2, respectively.
The bond distances and angles of interest are presented
in Table 2.
C37
In both structures, the molecules pack in the triclinic
P1ðZ ¼ 2Þ space group. The Cp ring is centrally bound
ꢀ
above the plane of the metal centre and the other ligands
are pointing down like the legs of a three-legged ‘‘bar
stool’’. The iron atom adopts a pseudo-octahedral coor-
dination geometry; the Cp ring occupying three coordi-
nation sites with the remaining three ligands oriented
perpendicular to each other. In 2a the packing is stabi-
lised by selective hydrogen interactions between the fluo-
rine and appropriate hydrogen atoms of the
neighbouring molecules (see Table 3).
Fig. 2. X-ray crystal structure of [CpFe(CO)(COMe)P(4-MeC6H4)3]
(2b) (50% probability). For the aromatic rings, the first digit refers to
the ring number while the second digit indicates the number of the
C-atom in the ring.
Table 2
˚
Selected interatomic bond distances (A) and angles (ꢁ) for [CpFe(CO)-
(COMe)P(4-FC6H4)3] (2a) and [CpFe(CO)(COMe)P(4-MeC6H4)3] (2b)
The Fe–P bond distances of 2a and 2b are, surpris-
2a
2b
˚
ingly, the same at 2.1932(8) A (see Table 2). Other sim-
˚
Distances (A)
ilarities between 2a and 2b, for example the Fe–Cp bond
Fe–C(1)
Fe–C(2)
Fe–P
C(1)–O(1)
C(2)–O(2)
C(2)–C(3)
Fe-centroid
1.748(3)
1.955(3)
2.1932(8)
1.150(3)
1.201(3)
1.521(4)
1.749(4)
1.744(3)
1.951(3)
2.1932(8)
1.149(3)
1.200(3)
1.538(4)
1.752(7)
C42
Fe
F3
C41
C2
C32
Angles (ꢁ)
O2
C(1)–Fe–C(2)
C(1)–Fe–P
C(2)–Fe–P
O(1)–C(1)–Fe
Cp–Fe–P
94.41(12)
91.69(9)
90.12(9)
174.3(3)
127.4(2)
124.7(1)
119.1(1)
94.77(12)
93.07(8)
89.03(8)
174.3(2)
125.9(1)
124.1(2)
120.7(2)
C31
P
C1
C11
O1
C3
Cp–Fe–C(1)
Cp–Fe–C(2)
C12
F1
C21
Torsion angle (ꢁ)
O(2)–C(2)–Fe-C1
C(1)–Fe–P–C(11)
C(1)–Fe–P–C(21)
C(1)–Fe–P–C(31)
C(2)–Fe–P–C(11)
C(2)–Fe–P–C(21)
C(2)–Fe–P–C(31)
156.5(3)
54.6(1)
ꢀ62.4(1)
175.2(1)
149.1(1)
32.0(1)
170.3(5)
ꢀ62.5(1)
174.5(1)
54.0(1)
C22
32.2(1)
ꢀ90.7(1)
F2
ꢀ90.4(1)
148.7(1)
Dihedral angles (ꢁ)
Fig. 1. X-ray crystal structure of [CpFe(CO)(COMe)P(4-FC6H4)3]
(2a) (50% probability). For the aromatic rings, the first digit refers to
the ring number while the second digit indicates the number of the
C-atom in the ring.
Plane 1 and 2a
8.1(2)
7.0(4)
a
Plane 1 is the plane of the Cp ring and plane 2 is described by P,
C(1) and C(2).

4164
N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
Table 3
of the triphenylphosphine analogues (Table 4 [14]),
˚
Inter- and intramolecular hydrogen bond distances (A) and angles (ꢁ)
for [CpFe(CO)(COMe)P(4-FC6H4)3] (2a) and [CpFe(CO)(COMe)P(4-
MeC6H4)3] (2b)
while the rest of the bond distances are similar. How-
ever, smaller p-acceptor properties of these ligands,
although not easily detected, should never be ignored
completely.
The acyl oxygen is orientated anti to the carbonyl
oxygen in both complexes with torsion angles of
156.5(3)ꢁ and 170.3(5)ꢁ for 2a and 2b, respectively. This
was probably due to the steric interaction between the
acyl ligand and the phenyl rings on the phosphine. A
similar observation was reported by Davies et al. [28]
for the P(C6H5)3 analogue.
D–Hꢂ ꢂ ꢂA
For 2a
C23–H23. . .O2a
C32–H32. . .O2
C33–H33. . .F2b
C42–H42. . .F1c
D–H
Hꢂ ꢂ ꢂA
Dꢂ ꢂ ꢂA
\(DHA)
0.93
0.93
0.93
0.93
2.57
2.39
2.54
2.52
3.420(4)
3.252(4)
3.383(3)
3.379(3)
153.0
154.0
150.9
153.3
For 2b
C13–H13. . .O2a
C26–H26. . .O2
0.93
0.93
2.49
2.45
3.370(4)
3.302(4)
157.0
153.1
a
1 + x,y,z.
b
3.3. Kinetic results
1 ꢀ x,1 ꢀ y,ꢀz.
c
ꢀ1 + x,1 + y,z.
The complex [CpFe(CO)2Me] (1) reacted with P(4-
FC6 H4)3 and P(4-MeC6H4)3 in acetonitrile to give the
corresponding acyl complexes [CpFe(CO)(COMe)P(4-
FC6H4)3] (2a) and [CpFe(CO)(COMe)P(4-MeC6H4)3]
(2b), as the unique metal complexes. Both the starting
materials and the products were fully characterised
and the reactions were spectroscopically quantitative
(IR, NMR).
distances (1.748(3) for 2a and 1.744(3) for 2b), are listed
in Table 2.
The most widely used method for determining ligand
steric behaviour at a metal centre is the calculation of
the Tolman cone angle [11], using an M–P bond distance
˚
˚
of 2.28 A, C–H bond distances of 0.97 A and a Van der
˚
Waals radius of 1.2 A for the H atoms. For this study
The kinetic runs were carried out under pseudo-first-
order conditions using a large excess of phosphine. The
reactions were followed at 50.0 ꢁC by monitoring the in-
crease in absorbance at ca. 360 nm in the UV–vis due to
product formation. In addition to the UV–vis kinetics,
the reactions were followed by IR spectroscopy at
58.0 ꢁC. The decrease in absorbance at 2006 and
1946 cmꢀ1 mCO for 1), as well as the increase in absor-
bance of the mCO for 2a and 2b, at 1918 and
1915 cmꢀ1, respectively, were monitored. The rates were
obtained as an average for the rate of the disappearance
of the two mCO absorption bands for 1, and the rate of
formation of the mCO absorption bands of the products.
The reaction for 2a was additionally followed using
19F NMR spectroscopy at 58 ꢁC and the spectral change
is presented in Fig. 3. The changes in the 19F NMR spec-
tra are characterised by the disappearance of the signal
actual Fe–P bond distances were used, yielding effective
cone angles hE of 152ꢁ for P(4-FC6H4)3 and 153ꢁ for
P(4-MeC6H4)3 [27]. The effective cone angles are slightly
larger than the Tolman cone angles (145ꢁ for both li-
gands [11]) due to the shorter M–P bond distance.
Since the effective angles hE of P(4-FC6H4)3 and P(4-
MeC6H4)3 are comparable, it is to be expected that their
respective steric environments are similar. This is exem-
plified in the Cp(centroid)–Fe–P angle, which remains
the same for both phosphine ligands.
In the QALE [14] hypothesis, P(C6H5)3 is classified as
a pure r-donor because it exhibits a comparatively long
Fe–P bond distance (see Table 4). The two ligands in 2a
and 2b can thus also be classified as pure r-donors, be-
cause their bond distances are similar to that of
P(C6H5)3. The Fe(CO) and Fe(COMe) bond distances
for complexes 2a and 2b are longer compared to that
Table 4
Structural correlations for the complexes of the type [CpFe(CO)(COMe)PR3] [14]
2a
2b
[CpFe(CO)(COMe)P(C6H5)3]
[Cp0Fe(CO)(COMe)P(C6H5)3]
Fe–P
2.1932(8)
1.748(3)
1.955(3)
1.748(4)
2.118(6)
2.1932(8)
1.744(3)
1.951(3)
1.751(7)
2.113(6)
2.202(2)
1.708(2)
1.917(8)
1.752(5)
2.118(11)
2.195(3)
1.694(10)
1.921(9)
1.746(8)
2.114(10)
Fe–C(O)
Fe–C(@O)
Fe–Cpa
Fe–Cpb
hE
152
d
153
d
159
e
159
e
Reference
Cp0 = MeC5H4.
a
The centroid of the Cp ring.
The average Fe–C(Cp) distances.
This work.
b
d
e
Reference [14].

N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
4165
Fig. 3. The 19F NMR spectral change during the reaction of [CpFe(CO)2Me] (1), (31 mM) with P(4-FC6H4)3, A, (120 mM) yielding
[CpFe(CO)(COMe)P(4-FC6H4)3] (2a) (B) in MeCN at 58.0 ꢁC.
for the non-coordinated phosphine (A) at d 114 ppm
and the formation of the acetyl complex 2a at d
112 ppm. The formation of the acetyl complex is accom-
panied by the formation of phosphine oxide, observed at
d 108 ppm (C). A kinetic plot of the integration of the
various peaks in the 19F NMR vs. time is presented in
Fig. 4. In agreement with these results, the 31P NMR
spectral change showed the formation of the acetyl com-
plex 2a at d 75.4 ppm with the uncoordinated phosphine
ligand resonating at d 8.5 ppm, while the corresponding
phosphine oxide was observed at d 27.5 ppm.
The reactions were first-order in the methyl complex
[CpFe(CO)2Me] (1), while the dependence of the rate
on phosphine concentration follows a non-linear behav-
iour. With increasing concentration of the phosphine,
the reactivity tends toward a limiting rate (saturation
kinetics) in agreement with Scheme 1. The rate constants
obtained in acetonitrile were k2 = (3.4 0.4) · 10ꢀ5 sꢀ1
for P(4-MeC6H4)3 and k2 = (2.2 0.2) · 10ꢀ6 sꢀ1 for
P(4-FC6H4)3. The same non-linear dependence on phos-
phine concentration was reported by Bassetti and
coworkers for the migratory carbonyl insertion of [g5-
C9H7Fe(CO)2Me] [16].
The general trend we found, that the P(4-MeC6H4)3
ligand is faster than the P(4-FC6H4)3 ligand is in broad
agreement with literature, though it is difficult to directly
compare our study with the literature results.
Upon changing the solvent from acetonitrile to
dichloromethane for P(4-MeC6H4)3, similar saturation
kinetics was observed at 30.0 ꢁC, with a rate of
k2 = (2.2 0.5) · 10ꢀ6 sꢀ1. If ÔnormalÕ activation energy
holds, a constant for acetonitrile at 30 ꢁC of ca. 7 · 10ꢀ6
sꢀ1 is calculated for the P(4-MeC6H4)3 complex. This is
ca. three times more rapid than in dichloromethane.
Fig. 4. Increase and decrease of the 19F NMR peak intensities for the reaction of [CpFe(CO)2Me] (1) (31 mM) with [P(4-FC6H4)3] (120 mM) yielding
[CpFe(CO)(COMe)P(4-FC6H4)3] (2a) in MeCN at 58 ꢁC (m = P(4-FC6H4)3, d = [CpFe(CO)(COMe)P(4-FC6H4)3], j = OP(4-FC6H4)3).

4166
N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
Thus we conclude that increased solvent donor ability/
polarity enhances the reaction rate. We can therefore
also not exclude nucleophilic assistance by the solvent,
or the formation of solvent-coordinated intermediates
for the overall process.
Bassetti and coworkers [16] studied the migratory car-
bonyl insertion of [g5-C9H7Fe(CO)2Me] using a series of
para-substituted phosphines P(4-XC6H4)3 (X = H, Cl, F,
OMe). However, they found no significant change of rate
with increasing donor-capacity of the ligand, the differ-
ences being within experimental errors for the system.
Besides the relatively small differences, they performed
their studies in toluene to minimise any solvent assisted
mechanisms. In addition they used the indenyl spectator
ligand, instead of cyclopentadienyl, which has the ability
of g3-coordination in intermediates, although they did
not find any evidence of a haptotropic shift.
1, and insertion product 2, the same amount of oxidation
is observed within 7 h. In five days, half of the amount of
the phosphine ligand used is oxidised (see Figs. 3 and 4
peaks A and C). This lead us to believe that either com-
plex 1 or its insertion product 2 catalyses the oxidation
of phosphine ligand in the present system. The formed
phosphine oxides appears to have no influence on the
CO migratory insertion reaction (see Fig. 4). This obser-
vation is in agreement with literature [16a].
In conclusion, the system reported here is compli-
cated by a number of factors. First of all, the oxidation
of phosphine ligand decreases the effective ligand con-
centration making kinetic results difficult to interpret.
In addition, proper comparison with literature reports
is not easy due to the small body of work performed
on the system. Where reports in literature exist, they
either report on systems with different spectator ligands
or phosphines that vary in both steric and electronic
properties. Finally, the rates of the reactions (in some
cases >10 days) make accurate kinetic measurements
difficult. To utilise the current system as a convenient
probe to evaluate electronic characteristics of Group
15 ligands, therefore requires significant modification
of reaction conditions, which is to be explored in future.
Nevertheless, the different reactants and products in
Scheme 1 were accurately identified and characterised
in this study, and the preliminary kinetics determined
indicates that this system can in principle be used as a
kinetic probe for evaluation of electronic and steric ef-
fects of Group 15 ligands.
The reaction of [CpFe(CO)2Me] with a series of PPh3,
PPh2 Me, PPhMe2 was reported by Bassetti and cowork-
ers [16c]. In this study, an increase in reaction rate was
observed with ligands of increasing basicity. However,
these phosphines are not iso-steric and it is difficult to
quantify the steric influence on the migratory CO inser-
tion reaction. The same study reports on the migratory
i
CO insertion in [CpFe(CO)2 Pr] complexes, but the pos-
sibility of a different mechanism cannot be excluded. In
general, the reaction rates of the migratory CO insertion
i
i
increase when going from Me to Pr. The Pr group is,
however, sterically significantly larger than the Me
group, and it also opens up mechanistic pathways involv-
ing agostic interactions with the b-hydrogens. It is
known from literature that a significant increase in rate
is seen on substituting Me groups for higher n-alkanes,
with the highest rate for an Et group [6].
Acknowledgements
One observation made during our studies is that all
excess phosphine ligand is oxidised as the reaction pro-
ceeds. Part of the products, 2a and 2b, also decomposes
in solution (see Fig. 4B). This decomposition product
was not analysed further. In an earlier study conducted
by Su and Wojcicki [29], it was shown that the thermal
reaction between [CpFe(CO)2R] (where R = Me or Et)
and P(C6H5)3 in hexane (69 ꢁC) and heptane (98 ꢁC)
produced the acetyl complex [CpFe(CO)(COR)P-
(C6H5)3]. Over an extended period of time this complex
undergoes decarbonylation to form [CpFe(CO)(R)P(C6-
H5)3]. In our studies we did not observe the decarbony-
lation product, [CpFe(CO)(Me)P(4-FC6H4)3] (see
Fig. 3), however we did find significant amounts of the
phosphine oxides.
Financial support from the Research Funds of the
University of the Free State and the University of
Johannesburg are gratefully acknowledged. Part of this
material is based on work supported by the South Afri-
can National Research Foundation [SA NRF, GUN
2053397 (University of Johannesburg) and 2068915
(Univeristy of the Free State)]. Any opinon, findings,
and conclusions or recommendations expressed in this
material are those of the authors and do not necessarily
reflect the views of the SA NRF. We thank Prof. D.
Levendis and Dr. D. Billing (University of the Witwa-
tersrand) for the use of their diffractometer.
References
In order to ascertain the origin of the phosphine oxide,
a blank reaction was performed which contained only the
ligand and the solvent under the same conditions as that
of the reaction. NMR spectra of the ligand showed a neg-
ligible degree of oxidation. Upon exposure to air it was
found that only ca. 7% of the phosphine ligand oxidised
after five days of heating. In the presence of alkyl complex
[1] L.S. Hegedus, Transition Metal in the Synthesis of Complex
Organic Molecules, University books, USA, 1994, p. 29.
[2] B. Cornils, W.A. Herrmann, Applied Homogeneous Catalysis
with Organometallic Compounds, vol. 2, Wiley-VCH, Weinheim,
1991, p. 521 (Chapter 3).
[3] (a) M. Green, D.J. Westlake, J. Chem. Soc. A (1971) 367;
(b) J.P. Bibler, A. Wojcicki, Inorg. Chem. 5 (1966) 889.

N.M. Makunya et al. / Journal of Organometallic Chemistry 690 (2005) 4159–4167
4167
[4] J.D. Cotton, M.M. Kroes, R.D. Markwell, E.A. Miles, J.
Organomet. Chem. 388 (1990) 133.
[5] M.J. Wax, R.G. Bergman, J. Am. Chem. Soc. 103 (1981) 7028.
[6] R. George, J.M. Anderson, J.R. Moss, J. Organomet. Chem. 505
(1995) 131.
[19] D.D. Perrin, W.L.F. Armarego, Purification of Laboratory
Chemicals, third ed., Butterworth-Heinemann Ltd, Oxford, 1988.
[20] R.J. Angelici, Synthesis and Techniques in Inorganic Chemistry,
second ed., University Science Books, Mill Valley, CA, 1986, p.
147.
[7] P.W.N.M. van Leeuwen, Homogeneous Catalysis: Understanding
the Art, Kluwer Academic Publishers, Netherlands, 2004, p. 10.
[8] (a) For an example of phosphite stabilised cobalt catalysts see: M.
Haumann, R. Meijboom, J.R. Moss, A. Roodt, Dalton Trans.
(2004) 1679;
(b) R. Meijboom, M. Haumann, A. Roodt, L. Damoense, Helv.
Chim. Acta 88 (2005) 676.
[9] A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M. Haumann,
P. Wasserscheid, Angew. Chem., Int. Ed. 44 (2005) 815.
[10] R.H. Crabtree, Organometallic Chemistry of the Transition
Metals, Wiley, New York, USA, 1988, p. 71.
[11] C.A. Tolman, Chem. Rev. 77 (1977) 313.
[12] B. Douglas, D. McDaniel, J. Alexander, Concepts and Models of
Inorganic Chemistry, third ed., Wiley, New York, USA, 1994, p.
570.
[13] M.M. Rahman, H.Y. Liu, K. Eriks, A. Prock, W.P. Giering,
Organometallics 8 (1989) 1.
[21] Bruker SAINT+. Version 6.02 (including XRPEP). Bruker AXS
Inc., Madison, Wisconsin, USA, 1999.
[22] G.M. Sheldrick, SHELXL97, University of Go¨ttingen, Germany,
1997.
[23] K. Brandenburg, M. Berndt, Diamond version 2.1e., Crystal
Impact GbR, Bonn, Germany, 2001.
[24] SCIENTIST for Windows, ver. 4.10.1998, MicroMath, Utah,
USA, 1990.
[25] (a) A.L. Fernandez, C. Reyes, A. Prock, W.P. Giering, Perkin 2
(2000) 1033;
(b) A.L. Fernandez, T.Y. Lee, C. Reyes, A. Prock, W.P. Giering,
Organometallics 17 (1998) 3169;
(c) A. Fernandez, C. Reyes, A. Prock, W.P. Giering, Organomet-
allics 17 (1998) 2503;
(d) D.C. Woska, M. Wilson, J. Bartholomew, K. Eriks, A. Prock,
W.P. Giering, Organometallics 11 (1992) 3343.
[26] (a) R.T. Aplin, J. Booth, R.G. Compton, S.G. Davies, S. Jones,
J.P. McNally, M.R. Metzler, W.C. Watkins, J. Chem. Soc.,
Perkin Trans. 2 (1999) 913;
[14] H.Y. Liu, L.L. Koh, K. Eriks, W.P. Giering, A. Prock, Acta
Cryst. C 46 (1990) 51.
[15] I.S. Buttler, F. Basolo, R.G. Pearson, Inorg. Chem. 6 (1967) 2074.
[16] (a) M. Bassetti, L. Mannina, D. Monti, Organometallics 13
(1994) 3293;
(b) S.G. Davies, S. Jones, M.R. Metzler, K. Yanada, R. Yanada,
J. Chem. Soc., Perkin Trans. 2 (1998) 1147;
(c) S.C. Case-Green, J.F. Costello, S.G. Davies, N. Heaton,
C.J.R. Hedgecock, V.M. Humphreys, M.R. Metzler, J.C. Prime,
J. Chem. Soc., Perkin Trans. 1 (1994) 933;
(b) M. Bassetti, D. Monti, J. Am. Chem. Soc. 115 (1993) 4658;
(c) M. Allevi, M. Bassetti, C. Lo Sterzo, D. Monti, J. Chem. Soc.,
Dalton Trans. (1996) 3527.
(d) S.G. Davies, M.R. Metzler, W.C. Watkins, R.G. Compton, J.
Booth, J.C. Elkund, J. Chem. Soc., Perkin Trans. 2 (1993) 1603.
[27] S. Otto, A. Roodt, J. Smith, Inorg. Chim. Acta 303 (2000)
295.
[28] S.G. Davies, J.I. Seeman, I.H. Williams, Tetrahedron Lett. 27
(1986) 619.
[17] (a) See for example: S. Otto, A. Roodt, Inorg. Chim. Acta 357
(2004) 1;
(b) A. Roodt, S. Otto, G. Steyl, Coord. Chem. Rev. 245 (2003)
125.
[18] D.F. Shriver, M.A. Drezdzon, The Manipulation of Air-Sensitive
Compounds, Wiley-Interscience, New York, 1986.
[29] S.R. Su, A. Wojcicki, J. Organomet. Chem. 27 (1971) 231.