Reductive elimination of halogens assisted by phosphine ligands in Fe(CO)4X2 (X = I, Br) complexes

Article abstract of DOI:10.1016/j.jorganchem.2006.05.041

Fe(CO)₄X₂ complexes [X = I (1), Br(1′)] react with phosphine ligands L (L = PMe₃, PEt₃, PMe₂Ph, PMePh₂, PPh₃) via a two-step mechanism: in the first step fac-Fe(CO)₃LX₂ complexes are formed; in the second step two parallel pathways, a and b, are observed; in pathway a, reductive elimination with formation of equimolar amounts of Fe(CO)₃L₂ (5) and phosphonium salts [LX]⁺X^- is observed; in pathway b, disubstituted dihalide complexes cis,trans,cis-Fe(CO)₂L₂X₂ are formed. The relative weights of pathways a and b depend on the basicity, steric hindrance and concentration of ligand L, on the nature of the halogen and on temperature. A radical mechanism which accounts for most of the experimental results is proposed.

Full text of DOI:10.1016/j.jorganchem.2006.05.041

Journal of Organometallic Chemistry 691 (2006) 3881–3888
www.elsevier.com/locate/jorganchem
Reductive elimination of halogens assisted by phosphine ligands
in Fe(CO)4X2 (X = I,Br) complexes
*
Gianfranco Bellachioma, Giuseppe Cardaci , Alceo Macchioni,
Chiara Venturi, Cristiano Zuccaccia
`
Dipartimento di Chimica, Universita di Perugia, via Elce di Sotto, 8, 06123 Perugia, Italy
Received 6 March 2006; received in revised form 17 May 2006; accepted 19 May 2006
Available online 3 June 2006
Abstract
Fe(CO)4X2 complexes [X = I (1), Br(10)] react with phosphine ligands L (L = PMe3, PEt3, PMe2Ph, PMePh2, PPh3) via a two-step
mechanism: in the first step fac-Fe(CO)3LX2 complexes are formed; in the second step two parallel pathways, a and b, are observed;
in pathway a, reductive elimination with formation of equimolar amounts of Fe(CO)3L2 (5) and phosphonium salts [LX]+Xꢀ is
observed; in pathway b, disubstituted dihalide complexes cis,trans,cis-Fe(CO)2L2X2 are formed. The relative weights of pathways a
and b depend on the basicity, steric hindrance and concentration of ligand L, on the nature of the halogen and on temperature. A radical
mechanism which accounts for most of the experimental results is proposed.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Reductive elimination; Carbonyl halide iron complexes; Phosphonium salts; Radical mechanism
1. Introduction
During our previous studies [5] on the reaction of
Fe(CO)4X2 with phosphine ligands to obtain cis,trans,cis-
FeCO)2L2X2 [L = PPh3, P(iso-Propyl)3; X = I, Br] the for-
mation of small quantities of Fe(CO)3L2 was observed. The
substitution reaction of Fe(CO)4X2 with phosphine ligands
was studied by Basolo a long time ago [6] and in this case
The reductive elimination of halogens in dihalide com-
plexes of transition metals, shown in Scheme 1 is thermo-
dynamically forbidden [1] since in this reaction two fairly
strong metal–halogen bonds are broken and one weak hal-
ogen–halogen bond is formed. The metal–ligand bond
energy, that can assist the reaction, is not sufficient to make
the reaction thermodynamically allowed [2]. In order for
the halide elimination to occur, it must be assisted by other
reactions that can counterbalance its endothermicity. A few
examples of reductive elimination of halogens are described
in the literature: in particular, the reaction of Fe(CO)4I2
with PPh3 to give Fe(CO)3(PPh3)2 in the presence of tin(II)
halides [3] and the reaction of polyhalide complexes of
tungsten W(CO)L2I3(@CNEt2) with PMe3 to give
[W(CO)L2(PMe3)2(@CNEt2)]+Iꢀ and [PMe3I]+Iꢀ [4].
too, the formation of
a very small quantity of
Fe(CO)3(PPh3)2 was observed.
In this paper, we report the results of our investigation
on the reaction of Fe(CO)4X2 (X = I, Br) with phosphine
ligands more basic than those studied by Basolo to high-
light the reductive elimination of halogens in Fe(CO)4X2
and to shed light on its mechanism.
2. Experimental
Unless otherwise stated, all operations were carried out
under nitrogen using standard Schlenk techniques. Com-
plexes Fe(CO)4X2 (X = I (1), Br (10)) were prepared as
described in [7]. Fe(CO)4PMe3 was prepared as described
in [8]. Phosphine ligands (PMe3, PMe2Ph, PMePh2, PPh3,
PEt3) were commercial products and were used without
*
Corresponding author. Tel.: +39 075 5855578; fax: +39 075 5855598.
E-mail address: cardchim@unipg.it (G. Cardaci).
0022-328X/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2006.05.041

3882
G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
X
by a lower case letter (a, PMe3; b, PEt3; c, PMe2Ph; d,
X
X
+
L
M
PMePh2; e, PPh3) and the halogens I and Br by an apex.
The CO stretching frequencies are given in Table 1. The
structure of the complexes was determined on the basis
of the number and intensity of the CO stretching bands.
+
L
M
X
Scheme 1.
2.1. Preparation of [PMe3I]+Iꢀ
further purification. Toluene was dehydrated with Na/K
alloy and distilled under nitrogen; diethylether (DE) was
purified by refluxing with NaOH pellets, distilled, refluxed
with Na and benzophenone and then freshly distilled under
nitrogen before use; acetonitrile (ACN) was purified as
described in [9]. 1H and 31P{H} NMR spectra were
recorded on a Bruker DRX 400 spectrometer; referencing
was relative to TMS (1H) and 85% H3PO4 (31P). Infrared
spectra were recorded on a Perkin–Elmer 1725 X spectro-
photometer. Elemental analyses were performed on a
Carlo Erba 1106 elemental microanalyzer.
Three grams of I2 were stirred in 100 mL of toluene. An
excess of PMe3 (3 g) was added and the solution was left to
react for 5 h. A yellow solid precipitated. The solid was
washed and dried. Yield 72%. Anal. Calc. for C3H9I2P:
C, 10.92; H, 2.75. Found: C, 10.85; H, 2.71%.
[PMe3I]+Iꢀ is easily hydrolyzed by H2O, giving
[PHMe3]+Iꢀ, which was crystallized as white crystals. H
1
2
3
NMR (CD3NO2): d, 1.99 (dd, JH–P = 15.7 Hz; JH–H
1
3
= 5.5 Hz, PMe3); 6.4 (dm, JH–P = 496.4 Hz; JH–H
= 5.6 Hz, H); 31P {1H} NMR (CD3NO2): d, ꢀ0.44 (s,
PMe3).
The structure and numbering of the complexes are given
in Chart 1 in which the different ligands are distinguished
L
L
L
CO
Fe
OC
OC
X
OC
OC
X
X
X
X
X
X
OC
OC
OC
X
Fe
L
Fe
L
Fe
CO
CO
CO
X= I (1)
X= Br (1')
X=I; L= PMe3 (2a)
L=PEt3 (2b)
X=I; L= PMe3 (3a)
L=PEt3 (3b)
X=I; L= PMe3 (3a)
L=PEt3 (3b)
L=PMe2Ph (2c)
L=PMePh2 (2d)
L=PPh3 (2e)
L=PMe2Ph (3c)
L=PMePh2 (3d)
L=PPh3 (3e)
X=Br; L=PMe3 (2a')
X=Br; L=PMe3 (3a')
L
L
L
I
L
I
I
L
OC
OC
Fe
CO
Fe
Fe
L
CO
I
OC
L
L
(6a)
(7a)
L= PMe3 (5a)
L=PEt3 (5b)
L=PMe2Ph (5c)
L=PMePh2 (5d)
L=PPh3 (5e)
Chart 1.
Table 1
CO stretching frequencies (cmꢀ1) of the complexes fac-Fe(CO)3LX2, cis,trans,cis-Fe(CO)2L2X2, all trans-Fe(CO)2L2X2 and Fe(CO)3L2 in toluene
L
X
fac-Fe(CO)3LX2
cis,trans,cis-Fe(CO)2L2X2
All trans-Fe(CO)2L2X2
Fe(CO)3L2
PMe3
PMe3
PMe2Ph
PMePh2
PPh3
Br
I
I
I
I
2096, 2055, 2022
2089, 2043, 2015
2089, 2045, 2022
2088, 2043, 2022
2087, 2041, 2026
2084, 2037, 2014
2022, 1970
2024, 1964
2023, 1971
2022, 1973
1868
1868
1874
1876
1974
PEt3
PEt3
I
Br
2007, 1956
2016,1983
1966
1865
1866

G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
3883
2.2. Preparation of fac-Fe(CO)3PMe3I2 (2a)
2.6. Preparation of cis,trans,cis-Fe(CO)2(PMe3)2Br2 (3a0)
Three grams of Fe(CO)4PMe3 [8] were dissolved in
100 mL of DE; 3.12 g of iodine were slowly added to this
stirred solution cooled to ꢀ15 ꢁC. The reaction was com-
plete in 15 min; the solution was filtered, concentrated up
to 20 mL and crystallized at ꢀ20 ꢁC. Complex 2a was
obtained as orange crystals. Yield 90%. Anal. Calc. for
C6H9FeI2O3P: C, 15.33; H, 1.93. Found: C, 15.45; H,
1.98%. mCO (CH2Cl2, cmꢀ1): 2091, 2046, 2022.
Five grams of Fe(CO)4Br2 were dissolved in 100 mL of
DE; 3.5 g of PMe3 (molar ratio 3/1) were added under stir-
ring at ꢀ15 ꢁC. The reaction was complete in 1 h. The solu-
tion was filtered and dried. Complex 3a0 was crystallized
from CH2Cl2–n-hexane at ꢀ15 ꢁC. Yield: 74%. Anal. Calc.
for C8H9FeBr2O2P2: C, 23.17; H, 2.19. Found: C, 23.35; H,
2.25%. mCO (toluene, cmꢀ1): 2022, 1970.
2.7. Photochemical preparation of all
trans-Fe(CO)2(PMe3)2I2 (4a)
2.3. Reaction between complex 2a and PMe3
(a) 0.15 g of complex 2a were dissolved in 10 mL of tol-
uene and introduced into a reactor thermostatted at
ꢀ20 ꢁC; 0.145 g of PMe3 were added (molar ratio 6/1).
The course of the reaction was followed by IR. The forma-
tion of complex 3a was observed but no Fe(CO)3(PMe3)2
was formed. After 18 h only 10% of complex 2a had
reacted. The solution was then heated up to 25 ꢁC and
the reaction was complete in 12 h.
(b) 0.15 g of complex 2a and 0.027 g of complex
[Fe(CO)4I2] (1) were dissolved in 10 mL of toluene, thermo-
statted at ꢀ20 ꢁC; 0.145 g of PMe3 were added. The reac-
tion was complete in 30 min with formation of 3a, 5a,
and [PMe3I]+Iꢀ.
Complex 3a (1.27 g) was dissolved in 50 mL of toluene.
The solution was introduced into a photochemical reactor
and irradiated with a medium pressure Hanovia lamp
through a filter to block the light with k < 350 nm. The
isomerization was complete in 10 min. The solution was
dried and complex 4a was crystallized from a CH2Cl2–n-
hexane solution as red crystals. Yield: 40%. mCO(toluene,
cmꢀ1): 1974. In the presence of an excess of PMe3 (molar
ratio 6/1) no photochemical transformation of complex
3a to complex 4a was observed, but a slow reaction
(ꢁ12 h) occurred that gave complex mer,trans-Fe(CO)L3I2
(7a) with a CO stretching band at 1956 cmꢀ1. This last
complex was characterized spectroscopically by NMR:
j2 + 4j
1H NMR (CD2Cl2): d, 1.95 (t,
J
= 7.2 Hz, PMe3
H–P
2
2.4. Preparation of cis,trans,cis-Fe(CO)2(PMe3)2I2 (3a)
trans), 1.85 (d, JH–P = 7.2 Hz, PMe3 cis). 31P {1H}
NMR: d = 42.4 (t, JP–P = 55.7 Hz, PMe3 trans), 0.4 (d,
2
Five grams of Fe(CO)4I2 were dissolved in 100 mL of
DE; 2.70 g of PMe3 (molar ratio 3/1) were added under
magnetic stirring at ꢀ15 ꢁC. The reaction was complete
in 30 min; the solution was filtered and dried. Complex
3a was crystallized from CH2Cl2–n-hexane solution at
ꢀ15 ꢁC as red–brown crystals. Yield: 75%. Anal. Calc.
for C8H9FeI2O2P2: C, 18.89; H, 1.78. Found: C, 18.95;
H, 1.81%. mCO (DE, jc2m+ꢀ41j): 2021, 1971. 1H NMR (CD2Cl2):
2JP–P = 55.2 Hz, PMe3 cis).
2.8. Thermal isomerization of complex 4a
One gram of complex 4a was dissolved in 50 mL of tol-
uene and 0.9 g of PMe3 were added at room temperature.
Complex 4a isomerized to complex 3a in 10 min.
d = 1.89 (tHarris
,
JH–P = 8.4 Hz, PMe3); 31P{H}
2.9. Thermal isomerization of complex 3b
NMR(CD2Cl2): d = 7.39 (s, PMe3).
In the presence of an excess of PMe3, complex 3a reacts
slowly at room temperature in toluene forming mer,cis-
0.5 g of cis,trans,cis-Fe(CO)2(PEt3)2I2 (3b) were dissolved
in 25 mL of toluene at 25 ꢁC. The solution remained
unchanged for 3 h; then 0.6 g of PEt3 (molar ratio 6/1) were
added. Under these conditions, complex 3b isomerized
slowly into complex 4b. The reaction was complete in 20 h.
Fe(CO)L3I2 (6a) (mCO = 1927 cmꢀ1) in 24 h. 1H NMR
j2 + 4j
(CD2Cl2): d = 1.96 (t,
J
= 8.8 Hz, PMe3), 1.90 (d,
HP
2JHP = 9.2 Hz, PMe3). 31P{1H} NMR (CD2Cl2): A2B sys-
tem: d = ꢀ0.97, ꢀ1.33, ꢀ3.28, ꢀ3.62, ꢀ3.67, ꢀ4.01;
Complex 4b shows a CO stretching band at 1966 cmꢀ1
.
2
mA = ꢀ1.15 ppm; mB = ꢀ3.64 ppm, JAB = 58.3 Hz.
2.10. Preparation of Fe(CO)3(PMe3)2 (5a)
2.5. Preparation of cis,trans,cis-Fe(CO)2(PEt3)2I2 (3b)
One gram of complex 3a was dissolved in 50 mL of
ACN. The solution was thermostatted at ꢀ18 ꢁC and stir-
red with an excess of sodium amalgam (1.2%) (molar ratio
6/1) under bubbling of carbon monoxide. The colour of the
solution initially changed from red to green and then to yel-
low. After 1 h the reaction was complete. The solution was
filtered and dried. The solid residue crystallized from an
ethyl alcohol–n-hexane mixture as yellow crystals and com-
pared with a specimen prepared by Fe2(CO)9 and PMe3 [8].
One gram of Fe(CO)4I2 was dissolved in 50 mL of tolu-
ene at room temperature; 0.7 g of PEt3 (molar ratio 2.5/1)
were added under stirring. The reaction was complete in
10 min. The solution was dried and complex 3b was crystal-
lized from CH2Cl2–n-hexane at ꢀ15 ꢁC as red brown crys-
tals. Yield 80%. Anal. Calc. for C14H30FeI2O2P2: C, 27.93;
H, 5.02. Found: C, 27.75; H, 4.95%. mCO (toluene, cmꢀ1):
2016, 1983.

3884
G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
Table 4
2.11. Photochemical reaction of complex 2a
Effect of the ligand basicity on the A and B values at 25 ꢁC with
[L] = 2.50 · 10ꢀ2 M for iodide complexes
(a) 1.2 g of complex 2a were dissolved in 50 mL of tolu-
ene and irradiated at room temperature with a medium
pressure Hanovia lamp, equipped with a filter to block
the ultraviolet radiation. A fast decomposition was
observed with formation of complex 4a (mCO = 1974 cmꢀ1).
The formation of an intermediate, showing a CO stretching
band at 2029 cmꢀ1, was also observed.
(b) 0.12 g of complex 2a were dissolved in 50 mL of tol-
uene. An excess of PMe3 was added (molar ratio 8/1). The
solution was irradiated as described in point a. A fast
quantitative reaction of complex 2a and its transformation
into complex 3a were observed.
a
L
pKa
Ab
Bc
PMe3
PMe2Ph
PMePh2
PPh3
Pet3
8.65
6.50
4.57
2.73
8.69
0.05
0.02
0
0
0.22
1
1.45
0.11
0
1
a
Values from [10].
A = [Fe(CO)3L2]/[Fe(CO)2L2I2].
B = Ratio between the absorbance of the low frequency CO stretching
b
c
of complex 3 and the absorbance of the intermediate frequency CO
stretching of complex 2.
Table 5
2.12. Effect of the basicity and concentration of phosphine
ligands on the reaction products
Effect of the halogen on the Aa values at various temperatures
([PMe3] = 11.70 · 10ꢀ2 M) in toluene
A (PMe3)
The reaction between Fe(CO)4X2 (X = I (1), Br (10)) and
L (PMe3, PEt3, PMe2Ph, PMePh2, PPh3) were followed at
various temperatures in the range ꢀ40 to +25 ꢁC and at
different concentrations of ligands in the range 1.5 · 10ꢀ1
to 10ꢀ2 M. In a typical experiment, 0.025 g of Fe(CO)4I2
were dissolved in 5 mL of toluene. The solution was
injected in a thermostatted and blackened reactor. PMe3
was then added and aliquots of the solution were with-
drawn at various times and analyzed by IR in the range
2100–1800 cmꢀ1 in order to determine their composition.
The concentrations of complexes 5 and 3 and their relative
ratios A (Tables 2–5) were obtained by the experimental
values of the extinction coefficients. The relative ratios B
of complexes 3 and 2 were measured by the ratio between
T (ꢁC)
ꢀ40
ꢀ20
0
I
Br
0.74
0.80
0.40
0.08
0.08
0.08
a
A = [Fe(CO)3L2]/[Fe(CO)2L2I2].
the absorbance of the low frequency CO stretching of com-
plex 3 and the absorbance of the intermediate frequency
CO stretching of complex 2, assuming the extinction coef-
ficients were similar for the different L.
The white precipitated was filtered, washed with toluene
and characterized as [PMe3I]+Iꢀ.
3. Results
Table 2
The substitution reaction of Fe(CO)4X2 [X = I (1); Br
(10)] with phosphine ligands L, having different basicity
and steric hindrance (L = PMe3, PEt3, PMe2Ph, PMePh2,
PPh3) [10], was studied in the temperature range ꢀ40 ꢁC
to +25 ꢁC. The reactions were complete in less of an hour
at ꢀ40 ꢁC and proceeded in two steps as shown in Scheme
2: the first step yielded the monosubstituted fac-
Fe(CO)3LX2 (2) complex; the second step proceeded via
two parallel pathways a and b; pathway a gave equimolar
amounts of Fe(CO)3L2 (5) and phosphonium salts
[LX]+Xꢀ by reductive elimination; pathway b gave disub-
stituted dihalide cis,trans,cis-Fe(CO)2L2X2 (3) complexes,
by carbon monoxide substitution.
Effect of the ligand concentration on the products of the reaction between
Fe(CO)4I2 and PMe3 at ꢀ20 ꢁC in toluene; [Fe(CO)4I2] ꢁ (1.35 0.15)
· 10ꢀ2
M
102 Æ [PMe3] (M)
103 Æ [Fe(CO)2
(PMe3)2I2]
(3a) (M)
103 Æ [Fe(CO)3
(PMe3)2]
Aa = [5a]/[3a]
(5a) (M)
2.50
5.04
8.59
11.70
14.80
a
7.70
7.70
8.40
8.12
8.05
3.37
4.50
5.92
6.46
6.73
0.44
0.63
0.70
0.80
0.83
A = [Fe(CO)3L2]/[Fe(CO)2L2I2].
A quantitative study was carried out for L = PMe3 and
Fe(CO)4I2. The characterization of the phosphonium salt
[PMe3I]+Iꢀ was achieved by elemental analysis and by
comparison with a specimen prepared by PMe3 and I2
[11]. The NMR characterization of [PMe3I]+Iꢀ in solution
was not possible due to its insolubility in the common
organic solvents. A 31P NMR spectrum is reported in the
literature for the solid state (ꢀ5.3 ppm) and for CDCl3
solution (80 ppm) [12]. We were not able to solubilize
[PMe3I]+Iꢀ in either CDCl3 or in other organic solvents;
when the solubilization occurs the species in solution are
Table 3
Effect of temperature and ligand concentration on the Aa values for the
reaction of Fe(CO)4I2 with PMe3 and PEt3
A(PMe3)
A(PEt3)
102 · [L] (M)
ꢀ40 ꢁC
ꢀ20 ꢁC
0 ꢁC
25 ꢁC
25 ꢁC
2.50
5.04
8.59
11.70
14.80
a
0.26
0.56
0.63
0.74
0.73
0.44
0.63
0.70
0.80
0.83
0.11
0.14
0.31
0.40
0.38
0.05
0.22
0.27
0.36
0.40
0.40
A = [Fe(CO)3L2]/[Fe(CO)2L2I2].

G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
3885
L
OC
[LX]+X-
CO
+
Fe
+2 L
OC
L
a
CO
Fe
L
OC
OC
X
X
(5)
OC
OC
+ L
Fe
L
X
X
b
CO
(2)
CO
(1)
X
X
+ L
OC
OC
Fe
L
(3)
Scheme 2.
the products of the reaction of [PMe3I]+Iꢀ with the solvent
(for example CH3OD) or with H2O contained in the sol-
vent. In fact [PMe3I]+Iꢀ is very sensitive to H2O with for-
mation of [PHMe3]+Iꢀ, which was isolated and completely
characterized (see Section 2). Owing to the high reactivity
of [PMe3I]+Iꢀ with H2O, the reaction between Fe(CO)4I2
and PMe3 was carried out in toluene dehydrated by Na/
K alloy.
The concentrations of complexes 2, 3 and 5 of Scheme 2
depend on the concentration, basicity and steric hindrance
of L, on temperature and on the halogen. The effect of
PMe3 concentrations at ꢀ20 ꢁC in toluene is given in Table
2, in which the ratios A between the concentrations of com-
plexes 5 and 3 are reported; the A values represent the rel-
ative weight of the reductive elimination with respect to the
substitution reaction. The A values increase with increasing
the concentration of PMe3 as shown in Fig. 1.
for the temperature of around ꢀ20ꢁC. The values of A for
PEt3 are significantly greater than for PMe3, indicating a
strong effect of the steric hindrance of PEt3, since the basi-
city of PMe3 and PEt3 is similar [10].
The effect of the basicity of the phosphine ligands on the
reductive elimination is evidenced by the increase of the A
values on increasing the pKa of the ligands (Table 4). How-
ever the basicity of the ligands also influences the forma-
tion of fac-Fe(CO)3LI2. On decreasing the basicity of the
ligands, the concentration of the monosubstituted com-
plexes increases until with PPh3 the reaction stops at the
first step, as indicated by the B values of Table 4, which
correspond to the ratio between the absorbances of cis,-
trans,cis-Fe(CO)2L2I2 and fac-Fe(CO)3LI2. The formation
of fac-Fe(CO)3LI2 is kinetically controlled: in fact, the
monosubstituted derivatives react very slowly with the
phosphine ligands, forming the disubstituted dihalide
derivatives 3, by a CO dissociation mechanism, as studied
by Basolo [6] for L = EPh3 ((E = P, As, Sb). The disubsti-
tuted dihalide complexes 3 are the stable thermodynamic
products of the reaction.
The effect of temperature on the value of A in the range
ꢀ40ꢁ to +25 ꢁC is given in Table 3 for various concentra-
tions of L (L = PMe3, PEt3). The trend shows a maximum
The effect of halogen is given in Table 5. The reductive
elimination is much easier with iodide than with bromide
as evidenced by the strong increase of the A values of
iodide; moreover the values of A of iodide are influenced
by temperature, while those of bromide are not influenced.
The course of the reaction of Scheme 2 shows that the
reductive elimination follows the formation of the monosub-
stituted fac-Fe(CO)3LI2 complex; however, by starting from
fac-Fe(CO)3LI2, prepared by reaction of Fe(CO)4PMe3 and
iodine (see Section 2), only the formation of cis,trans,cis-
Fe(CO)2L2I2 via the very slow CO dissociation is observed.
The addition of small amounts of Fe(CO)4I2 to the solution
of fac-Fe(CO)3LI2 and PMe3 increases the reaction rate and
activates the elimination reaction. Then, during the first step
of the reaction, intermediate species, which catalyze the sec-
ond step of the reaction and promote the reductive elimina-
tion, must be formed.
9
8
7
6
5
4
2
4
6
8
10
12
14
16
102.[PMe3]
In order to obtain information about these intermediates,
photochemical reactions were carried out on the reaction
Fig. 1. Molar ratio of 5a and 3a complexes vs. [PMe3] for the reaction of
Scheme 2 in toluene at ꢀ20 ꢁC. [Fe(CO)4I2] = (1.35 0.15) · 10ꢀ2 M.

3886
G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
between fac-Fe(CO)3LI2 and PMe3. In the presence of stoi-
chiometric amount of PMe3 formation of all trans-
Fe(CO)2L2I2 (4a) complex was observed; in the presence of
an excess of PMe3, cis,trans,cis-Fe(CO)2L2I2 (3a) was
obtained. Reductive elimination was never observed. In the
absence of PMe3 cis,trans,cis-Fe(CO)2L2I2 is not photo-
chemically stable and isomerizes to all trans-Fe(CO)2L2I2.
This last complex in the presence of PMe3 isomerizes to cis,-
trans,cis-Fe(CO)2L2I2. Complex cis,trans,cis-Fe(CO)2L2I2
(L = PEt3) is not thermally stable and in the presence of an
excess of the ligand it slowly isomerizes to all trans-
Fe(CO)2L2I2.
mer,cis-Fe(CO)L3I2, in which CO is trans to iodide, as
the most probable structure of complex 6a with respect
to the structure fac,cis-Fe(CO)L3I2 in which CO is trans
to L.
Photochemical isomerization of cis to trans dihalide
complexes has often been observed in iron [14] and ruthe-
nium [15,16] complexes; similarly, cis to trans photochem-
ical isomerizations of alkyl halide complexes of iron,
ruthenium and osmium have been described previously
[17]; thermal trans to cis isomerization, catalyzed by phos-
phine ligands have also been described in ruthenium com-
plexes and its mechanism has been discussed [18]. Thermal
cis to trans isomerization, catalyzed by phosphine ligand as
the isomerization of cis,trans,cis-Fe(CO)2(PEt3)2I2 to all
trans-Fe(CO)2(PEt3)2I2, is uncommon and it is due to the
thermodynamic stabilization of the trans dihalide structure
owing to the steric hindrance of the PEt3 ligand.
Both cis,trans,cis-Fe(CO)2L2I2 (3a) and all trans-
Fe(CO)2L2I2 (4a) react with PMe3 as indicated in Scheme
3 to give trisubstituted complexes.
Complex 3a reacts with PMe3 to give mer,cis-
Fe(CO)L3I2 (6a) which isomerizes thermally in the presence
of L to mer,trans-Fe(CO)L3I2 (7a). Complex 7a is formed
from complex 4a photochemically. The structures of com-
4. Discussion
1
plexes 6a and 7a were assigned on the basis of H NMR
1
and 31P{1H} NMR spectra. The H NMR spectra show
4.1. Thermodynamics
a Harris triplet and a doublet for both complexes indicat-
ing the presence of two equivalent phosphine ligands.
The 31P{1H} NMR spectra are different: complex 7a shows
a triplet and a doublet corresponding to an A2X structure,
while complex 6a shows an A2B structure. The coupling
The most important aspect of the reaction studied in this
work is the reductive elimination of halogens. The course
of the reaction (Scheme 2) indicates that the reductive elim-
ination occurs during the second step of the reaction. On
the basis of the literature information, a few semiquantita-
tive considerations can be made about the thermodynamics
of the reductive elimination of halogens in iron complexes.
With reference to Scheme 4: and considering that the
Fe–I bond strength of fac-Fe(CO)3LI2 is higher or equal
to the Fe–I bond strength in Fe(CO)4I2, measured by
Connor and coworkers [EFe–I = 177 kJ] [19] and using
for I2 EI–I = 156 kJ [20], the enthalpy of the reaction of
Scheme 4 (DH4) is in kJ:
2
constants JPP are ꢂ 55 Hz for complex 7a and 58.3 Hz
for complex 6a, in agreement with a cis structure between
A and X and A and B, respectively [13]. The structures
of trisubstituted derivatives in agreement with these spec-
troscopic properties are fac,cis-Fe(CO)L3I2, mer,cis-
Fe(CO)L3I2 and mer,trans-Fe(CO)L3I2. While it is easy to
assign the structure mer,trans-Fe(CO)L3I2 to complex 7a,
which is also obtained photochemically, it is difficult to
choose between the fac,cis-Fe(CO)L3I2 and the mer,cis-
Fe(CO)L3I2 structures for complex 6a. The lower CO
stretching frequency of complex 6a suggests the structure
DH4 P 2EFe–I ꢀEFe–L ꢀEI–I ¼ 2:177ꢀEFe–L ꢀ156 ¼ 198ꢀEFe–L
The reaction of Scheme 4 is exothermic if EFe-L
> 198 kJ. Unfortunately, EFe-L are unknown in the litera-
ture; in general, however, the Metal-L bond strengths are
much lower than 198 kJ [2a]. Recently, the Fe–PH3 bond
strength in Fe(CO)4PH3 was calculated as 169 kJ [21].
Since the entropic effect is negligible with respect to the
enthalpic one when covalent bonds are involved [2], the
DH4 value confirms that the reaction of Scheme 4 is not
thermodynamically allowed; it can be made thermodynam-
ically allowed by coupling with the reaction of quaterniza-
tion of L with I2:
L
L
I
OC
I
OC
OC
hν
Fe
Fe
CO
+L
I
I
L
L
(4a)
(3a)
hν
+L
+L
L
L
L
L
I
L
I
L
L
OC
OC
OC
OC
I
I
Fe
Fe
L
+
Fe
CO
I2
+
CO
I
Fe
L
I
OC
L
L
CO
(6a)
(7a)
Scheme 3.
Scheme 4.

G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
3887
L þ I2 ! ½LIꢃþIꢀ
Fe(CO)4I + I
(a)
Fe(CO)4I2
The enthalpy of the quaternization reaction is unknown;
however the enthalpy of the reaction between PPh3 and
CH3I is ꢀ137.2 kJ [22] and that of the reaction between
PI3 and I2 to give [PI4]+Iꢀ was calculated to be ꢀ70 kJ
[23]. Therefore it is reasonable that the quaternization reac-
tion is exothermic enough to make the elimination reaction
(pathway a of Scheme 2) thermodynamically allowed. To
our knowledge only one example of this type of coupling
is described in the literature and concerns the reaction be-
tween W(CO)L2I3(=CNEt2) with PMe3 to give [W(CO)L2-
(PMe3)2(=CNEt2)]+Iꢀ and [PMe3I]+Iꢀ [4].
(b)
(c)
Fe(CO)4LI
Fe(CO)4I
+ L
Fe(CO)3LI + CO
Fe(CO)4LI
(d)
Fe(CO)3L2I
Fe(CO)3LI + L
Fe(CO)3L2I
F e(CO)2L2I + CO
(e)
(f)
[Fe(CO)3L2] ++ I-
Fe(CO)3L2I
The substitution of iodine with bromine strengthens the
Fe–X bond [24] and, although the Br–Br bond is stronger
than the I–I bond, the enthalpy of the reaction of Scheme 4
with bromine is less exothermic than that with iodine. This
is experimentally confirmed by the A values for X = Br
(Table 5) which are about 10 times lower than those for
X = I.
(g)
(h)
(i)
Fe(CO)3LI + I
Fe(CO)2L2I + I
Fe(CO)3L2] + + LI
Fe(CO)3LI2
Fe(CO)2L2I2
Fe(CO)3L2 + [LI]+
Scheme 5.
by Connelly and co-workers [26] in the oxidative elimina-
tion of Fe(CO)3(PPh3)2 by I2; radical species [Fe(CO)3L2]+
were isolated [27] and their reactivity with nucleophiles and
radicals [28] have been widely described; radicals
Fe(CO)2L2X (X = Br, I) were recently isolated as interme-
diates in the monoelectron reduction of Fe(CO)2L2X2
[29,30]; they are very stable in the absence of monoelec-
tron-reducing species or radicals. The reaction between
17-electron radicals and nucleophiles to give 19-electron
radicals is well documented in the literature for a few met-
als and, in particular, for iron [27] supporting equilibria 5b,
5c, 5d and 5e.
5. Mechanism
Another important aspect of the elimination reaction is
the mechanism of the coupling between the reaction of
Scheme 4 and the quaternization reaction. It is of interest
to emphasize that the first and second steps of Scheme 2
are not completely distinct. In fact, the reductive elimination
does not occur by starting from fac-Fe(CO)3(PMe3)I2: in this
latter case only the substitution of carbon monoxide to cis,-
trans,cis-Fe(CO)2(PMe3)2I2 (3a) is observed, as described by
Basolo [6]. Similarly the photochemical reaction between
fac-Fe(CO)3(PMe3)I2 and PMe3 only gives complex 3a.
However the addition of a small quantity of Fe(CO)4I2 to
the toluene solution of fac-Fe(CO)3(PMe3)I2 and PMe3 acti-
vates the reductive elimination reaction. This suggests that
the reductive elimination occurs at the level of intermediates,
which can also activate fac-Fe(CO)3(PMe3)I2. This is reason-
able considering that fac-Fe(CO)3(PMe3)I2 is not a thermo-
dynamic but a kinetic product of the reaction.
Scheme 5 qualitatively explains most of our experimen-
tal observations: in fact, the L basicity acts on the position
of equilibria 5b and 5d, influencing the relative concentra-
tions of fac-Fe(CO)3LI2 and cis,trans,cis-Fe(CO)2L2I2.
18
PMe2Ph
16
The Fe–I bond energy of Fe(CO)4I2 (EFe–I = 177 kJ) [19]
is very near to the I–I bond energy (EI–I = 156 kJ) [20]; so it
is easy to obtain a 17-electron radical Fe(CO)4I by ther-
mally breaking one of the Fe–I bonds. The 17-electron spe-
cies can activate the steps of Scheme 5, which explain most
of the experimental observations.
The 17-electron radicals, which gave the end products
fac-Fe(CO)3LI2, cis,trans,cis-Fe(CO)2L2I2 and Fe(CO)3L2
by reactions 5g, 5h and 5i are Fe(CO)3LI, Fe(CO)2L2I
and [Fe(CO)3L2]+, respectively. These three species are cor-
related by equilibria 5d, 5e and 5f and the concentrations
of the end products depend on the values of these equilib-
rium constants. Radical Fe(CO)4I may be responsible for
the activation of fac-Fe(CO)3LI2 via Fe–I–Fe bridge that
is easily formed in transition metal halides [25].
14
PMe3
12
10
8
PEt3
6
50
55
60
65
70
75
80
10.Fe-L relative bond energy (kJ)
Fig. 2. ꢀlogA = ꢀlog[5]/[3] at 25 ꢁC in toluene vs. relative Fe–L bond
energy of Fe(CO)3L2, assuming the energy of Fe–PPh3 bond to be zero.
Relative Fe–L bond energies from [31].
Some steps suggested in Scheme 5 have already been
proposed in the literature: equilibrium 5f was observed

3888
G. Bellachioma et al. / Journal of Organometallic Chemistry 691 (2006) 3881–3888
(b) S.G. Lias, J.E. Bartness, J.F. Liebman, J.L. Holmes, R.P.
Levin, W.G. Mallard, J. Phys. Chem. Ref. Data 17 (1988), Suppl. 1;
S.G. Lias, J.F. Liebman, R.D. Levin, J. Phys. Chem. Ref. Data 13
(1984) 695;
(c) H.-Y. Liu, K. Eriks, A. Prock, W.P. Giering, Organometallics 9
(1990) 1758;
With the most basic ligands PMe3 and PEt3 the ionization
of the Fe–I bond is easier and the elimination reaction 5f is
activated; then the ratio A between a and b reaction path-
ways of Scheme 2 depends prevalently on the L basicity.
The higher value of A of PEt3 can be explained by its
higher steric hindrance, which influences the Fe–L bond
energy and favours the ionization of step 5f. Fe–L bond
energies are unknown; however their relative values were
measured by Nolan and co-workers for the Fe(CO)3L2
complexes [31]. The plot of logA vs. Fe–L bond energies
for L = PMe2Ph, PMe3, PEt3, obtained by assuming Fe–
PPh3 equal zero, is shown in Fig. 2. It is clearly linear
and suggests that the higher value of A for PEt3 is due to
a higher Fe–PEt3 bond energy.
Reaction 5i requires the presence of LI radicals in solu-
tion. There is no experimental evidence of the presence of
this radical in solution. LX radicals were proposed in the
literature to explain the chain reaction of PPh3 and bromo-
form [32] and PPh3Me radical was hypothesized in the pho-
tochemical reaction of PPh3 and CpW(CO)2Me [33]. R3PX
(X = halides) radicals were identified by ESR spectroscopy
[34]. Although no evidence of the presence of PR3I in solu-
tion was obtained, it is reasonable to propose their role to
explain the formation of the cation [PR3I]+ through the
reaction of PR3I radical and [Fe(CO)3L2]+ (reaction 5i).
Md.M. Rahman, H.-Y. Liu, K. Eriks, A. Prock, W.P. Giering,
Organometallics 8 (1989) 1.
[11] J. Von Goubeau, R. Baumgartner, Zeit. Elektrochem. 64 (1960) 598.
[12] N. Bricklebank, S.M. Godfrey, A.G. Mackie, C.A. McAuliffe, R.G.
Pritchard, P.J. Kobryn, J. Chem. Soc., Dalton Trans. (1993) 101.
[13] P.S. Pregosin, R.W. Kunz, 31P and 13C NMR Transition Metal
Phosphine Complexes, Springer Verlag, New York, 1979, pp. 28–34.
[14] P.S. Braterman, Metal Carbonyl Spectra, Academic Press, London,
1975.
[15] M. Pankowski, M. Bigorgne, J. Organomet. Chem. 125 (1977) 231;
M. Pankowski, M. Bigorgne, J. Organomet. Chem. 19 (1969) 393.
[16] D.W. Krassowski, J.H. Nelson, K.R. Bower, D. Hauerstein, R.A.
Jacobson, Inorg. Chem. 27 (1988) 4294.
[17] G. Bellachioma, G. Cardaci, C. Jablowski, A. Macchioni, G.
Reichenbach, Inorg. Chem. 32 (1993) 2404;
G. Bellachioma, G. Cardaci, A. Macchioni, G. Reichenbach, J.
Organomet. Chem. 540 (1997) 7;
M. Aubert, G. Bellachioma, G. Cardaci, A. Macchioni, G. Reichen-
bach, J. Chem. Soc., Dalton Trans. (1997) 1759.
[18] C.F.J. Barnard, J.A. Daniels, R.J. Mawby, J. Chem. Soc., Dalton
Trans. (1979) 1331.
[19] J.A. Connor, C.P. Demain, H.A. Skinner, M.T. Zafarani-Moattar, J.
Organomet. Chem. 170 (1979) 117.
[20] R.C. West, Handbook of Chemistry and Physics, CRC Press, Boca
Baton, 1986.
[21] T.P.M. Goumans, A.W. Ehlers, M.C. van Hemert, A. Rosa, E.-J.
Baerends, K. Lammertsma, J. Am. Chem. Soc. 125 (2003) 3558.
[22] Y. Kondo, A. Zanka, S. Kusabayashi, J. Chem. Soc., Perkin Trans.
II (1985) 827.
Acknowledgements
This work was supported by a grant from the Ministero
`
dell’Istruzione, dell’Universita e della Ricerca (MIUR,
Rome, Italy) PRIN 2004.
[23] M. Gonsior, I. Kroning, L. Muller, I. Raabe, M. Jansen, L. van
Wullen, Chem. Eur. J. 8 (2002) 4475.
[24] T.L. Allen, J. Chem. Phys. 26 (1957) 1644;
R.A. Walton, Coord. Chem. Rev. 21 (1976) 63.
References
[25] J.L. Davidson, D.W.A. Sharpe, J. Chem. Soc., Dalton Trans. (1972)
107;
G. Bor, J. Organomet. Chem. 94 (1975) 181;
[1] R.H. Crabtreee, The Organometallic Chemistry of the Transition
Metals, third ed., Wiley Interscience, New York, 2001;
K.I. Goldberg, J.Y. Yam, E.M. Breitung, J. Am. Chem. Soc. 117
(1995) 6889.
H. Alper, A.S.K. Chan, J. Am. Chem. Soc. 95 (1975) 4905.
[26] P.K. Baker, N.G. Connelly, B.M.R. Jones, J.P. Maher, K.R. Somers,
J. Chem. Soc., Dalton Trans. (1980) 579.
[27] W.C. Trogler, in: W.C. Trogler (Ed.), Organometallic Radical
Processes, Elsevier, Amsterdam, 1990.
[2] (a) P.B. Dias, M.E. Minas de Piedade, J.A. Martnho Simoes, Coord.
Chem. Rev. 135/136 (1994) 737;
[28] M.J. Therien, W.C. Trogler, J. Am. Chem. Soc. 108 (1986) 3687;
M.J. Therien, C.-L. Ni, F.C. Anson, J.G. Osteryoung, W.C. Trogler,
J. Am. Chem. Soc. 108 (1986) 4037.
[29] H. Kandler, C. Gauss, W. Bidell, S. Rosenberg, T. Burgi, I.
Eremenko, D. Veghini, O. Orama, P. Burger, H. Berke, Chem. Eur.
J. 8 (1995) 514.
(b) J.A. Martinho Simoes, J.L. Beauchamp, Chem. Rev. 90 (1990)
629;
(c) S.J. Blanksby, G.E. Ellison, Acc. Chem. Res. 36 (2003) 255.
[3] J. Fortune, A.R. Manning, Inorg. Chem. 19 (1980) 2590.
[4] A.L. Filippou, E.O. Fischer, H.G. Alt, J. Organomet. Chem. 344
(1988) 215.
[30] C. Venturi, G. Bellachioma, G. Cardaci, A. Macchioni, C. Zuccaccia,
Inorg. Chim. Acta 358 (2005) 3815.
[5] G. Cardaci, G. Bellachioma, P.F. Zanazzi, Organometallics 7 (1988)
172.
[31] C. Li, E.D. Stevens, S.P. Nolan, Organometallics 14 (1995) 3791;
C. Li, S.P. Nolan, Organometallics 14 (1995) 1327;
L. Luo, S.P. Nolan, Inorg. Chem. 32 (1993) 2410.
[32] F. Ramirez, N. McKelvie, J. Am. Chem. Soc. 79 (1957) 5829.
[33] A.S. Goldman, D.R. Tyler, J. Am. Chem. Soc. 108 (1986) 89.
[34] M.C.R. Symons, R.L. Pextersen, J. Chem. Soc., Faraday Trans. II 75
(1979) 210;
[6] I.A. Cohen, F. Basolo, J. Inorg. Nucl. Chem. 23 (1966) 511.
[7] W. Hieber, G. Bader, Z. Anorg. Allg. Chem. 190 (1930) 193.
[8] G. Reichenbach, G. Cardaci, G. Bellachioma, J. Chem. Soc., Dalton
Trans. (1982) 847.
[9] J.F. Coetzee, G.P. Cunningham, D.K. McGuire, G.R. Padmanab-
ham, Anal. Chem. 34 (1962) 1139.
[10] (a) C.A. Streuli, Anal. Chem. 32 (1960) 985;
W.R. Henderson Jr., C.A. Streuli, J. Am. Chem. Soc. 82 (1960)
5791;
R.A.J. Janssen, M.H.W. Sonnemans, H.M. Buck, J. Am. Chem. Soc.
108 (1986) 6145.