Preparation, spectroscopic properties, and crystal structures of Fe2(CO)6(μ-CO)(μ-CF2)2, Fe2(CO)6(μ-CO)2(μ-CF2), and Fe2(CO)6(μ-CF2)(PPh3)2 - Theoretical studies of methylenic vs. carbonyl bridges in diiron complexes

Article abstract of DOI:10.1002/1521-3749(200108)627:8<1859::AID-ZAAC1859>3.0.CO;2-P

The reaction of Na₂[Fe(CO)₄] with Br₂CF₂ in n-pentane generates a mixture of the compounds (CO)₃Fe(μ-CO)_3-n(μ-CF₂) _nFe(CO)₃ (2, n = 2; 3, n = 1) in low yields with 3 as the main product. 3 is obtained free from 2 by reacting Br₂CF₂ with Na₂[Fe₂(CO)₈]. The non-isolable monomeric complex (CO)₄Fe=CF₂ (1) can probably considered as the precursor for 2. 3 reacts with PPh₃ with replacement of two CO ligands to form Fe₂(CO)₆(μ-CF₂)(PPh₃)₂ (4). The complexes 2-4 were characterized by single crystal X-ray diffraction. While the structure of 2 is strictly similar to that of Fe₂(CO)₉, the structure of 3 can better be described as a re suiting from superposition of the two enantiomers 3a and 3b with two semibridging CO groups. Quantum chemical DFT calculations for the series (CO)₃Fe(μCO)_3-n(μ-CF₂) _nFe(CO)₃ (n = 0, 1, 2, 3) as well as for the corresponding (μ-CH₂) derivatives indicate that the progressively larger σ donor and π acceptor properties for the bridging ligands, in the order CO < CF₂ < CH₂, favor a stronger Fe-Fe bond. WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001.

Full text of DOI:10.1002/1521-3749(200108)627:8<1859::AID-ZAAC1859>3.0.CO;2-P

Preparation, Spectroscopic Properties, and Crystal Structures
of Fe2ꢀCO)6ꢀl-CO)ꢀl-CF2)2, Fe2ꢀCO)6ꢀl-CO)2ꢀl-CF2),
and Fe2ꢀCO)6ꢀl-CF2)ꢀPPh3)2 ± Theoretical Studies of Methylenic vs.
Carbonyl Bridges in Diiron Complexes
,a
a
b
,b
W.Petz* , F.Weller , A.Barthel , C.Mealli c, and J.Reinhold**
a Marburg, Fachbereich Chemie der Philipps-UniversitaÈt
b Leipzig, Wilhelm-Ostwald-Institut fuÈr Physikalische und Theoretische Chemie der UniversitaÈt
c Firenze/Italy, Istituto per lo Studio della Stereochimica ed Energetica dei Composti di Coordinazione, CNR
Received April 27th, 2001.
Ad memoriam Egon Wiberg
Abstract. The reaction ofNa 2[Fe1CO)4] with Br2CF2 in n-
sulting from superposition of the two enantiomers 3 a and
pentane generates a mixture ofthe compounds 1CO) Fe1l-
3 b with two semibridging CO groups. Quantum
chemical DFT calculations for the series 1CO)3Fe1l-
CO)3±n1l-CF2)nFe1CO)3 1n = 0, 1, 2, 3) as well as for the cor-
responding 1l-CH2) derivatives indicate that the progres-
sively larger r donor and p acceptor properties for the brid-
ging ligands, in the order CO < CF2 < CH2, favor a stronger
Fe±Fe bond.
3
CO)3±n1l-CF2)nFe1CO)3 12, n = 2; 3, n = 1) in low yields with
3 as the main product. 3 is obtained free from 2 by reacting
Br2CF2 with Na2[Fe21CO)8]. The non-isolable monomeric
complex 1CO)4Fe=CF2 11) can probably considered as the
precursor for 2. 3 reacts with PPh3 with replacement oftwo
CO ligands to form Fe21CO)61l-CF2)1PPh3)2 14). The com-
plexes 2±4 were characterized by single crystal X-ray diffrac-
tion. While the structure of 2 is strictly similar to that of
Fe21CO)9, the structure of 3 can better be described as a re-
Keywords: Iron carbonyl; Carbondifluoride ligand; DFT cal-
culations; Crystal structure
Darstellung, spektroskopische Eigenschaften und Kristallstrukturen von
Fe2ꢀCO)6ꢀl-CO)ꢀl-CF2)2, Fe2ꢀCO)6ꢀl-CO)2ꢀl-CF2) und Fe2ꢀCO)6ꢀl-CF2)ꢀPPh3)2
± Theoretische Untersuchung uÈber Methylen- vs. Carbonyl-BruÈcken in Dieisen
Komplexen
InhaltsuÈbersicht. Aus der Reaktion von Na2[Fe1CO)4] mit
Br2CF2 in n-Pentan laÈût sich ein Gemisch der Komplexe
1CO)3Fe1l-CO)3±n1l-CF2)nFe1CO)3 12, n = 2; 3, n = 1) in
maÈûigen Ausbeuten isolieren, wobei 3 das Hauptprodukt ist.
Man erhaÈlt 3 auch frei von 2 bei der Umsetzung von Br2CF2
mit Na2[Fe21CO)8]. Der nicht isolierbare monomere Kom-
plex 1CO)4Fe=CF2 11) ist vermutlich als Vorstufe fuÈr 2 auf-
zufassen. 3 reagiert mit PPh3 unter VerdraÈngung von zwei
CO-Liganden zu Fe21CO)61l-CF2)1PPh3)2 14) ab. Die Kom-
plexe 2±4 sind durch Kristallstrukturanalysen charakterisiert.
WaÈhrend die Struktur von 2 der von Fe21CO)9 aÈhnlich ist,
laÈût sich die von 3 besser als Ûberlagerung der Enantiome-
ren 3 a und 3 b mit zwei halbverbruÈckenden CO-Gruppen
beschreiben. Quantenchemische DFT-Rechnungen fuÈr die
Reihe 1CO)3Fe1l-CO)3-n1l-CF2)nFe1CO)3 1n = 0, 1, 2, 3) so-
wie fuÈr die entsprechenden 1l-CH2)-Verbindungen zeigen,
daû die fortschreitend staÈrkeren r-Donor- und p-Acceptor-
eigenschaften der BruÈckenliganden in der Reihe CO < CF2 <
CH2 eine staÈrkere Fe±Fe-Bindung favorisieren.
Introduction
* Prof. Dr. Wolfgang Petz,
Fachbereich Chemie der Philipps-UniversitaÈt,
D-35032 Marburg,
Fax: ++49-64 21-2 82 56 53
E-mail: petz@mailer.uni-marburg.de
** Prof. Dr. Joachim Reinhold,
Wilhelm Ostwald-Institut fuÈr Physikalische und Theoretische
Chemie der UniversitaÈt Leipzig,
D-04103 Leipzig,
Transition metal complexes with the difluorocarbene
ligand, CF2, are known since many years and an excel-
lent review has appeared summarizing the relevant
preparative and theoretical studies on this field of
chemistry [1]. Similar to CO and some related alkyli-
denes the CF2 ligand is coordinated either in a termi-
nal or in a bridging manner. Terminal CF2 complexes
concentrate on compounds ofmolybdenum, the iron
triad, and iridium. A few complexes with the CF2 li-
Fax: ++49-3 41-9 73 63 99
E-mail: reinhold@quant1.chemie.uni-leipzig.de
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869 Ó WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001 0044±2313/01/6271859±1869 $ 17.50+.50/0 1859

W. Petz, F. Weller, A. Barthel, C. Mealli, J. Reinhold
gand bonded in a bridging manner are known from
manganese [2], iron [3], cobalt, and iridium [2, 3]. It
was concluded, that dihalogencarbene complexes are
links between the typical heteroatom stabilized elec-
trophilic ªFischerº carbene complexes and the nucleo-
philic alkylidene compounds ofthe ªSchrockº type.
Thus, depending on the further substituents and the
oxidation state ofthe central metal atom, the coordi-
nated CF2 ligand exhibits nucleophilic or electrophilic
behavior. The ground state ofthe rfee CF ligand is a
2
singlet [4] similar to that ofheteroatom stabilized car-
benes, while the ground state ofalkylidenes is a triplet
state [5].
The access to difluorocarbene complexes with a
terminal ligand follows two main pathways. One route
starts from the corresponding transition metal CF3
complexes reacting with fluoride abstracting agents.
The other route comprises transfer of CF2 from
M1CF3)2 1M = Hg, Cd) to low valent transition metal
compounds [1]. In one case the thermal decomposi-
tion ofa complex with the Ir1COCF 2Cl) fragment
leads to the formation of the corresponding CF2 com-
plex [6].
Compounds with a bridging CF2 ligand originate
from various sources including the cheap and easily
available dibromodifluoromethane. This compound is
well known in organic chemistry to generate the di-
fluorocarbene species by reacting with elemental lead
[7, 8] and was also reported to react with neutral car-
bonyl compounds ofCo and Fe under photochemical
conditions to produce l-CF2 complexes [3].
Scheme 1
other source for 2 would be the reaction ofBr CF2
2
with a reduced species of 3, but we were not able to
increase the yield of 2 upon reducing 3 and again re-
acting with Br2CF2.
We suppose, that Br2CF2 acts not only as a source
for CF2 but also as an oxidizing reagent towards the
dianion [Fe1CO)4]2±. The oxidation 1geminal dihalides
such as Br2CF2, Cl2CS or even Cl2CH2 often mimic
halogen) generates first [Fe21CO)8]2± which also pro-
duces 3 with Br2CF2 so that 3 is the main product.
Fe1CO)5 always forms when polynuclear carbonylates
ofiron are exposed to oxidizing agencies.
While 3 is the main product, the portion of 2
amounts only about 5% and the overall yield ofthe
mixture varies between 12 and 23%. The material is
separated by vacuum sublimation or by low tempera-
ture crystallization from n-pentane. The complex 3 is
obtained free from 2 ifthe reaction is carried out un-
der similar conditions with Na2[Fe21CO)8]; the corre-
sponding Li-salt [10] leads to the same results. So we
believe, that the dianion is the main source for the for-
mation of 3.
This paper describes the access to the compounds
Fe21CO)71l-CF2)2, Fe21CO)81l-CF2) and Fe21CO)61l-
CF2)1PPh3)2 using Br2CF2 as CF2 source [9]. Further-
more, theoretical studies ofthe series ofcompounds
ofthe type 1CO) 3Fe1l-CO)3±n1l-CF2)nFe1CO)3
1n = 1, 2, 3) are presented in comparison to the corre-
sponding 1l-CH2) compounds.
The choice ofthe solvent is very important for reac-
tion and isolation. Thus, ifthe reaction is carried out
in THF solution only traces of 2 and 3 are formed as
shown by 19F NMR spectroscopy. Br2CF2 as solvent,
however, leads to the same results as found in n-pen-
tane solution. The relative amounts of 2 and 3 are in-
dependent ofthe reaction temperature and do also
not change upon reverse addition ofthe educts. Below
±20 °C no reaction is observed; it starts above this
temperature indicating by forming a brownish solu-
tion. Attempts to separate 2 and 3 by column chroma-
tography have failed.
Results and Discussion
1
Preparation
Addition ofan excess ofBr 2CF2 to a suspension of
Collmans reagent in n-pentane leads to a slight warm-
ing ofthe mixture and ofrmation ofa dark greenish
brown solution and a dark brown to black solid. A
mixture ofthe compounds 2 and 3 is isolated from the
residue ofthe pentane solution upon sublimation. The
formation of both compounds during the reaction can
be rationalized from the following reactions summar-
ized in scheme 1.
The expected monomeric species 1 could not be iso-
lated; however, the presence of 2 is an indirect proof
for the formation of 1 being formed in the first step of
the reaction. 1 is probably unstable under the reaction
conditions and reacts either with itselfto give 2 or
with Fe1CO)5 to produce additional amounts of 3. An-
Two ofthe carbonyl groups of 3 can be replaced by
phosphine ligands. Ifa pentane solution of 3 is added
at room temperature to a solution oftwo equivalents
ofPPh in n-pentane immediately brown crystals of 4
3
begin to separate, and within a few minutes quantita-
tive yield is obtained. The crystals were directly suita-
ble for X-ray analysis.
1860
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869

About Fe21CO)61l-CO)1l-CF2)2, Fe21CO)61l-CO)21l-CF2), and Fe21CO)61l-CF2)1PPh3)2
3 + 2 PPh3 ® 1CO)31Ph3P)Fe1l-CF2)Fe1PPh3)1CO)3 + 2 CO
plays two doublets at 61 and 22 ppm according to
fluorine atoms in different chemical environments; the
fluorine atoms of 3 produces a singlet at 51 ppm. In
the 13C NMR spectrum of 3 the CF2 carbon atom ap-
4
11)
4 is stable as a solid under an argon atmosphere but
unstable in solution; in THF or in toluene slow de-
composition occurs to produce 1CO)3Fe1PPh3)2 along
with traces of1CO) 4FePPh3 1see spectroscopic re-
sults). Attempts to prepare 4 in THF or toluene lead
to a mixture of 4 with the tri- and tetracarbonyl iron
1
pears at 219 ppm with JC,F = 384 Hz typical for brid-
ging CF2 ligands [2, 3, 14]. The CO groups of 3 exhibit
a single resonance which, at ±40 °C splits into a triplet
due to the coupling with the fluorine atoms. This indi-
cates scrambling within the NMR time scale. Unfortu-
nately, the spectra could not be recorded at lower
temperatures due to the crystallization ofthe species.
phosphine compounds. The remaining ofthe CF
2
group could not yet be cleared up 1see 19F NMR spec-
troscopy).
Replacement oftwo CO groups by PPh causes the
3
Other phosphines such as P1NMe2)3 or dppe also
react with 3, but the isolated products are not totally
characterized as yet. Attempts to reduce 3 with alkali
metals have failed. Further studies on the chemistry of
2 and 3 are presently in progress.
19F NMR signal in 4 to shift from 51 ppm to 42 ppm
indicating a better shielding ofthe lfuorine atoms un-
der the influence of the phosphine ligands. In solution
4 shows highly dynamic behavior; the two differently
bonded phosphine ligands give only one 31P NMR sig-
nal 1JP, F = 8.4 Hz) and the four chemically different
types ofCO groups ofund in the solid state are also
condensed to one signal in the solution 13C NMR
spectrum down to ±40 °C. Again, crystallization did
not allow to work at lower temperatures. These results
indicate a rapid scrambling ofthe ligands in the NMR
time scale but a description ofthe actuel mechanism
can only be speculative at this time. Whether bridging
CO or PPh3 groups are involved as intermediates or
the scrambling takes place isolated at each iron atom
is a still open question.
2
Spectroscopy
The infrared spectrum of 2 exhibits strong bands in
the typical region for terminal CO groups and one
sharp and intense band at 1879 cm±1 which can be as-
signed to the single bridging carbonyl ligand. Two
medium intense bands at 1091 and 1046 cm±1 belong
to stretching vibrations ofthe CF 2 group. The IR spec-
trum of 3 does not clearly announce the presence of
bridging CO ligands; the bands ofthe terminal CO
groups are only accompanied by a shoulder on the
low frequency side of the carbonyl bands, which can
be attributed to semibridging CO groups and the ap-
pearance ofthe bands is independent rfom the polar-
ity ofthe solvent. The m1CF) vibrations of 3 are shifted
to 1034 and 997 cm±1. High values ofthe m1CO)
stretching frequencies in 2 and 3 suggest that the p-ac-
As mentioned above, solutions of 4 at room tem-
perature are not stable and a slow decomposition is
observed with formation of the known complex
1CO)3Fe1PPh3)2 and less amounts of1CO) 4FePPh3.
After about 200 h in THF solution, the original 19F
NMR signal of 4 has disappeared and new signals ap-
pear at 99 1main signal), 52, 47, 7 and ±45 ppm. A sim-
ilar evolution is observed in the 31P NMR spectrum
which shows signals ofthe phosphine species along
with bands at 63, 49, and 41 ppm. The latter have not
been identified as yet.
The slow decomposition of 4 in solution leads to
the suggestion that concentration ofthe phosphine li-
gands at one iron atom may be an irreversible process
leading to splitting off 1CO)3Fe1PPh3)2. Such an inter-
mediate is probably not involved in the scrambling
process.
ceptor ability ofthe CF ligand is substantially higher
2
than that ofa CO ligand as shown by comparison with
the m1CO) frequencies in Fe21CO)9 12018, 1829 cm±1).
The IR spectrum of 4 exhibits only bands oftermi-
nal CO groups, the center ofwhich, however, has
shifted to lower frequencies relative to the CO vibra-
tions of 2 and 3 according to the more electron donat-
ing properties ofthe phosphines. In the region ofthe
m1CF) vibrations medium intense bands at 997 and
937 cm±1 can be assigned to the CF2 ligand, similarly
shifted to lower frequencies like the carbonyl bands.
Relative to matrix isolated free CF2 with the
ms1FCF) at 1102 and mas1FCF) at 1221 cm±1 the two CF
vibrations ofthe complexes 2 to 4 are shifted to lower
frequencies [11] and we suppose that in the complexes
similarly the bands at higher frequencies can be as-
signed to the asymmetric CF2 vibration.
3
Crystal Structures
3.1 General remarks
The structures ofthe compounds 2, 3, and 4 could be
confirmed by single-crystal X-ray analyses. Suitable
crystals of 2 and 3 were obtained by slow evaporation
ofthe solvent rfom a solution in n-pentane. Crystals
The CF2 ligand is easily characterized by 19F and
13C NMR spectroscopy. While terminal ligands exhibit
strongly deshielded fluorine and carbon atoms with
shifts to about 160 1F) and 270 1C) ppm [12, 13] a shift
to higher field is observed for both nuclei when ar-
ranged in a bridge. The 19F NMR spectrum of 2 dis-
of 4 were obtained by addition ofa solution ofPPh
3
in n-pentane to a solution of 3 in the same solvent.
ORTEP views ofthe molecules are depicted in Fig-
ures 1 to 4; details ofthe structure determination are
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869
1861

W. Petz, F. Weller, A. Barthel, C. Mealli, J. Reinhold
Table 1 Experimental data for the X-ray studies of the complexes 2, 3 and 4
2
3a
3b
4
empirical formula
formula weight
crystal size/mm3
crystal system
C9F4Fe2O7
407.79
0.18 ´ 0.34 ´ 0.34
orthorhombic
Pnma
C9F2Fe2O8
385.79
0.08 ´ 0.17 ´ 0.41
monoclinic
P21/m
C9F2Fe2O8
385.79
0.08 ´ 0.17 ´ 0.41
monoclinic
P21/m
C43H30F2Fe2O6P2
854.31
0.35 ´ 0.27 ´ 0.04
triclinic
space group
P1
unit cell dimensions/pm
a = 895.712)
b = 1113.112)
c = 1237.814)
a = 625.911)
b = 1199.812)
c = 847.511)
a = 625.911)
b = 1199.812)
c = 847.511)
a = 1059.211)
b = 1361.411)
c = 1436.411)
a = 97.113111)
b = 110.008111)
c = 93.562ꢀ12)
2
1.478
1919.113)
8000 rflns
21312)
b = 107.180110)
b = 107.180110)
Z
4
2
2
dc/gcm±3
2.195
1234.116)
2.107
608.311)
2000 rflns
20312)
376
2.107
608.011)
2000 rflns
20312)
376
volume 1´ 10±30 m3)
cell determination
temperature/K
F1000)
29312)
792
872
diffractometer
wave length/pm
CAD4 1Enraf-Nonius)
154.178
IPDS 1Stoe)
71.073
IPDS 1Stoe)
71.073
IPDS 1Stoe)
71.073
theta range for data collection 5.34 to 74.79°
2.52 to 28.00°
±8 < h < 8, ±15 < k < 15,
±10 < l < 11
phi-scans
2.52 to 28.00°
±8 < h < 8, ±15 < k < 15,
±10 < l < 11
phi-scans
2.06 to 26.97°
±13 < h < 13, ±17 < k < 17,
±18 < l < 18
phi-scans
index range
0 < h < 11, ±13 < k < 0,
±15 < l < 0
scan method
omega scans
control reflections and decay
reflections collected
independent
observed [I > 2r1I)]
rfls used for refinement
completeness
largest diff. peak and hole
solution
refinement
2 rfls all 120 min, 0%
1329
1329 [Rint = 0.000]
1121
5544
5544
21041
7729 [Rint = 0.0646]
4253
1496 [Rint = 0.0302]
1265
1496
97.3% 1to theta = 28.00°)
0.701 and ±1.442 e ´ A±3
direct/difmap
1496 [Rint = 0.0302]
1265
1496
97.3% 1to theta = 28.00°)
0.492 and ±0.406 e ´ A±3
direct/difmap
1329
7729
99.6% 1to theta = 74.79°)
0.621 and ±0.478 e ´ A±3
direct/difmap
92.4% 1to theta = 26.97°)
0.740 and ±0.388 e ´ A±3
direct/difmap
full matrix least squares on F2 full matrix least squares on F2 full matrix least squares on F2 full matrix least squares on F2
geom, mixed
hydrogen atoms
programs used
SHELXS-97 1Sheldrick, 1990) SHELXS-97 1Sheldrick, 1990) SHELXS-97 1Sheldrick, 1990) SHELXS-97 1Sheldrick, 1990)
SHELXL-97 1Sheldrick, 1997) SHELXL-97 1Sheldrick, 1997) SHELXL-97 1Sheldrick, 1997) SHELXL-97 1Sheldrick, 1997)
SHELXTL
1329/0/112
1.117
wR2 = 0.1112
R1 = 0.0364
SHELXTL
1496/0/106
1.097
wR2 = 0.1247
R1 = 0.0465
SHELXTL
1496/10/181
0.981
wR2 = 0.0704
R1 = 0.0273
SHELXTL
7729/0/430
0.838
wR2 = 0.1137
R1 = 0.0455
data/restraints/parameters
goodness-of-fit on F2
R index 1all data)
R index conventional
[I > 2r1I)]
a)
refined mediated, b) refined disordered
collected in Table 1; bond lengths and angles are sum-
marized in Table 3 to 5. Other crystal structures of
compounds containing a bridging CF2 ligand have
been reported for 1CO)4Mn1l-CF2)2Mn1CO)4 [2] and
the heterobimetallic complex [Cp1CO)Fe1l-CF2)1l-
CO)IrCl1CO)1PMe2Ph)]BF4 [14].
3.2 Molecular structure of 2
The structure of 2 is shown in Fig. 1. The molecule
can be regarded as a derivative ofFe 1CO)9 in which
2
two bridging CO groups are replaced by CF2 ligands.
The Fe±Fe distance of247 pm is about 6 pm shorter
than in Fe21CO)9 [15] and, to our knowledge, is the
shortest one found in these types of compounds. A
comparison with the average structure of 3 reveals
that more CF2 groups in place ofthe CO ones afvor
the closeness ofthe metal atoms 1see theoretical part).
The single bridging CO ligand along with the metal
atoms and two terminal carbonyl groups are located
on the crystallographic mirror plane ofthe molecule.
Fig.1 Molecular structure of 2 with atomic numbering
scheme; thermal ellipsoids are shown at the 50% probability
level.
By ignoring the Fe±Fe bond, the coordination ofeach
metal atom is nearly octahedral with the two octahe-
dra sharing one face. The F±C±F angle of 103.9° is
1862
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869

About Fe21CO)61l-CO)1l-CF2)2, Fe21CO)61l-CO)21l-CF2), and Fe21CO)61l-CF2)1PPh3)2
similar to those found in the other two known struc-
tures ofcompounds with bridging CF 2 ligands [2, 14].
Surprisingly, the dihedral angle between the planes
formed by Fe11), C16), and Fe12) and Fe11), C16a),
and Fe12) deviates from 120° 1111° is recorded). This
might be due to either a repulsion between the brid-
ging CO and CF2 groups or attraction between the
CF2 groups themselves. Remarkably, the distance be-
tween the F11) and F11a) atoms from the different
CF2 groups is 279 pm, well below the sum ofthe van
der Waals radii of294 pm.
The eight atoms ofthe bridging groups almost lie
all in the molecular pseudo mirror plane. The devia-
tions ofthe atoms rfom their best plane are given in
Table 2.
The relatively large deviation ofthe carbonyl car-
bon atom from this plane is also expressed by differ-
ent distances to the iron atoms 1196 and 203 pm) indi-
cating a slight tendency to a semibridging species,
which, however, is not feasible in the presence of the
two symmetrically arranged CF2 bridges. The angles
Fe±C1O)±Fe and the Fe±C1F2)±Fe comprise 76.5° and
76.6°, respectively, and are closely related to those in
Fe21CO)9 with 77.6° [15].
3.3 Molecular structure of 3
The crystal structure of 3 shows an unusual elongation
ofthe thermal ellipsoids ofthe iron atoms perpendicu-
lar to the crystallographic mirror plane as well as a
similar elongation with the bridging carbonyl carbon
atoms parallel to this plane as depicted in Fig. 2. The
latter features are independent from the absorption
correction and from the data collection technique. In
fact, the data were measured twice on different dif-
fractometers having different diffraction geometries.
Moreover, merohedral twinning can be excluded on
the basis of the crystal's metrics and as in the diffrac-
tion pattern ofthe IPDS measurement there were no
signs ofadditional individuals which could suggest a
non-merohedral twin. Ultimatively, it may be reason-
ably hypothesized that 3 is an average structure from
two less symmetric species, namely it results from the
superposition ofthe enantiomers 3 a and 3 b with two
semibridging CO ligands in place ofthe symmetrically
bridging ones as represented in Fig. 3. Thus, 3 can be
considered as the racemic mixture. Calculation ofa
split model for all atoms except O15) and C16) im-
proved considerably the quality ofthe reifnement and
the plausibility ofthe thermal ellipsoids. We also re-
strained all ofthe terminal Fe±C and C±O distances
as well as the C±F ones to be consistently equal 1See
table 4).
Table 2 Deviation ofthe bridging atoms in 2 from the best
plane 1in pm)
C15)
O15)
C16)
F11)
±4.14
+0.92
+1.81
±1.99
F12)
+1.80
+1.81
±1.99
+1.80
C16 a)
F11 a)
F12 a)
Table 3 Distances/pm and angles/° in 2
Fe11)±C11)
Fe11)±C12 a)
Fe11)±C12)
Fe11)±C15)
Fe11)±C16 a)
Fe11)±C16)
Fe11)±Fe12)
Fe12)±C13)
Fe12)±C14)
Fe12)±C14 a)
181.415)
183.314)
183.314)
196.115)
200.814)
200.814)
246.87111)
180.716)
184.014)
184.014)
Fe12)±C16)
Fe12)±C16 a)
Fe12)±C15)
C11)±O11)
C12)±O12)
C13)±O13)
C14)±O14)
C15)±O15)
C16)±F11)
C16)±F12)
197.614)
197.614)
202.716)
112.717)
112.214)
114.217)
112.315)
115.016)
135.114)
135.314)
The enantiomers 3 a and 3 b can be considered as
intermediates on the way from a singly bridged struc-
ture as proposed for Os21CO)9 to the triply bridged
structure recorded for Fe21CO)9. The lack ofclear
bridging m1CO) vibrations in the IR spectrum, as in
contrast recorded for 2, supports this view.
C11)±Fe11)±C12 a)
C11)±Fe11)±C12)
C12 a)±Fe11)±C12)
C11)±Fe11)±C15)
C12 a)±Fe11)±C15)
C12)±Fe11)±C15)
C11)±Fe11)±C16 a)
C12 a)±Fe11)±C16 a) 91.37115)
C12)±Fe11)±C16 a) 170.66115)
C15)±Fe11)±C16 a)
C11)±Fe11)±C16)
93.51115)
93.51115)
97.112)
C14)±Fe12)±C16)
C14 a)±Fe12)±C16) 170.19116)
C13)±Fe12)±C16 a) 91.38117)
C14)±Fe12)±C16 a) 170.19116)
C14 a)±Fe12)±C16 a) 91.14116)
C16)±Fe12)±C16 a)
C13)±Fe12)±C15)
C14)±Fe12)±C15)
C14 a)±Fe12)±C15)
C16)±Fe12)±C15)
C16 a)±Fe12)±C15)
C13)±Fe12)±Fe11)
C14)±Fe12)±Fe11)
91.14116)
177.912)
87.87116)
87.87116)
89.94116)
81.512)
178.512)
85.66116)
85.66116)
87.52115)
87.52115)
127.98118)
117.95112)
88.46117)
89.94116)
C12 a)±Fe11)±C16) 170.67115)
C12)±Fe11)±C16)
C15)±Fe11)±C16)
C16 a)±Fe11)±C16)
C11)±Fe11)±Fe12)
91.37115)
88.46117)
79.912)
C14 a)±Fe12)±Fe11) 117.95112)
C16)±Fe12)±Fe11)
C16 a)±Fe12)±Fe11)
C15)±Fe12)±Fe11)
O11)±C11)±Fe11)
O12)±C12)±Fe11)
O13)±C13)±Fe12)
O14)±C14)±Fe12)
O15)±C15)±Fe11)
O15)±C15)±Fe12)
Fe11)±C15)±Fe12)
Fe12)±C16)±Fe11)
F12)±C16)±Fe12)
F12)±C16)±Fe11)
52.30110)
52.30110)
50.57115)
179.013)
178.613)
178.615)
178.414)
145.015)
138.515)
76.512)
124.94116)
C12 a)±Fe11)±Fe12) 120.36111)
C12)±Fe11)±Fe12)
C15)±Fe11)±Fe12)
C16 a)±Fe11)±Fe12)
C16)±Fe11)±Fe12)
C13)±Fe12)±C14)
C13)±Fe12)±C14 a)
C14)±Fe12)±C14 a)
C13)±Fe12)±C16)
F12)±C16)±F11)
120.36111)
52.98117)
51.12110)
51.12110)
95.3117)
95.3117)
95.412)
91.38117)
103.013)
120.512)
118.412)
76.58112)
119.413)
118.712)
Fig.2 Crystal structure of 3, average structure; with atomic
numbering scheme; thermal ellipsoids are shown at the 50%
probability level.
F11)±C16)±Fe12)
F11)±C16)±Fe11)
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869
1863

W. Petz, F. Weller, A. Barthel, C. Mealli, J. Reinhold
Table 4 Distances/pm and angles/° in 3 and in 3 a, b
average structure of 3
disordered structure 3 a, b
Fe11)±C121 a)
Fe11)±C11 a)
Fe11)±C16)
Fe11)±C151 a),
Fe11)±C152)
Fe11)±C12)
Fe11)±C11)
Fe11)±C16)
Fe11)±C15)
181.014)
180.816)
183.612)
198.69117)
246.116),
184.816)
258.1616)
180.215)
183.413)
198.75119)
248.916),
183.016)
112.913)
114.915),
114.115)
113.313)
114.815),
115.015)
136.714),
137.014)
111.816),
112.316)
183.614)
197.614)
215.215)
Fe11)±Fe12)
Fe12)±C14)
Fe12)±C13)
Fe12)±C16)
Fe12)±C15)
254.0519)
181.715)
183.415)
197.314)
215.315)
Fe11)±Fe12)
Fe12)±C142 a)
Fe12)±C13 a)
Fe12)±C16)
Fe12)±C152),
Fe12)±C151 a)
C11 a)±O11 a)
C121 a)±O121 a),
C122)±O122)
C13 a)±O13 a)
C141)±O141),
C142)±O142)
C16)±F11),
C11)±O11)
C12)±O12)
112.715)
112.816)
C13)±O13)
C14)±O14)
113.315)
113.115)
C16)±F
135.513)
109.115)
C16)±F12 a)
C151 a)±O15 a),
C152)±O15)
C15)±O15)
C12)±Fe11)±C12 a) 94.212)
C121 a)±Fe11)±C122)
C121 a)±Fe11)±C11 a)
C121 a)±Fe11)±C16)
C11 a)±Fe11)±C16)
C121 a)±Fe11)±C151 a)
C122)±Fe11)±C151 a)
C11 a)±Fe11)±C151 a)
C16)±Fe11)±C151 a)
C152)±Fe11)±C151 a)
C122)±Fe11)±Fe12),
C121 a)±Fe11)±Fe12)
C11 a)±Fe11)±Fe12)
C16)±Fe11)±Fe12)
C152)±Fe11)±Fe12),
C151 a)±Fe11)±Fe12)
C141)±Fe12)±C142 a)
C142 a)±Fe12)±C13 a)
C14 a)±Fe12)±C16)
C13 a)±Fe12)±C16)
C141)±Fe12)±C152),
C142 a)±Fe12)±C151 a)
C142 a)±Fe12)±C152),
C141)±Fe12)±C151 a)
C13 a)±Fe12)±C152),
C13 a)±Fe12)±C151 a)
C16)±Fe12)±C152)
94.317)
92.118)
87.716)
166.814)
79.417)
167.617)
92.115)
74.86112)
95.612)
133.215)
110.014)
118.717)
49.512)
66.012)
42.512)
95.413)
99.316)
95.716)
165.014)
95.518)
91.017)
168.815)
173.514)
90.315)
88.417)
74.812)
C12)±Fe11)±C11)
C12)±Fe11)±C16)
95.71114)
91.35113)
C11)±Fe11)±C16) 169.62118)
C12)±Fe11)±C15) 85.11117)
C12 a)±Fe11)±C15) 173.95116)
C11)±Fe11)±C15)
C16)±Fe11)±C15)
90.35115)
82.66114)
C15)±Fe11)±C15 a) 95.013)
C12)±Fe11)±Fe12) 122.39110)
Fig.3 Crystal structures ofthe separated enantiomers 3 a
and 3 b; with atomic numbering scheme; thermal ellipsoids
are shown at the 50% probability level.
C11)±Fe11)±Fe12) 119.71113)
C16)±Fe11)±Fe12)
C15)±Fe11)±Fe12)
49.91112)
53.84114)
C14)±Fe12)±C14 a) 95.112)
The most dramatic effect on dealing with the enan-
tiomers 3 a and 3 b concerns the two possible angles
O15)±C15)±Fe11) which, from the unique 144° in the
average structure 3, become very different. Thus,
O15)±C15)±Fe11) is as close as 129° 1semibridging
mode) and O15)±C15)±Fe12) is as open as 160° 1termi-
nal mode). Also, the unusually long Fe±C15) average
distances of215 pm observed in 3 subdivides into a
short 1183 pm) and long 1249 pm) separation to the
first and the second iron atom, respectively. This se-
paration also causes the Fe±Fe distance to elongate
slightly from 254 pm in 3 into 258 pm in 3 a and 3 b.
With respect to 2 the Fe±C1F2)±Fe angle has increased
to 80° 13) or 81° 13 a, b). In Table 4 distances and an-
gles ofthe average and the separated structures are
compared. The structural parameters for the semibrid-
ging CO groups in 3 a or 3 b are consistent with those
C14)±Fe12)±C13)
C14)±Fe12)±C16)
95.87115)
91.72113)
C13)±Fe12)±C16) 168.75118)
C14)±Fe12)±C15) 84.73117)
C14 a)±Fe12)±C15) 174.42117)
C13)±Fe12)±C15)
C16)±Fe12)±C15)
89.69115)
82.72114)
C15)±Fe12)±C15 a) 94.913)
C14)±Fe12)±Fe11) 122.42111)
C13)±Fe12)±Fe11) 118.73114)
C16)±Fe12)±Fe11)
C15)±Fe12)±Fe11)
50.02111)
55.83114)
C16)±Fe12)±Fe11)
C152)±Fe12)±Fe11),
C151 a)±Fe12)±Fe11)
O11 a)±C11 a)±Fe11)
O13 a)±C13 a)±Fe12)
O121 a)±C121 a)±Fe11)
O142 a)±C142 a)±Fe12)
F11)±C16)±F12 a)
49.512)
42.718),
65.216)
O11)±C11)±Fe11) 175.514)
O13)±C13)±Fe12) 177.114)
O12)±C12)±Fe11) 177.813)
O14)±C14)±Fe12) 178.714)
175.316)
175.614)
177.9119)
178.0113)
102.114)
119.914),
117.214)
119.414),
117.713)
81.0117)
129.413)
158.113)
128.717)
160.014)
71.312)
F±C16)±F1a)
F±C16)±Fe12)
100.913)
119.312)
F11)±C16)±Fe12),
F12 a)±C16)±Fe12)
F11)±C16)±Fe11),
found
in
[Cp1CO)Fe1l-CO)1l-CF2)IrCl1CO)-
1PMe2Ph)]BF4 [14], and Fe21CO)7dipy [16] which con-
tain clearly defined semibridging CO groups. In the
latter, however, the groups connect different metal
atoms or metal atoms with different substituents.
Semibridging CO groups serve to remove electron
density from the more negatively charged part of the
molecule and are not found in dinuclear compounds
with two identical molecule halves [16].
F±C16)±Fe11)
119.06119)
80.07115)
F12 a)±C16)±Fe11)
Fe12)±C16)±Fe11)
O15 a)±C151 a)±Fe11),
O15 a)±C151 a)±Fe12)
O15)±C152)±Fe12),
O15)±C152)±Fe11)
Fe11)±C152)±Fe12),
Fe11)±C151 a)±Fe12)
Fe12)±C16)±Fe11)
O15)±C15)±Fe11) 144.114)
O15)±C15)±Fe12) 143.415)
Fe11)±C15)±Fe12)
72.33114)
72.312)
1864
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869

About Fe21CO)61l-CO)1l-CF2)2, Fe21CO)61l-CO)21l-CF2), and Fe21CO)61l-CF2)1PPh3)2
Table 5 Distances/pm and angles/° in 4
Fe11)±C13)
Fe11)±C11)
175.516)
179.615)
P11)±C1131)
184.012)
184.31119)
184.33117)
184.64119)
114.915)
115.915)
115.115)
115.115)
114.614)
114.715)
137.814)
139.115)
P12)±C1231)
P12)±C1221)
P12)±C1211)
C11)±O11)
C12)±O12)
C13)±O13)
C14)±O14)
C15)±O15)
C16)±O16)
C17)±F11)
C17)±F12)
Fe11)±C12)
Fe11)±C17)
Fe11)±P11)
Fe11)±Fe12)
Fe12)±C16)
Fe12)±C15)
Fe12)±C14)
Fe12)±C17)
Fe12)±P12)
P11)±C1121)
P11)±C1111)
179.714)
198.414)
230.89111)
270.65111)
179.314)
179.414)
179.914)
197.415)
223.53111)
183.512)
183.712)
C13)±Fe11)±C11) 90.412)
C13)±Fe11)±C12) 88.112)
C11)±Fe11)±C12) 173.09118)
C13)±Fe11)±C17) 95.012)
C11)±Fe11)±C17) 93.05118)
C12)±Fe11)±C17) 93.81117)
C16)±Fe12)±C15)
C16)±Fe12)±C14)
C15)±Fe12)±C14)
C16)±Fe12)±C17)
C15)±Fe12)±C17)
C14)±Fe12)±C17)
C16)±Fe12)±P12)
C15)±Fe12)±P12)
C14)±Fe12)±P12)
C17)±Fe12)±P12)
C16)±Fe12)±Fe11)
C15)±Fe12)±Fe11)
C14)±Fe12)±Fe11)
C17)±Fe12)±Fe11)
P12)±Fe12)±Fe11)
F11)±C17)±F12)
F11)±C17)±Fe12)
F12)±C17)±Fe12)
F11)±C17)±Fe11)
F12)±C17)±Fe11)
Fe12)±C17)±Fe11)
94.18117)
91.32119)
174.512)
154.37118)
87.47117)
87.29119)
102.56113)
88.05113)
91.39114)
103.06112)
107.44113)
87.73113)
90.01114)
47.00112)
149.9214)
101.213)
C13)±Fe11)±P11)
C11)±Fe11)±P11)
C12)±Fe11)±P11)
C17)±Fe11)±P11)
102.24115)
84.85114)
88.87112)
162.66113)
Fig.4 Crystal structure of 4; with atomic numbering
scheme; thermal ellipsoids are shown at the 50% probability
level.
C13)±Fe11)±Fe12) 141.43114)
C11)±Fe11)±Fe12) 95.23114)
C12)±Fe11)±Fe12) 90.13113)
C17)±Fe11)±Fe12) 46.71113)
P11)±Fe11)±Fe12) 116.2414)
O11)±C11)±Fe11) 173.614)
O12)±C12)±Fe11) 173.914)
O13)±C13)±Fe11) 177.314)
O14)±C14)±Fe12) 177.214)
O15)±C15)±Fe12) 176.613)
O16)±C16)±Fe12) 179.114)
3.4 Molecular structure of 4
The replacement oftwo terminal CO groups in 2 by
PPh3 affects the molecular structure dramatically as
depicted in Fig. 4. The remaining six CO groups are
now bonded in a terminal manner and the structure
can be viewed as derived from the structure proposed
for Os21CO)9 [15] which carries only one CO bridge.
The crystal structure of 4 is the first example of a
compound where a CF2 bridge is not accompanied by
an additional bridging group. The Fe±Fe distance has
118.613)
119.013)
116.213)
116.413)
86.29116)
happens, we consider not only difluormethylene but
involve also the parent methylene CH2 1see
Scheme 1). The analysis is closely related to that given
recently for bridging alkylindium1I) groups [21].
increased to
a value similar to that found in
[Fe21CO)8]2± [17] and in Fe21CO)81l-E) compounds
with a single bridging ligand 1E = C=CF2, 268 pm [18];
E = SiR2, 272 pm [19]). Probably for sterical reasons,
the two PPh3 groups are not arranged symmetrically;
one is in cis position and the other one in trans posi-
tion to the bridging CF2 ligand.
Scheme 2
The Fe11)±C17)±Fe12) angle at the CF2 ligand is
more open than in 2 and 3, and is similar to the re-
lated angle in compounds such as Fe21CO)7-
{1Ph2P)2CH2} with a single carbonyl bridge [20]. This
Density functional theory 1DFT) optimizations ap-
plying the B3LYP functionals [22] were performed
using the Gaussian98 package [23]. Effective core po-
tentials with Ne core in connection with the corre-
sponding valence basis set was employed for Fe [24],
whereas the standard 6-31G* basis set was used for
the other atoms. The optimizations were carried out
keeping certain symmetries for the structures. For
n = 0 and n = 3, this is D3h symmetry. For n = 1 and
n = 2, we used C2v symmetry. The optimized structures
for the cases n = 2 and n = 3 are determined as mini-
ma on the potential surface. In the cases n = 1, fre-
quency calculations yield an imaginary frequency of
94i and 108i for L = CF2 and L = CH2, respectively.
The vibration corresponding to the imaginary fre-
quency indicates an asymmetric distortion ofthe two
bridging carbonyl ligands. Geometry optimizations of
emphasizes the similar behavior ofCO and CF li-
gands.
2
In comparison to olefine complexes the structure of
4 can also be viewed as a coordination ofthe carbene
complex Ph3P1CO)3Fe=CF2 with its double bond in
the equatorial position ofthe 16 electron rfagment
Ph3P1CO)3Fe. Also the formulation of 4 as a bis-ferra-
cyclopropane could be possibly taken into account.
4
Theoretical Analysis
In this section we discuss the alteration ofthe electro-
nic, structural and bonding properties caused by the
stepwise replacement ofbridging CO by CF groups
2
in diiron carbonyls. To get a deeper insight into what
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869
1865

W. Petz, F. Weller, A. Barthel, C. Mealli, J. Reinhold
these systems in lower symmetry lead to singly
bridged structures. Otherwise, the experiments indi-
cate the semibridged structure 3 a, b for L = CF2 and a
symmetrically triple-bridged structure for L = CH2
[25]. In agreement with the strategy used in ref. [21],
we consider in the following only symmetrically triple-
bridged systems for the sake of systematic and direct
comparisons within the series from n = 0 to n = 3.
Thus, for n = 1 with L = CF2, we refer to the average
structure 3 instead ofthe disordered structure 3 a, b
1Table 4).
Tables 8 and 9, selected population values are col-
lected. The total populations at the bridges are sepa-
rated into a r and a p part 1given in brackets). Only
the contributions ofthe p atomic orbitals 1AOs) out
p
of the plane ofthe bridging ligands are involved in
the backbonding and are, therefore, included into the
p part. The latter is contributed from all of the AOs
ofthe bridging CO and CF ligands, whereas a single
2
carbon pp orbital is involved in case of CH2. The con-
tributions from the remainingAOs are all collected in
the r part, includingthe CO p orbitals lying in the
p
Selected structural parameters resulting from the
geometry optimizations are presented in Table 6 and
Table 7 for L = CF2 and L = CH2, respectively. The ex-
perimental values available are given in parenthesis.
The agreement appears generally good. Opposite to
the trend found for InR substitution in Fe21CO)9 [21],
there is a systematic decrease ofthe Fe±Fe distance in
going from n = 0 to n = 3. This effect is significantly
stronger for CH2 than for CF2. Since the metal-bridge
distances remain nearly constant, the angles at the
bridges decrease correspondingly. For both the CF2
and the CH2 substitutions, the calculated terminal
bond distances 1Fe±C as well as C±O) do not change
significantly within the series.
plane of the bridging ligands.
In both series, the population at the iron centers de-
creases in going from n = 0 to n = 3. The same applies
to the terminal carbonyl ligands as well. Consequently,
the total population ofthe bridging system increases
1as referred to the standard populations). However,
the distinction between the r and p populations ofthe
bridges indicate that their p acceptor capability is ex-
erted more eciently than the r donor one. Thus, in
going from n = 0 to n = 3, the bridging moiety re-
ceives more electron density from the metals than it
transfers to them. It turns out that these charge trans-
fer e€ects are stronger for CH2 than for CF2 substitu-
In order to find out the electronic reasons for the
structural peculiarities pointed out, a natural bond or-
der 1NBO) population analysis [26] was performed. In
Table 8 NBO Population Values for 1CO)3Fe1l-CO)3±n1l-
CF2)nFe1CO)3 1n = 0, 1, 2, 3)a,b,c)
n = 0
n = 1
n = 2
n = 3
Fe
CF2
16.577
16.563
24.112
16.549
24.099
16.533
24.075
Table 6 Selected Structural Parameters 1pm, deg) for
1CO)3Fe1l-CO)3±n1l-CF2)nFe1CO)3 1n = 0, 1, 2, 3)a,b)
ꢀ19.361/4.751) ꢀ19.369/4.730) ꢀ19.384/4.691)
13.972 13.951
CObr
13.985
n = 0
n = 1
n = 2
n = 3
248.2
ꢀ11.532/2.453) ꢀ11.518/2.454) ꢀ11.529/2.422)
Rbridges 41.955 52.056 62.149
ꢀ34.596/7.359) ꢀ42.397/9.659) ꢀ50.267/11.882) ꢀ58.152/14.073)
DRbridges ±0.045 0.056 0.149 0.225
ꢀ±1.404/1.359) ꢀ±1.603/1.659) ꢀ±1.733/1.882) ꢀ±1.848/2.073)
72.225
Fe±Fe
252.6 ꢀ252.3)
200.8 ꢀ201.6)
251.5 ꢀ254.1)
200.7 ꢀ215.3)
201.1 ꢀ197.5)
116.7 ꢀ109,1)
135.8 ꢀ135.5)
181.5/ꢀ182.1)
114.6/ꢀ113.0)
77.4 ꢀ72.3)
249.8 ꢀ246.9)
200.8 ꢀ199.5)
201.0 ꢀ199.1)
116.6 ꢀ115.0)
135.5 ꢀ135.2)
181.5/ꢀ182.8)
114.5/ꢀ112.9)
76.8 ꢀ76.5)
Fe±Cbr1O)
Fe±Cbr1F)
C±O
C±F
Fe±Cterm
201.4
117.0 ꢀ117.6)
COterm
DCOterm ±0.185
13.815
13.796/13.816 13.803/13.772 13.785
±0.204/±0.184 ±0.197/±0.228 ±0.215
135.2
181.4
114.5
181.6 ꢀ183.8)
114.7 ꢀ115.6)
79.0 ꢀ77.6)
C
term±Oterm
a)
b)
Averaged values for n = 1 and n = 2. r and p components 1see text)
Fe±Cbr1O)±Fe
Fe±Cbr1F)±Fe
c)
in parenthesis. The standard populations of the free fragments are: Fe:
16.000, CF2: 24.000 120.000/4.000), CO: 14.000 112.000/2.000)
77.6 ꢀ80.1)
77.0 ꢀ76.6)
76.1
a)
b)
Averaged values for n = 1 and n = 2.
Experimental values 1in par-
Table 9 NBO Population Values for 1CO)3Fe1l-CO)3±n1l-
CH2)nFe1CO)3 1n = 0, 1, 2, 3)a,b,c)
enthesis) from ref. [15] 1n = 0), from the average structure of 3 in Table 4
1n = 1) and from Table 3 1n = 2).
Table 7 Selected Structural Parameters 1pm, deg) for
1CO)3Fe1l-CO)3±n1l-CH2)nFe1CO)3 1n = 0, 1, 2, 3)a,b)
n = 0
n = 1
n = 2
n = 3
Fe
16.577
16.527
16.476
16.419
CH2
8.176
ꢀ7.238/0.938)
13.972
8.166
ꢀ7.248/0.918)
13.938
8.130
ꢀ7.271/0.859)
n = 0
n = 1
n = 2
n = 3
241.0
CObr
13.985
Fe±Fe
252.6 ꢀ252.3)
200.8 ꢀ201.6)
250.4 ꢀ250.4)
200.7 ꢀ201.5)
202.2 ꢀ201.5)
116.8
109.2
180.8 ꢀ182.3)
114.8
246.3
201.0
201.3
116.4
109.2
181.1/177.9
114.8
75.6
ꢀ11.532/2.453) ꢀ11.527/2.445) ꢀ11.552/2.386)
Fe±Cbr1O)
Fe±Cbr1H)
C±O
C±H
Fe±Cterm
Rbridges 41.955 36.120 30.270
24.390
200.7
ꢀ34.596/7.359) ꢀ30.292/5.828) ꢀ26.048/4.222) ꢀ21.813/2.577)
DRbridges ±0.045 0.120 0.270 0.390
ꢀ±1.404/1.359) ꢀ±1.708/1.828) ꢀ±1.952/2.222) ꢀ±2.187/2.577)
COterm
DCOterm ±0.185
117.0 ꢀ117.6)
109.2
180.2
114.9
181.6 ꢀ183.8)
114.7 ꢀ115.6)
79.0 ꢀ77.6)
13.815
13.801/13.811 13.808/13.773 13.795
±0.199/±0.189 ±0.192/±0.227 ±0.205
C
term±Oterm
Fe±Cbr1O)±Fe
Fe±Cbr1H)±Fe
77.2 ꢀ76.6)
76.5 ꢀ76.6)
75.4
73.8
a)
b)
Averaged values for n = 1 and n = 2. r and p components 1see text)
c)
a)
b)
in parenthesis. The standard populations of the free fragments are: Fe:
16.000, CH2: 8.000 18.000/0.000), CO: 14.000 112.000/2.000)
Averaged values for n = 1 and n = 2.
enthesis) from ref. [15] 1n = 0) and from ref. [25] 1n = 1).
Experimental values 1in par-
1866
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869

About Fe21CO)61l-CO)1l-CF2)2, Fe21CO)61l-CO)21l-CF2), and Fe21CO)61l-CF2)1PPh3)2
ents. The order both for the r donor and p acceptor
abilities is CO < CF2 < CH2. This is also consistent with
the p acceptor order ꢀCO < CF2) estimated from the IR
spectroscopy ꢀsee Section 2).
Fig.5 Energetic levels [a.u.] ofthe two highest occupied
molecular orbitals for 1CO)3Fe1l-CO)3-n1l-L)nFe1CO)3
1L = CH2, CF2; n = 0, 1, 2, 3)
Scheme 3
The relations above can also be inferred from the
relative energy position ofthe r donor and the p ac-
ceptor orbitals ofthe free ligand molecules. The lower
in energy is the MO with p accepting capabilities, the
larger is its functionality. In the case of CH2, the ac-
ceptor orbital is a pure pp orbital located at the car-
antibonding in character. This causes a repulsive inter-
action between the two metal centers, as it has been
pointed out before [27, 28]. In going from n = 0 to
n = 3, the increase ofthe acceptor ability ofthe brid-
ging moiety implies a decrease ofelectron density at
the metal centers. The latter fact is attributed to the
reduced iron contribution to the two highest occupied
MOs. Consequently, the metal±metal antibonding
character of these MOs becomes less effective. This
reduces the intermetallic repulsion, which leads to a
decrease ofthe Fe±Fe distance through the series.
In Figure 5, the energetic stabilization ofthe highest
occupied complex MOs is shown. The MOs which de-
scribe the backdonation to CH2 are the lowest in en-
ergy, those describing the interaction between CO p*
orbitals and the iron centers are the highest. This
agrees with the p acceptor order CO < CF2 < CH2 for
the individual ligands. The average ofthe respective
orbital energies steadily decreases through the series,
more so for the CH2 than for CF2 bridges.
bon atom. The p electron system ofCF is ofhetero-
2
allylic type 1see Scheme 3), the actual acceptor MO
having carbon-fluorine antibonding character. Be-
cause of the significant electronegativity difference
between carbon and fluorine atoms, the latter MO re-
ceives relatively small fluorine contributions and it is
only slightly destabilized compared to the pure carbon
pp orbital ofCH . The corresponding p* acceptor MO
2
ofCO lies highest in energy because it carries a more
significant contribution from the oxygen atom. Thus,
the energy order ofthe p acceptor orbitals ofthe
bridges is consistent with the different p acceptor be-
havior ascertained for the dimeric complexes.
Scheme 4
Scheme 5
Based on these findings, the decrease of the Fe±Fe
distance within the two series can be explained. We
consider the two highest occupied MOs ofthe symme-
trically bridged complexes 1e@ in D3h, a2 and b2 in C2v
symmetry). These orbitals are commonly considered
to be responsible for the main part of the acceptor in-
teraction between the bridging ligands and the metal
centers. As shown Scheme 4 these levels are at the
same time metal-bridge bonding and metal±metal
Based on the above considerations, the decrease of
the Fe±Fe distance could be ascribed mainly to the or-
bital interactions represented by the two highest occu-
pied MOs. However, it is also interesting to check if
qualitative MO theory and the fragment molecular or-
bital 1FMO) analysis can add anything to our under-
standing ofthe experimental and DFT trends. To the
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869
1867

W. Petz, F. Weller, A. Barthel, C. Mealli, J. Reinhold
pared by addition ofFe1CO) to a suspension ofCollman's
reagent in THF.
purpose, we have exploited the EHMO method and
graphic capabilities ofthe CACAO package [29]. For
instance, molecular orbital overlap population
1MOOP) diagrams are very useful to see how the
strength ofa certain bond 1e. g. Fe±Fe) builds up by
populating progressively the MOs up to the HOMOs.
From a comparison of the three different D3h systems
with n = 3 and a fixed Fe±Fe distance of 260 pm, it is
confirmed that the Fe±Fe bond strength decreases
dramatically when the HOMOs are populated. Thus,
the final Fe±Fe overlap population is close to zero,
i.e.: +0.08, +0.03 and ±0.03 for the CH2, CF2, and CO
systems, respectively. The latter values, however, can-
not be justified only in terms of the different Fe±Fe
repulsion implicit in the HOMOs. In fact, the given
trends exist already before the HOMOs are populated
and their origin is attributable to another critical
backdonation, namely that between the out-of-phase
5
Reaction of Na2[FeꢀCO)4] ´ 1.5 dioxane with Br2CF2:
A
suspension of2.00 g Na [Fe1CO)4] ´ 1.5 dioxane 15.80 mmol)
2
in about 40 ml n-pentane was cooled to 0 °C. To this mixture
excess Br2CF2 11 ml) was added under vigorous stirring.
After stirring for 1 h at 0 °C and an additional h at room
temperature the mixture was filtered and the solution evapo-
rated to dryness. The reaction tube with the residue was at-
tached to a vacuum line and all volatile material was sub-
limed at room temperature/10±3 torr into a Schlenck tube,
which was cooled with liquid air. After 3 h 250 mg of a mix-
ture of 2 and 3 were obtained. From 19F NMR spectroscopy
2 and 3 are present a 1 : 20 ratio, respectively; yield 22.5%.
The yields ofseveral runs varied between about 12 and 23%.
Crystals were grown by slow blowing argon over a surface
ofa n-pentane solution ofthe mixture of 2 and 3. The crys-
tals were separated mechanically.
Compound 2: Orange red crystals.
2
IR 1KBr, Nujol): 2097, 2064, 2041, 2021 1vs, CO), 1879 1s, l-CO), 1091 1s,
combination ofz orbitals and the a2@ combination of
l-CF2), 1046 1s, l-CF2), 739 1s), 720 1s), 694 1m), 627 1sh), 617 1m), 571
2
1s) cm±1
.
19F NMR 1toluene): d = 60.82, 21.67, JF,F = 30.13 Hz. Mass spec-
p acceptor orbitals at the bridges 1see Scheme 5). In
previous articles concerning Fe21CO)9 [28], it was sta-
ted that the latter interaction is critically small but es-
sential to admit the existence ofa direct metal-metal
bond. With the CF2 and CH2 ligands, the energy gap
between the two interacting a2@ FMOs is progressively
reduced. Accordingly, the overlap population between
the latter increases in the order CO < CF2 < CH2
10.13, 0.17 and 0.23, respectively). By contrast, the
trum 1m/z): 408 [M] 15%), 380 [M±CO] 17%), 341 14%), 329 17%), 324
[M±3 CO] 11%), 323 111%), 313 16%), 296 [M±4 CO] 19%), 285 14%),
268 [M±5 CO] 17%), 257 16%), 240 [M±6 CO] 126%), 212 [M±7 CO]
156%), 201 115%), 199 [1CO)2Fe2CF] 122%), 193 [Fe2CF2CF] 112%), 190
[1CO)Fe2CF2] 132%), 183 [1CO)CFe2CF] 112%), 174 [Fe21CF)2] 139%),
171 [1CO)Fe2CF] 111%), 162 [Fe2CF2] 138%), 154 [1CO)Fe1CF2)1CFH)]
111%), 145 [F2Fe1CF2)H] 125%), 142 114%), 135 [1CO)Fe1CF2)1CH)]
115%), 131 [Fe2F] 114%), 118 [Fe1CF)2] 153%), 112 [Fe1CO)2, Fe2]
119%), 106 [FeCF2] 118%), 99 [FeC1CF)] 132%), 87 [FeCF] 140%), 84
[Fe1CO)] 145%), 81 [CF2CF] 136%), 77 130%), 75 [FeF] 130%), 68 [FeC]
1100%).
overlap population ofa @ type almost vanishes in the
2
Compound 3: pale yellow crystals.
C9F2Fe2O8 1385.79); C 27.23 1Calcd. C, 28.02).
analogous system with three bridging InR groups [21]
due to the significant electropositivity of the indium
atoms. In this case, not only the metal-bridge bonding
is weakened but also the impossibility for the z2 orbi-
tals to discard part oftheir electron density into the
bridges keeps the repulsion between the metals high;
in this manner, the long Fe±Fe separation ofca.
300 pm can be rationalized.
In conclusion, the comparison between systems
which contain two terminal Fe1CO)3 fragments and
three bridging groups of different nature nicely con-
firms the idea that direct metal±metal bonding builds
up from several different components.
IR 1KBr, Nujol): 2066, 2039, 2016, 1950 sh 1vs, CO), 1034 1s, CF2), 997 1s,
CF2), 700 s, 598 m, 565 s cm±1
.
13C NMR 1Toluol-d8): d = 218.98 1t,
3
1JC,F = 383.63 Hz), 207.44 1t, CO, JC,F = 4.7 Hz). 19F NMR 1Toluol-d8):
d = 50.60 1s). Mass spectrum 1m/z): 386 [M] 12%), 370 [M±O] 10.5%), 358
[M±CO] 10.8%), 342 [M±CO±O] 12.1%), 330 [M±2 CO] 16%), 302 [M±
3 CO] 11.3%), 274 [M±4 CO] 10.8%), 258 [M±3 CO±O] 11.4%), 246 [M±
5 CO] 13.6%), 218 [M-6 CO or 1] 118%), 207 18.3%), 199 [1-F] 13.4%),
196 [Fe1CO)5] 12%), 190 [M±7 CO] 133%), 188 13.3%), 171 [1±CO±F]
15%), 168 [Fe1CO)4] 15%), 162 [1±2 CO or Fe2=CF2] 126%), 143
[1±2 CO±F] 111%), 140 [Fe1CO)3] 15%), 136 14), 134 [1±3 CO] 13%), 131
[1CO)2FeF] 17%), 127 15%), 124 [1CO)2FeC] 14%), 112 [Fe2 or Fe1CO)2]
116%), 106 [FeCF2] 14.3%), 96 [1CO)FeC] 14%), 87 [FeCF] 111%), 85
[F3CO] 124%), 84 [FeCO] 152%), 81 131%), 75 [FeF] 18%), 68 [FeC]
191%), 66 [F2CO] 16%), 56 [Fe] 1100%), 47 [FCO] 110%), 44 [CO2]
141%).
Reaction of Na2[Fe2ꢀCO)8] with Br2CF2: 3.23 g Na2-
[Fe21CO)8] ´ 4 THF 14.82 mmol) was treated in a silmilar
manner as described above with excess Br2CF2 in n-pentane;
yellow crystals of 3: 220 mg 10.57 mmol); yield 12%.
5
Experimental Section
General Considerations. All operations were carried out un-
der an argon atmosphere in dried and degassed solvents
using Schlenk techniques. IR spectra 1in Nujol) were run on
a Nicolet 510 spectrometer. 13C NMR spectra were recorded
on a Bruker AC 300 instrument using SiMe4 1d = 0.00 ppm)
as the external standard. 19F NMR and 31P NMR spectra
were run on a Bruker ARX 200 spectrometer. Mass spectra
were obtained with a CH 7 instrument from MAT 1Bremen).
Elemental analyses were performed by the analytical service
ofthe Fachbereich Chemie der UniversitaÈt Marburg 1Ger-
many). Comercial available Na2[Fe1CO)4] ´ 1.5 dioxane
1Collman's reagent, Aldrich) and Br2CF2 1ABCR) were
used without further purification. Na2[Fe21CO)8] was pre-
Fe2ꢀCO)6ꢀl-CF2)ꢀPPh3)2 ꢀ4): To a solution of410 mg
3
11.06 mmol) in about 40 ml n-pentane at room temperature
was poured a solution of560 mg PPh 12.14 mmol) in n-pen-
3
tane. Within a few minutes brown crystals of 4 separated.
The crystals of 4 were filtered through a G3 frit and dried in
vacuum; yield 815 mg 190%).
C43H30F2Fe2O6P2 1854.31); C 59.67 1calcd. 60.45), H 3.76
1calcd. 3.54)%.
IR 1Nujol): m1CO) = 2054 w, 2000 vs, 1987 vs, 1966 vs, 1933 vs, 1892 w,
1867 w, m1CF) = 997, 937 cm±1
.
13C NMR 1THF-d8): d = 216.18 1CO),
1
216.40 1CF2, JC,F = 246.5 Hz). 31P NMR 1THF-d8): d = 55.75 13JP, F
8.4 Hz). 19F NMR 1THF-d8): d = 41.50.
=
1868
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869

About Fe21CO)61l-CO)1l-CF2)2, Fe21CO)61l-CO)21l-CF2), and Fe21CO)61l-CF2)1PPh3)2
[13] D. L. Reger, M. D. Dukes, J. Organomet. Chem. 1978,
X-ray structural investigations
153, 67.
Single crystals ofthe three compounds were grown as de-
scribed above. A crystal of 2, forming plates, was mounted in
a glass capillary 0.3 mm Æ with the aid ofsome high vacuum
grease and was measured on a CAD4 Diffractometer 1Enraf-
Nonius) using CuKa radiation at room temperature. 3 and 4,
which form needles and platelets, respectively, were dipped
into oil, mounted on glass threads and measured on a IPDS
1Stoe) at ±70 °C 13) and ±60 °C 14). After data reduction all
data were subjected to psiscan correction. Solution ofthe
structures was effected by direct methods and improvement
ofthe models was achieved by successive reifnement cycles
and difference Fourier syntheses [30]. Disorder in 3 was effec-
tively accounted for by a model with split positions for all
atoms except O15) and C16). Details ofthe structures, struc-
ture solutions and convergence are shown in table 1. The
crystallographic data have been deposited as supplementary
publications no. CCDC-166454 12), 166455 13 a), 166456 13 b),
and 166457 14) at the Cambridge Crystallographic Data
Centre. Copies of the data can be ordered free from CCDC,
12 Union Road, Cambridge CB2 1EZ, UK 1e-mail:
deposit@ccdc.cam.ac.uk).
[14] A. M. Crespi, M. Sabat, D. F. Shriver, Inorg. Chem.
1988, 27, 812.
[15] F. A. Cotton, J. M. Troup, J. Chem. Soc. Dalton Trans.
1974, 800.
[16] a) F. A. Cotton, J. M. Troup, J. Am. Chem. Soc. 1974,
96, 1233; b) M. Dela Varga Ortiz, R. Costa, R. Reina,
personal communication.
[17] a) W. Petz, F. Weller, Z. Kristallogr. 1997, 212, 157; b)
H. B. Chin, M. B. Smith, R. D. Wilson, R. Bau, J. Am.
Chem. Soc. 1974, 96, 5285; c) H. Deng, S. G. Shore. In-
org. Chem. 1992, 31, 2289.
[18] W. Schulze, K. Seppelt, Inorg. Chem. 1988, 27, 3872.
[19] H. Tobita, I. Shinagawa, S. Ohnuki, M. Abe, H. Izumi,
H. Ogino, J. Organomet. Chem. 1994, 473, 187.
[20] F. A. Cotton, J. M. Troup, J. Am. Chem. Soc. 1974, 96,
4422.
[21] A. Barthel, C. Mealli, W. Uhl, J. Reinhold, Organome-
tallics 2001, 20, 786.
[22] 1a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648; 1b)
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
[23] 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. Dapp-
rich, 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. Clif-
ford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui,
K. Morokuma, D. K. Malick, A. D. Rabuck, K. Ragha-
vachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Ko-
maromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gon-
zalez, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez,
M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian
98, Revision A.3; Gaussian Inc.: Pittsburgh PA, 1998.
[24] P. J. Hay, W. R. J. Wadt, J. Chem. Phys. 1985, 82, 299.
[25] 1a) C. E. Sumner, P. E. Riley, R. E. Davis, R. Pettit, J.
Am. Chem. Soc. 1980, 102, 1752; 1b) C. E. Summer Jr.,
J. A. Collier, R. Pettit, Organometallics 1982, 1, 1350.
[26] E. D. Glendening, A. E. Reed, J. E. Carpenter, F. Wein-
hold, NBO Version 3.1.
We thank the Deutsche Forschungsgemeinschaft, the Fonds
der Chemischen Industrie and the Naturwissenschaftlich-
Theoretisches Zentrum der UniversitaÈt Leipzig for financial
support. We are also grateful to the Max-Planck-Society,
Munich, Germany, for financial support. A. B. thanks the
Deutsche Forschungsgemeinschaft for support through the
Graduate College ªPhysical Chemistry ofInterfacesº.
References
[1] P. J. Brothers, W. R. Roper, Chem. Rev. 1988, 88, 1293.
[2] W. Schulze, H. Hartl, K. Seppelt, Angew. Chem. 1986,
98, 189; Angew. Chem. Int. Ed. Engl. 1986, 25, 185.
[3] a) F. Seel, G. V. RoÈschenthaler, Z. Anorg. Allg. Chem.
1971, 386, 297; b) F. Seel, G. V. RoÈschenthaler, Angew.
Chem. 1970, 82, 182; Angew. Chem. Int. Ed. Engl. 1970,
9, 166.
[4] a) S. Koda, Chem. Phys. 1982, 66, 383; b) E. A. Carter,
W. A. Goddard, III, J Phys. Chem. 1988, 88, 1752; c)
F. X. Powell, D. R. Lide, Jr., J. Chem. Phys. 1966, 45,
1067. c) C. W. Mathews, J. Chem. Phys. 1966, 45, 1068.
[5] D. G. Leopold, K. K. Murray, W. C. Lineberger, J.
Chem. Phys. 1984, 81, 1048.
[6] P. J. Brothers, A. K. Burrell, G. R. Clark, C. E. F. Rick-
ard, W. R. Roper, J. Organomet. Chem. 1990, 394, 615.
[7] H. P. Fritz, W. Kornrumpf, Z. Naturforsch. 1981, 36 b,
1375.
[8] M. S. Novikov, A. F. Khlebnikov, A. Krebs, R. R. Kosti-
kov, Eur. J. Org. Chem. 1998, 133.
[27] R. H. Summerville, R. Hoffmann, J. Am. Chem. Soc.
1979, 101, 3821.
[28] 1a) C. Mealli, D. M. Proserpio, J. Chem. Educ. 1990,
386, 203; 1b) J. Reinhold, E. Hunstock, C. Mealli, New J.
Chem. 1994, 18, 465; 1c) E. Hunstock, C. Mealli,
M. J. Calhorda, J. Reinhold, Inorg. Chem. 1999, 38,
5053.
[29] 1a) C. Mealli, D. M. Proserpio, J. Chem. Educ. 1990, 67,
399; 1b) C. Mealli, A. Ienco, D. M. Prosperio, Book of
Abstracts of the XXXIII ICCC; Florence, 1998, 510.
[30] G. M. Sheldrick, SHELXS-97 and SHELXL-97, Pro-
grams for the Solution and Refinement of Crystal
Structures, GoÈttingen, 1997.
[9] W. Petz, F. Weller. Book of Abstracts of the XXXIII
FECHEM ICCC; Lisboa, 1999, 510.
[10] B. NeumuÈller, W. Petz, Organometallics 2001, 20, 163.
[11] C. KoÈtting, W. Sander, M. Senzlober, H. BuÈrger, Chem.
Eur. J. 1998, 4, 1611.
[12] J. D. Koola, D. M. Roddick, Organometallics 1991, 10,
591.
Z. Anorg. Allg. Chem. 2001, 627, 1859±1869
1869