Bonding in `closed,' open, and half-open ferrocenes: New insight from structural and Mo?ssbauer spectroscopic studies

Article abstract of DOI:10.1016/S0022-328X(01)00901-9

A 17 electron edge-bridged open metallocene, bis(6,6-dimethylcyclohexadienyl)iron cation, which had previously been observed electrochemically, has been isolated and characterized. M?ssbauer spectroscopic data have been obtained for this compound, and for various neutral 18 electron open, half-open, and `closed' ferrocenes, and are in accord with the previous indications that much greater metal-ligand orbital mixing occurs for pentadienyl, as opposed to cyclopentadienyl, ligands. Pertinent structural data have also been obtained for Fe(C₅H₅)(2,4-C₇H₁₁) and Fe(c-C₇H₉)₂ (C₇H₁₁=dimethylpentadienyl; c-C₇H₉=cycloheptadienyl) in order to aid in comparisons between metal-pentadienyl and metal-cyclopentadienyl bonding.

Full text of DOI:10.1016/S0022-328X(01)00901-9

Journal of Organometallic Chemistry 637–639 (2001) 172–181
www.elsevier.com/locate/jorganchem
Bonding in ‘closed,’ open, and half-open ferrocenes: new insight
from structural and Mo¨ssbauer spectroscopic studies
Rehan Basta a, David R. Wilson a,1, Huairang Ma a, Atta M. Arif a, Rolfe H. Herber b,*,
Richard D. Ernst a,*
a Department of Chemistry, Uni6ersity of Utah, 315 S, 1400 E, Room 2020, Salt Lake City, UT 84112-0850, USA
b Racah Institute of Physics, The Hebrew Uni6ersity of Jerusalem, 91904 Jerusalem, Israel
Received 20 December 2000; received in revised form 22 February 2001; accepted 22 February 2001
Abstract
A 17 electron edge-bridged open metallocene, bis(6,6-dimethylcyclohexadienyl)iron cation, which had previously been observed
electrochemically, has been isolated and characterized. Mo¨ssbauer spectroscopic data have been obtained for this compound, and
for various neutral 18 electron open, half-open, and ‘closed’ ferrocenes, and are in accord with the previous indications that much
greater metal–ligand orbital mixing occurs for pentadienyl, as opposed to cyclopentadienyl, ligands. Pertinent structural data have
also been obtained for Fe(C5H5)(2,4-C7H11) and Fe(c-C7H9)2 (C7H11=dimethylpentadienyl; c-C7H9=cycloheptadienyl) in order
to aid in comparisons between metal–pentadienyl and metal–cyclopentadienyl bonding. © 2001 Elsevier Science B.V. All rights
reserved.
Keywords: Metallocenes; Mo¨ssbauer spectroscopy; Ferrocene half-sandwich
1. Introduction
award of the 1973 Nobel Prize to Professors Fischer
and Wilkinson [5].
The discovery of ferrocene (1) and the elucidation of
its structure was one of those rare events that revolu-
tionized the entire field of organometallic chemistry,
and after a period of 50 years, has led to many practical
applications for metal cyclopentadienyl (Cp) complexes,
including very large scale polymerization processes [1],
applications in organic synthesis [2], as well as more
specialized applications (e.g. low dimensional conduc-
tors and superconductors, redox tunable auxiliaries in
molecular composites, luminescent, fluorescent, and
photoreflective materials [3], in addition to the use of
ferrocenes themselves as burn enhancers [4]). The im-
portance of the discovery of ferrocene, and related
p-complexes, was recognized appropriately by the
Much of the interest in ferrocene derives from its
high thermal stability, and the strong bonding generally
found between transition metal centers and the Cp
ligand, which can be exemplified by the existence of a
whole series of thermally stable metallocenes from the
15 electron vanadocene to the 20 electron nickelocene
[6]. As a result, Cp is often described as a ‘stabilizing
ligand.’ Recently it has become evident that open pen-
tadienyl ligands, e.g. C5H7 (2) and 2,4-C7H11 (3), can be
even more strongly bound ligands than Cp [7], and a
* Corresponding authors. Tel.: +1-801-581-6681; fax: +1-801-
581-8433 (R.D.E.).
E-mail
addresses:
herber@vms.huji.ac.il
(R.H.
Herber),
ernst@chem.utah.edu (R.D. Ernst).
1 Present address. The Dow Chemical Company, 1776 Building,
Chemical Sciences Capability, Midland, MI 48674, USA.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)00901-9

R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
173
variety of open (4) and half-open (5) metallocenes can be
isolated [8]. Since the p-molecular orbitals of pentadienyl
anions in the U conformation are quite analogous to
those of Cp [9], it has become of interest to probe the
electronic natures of metal–pentadienyl interactions rel-
ative to those of their more familiar Cp counterparts. As
Mo¨ssbauer spectroscopy yields significant information
pertaining to the populations of metal-based orbitals, it
provides substantial, important information on the rela-
tive extents of metal–ligand bonding interactions,
thereby aiding in the efforts to gain a better understand-
ing of open versus ‘closed’ pentadienyl ligands [10].
Moreover, in a historical context it is appropriate to note
that ferrocene was the first organometallic compound to
which 57Fe Mo¨ssbauer spectroscopy was applied in the
elucidation of the electronic environment of the metal
atom. Herein we report Mo¨ssbauer spectroscopic data
for various open, ‘closed,’ and half-open ferrocenes, as
well as pertinent structural studies to assist in spectro-
scopic correlations and in bonding comparisons.
idation of the Fe(6,6-dmch)2 complex had been reported
by DiMauro and Wolczanski [15], and subsequent stud-
ies by LeSuer and Geiger have revealed the cation to be
fairly stable in solution even at room temperature [16].
As a result, we have found it possible to isolate the cation
as a tetraphenylborate salt. The stability of this cation
is quite remarkable compared to other 17 electron open
ferrocene cations [17], which could only be reversibly
generated electrochemically using rapid scan rates at low
temperatures. Thus, Fe(6,6-dmch)+2 actually appears to
be the first clear example of an isolable 17 electron open
metallocene, although DiMauro and Wolczanski have
provided some evidence for the existence of Mn(6,6-
dmch)2 [13].
Although structural comparisons between ferrocene
and various open ferrocenes have been made [8], such
comparisons suffer from complications due to the differ-
ing symmetries and steric environments. These complica-
tions are nicely removed, however, in a half-open
ferrocene complex such as Fe(C5H5)(2,4-C7H11), which
then allows for direct and more valid comparisons to be
made between the two ligands. As a result, a structural
determination of this complex was undertaken. The solid
state structure of Fe(C5H5)(2,4-C7H11) is consistent with
the spectroscopic data (see Section 4), and is presented
in Fig. 1, while pertinent bonding parameters are listed
in Table 1. The molecule lies on a crystallographic mirror
plane, resulting in an ideally eclipsed conformation (9)
as opposed to the staggered alternative (10). There is,
however, significant thermal motion associated with the
C5H5 ligand, and it is therefore possible that the ground
state conformation might actually not quite be eclipsed,
leading to a superposition of two slightly staggered
images in the solid state. As in the case of ferrocene, it
must be expected that any difference in energy for these
conformations should be rather small [18].
2. Results and discussion
2.1. Synthetic and structural data
The syntheses of the pentadienyl, 6,6-dmch (6), c-C7H9
(7), and c-C8H11 (8) compounds studied here have
generally been described previously [11], although in the
case of Fe(c-C7H9)2 [12], a new and more straightforward
route utilizing the cycloheptadienyl anion was utilized,
similar to the routes previously developed for its 6,6-
dmch [13] and c-C8H11 [14] analogues (dmch=dimethyl-
cyclohexadienyl; c-C7H9=cycloheptadienyl; c-C8H11=
cyclooctadienyl). Notably, reversible electrochemical ox-

174
R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
C5H5 in 9). Additionally, in 11 there could be greater
electronic competition between the two 2,4-C7H11 lig-
ands, which generally appear to be significantly
stronger accepting ligands than C5H5 [10].
For the C5H5 ligand in 9, one finds an average FeꢀC
,
distance of 2.063(5) A, indistinguishable from that of
,
2.064(3) A found for ferrocene by gas phase electron
diffraction [19]. However, it is possible that the value in
9 could suffer a slight systematic shortening due to
librational effects. For the pentadienyl ligand, there are
modest differences in the FeꢀC(1,2,3) distances, follow-
ing a similar trend as in Fe(2,4-C7H11)2 (11), in which
the formally uncharged (2,4) positions are closest to Fe,
while the terminal (1,5) positions are farthest. The
overall average FeꢀC distance for the 2,4-C7H11 ligand
,
in 9 is 2.060(10,1) A [20], significantly shorter than the
,
average of 2.089 A for Fe(2,4-C7H11)2 [21]. The length-
ening of the FeꢀC bonds in 11 has been attributed to
the appreciable steric crowding that results from the
long C1ꢀC1% separation for the 2,4-C7H11 ligand (2.746
,
,
A for 9 and 2.785(5) A for 11), which then requires a
much closer approach of the ligand plane to the metal
Several additional observations can also be made
concerning the ligand planes. On average, the C5H5
substituents tilt down toward the iron center by 2.4°,
close to the value of 3.7(9)° observed for ferrocene [19].
The reason for the tilt has been proposed to be an
attempt of the ligand to improve orbital overlap with
the relatively small iron center (12). On this basis one
would then expect to observe even greater tilts for the
2,4-C7H11 substituents, and indeed this is observed. The
values for the ‘downward oriented’ C(1–3) substituents
are, respectively, 17.6, 12.3, and 6.8°. For the internal
(endo) substituents on C(1), the tilt is 46.2° in the
opposite direction. The much greater tilt here could
arise from an attempt to rehybridize the orbitals on C1
more toward sp3, and to an attempt to reduce steric
interactions between the two opposing substituents
(Hn/Hn in 2 or 3) [21,22]. The greater upward tilt of Hn
would then either initiate or at least be accompanied by
a significant downward tilt of Hx, and one then ob-
serves lesser degrees of tilting as one moves to C2 and
C3. Finally, there is a tilt of 6.6(1)° between the two
ligand planes, which can be compared to the value of
15.0° for 11.
,
,
center (cf. 1.420 A for 2,4-C7H11 versus 1.687 A for
Fig. 1. Perspective view of the half-open ferrocene, Fe(C5H5)(2,4-
C7H11).
Table 1
Bonding parameters for Fe(C5H5)(2,4-C7H11
)
,
Bond length (A)
FeꢀC(1)
FeꢀC(2)
FeꢀC(3)
FeꢀC(5)
FeꢀC(6)
FeꢀC(7)
2.076(1)
2.045(1)
2.059(2)
2.059(2)
2.057(2)
2.071(2)
C(1)ꢀC(2)
C(2)ꢀC(3)
C(2)ꢀC(4)
C(5)ꢀC(6)
C(6)ꢀC(7)
C(7)ꢀC(7%)
1.424(2)
1.425(2)
1.511(2)
1.406(3)
1.398(3)
1.39(4)
Bond angles (°)
C(1)ꢀC(2)ꢀC(3)
C(1)ꢀC(2)ꢀC(4)
C(2)ꢀC(3)ꢀC(2%)
C(3)ꢀC(2)ꢀC(4)
In order to learn more about the bonding in the
edge-bridged analogues, a single crystal diffraction
study of Fe(c-C7H9)2 was carried out. Quite unusually,
but not without precedent, four independent molecules
121.37(13)
120.63(13)
125.37(17)
117.54(12)
C(5)ꢀC(6)ꢀC(7)
C(6)ꢀC(7)ꢀC(7%)
C(6)ꢀC(5)ꢀC(6%)
108.2(2)
108.3(1)
106.9(2)

R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
175
Table 2
third molecule (carbon atoms designated with B’s), for
which a number of parameters (e.g. FeꢀC(1,4,5))
showed clear deviations from the values found in the
other three molecules. That the deviating parameters
could be considered less reliable can be supported by an
examination of thermal parameters for the various
C(1–5) atoms. The U values for these atoms in the
three similar molecules range from 28(1) to 50(1), while
in the unique molecule these range from 42(1) to 70(1).
Therefore, in the subsequent discussion, the average
values quoted for Fe(c-C7H9)2 will be derived only from
the three similar molecules.
To begin with, the molecule is found to adopt a
nearly gauche-eclipsed conformation (13, Fig. 2), hav-
ing a conformation angle of 53.9°, close to the ideal
value of 60° relative to a syn-eclipsed structure (0°, 14).
As the conformation angles for non-edge bridged open
ferrocenes have been found to be much closer to 60°
[21,23], the small deviation here may be attributed to
the presence of the C2H4 bridge, which leads to a
Bonding parameters for Fe(c-C7H9)2
A
B
C
,
Bond length (A)
FeꢀC(1)
FeꢀC(2)
FeꢀC(3)
FeꢀC(4)
FeꢀC(5)
FeꢀC(8)
FeꢀC(9)
FeꢀC(10)
FeꢀC(11)
FeꢀC(12)
C(1)ꢀC(2)
C(2)ꢀC(3)
C(3)ꢀC(4)
C(4)ꢀC(5)
C(8)ꢀC(9)
C(9)ꢀC(10)
C(10)ꢀC(11)
C(11)ꢀC(12)
2.071(2)
2.038(2)
2.064(2)
2.038(2)
2.126(2)
2.070(2)
2.041(2)
2.071(2)
2.039(2)
2.126(2)
1.418(3)
1.416(4)
1.395(4)
1.393(3)
1.430(3)
1.408(3)
1.403(3)
1.398(3)
2.077(2)
2.041(2)
2.067(2)
2.036(2)
2.128(2)
2.073(2)
2.044(2)
2.066(2)
2.038(2)
2.128(2)
1.413(3)
1.419(3)
1.404(3)
1.404(3)
1.416(3)
1.413(4)
1.396(3)
1.407(3)
2.057(2)
2.039(2)
2.066(2)
2.027(2)
2.113(2)
2.052(2)
2.042(2)
2.072(2)
2.033(2)
2.109(2)
1.416(4)
1.397(5)
1.378(5)
1.390(4)
1.411(3)
1.410(3)
1.386(4)
1.389(4)
2.072(2)
2.037(2)
2.072(2)
2.038(2)
2.126(2)
2.074(2)
2.044(2)
2.065(2)
2.037(2)
2.128(2)
1.421(3)
1.419(3)
1.408(3)
1.404(3)
1.418(3)
1.417(3)
1.407(3)
1.402(3)
Bond angles (°)
C(1)ꢀC(2)ꢀC(3)
C(2)ꢀC(3)ꢀC(4)
C(3)ꢀC(4)ꢀC(5)
C(8)ꢀC(9)ꢀC(10)
C(9)ꢀC(10)ꢀC(11)
,
,
relatively short C(1–5) separation, 2.715 A (cf. 2.785 A
for Fe(2,4-C7H11)2, conformation angle 59.7°, and 2.706
,
126.6(2)
122.8(2)
119.2(2)
126.9(2)
122.8(2)
126.4(2)
122.7(2)
119.8(2)
126.5(2)
123.0(2)
119.6(2)
125.9(2)
123.2(2)
120.0(3)
125.9(2)
123.0(2)
119.6(2)
126.7(2)
122.4(2)
119.5(2)
126.2(2)
122.9(2)
119.2(2)
A for Fe(2,3,4-C8H13)2, conformation angle 55.1°). The
relative FeꢀC bond distances (FeꢀC(1,8), C(2,9),
C(3,10), C(4,11), C(5,12), respectively: 2.068(9),
C(10)ꢀC(11)ꢀC(12) 119.1(2)
,
2.041(3), 2.068(3), 2.036(4), and 2.123(8) A) follow a
pattern similar to those observed for other open fer-
rocenes, with the shortest distances being found for the
formally uncharged carbon atoms in the 2 and 4 posi-
tions [21,23].
Fig. 2. Solid state structure of Fe(c-C7H9)2. While only one of the
four independent molecules is shown, the parameters of the other
three are similar.
The dienyl fragments are not quite planar, with the
carbon atom deviations ranging from an average of
,
were contained in the triclinic unit cell. However, a
close examination revealed that the structural parame-
ters for these molecules were nearly indistinguishable
(Table 2), and thus solid state packing effects appear to
be relatively minimal. The one exception was for the
−0.064 A for the C(5,12) atoms to an average of 0.086
,
A for the C(4,11) atoms, a positive deviation reflecting
a tilt toward the Fe center. However, the average
C(1–4) and C(8–11) torsion angle is 1.8°, while that for
the C(2–5) and C(9–12) fragments is 13.1°, reflecting a

176
R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
deviation of the C(5,12) atoms out of the plane, away
from the iron center and the opposing dienyl ligand.
This is consistent with the longest FeꢀC bonds involv-
ing C(5) and C(12), and may in part reflect the fact that
C(5,12) engage in near-eclipsing interactions with the
other dienyl ligand whereas C(1,8) reside by the other
ligands’ open edges. Related to these parameters is the
tilt between the two dienyl ligands, which averages
13.9°.
However, the near planarity of the C(1–4) and C(8–
11) fragments actually contrasts with the situations for
other open ferrocenes, including Fe(2,4-C7H11),
Fe(2,3,4-C8H13)2, and Fe(c-C8H11)2. It seems possible
that this could be brought about by the particular edge
bridge in this case, which is notably asymmetric. Thus,
the C3ꢀC2ꢀC1ꢀC7 and C10ꢀC9ꢀC8ꢀC14 torsion angles
average 53.8° versus 46.8° for their C3ꢀC4ꢀC5ꢀC6 and
C15ꢀC11ꢀC12ꢀC13 counterparts. This leads to an even
larger difference in the C2ꢀC1ꢀC7ꢀC6 and C9ꢀ
C8ꢀC14ꢀC13 versus C4ꢀC5ꢀC6ꢀC7 and C11ꢀC12ꢀ
C13ꢀC14 torsion angles, 35.9° versus 91.0°, respec-
tively. This asymmetry is likely favored by the avoid-
ance of eclipsing orientations of the hydrogen atoms on
C6 and C7 [24]. Although there are additional parame-
ters which deserve discussion, this must be postponed
until publication of the structural details of Fe(6,6-
dmch)2.
dependencies of these parameters, as well as the an-
isotropy of the msav (Gol’danskii–Karyagin effect) can
be elucidated from a series of spectra at different tem-
peratures. The values of these parameters and their
significance will be discussed for the individual com-
pounds in the following section.
2.2.1. Bis(cycloheptadienyl)iron
A typical resonance spectrum of Fe(c-C7H9)2 is
shown in Fig. 3, and the IS and QS parameters at 90 K
are summarized in Table 3. The IS at 90 K is signifi-
cantly smaller than that previously reported [25] for
neutral ferrocenoid iron complexes, as well as for the
bis pentadienyl complexes previously discussed by Ernst
et al. [11]. The smaller IS implies a larger s-electron
density at the metal atom nucleus, suggesting that a
larger ligand–metal interaction is obtained in the hep-
tadienyl compound. The temperature dependence of the
IS is not well represented by a linear regression, and the
data show significant curvature over the temperature
range 905T5320 K; hence it is not possible to extract
a meaningful value of the effective vibrating mass, Meff
[26], from these results, in contrast to data to be
discussed below. The QS shows a modest decrease with
increasing temperature, consistent with the effects of
thermal expansion as discussed earlier [27]. The temper-
ature dependence of fa — as extracted from the T-de-
pendent area under the resonance curve — is
summarized graphically in Fig. 4 which evidences a
reasonably linear behavior at low temperatures, but
departs from this linearity at T\ ꢀ270 K. Since this
compound has a reported [12,28] TMP of 36491 K,
these observations imply a large increase in the msav of
the iron atom (on warming) at temperatures well below
the melting point. This type of behavior has been
observed [29] in other ferrocene related solids, and can
be associated with the onset of ring motion in these
matrices. A differential scanning calorimetry study
shows a sharp endotherm (on warming) at 326 K, and
a smaller one at 330 K, as well as an exotherm (on
cooling) at 274 K, consistent with the Mo¨ssbauer data
of Fig. 4.
2.2. Mo¨ssbauer spectroscopy
The gamma ray resonance spectra of the neutral
compounds consist of two well separated components.
From such spectra it is possible to extract the isomer
shift (IS), related to the electron density around the
metal atom nucleus, the quadrupole splitting (QS),
related to the symmetry of the charge distribution
around the metal atom, and the recoil-free fraction ( fa),
related to the mean-square-amplitude of vibration
(msav) of the metal atom. In addition, the temperature
2.2.2. Bis(6,6%-dimethylcyclohexadienyl)iron
The synthesis and plausible structure based on IR
inferences of Fe(6,6-dmch)2 have been discussed by
DiMauro and Wolczanski [13], who proposed an anti-
eclipsed C2h geometry in accord with calculations by
Gleiter and coworkers [30]. The IS and QS parameters
for this compound are included in Table 3 in addition
to the other parameters extracted from the Mo¨ssbauer
data, and are unexceptional. The IS is nearly linear
over the whole temperature range (90–305 K) and is
shown in Fig. 5. Above 150 K, the slope, d(IS)/dT, is
−(4.2090.09)×10−4 mm s−1 K−1, with a correla-
tion coefficient of 0.997 for 17 data points, from which
Fig. 3. Mo¨ssbauer spectrum of Fe(c-C7H9)2 at 110 K. The velocity
scale is defined relative to the centroid of a a-Fe spectrum at room
temperature.

R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
177
Table 3
Summary of Mo¨ssbauer data for the compounds discussed in the text
Compound
IS(90) (mm s−1
)
QS(90) (mm s−1
)
−d(IS)/dT
Meff (Da)
−dln(A)/dT
(K−1×103)
qM (K)
(mm s−1 K×104)
Fe(C5H5)2
Fe(C5Me5)2
Fe(C5H7)2
Fe(2-C6H9)2
Fe(3-C6H9)2
Fe(2,3-C7H11 2
Fe(2,4-C7H11 2
Fe(2,3,4-C8H13 2
Fe(6,6-dmch)2
Fe(c-C7H9)2
Fe(c-C8H11 2
Fe(C5H5)(2,4-C7H11
Fe(6,6-dmch)+2
Fe(C5H5)+2
0.531(3)
0.492(3)
0.475(10)
0.458(5)
0.478(5)
0.457(5)
0.494(7)
0.465(7)
0.488(1)
0.409(3)
0.452(1)
0.470(5)
2.419(1)
2.473(17)
1.402(15)
1.206(18)
1.255(41)
1.261(6)
1.516(10)
1.404(5)
1.143(1)
1.207(1)
1.535(1)
1.946(10)
0.892 f, 0.878 g
0.1–0.6
3.29(7)
3.17(13)
3.52
3.84
126
131
8.50(42)
7.97(40)
5.13(22)
7.29(35)
5.49(41)
7.94(46)
5.48(25)
7.22(20)
6.52(36) d
5.89(28)
7.20(26)
85
86
113(7)
99(7)
117(11)
73.5
a
118(9)
108(9)
103(8)
182(15)
116(9)
104(5)
116 b, 95 c
a
a
4.06
2.30
3.60
4.04
a
a
)
)
111
)
102
3.58(11) b, 4.38(15) c
112 e
Not linear
4.13(15)
3.09
)
100
111(8)
103
103
)
5.41(35)
h
0.578 f, 0.457 g
ꢀ0.4
6.4
a Data from Ref. [11], corrected to 90 K.
b 90–298 K.
c 150–298 K.
d 90–190 K.
e From high T slopes.
f At 110 K.
g At 300 K.
h Spin–spin relaxation is fast even at 90 K.
2.2.3. Bis(6,6%-dimethylcyclohexadienyl)iron cation
Meff is calculated to be 99 Da. The difference between
Meff and the ‘bare’ atom mass of 57 Da is due to the
covalency of the Fe–ligand interaction. In this context,
it should be noted that Brougham et al. [31], in their
NMR and Mo¨ssbauer study of cyclohexadienyl and
cycloheptadienyl complexes of the Fe(CO)3 fragment,
report Meff=109 and 117 Da, respectively, although
(as noted by these authors) these values are based only
on 78 and 295 K data — see footnotes b and c of
Table 3.
The temperature dependence of the area under the
resonance curve, d(ln A)/dT, is shown graphically in
Fig. 6. At low temperatures this dependence is nearly
linear, but deviates dramatically from this dependence
at temperatures above ꢀ270 K, and becomes experi-
mentally unobservable at temperatures above ꢀ305 K.
This T-dependence is completely reversible since the
warming regime data (open circles) and cooling regime
data (filled circles) are essentially superimposable. As
Brougham et al. have noted from their solid state NMR
data of the tri(carbonyl)iron complexes, there is a phase
transition between 253 and 263 K, and 329 and 341 K,
respectively, for the two complexes studied by them.
They ascribe these transitions to the onset of ring
rotation, and it is probable that the recoil-free fraction
behavior shown in Fig. 6 is, in fact, a reflection of a
similar kind of ring motion, which dramatically in-
creases the msav of the iron atom in Fe(6,6-dmch)2, and
hence leads to a marked decrease in the resonance
effect above ꢀ305 K.
A typical Mo¨ssbauer spectrum of Fe(6,6-dmch)+2 at
110 K is shown in Fig. 7 and hyperfine parameters at
90 K are included in Table 3. The spectrum is charac-
terized by two broad resonance lines of unequal inten-
sity. Such spectra arise from relaxation effects involving
the paramagnetic metal center. In contrast to other
ferrocenium complexes reported [32] in the literature,
which show a strong temperature dependence of the
hyperfine and relaxation parameters, in the case of
Fe(6,6-dmch)+2 the spectra are insensitive to T over the
range 905T5300 K. This observation leads to the
conclusion that the paramagnetic relaxation is of the
spin–spin type (rather than spin–lattice relaxation as
was inferred for the ferrocenium complexes reported
Fig. 4. Logarithm of the area under the resonance curve for Fe(c-
C7H9)2 as a function of temperature.

178
R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
earlier [32b]), and that this relaxation is fast on the
Mo¨ssbauer time scale, even at 90 K.
2.2.4. Bis(cyclooctadienyl)iron
The hyperfine parameters of Fe(c-C8H11)2 at 90 K
and the parameters derived therefrom are included in
Table 3 from which it is noted that the IS(90) is
intermediate between those of its six- and seven-mem-
bered ring analogues, while the QS(90) is significantly
larger than that reported for the other open sandwich
compounds of this study. Both the IS and ln(A) values
are reasonably well represented by a linear regression
over the temperature range 905T5300 K, permitting
a calculation of Meff=100 Da, and a lattice tempera-
ture [26], qM=103 K. The latter value is comparable to
that calculated for Fe(6,6-dmch)2 (vide supra) and
other related iron organometallics, and justifies the use
of the ‘high temperature’ approximation in the calcula-
tion of Meff. The QS parameter has a negative slope
over the above temperature interval, as expected from
covalent solid thermal expansion considerations, and
none of the present data [IS(T), QS(T), and ln(A(T))]
indicate evidence for any phase transitions over this
interval. It is interesting to note in this context that in
the iron tricarbonyl complexes reported by Brougham
et al. [31], the onset of the ring re-orientation in the
heptadienyl complex occurs some 76–78° higher than it
does in the hexadienyl homologue. The same relative
order may also obtain in the case of Fe(6,6-dmch)2 and
Fe(c-C8H11)2 (whether in solid or solution [24] state),
but in the absence of high temperature (\300 K)
Mo¨ssbauer data for the latter complex this conclusion
must, at the present time, be considered speculative.
A number of interesting correlations may be drawn
from the IS and QS values. First, for the neutral
complexes one can observe that the IS values for the
pentadienyl compounds are all significantly smaller
than the value for ferrocene, reflecting a greater s
electron density at the iron nuclei in the pentadienyl
complexes. This trend can be attributed readily to the
greater electron withdrawing ability of pentadienyl lig-
ands, which would lead to greater positive charge on
their iron centers, thereby causing a contraction of the
metal orbitals, and greater s-electron density. Similarly,
the neutral pentadienyl compounds are characterized
by much lower QS values relative to ferrocene, a differ-
ence which has been attributed to a much higher degree
of metal–ligand orbital mixing [8]. Theoretical studies
indeed reveal enhanced metal–ligand mixing for penta-
dienyl [9,30,33], and the resulting increases in dxz and
dyz orbital populations, and decreases in dxy and dx
Fig. 5. Temperature dependence of the isomer shift for Fe(6,6-
dmch)2. The open circles refer to the warming regime, the closed
circles to the cooling regime, illustrating the reversibility of the
spectral data.
Fig. 6. Logarithm of the area under the resonance curve for Fe(6,6-
dmch)2 as a function of temperature. The straight line is meant as a
guide.
2
2
populations, can clearly account for the substan−tiay l
decreases in the QS values. Concerning the greater
metal–pentadienyl orbital mixing, two mechanisms can
be considered, one being the inherent differences in
ligand orbital energies. In addition, however, it may be
possible that the lower symmetry of the open fer-
Fig. 7. Mo¨ssbauer spectrum of Fe(6,6-dmch)+2 at 110 K. The velocity
scale is defined relative to the centroid of a a-Fe spectrum at room
temperature.

R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
179
3. Conclusions
rocenes, ideally C2 versus at least D5 for ferrocene,
could actually be responsible. A distinction between
these possibilities can be made using the QS data for
the half-open ferrocene, Fe(C5H5)(2,4-C7H11). Should
the orbital mixings be driven by the relative ligand
orbital energies, one would expect this compound to
have a QS value intermediate between those of fer-
rocene and Fe(2,4-C7H11)2. On the other hand, should
symmetry be dominant, one would expect this low
symmetry (Cs) compound to be characterized by a
QS value similar to that of Fe(2,4-C7H11)2. That the
QS value for the half-open ferrocene is almost exactly
equal to the average of the QS value for ferrocene
and Fe(2,4-C7H11)2 can be taken as a clear indication
that it is the inherent ligand orbital energies, rather
than the compound symmetries, which are important.
Thus, the lower degree of resonance stabilization for
pentadienyl versus Cp anions leads to the former hav-
ing, on average, higher energy filled orbitals, which
then serve as better electron donors, and lower energy
empty orbitals, which then serve as better electron
acceptors.
Although the general reductions in IS and QS val-
ues for the pentadienyl compounds may readily be
accounted for, the differences in values for individual
pentadienyl compounds generally seem to defy expla-
nation. One does see an increase in QS value for the
edge-bridged open ferrocenes as the bridge becomes
longer, but their IS values follow no such trend. Simi-
lar confusing differences are found for the other open
ferrocenes upon the addition of methyl groups in var-
ious locations (Table 3).
A final point of interest relates to the effects of
oxidation on the IS and QS values. While oxidation
of ferrocene leads to a small decrease in the IS value,
attributable to a contraction of the metal s orbitals
[34], one observes just the opposite change from the
oxidation of Fe(6,6-dmch)2. Although it is possible
that an increased degree of ligand-to-metal s dona-
tion could lead to the higher s electron density, it
would be desirable to have some theoretical substanti-
ation for such a proposal. Concerning the QS value,
oxidation of Fe(6,6-dmch)2 led to a small but notice-
able decrease, whereas for ferrocene oxidation led to
nearly a complete collapse of the QS value — from
ca. 2.4 mm s−1 to ca. 0.1–0.6 mm s−1, depending on
the counterion. As a result, Fe(6,6-dmch)+2 provides
the only example in which an open ferrocene actually
possesses a higher QS value than the ferrocene ana-
logue itself. It would certainly be of interest to com-
pare the changes in iron d orbital populations that
accompany the oxidation of Fe(6,6-dmch)2 with the
corresponding changes that occur for ferrocene, and
it is hoped that appropriate studies along these lines
will be pursued shortly.
The Mo¨ssbauer and structural data reported herein
have provided some interesting insights into the differ-
ences in bonding for Cp and pentadienyl ligands. First,
much lower quadrapole splitting values were observed
for the 18 electron pentadienyl complexes as compared
to ferrocene, which provides credible confirmation of
theoretical results indicating much greater covalency
(metal–ligand orbital mixing) for the open ligands.
Furthermore, the fact that the QS value for the half-
open ferrocene was found to be intermediate between
those of ferrocene and Fe(2,4-C7H11)2 has allowed us to
attribute the greater covalency to inherent ligand or-
bital energy differences rather than to a symmetry
effect. While the greater orbital mixing in the open
ferrocenes would be expected to lead to stronger Fe–
ligand bonding, earlier structural data for Fe(2,4-
C7H11)2 were not in accord with this view, an
observation that had been ascribed to both steric and
overlap problems for open dienyl bonding with the
small iron center [7,8]. The fact that the structural data
for Fe(C5H5)(2,4-C7H11) reveal nearly identical FeꢀC
distances for the two ligands confirms this conclusion,
especially considering that for earlier metal analogues
there is a clear structural preference for the open lig-
ands [10].
Mo¨ssbauer data for the edge-bridged compounds
have demonstrated that these species are also character-
ized by greater metal–ligand orbital mixings than fer-
rocene. The presence of edge bridges in these
compounds has, however, been found to lead to pro-
found differences relative to the non-bridged open fer-
rocenes, including the alteration of conformational
preferences, and apparent ligand rotation even in the
solid state, possibly as a result of the stabilization of the
as yet unobserved anti-eclipsed conformation. The pres-
ence of an edge bridge has also been found to allow for
the isolation of the first 17 electron open ferrocene
cation, whose Mo¨ssbauer parameters suggest interesting
electronic properties. While the current studies have
therefore provided answers to some fundamental ques-
tions concerning these compounds, there clearly remain
a number of additional issues that need to be ad-
dressed, especially for the Fe(dmch)2 cation, and efforts
along these lines are continuing.
4. Experimental
All reactions were carried out under N2 atmosphere
in Schlenk apparatus. Ether and hydrocarbon solvents
were distilled from sodium–benzophenone under N2
atmosphere. Published procedures were employed for
the synthesis of 1,3-cycloheptadiene [35] and Fe(c-
C8H11)2 [14], while K(c-C7H9) was prepared according

180
R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
to a general method for pentadienyl anions [36]. Ele-
mental analysis was obtained from E and R Labora-
tory. Single crystals of Fe(C5H5)(2,4-C7H11) and
Fe(c-C7H9)2 were obtained by slowly cooling concen-
trated hydrocarbon solutions to −30 °C. Conditions
for their X-ray data collections are summarized in
Table 4. Both structures were solved straightforwardly
using direct methods, and subsequent difference
Fourier maps revealed all remaining atoms. Methyl
group hydrogen atom locations were idealized, while
the coordinates of all other hydrogen atoms were
refined. All calculations employed the SHELXTL pro-
gram package, and additional structural details may be
obtained from the Cambridge Crystallographic Data
Centre (see Section 5).
Mo¨ssbauer spectra were acquired and evaluated as
described previously [23a,25,27], using a ꢀ50 mCi
source of 57Co(Rh) in conjunction with a fast propor-
tional counter and associated electronics. Typically an
excess of 106 counts per channel (of 256) was acquired
at each temperature, which was held constant to 90.5
K. All IS are reported with respect to the centroid of an
a-Fe spectrum at room temperature (r.t.), which was
also used to calibrate the velocity scale. The DSC data
referred to in the text were acquired at heating (cooling)
rates of 5 K min−1 using a Mettler Toledo Model
DSC30 differential scanning calorimeter.
4.1. Bis(cycloheptadienyl)iron, Fe(c-C7H9)2
This compound has previously been obtained from
the reduction of FeCl3 by a Grignard reagent in the
presence of 1,3-cycloheptadiene [12]. The following al-
ternative approach requires less diene and leads to
higher yields of product. To a slurry of FeCl2 (0.458 g,
3.61 mmol) in 30 ml of THF at −78 °C was slowly
added a solution of K(c-C7H9) (1.40 g, 10.8 mmol) in
THF. After the addition was complete, the solution was
allowed to warm to r.t. and was then stirred for 2 h.
The solvent was removed in vacuo, leaving a dark
brown residue, which was extracted using ca. 60 ml of
hexane. The solution was filtered through a Celite pad
on a coarse frit. The product was crystallized by con-
centration of the orange–brown filtrate in vacuo to ca.
20 ml and placement into a −30 °C freezer. After
removal of the supernatant from the crystalline
product, sublimation (50 °C, 10−2 Torr) afforded 420
mg (48% yield) of orange–red crystals. Spectroscopic
data were consistent with those published previously
[12].
4.2. Bis(6,6-dimethylcyclohexadienyl)iron
tetraphenylborate, [Fe(6,6-Me2C6H5)2]+[B(C6H5)4]−
To a solution of Fe(6,6-Me2C6H5)2 ([13], 200 mg,
0.740 mmol) in 20 ml of THF under N2 at 0 °C was
added solid [Fe(C5H5)2]BF4 [37] (200 mg, 0.731 mmol).
Within a few minutes a green solution had resulted.
After the solution had been stirred for 2 h, the solvent
was removed in vacuo. The green residue was washed
with pentane to remove ferrocene, and thereafter 20 ml
of THF was added to the green residue, along with
solid NaB(C6H5)4 (250 mg, 0.730 mmol). After further
stirring for 2 h at 0 °C the product had precipitated
from the solution, and it was then collected by filtration
through a coarse frit. The solid green product was
dissolved in ca. 20 ml of MeCN. The product could be
crystallized at −20 °C. After removal of the superna-
tant, the compound was dried in vacuo and isolated as
a green crystalline solid (0.24–0.28 g, 55–65% yield).
Anal. Found: C, 81.22; H, 7.34. Calc. for C40H42BFe:
C, 81.51; H, 7.18%.
Table 4
Crystal data and structure refinement parameters for Fe(C5H5)(2,4-
C7H11) and Fe(c-C7H9)2.
Empirical formula
Formula weight
Crystal system
Space group
Unit cell dimensions
FeC12H16
216.10
Orthorhombic
Pnma
FeC14H18
242.13
Triclinic
(
P1
,
a (A)
5.9428(2)
13.1192(7)
12.7727(6)
90
13.8790(4)
14.1131(3)
15.0322(5)
62.0902(14)
63.0398(9)
,
b (A)
,
c (A)
h (°)
i (°)
90
k (°)
90
995.82(8)
4
89.3141(16)
2242.99(11)
8
3
,
V (A )
Z
Crystal color, habit
Crystal size (mm)
Temperature (K)
Diffractometer
Red prism
0.20×0.13×0.10
200
Nonius Kappa
CCD
0.71073
3161
1784
Red–orange prism
0.26×0.23×0.14
200
4.3. (Cyclopentadienyl)(2,4-dimethylpentadienyl)iron,
Fe(C5H5)(2,4-C7H11)
Nonius Kappa CCD
This compound was prepared as previously described
[16]. Anal. Found: C, 66.79; H, 7.54. Calc. for
C12H16Fe: C, 66.70; H, 7.46%.
,
Radiation u (A)
0.71073
15 620
13 032
Data collected
Observed unique
data
1H-NMR (benzene-d6, ambient): l 5.19 (s, 1H), 3.88
(s, 5H), 2.57 (s, 2H), 1.76 (s, 6H), −0.76 (s, 2H).
13C-NMR (benzene-d6, ambient): l 93.1 (d, J=160
Hz), 90.8 (s), 75.6 (d, Cp, J=176 Hz), 43.5 (t, J=155
Hz), 27.2 (q, J=127 Hz).
R(F) a
0.0320
0.0720
0.0373 (2|)
0.0767
R(wF2) a
a Quantity minimized=R(wF2)=ꢀ[w(F2o−Fc2)2]/ꢀ[(wFo2)2]1/2; R=
ꢀD/ꢀ(ꢁFoꢁ), D=ꢁ(ꢁFoꢁ−ꢁFcꢁ)ꢁ.

R. Basta et al. / Journal of Organometallic Chemistry 637–639 (2001) 172–181
181
[12] J. Mu¨ller, B. Mertschenk, Chem. Ber. 105 (1972) 3346.
EIMS (17 eV); m/z (relative intensity): 216 (27), 215
(15), 214 (100), 148 (33), 134 (15), 121 (33), 56 (43).
[13] P.T. DiMauro, P.T. Wolczanski, Organometallics 6 (1987) 1947.
[14] V. Kulsomphob, R. Tomaszewski, G.P.A. Yap, L.M. Liable-
Sands, A.L. Rheingold, R.D. Ernst, J. Chem. Soc. Dalton Trans.
(1999) 3995.
[15] P.T. DiMauro, P.T. Wolczanski, Polyhedron 14 (1995) 149.
[16] R. LeSuer, W.E. Geiger, unpublished results.
[17] Ch. Elschenbroich, E. Bilger, R.D. Ernst, D.R. Wilson, M.S.
Kralik, Organometallics 4 (1985) 2068.
[18] (a) S. Carter, J.N. Murrell, J. Organometal. Chem. 192 (1980)
399;
(b) D.L. Clark, J.C. Gordon, J.T. McFarlan, R.L. Vincent-Hol-
lis, J.G. Watkin, B.D. Zwick, Inorg. Chim. Acta 244 (1996) 269.
[19] (a) R.K. Bohn, A. Haaland, J. Organomet. Chem. 5 (1966) 470;
(b) A. Haaland, J.E. Nilsson, Acta Chem. Scand. 22 (1968) 2653.
[20] The first value in parenthesis represents the variation in values
being averaged, while the second value represents the uncertainty
in the average value.
5. Supplementary material
Crystallographic data for structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC nos. 153826 and 153827 for com-
pounds Fe(C5H5)(2,4-C7H11) and Fe(c-C7H9)2. Copies
of this information may be obtained free of charge
from The Director, CCDC, 12 Union Road, Cam-
bridge, CB2 1EZ, UK (Fax: +44-1223-336033; e-mail:
deposit@ccdc.cam.ac.uk or www.ccdc.cam.ac.uk).
[21] D.R. Wilson, R.D. Ernst, T.H. Cymbaluk, Organometallics 2
(1983) 1220.
[22] (a) R. Hoffmann, P. Hofmann, J. Am. Chem. Soc. 98 (1976) 598;
(b) R.D. Ernst, T.H. Cymbaluk, Organometallics 1 (1982) 708.
[23] (a) J.-C. Han, J.P. Hutchinson, R.D. Ernst, J. Organomet.
Chem. 321 (1987) 389;
(b) W. Trakarnpruk, A.M. Arif, R.D. Ernst, J. Organomet.
Chem. 485 (1995) 25.
[24] D.A. Brown, N.J. Fitzpatrick, M.A. McGinn, J. Chem. Soc.
Dalton Trans. (1986) 701.
Acknowledgements
Partial support of this research by the NSF and
University of Utah is gratefully acknowledged. We
would like to thank Mr Bob LeSuer and Professor
William Geiger for information and suggestions regard-
ing the stability of the Fe(dmch)2 cation.
[25] (a) R.H. Herber, I. Gattinger, F.H. Ko¨hler, Inorg. Chem. 39
(2000) 851;
(b) R.H. Herber, B. Bildstein, P. Denifl, H. Schottenberger,
Inorg. Chem. 36 (1997) 3586 (and references therein).
[26] R.H. Herber, in: R.H. Herber (Ed.), Chemical Mo¨ssbauer Spec-
troscopy, Plenum Press, New York, 1984, pp. 199–216.
[27] R.H. Herber, Inorg. Chim. Acta 291 (1999) 74.
[28] E.O. Fischer, J. Mu¨ller, J. Organomet. Chem. 1 (1963) 89.
[29] (a) R.H. Herber, I. Nowik, Hyperfine Interact. 126 (2000) 127;
(b) H. Schottenberger, M.R. Buchmeiser, R.H. Herber, J.
Organomet. Chem. 612 (2000) 1.
References
[1] (a) A. McKnight, R.M. Waymouth, Chem. Rev. 88 (1988) 2587;
(b) M. Bochmann, J. Chem. Soc. Dalton Trans. (1996) 255;
(c) H.-H. Brintzinger, D. Fischer, R. Mu¨lhaupt, B. Rieger, R.M.
Waymouth, Angew. Chem. Int. Ed. Engl. 34 (1995) 1143;
(d) D. Veghini, L.M. Henling, T.J. Burkhardt, J.E. Bercaw, J.
Am. Chem. Soc. 121 (1999) 564;
(e) F.J. Karol, C. Wu, W.T. Reichle, N.J. Maraschin, J. Catal.
60 (1979) 68.
[2] (a) A. Toqni, Angew. Chem. Int. Ed. Engl. 35 (1996) 1475;
(b) in: A. Toqni, T. Hayashi (Eds.), Ferrocenes. Homogeneous
Catalysis, Organic Synthesis, Materials Science, Weinheim,
VCH, 1995.
[30] (a) M.C. Bo¨hm, M. Eckert-Maksic´, R.D. Ernst, D.R. Wilson, R.
Gleiter, J. Am. Chem. Soc. 104 (1982) 2699;
(b) R. Gleiter, M.C. Bo¨hm, R.D. Ernst, Electron. Spectrosc.
Relat. Phenom. 22 (1984) 269.
[31] D.F. Brougham, P.J. Barrie, G.E. Hawkes, I. Abramams, M.
Motevalli, D.A. Brown, G.J. Long, Inorg. Chem. 35 (1996) 5595.
[32] (a) R.J. Webb, M.D. Lowery, Y. Shiomi, M. Sorai, R.J. Wit-
tebort, D.N. Hendrickson, Inorg. Chem. 31 (1992) 5211 (and
references therein);
(b) R.H. Herber, T.P. Hanusa, Hyperfine Interact. 108 (1997)
563.
[33] R.M. Kowaleski, F. Basolo, J.H. Osborne, W.C. Trogler,
Organometallics 7 (1988) 1425.
[34] R.H. Herber, in: M. Tsutsui (Ed.), Characterization of
Organometallic Compounds. Part 1, Interscience, New York,
1969, p. 326ff.
[35] A.P. Ter Borg, A.F. Bickel, Recueil 80 (1961) 1229.
[36] D.R. Wilson, L. Stahl, R.D. Ernst, Organomet. Synth. 3 (1986)
136.
[3] H. Schottenberger, K. Wurst, M.R. Buchmeiser, J. Organomet.
Chem. 584 (1999) 301 (and references therein).
[4] J.A. Raloff, Sci. News 144 (1993) 223.
[5] D. Seyferth, A. Davison, Science 182 (1973) 699.
[6] (a) A. Haaland, Acc. Chem. Res. 12 (1979) 415;
(b) A. Toqni, R. Halterman (Eds.), Metallocenes. Synthesis,
Reactivity, Applications, vol. 1 and 2, VCH, Weinheim, 1998.
[7] (a) R.D. Ernst, Acc. Chem. Res. 18 (1985) 56;
(b) R.D. Ernst, Struct. Bonding (Berlin) 57 (1984) 1.
[8] R.D. Ernst, Chem. Rev. 88 (1988) 1255.
[9] I. Hyla-Kryspin, T.E. Waldman, E. Mele´ndez, W. Trakarnpruk,
A.M. Arif, M.L. Ziegler, R.D. Ernst, R. Gleiter, Organometal-
lics 14 (1995) 5030.
[10] R.D. Ernst, Comments Inorg. Chem. 21 (1999) 285.
[11] R.D. Ernst, D.R. Wilson, R.H. Herber, J. Am. Chem. Soc. 106
(1984) 1646.
[37] D.N. Hendrickson, Y.S. Sohn, H.B. Gray, Inorg. Chem. 10
(1971) 1559.