Effects of metal atom motion and ring rotation in iron organometallics. Synthesis of isotopically labelled nonamethyl [57Fe]ferrocene

Article abstract of DOI:10.1016/S0022-328X(03)00778-2

Temperature dependent ⁵⁷Fe M?ssbauer effect studies on nonamethyl ferrocene (NMF), have been carried out on isotopically enriched material over the temperature range 90< T <290 K. These measurements have led to the elucidation of the dynamical behaviour of the iron atom and reflect the onset of ring rotation and libration. These results are compared to corresponding data for ferrocene and related compounds. A convenient small scale synthesis suited for the preparation of ⁵⁷Fe-enriched nonamethylferrocene which allows its easy separation in pure form, is also presented. With the exception of nonamethylferrocene, which is unresolvable due to its lattice-neighbour related conformational ring disorders, X-ray structure determinations of all new compounds of the reaction sequence, and of the starting compound hydroxymethyloctamethylferrocene (octamethylferrocenyl methanol) have been performed.

Full text of DOI:10.1016/S0022-328X(03)00778-2

Journal of Organometallic Chemistry 682 (2003) 163ꢀ171
/
www.elsevier.com/locate/jorganchem
Effects of metal atom motion and ring rotation in iron
organometallics. Synthesis of isotopically labelled nonamethyl
[57Fe]ferrocene
Rolfe H. Herber a,*, Israel Nowik a, Herwig Schottenberger b, Klaus Wurst b,
Norbert Schuler b, Adrian Gallus Mueller b
a Racah Institute of Physics, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel
b Institut fur Allgemeine, Anorganische und Theoretische Chemie, Leopold Franzens Universitat, Innrain 52a, A-6020 Innsbruck, Austria
¨
¨
Received 8 April 2003; received in revised form 17 June 2003; accepted 17 June 2003
Abstract
Temperature dependent 57Fe Mossbauer effect studies on nonamethyl ferrocene (NMF), have been carried out on isotopically
¨
enriched material over the temperature range 90B
/
T B290 K. These measurements have led to the elucidation of the dynamical
/
behaviour of the iron atom and reflect the onset of ring rotation and libration. These results are compared to corresponding data for
ferrocene and related compounds. A convenient small scale synthesis suited for the preparation of 57Fe-enriched nonamethylferro-
cene which allows its easy separation in pure form, is also presented. With the exception of nonamethylferrocene, which is
unresolvable due to its lattice-neighbour related conformational ring disorders, X-ray structure determinations of all new
compounds of the reaction sequence, and of the starting compound hydroxymethyloctamethylferrocene (octamethylferrocenyl
methanol) have been performed.
# 2003 Elsevier B.V. All rights reserved.
Keywords: Octamethylferrocene derivatives; 57Iron labelled; X-ray structures; Mossbauer spectroscopy
¨
1. Introduction
derived from octamethyl ferrocene [h5-(Cp(Me)4]2Fe
(OMF), the recoil-free fraction (f) observed in a
Mossbauer spectrum effectively became unobservable
¨
at temperatures well below the melting point. Since the
Temperature-dependent 57Fe Mossbauer effect spec-
¨
troscopy has proven itself to be a powerful technique in
the elucidation of the details of the metal atom motion
in iron organometallics. In particular, recent studies
area under the resonance curve in a Mossbauer spec-
¨
trum is proportional to the recoil-free fraction, and the
latter is directly related to the mean-square-amplitude of
vibration (msav, Žx2) of the metal atom by the
relationship
from these laboratories [1ꢀ3] have shown that the
/
details of this motion can be related to the molecular
level architecture of ferrocene-related solids, and can
lead to a better understanding of the macroscopic
properties of these compounds which have found
considerable use in opto-electronics and chemical sensor
applications. An unexpected observation relative to the
metal atom motion in ferrocene-related solids was the
demonstration [4,5] that in a number of such solids
f ꢁexp(ꢂk2Žx2)
where k is the wave number of the Mossbauer g-ray, it
¨
follows that the large reduction in the resonance effect
must be due to a large increase of Žx2 with increasing
temperature. In the case of octamethylferrocene com-
pounds where one proton of a Cp ring may be
substituted by an additional small group such as Ã
/
CÅ
/
* Corresponding authors. Tel.: ꢃ
6586-347; Tel: ꢃ43-512-507-5120.
E-mail addresses: herber@vms.huji.ac.il
herwig.schottenberger@uibk.ac.at (H. Schottenberger).
/
972-2-6584-347; fax: ꢃ
/
972-2-
CH or ÃCHÄCH2, the onset of ring rotation results in
/
/
/
both an increase of the ring-ring distance as well as a
librational motion of one of the rings with respect to the
(R.H.
Herber),
0022-328X/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00778-2

164
R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ171
/
other. This ‘‘gear-wheel’’ effect leads to a marked
increase of Žx2, and hence to a pronounced decrease
in the resonance effect with increasing temperature. In
Table 1
Summary of Mossbauer data for nonamethyl[57Fe]ferrocene (4)
¨
Parameter
Value
T Range Multiplier Units
the case of octamethylferrocene (TMP
‘‘vibrational anomaly’’ occurs at Â349 K, that is at a
reduced temperature TRꢁT/TMP of 0.81 [4]. In the case
of ethynyl(octamethyl)ferrocene [6] (TMP 246 K), this
anomaly occurs at TRÂ0.56 [5], while in ethenyl(octa-
methyl)ferrocene (TMP 432 K), this anomaly occurs at
TRÂ0.53 [7]. Furthermore, it is interesting to note that
ꢁ
/
431 K), the
IS (90 K)
QS (90 K)
0.5070.006
2.4640.006
mm sꢂ1
/
mm sꢂ1
/
ꢂ
/
dIS/dT
Meff
Area ratio (R) 1.000.01
4.1610.065 200ꢂ
1001.6 200ꢀ290
90ꢀ290
/
290 10ꢂ4
mm sꢂ1 Kꢂ1
daltons
ꢁ
/
/
/
/
ꢁ
/
/
in a recent quasi-elastic nuclear forward-scattering
experiment on ethynyl(octamethyl)ferrocene reported
by Asthalter et al. [8] a very similar dynamical behaviour
of the iron atom was reported. In that study, a
‘‘hysteresis’’ of about 25 K in the metal atom motion
was documented in detail. This hysteresis, which is also
observed in the 57Fe Mossbauer experiments*
/
albeit
¨
appears to be characteristic of the
somewhat smaller*
/
phase transition involving the onset on warming ( and
cessation on cooling) of ring rotation in this solid.
A previous study [9] of the metal atom motion in
nonamethylferrocene (NMF), [h5-Cp(CH3)5]Fe[h5-
(Cp(CH3)4], showed that a similar effect to that noted
in OMF was also present here, although its onset was at
a significantly lower temperature. A detailed study of
NMF in the temperature range near the anomaly was
hampered by the very small resonance effect observed,
which in turn requires long data acquisition times. In
order to overcome this limitation, a sample of NMF
enriched in 57Fe (to 95%, while the natural abundance is
2.21%) was prepared as discussed below, and subjected
Fig. 1. Temperature dependence of the IS parameter for NMF. The
slope of the high temperature region of this data set (solid line) is ꢂ
(4.160.06)ꢄ 100 daltons. The
10ꢂ4 mm sꢂ1 Kꢂ1 from which Meff
open circles represent data acquired in a warming mode between data
points, while the full circles refer to cooling data, showing the
reversibility of the data with temperature.
/
/
Â/
The temperature dependence of the quadrupole split-
ting (QS) is shown in Fig. 2. In contrast, there is a
to a detailed temperature-dependent Mossbauer study in
¨
both warming and cooling regimes. The results of this
study are detailed herein.
significant change in slope (dQS/dT) at Â210 K. The
/
low temperature data show that QS is not a sensitive
function of temperature as has been previously noted in
the literature [11]. The small negative temperature
dependence in this region can be readily accounted for
by thermal expansion of the covalent lattice, as has been
noted previously. Above the transition point, this
temperature dependence becomes significantly more
negative. In this context it is appropriate to note that
spectra were acquired both in a warming and cooling
2. Results and discussion
As is true of all previously examined diamagnetic
ferrocenyl compounds, the Mossbauer spectra over the
¨
temperature range used in the present study (905
/
T 5
/
273 K) consist of well resolved doublets corresponding
to a single iron site in the structure. The line widths in a
temperature region where saturation effects can be
ignored (vide infra) are approximately 0.23 mm sꢂ1. A
summary of the hyperfine parameters and related
quantities is given in Table 1. The temperature depen-
dence of the isomer shifts (IS) is summarized in Fig. 1.
In the high temperature limit, that is above Â
this parameter is well fitted by a linear regression with a
slope of ꢂ(4.160.07)ꢄ
10ꢂ4 mm sꢂ1 Kꢂ1 (correlation
coefficient 0.98 for 25 data points) leading to an effective
/200 K,
/
/
vibrating mass Meff
ꢁ1001.6 daltons [10], indicative of
/
strong covalency in the Fe-h5 Cp bonding. As is seen
from the figure, there is no significant discontinuity in
dIS/dT over the entire temperature range.
Fig. 2. Temperature dependence of the QS parameter for NMF.

R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ
/171
165
mode, and there is no evidence for non-reversibility in
these data.
ferrocene is also operational in decamethyl ferrocene
[h5-Cp(CH3)5]2Fe where, at sufficiently high tempera-
tures, only rotational motion of the two rings is
presumed to be present [9].
The temperature dependence of the recoil-free frac-
tion, as reflected in the area under the resonance curve,
is summarized in Fig. 3 in which the abscissa is
expressed in terms of TR, the ratio of the experimental
temperature to the melting point of NMF, 4812 K.
Expressing the abscissa in terms of TR not only
emphasizes the early onset of the recoil-free fraction
anomaly in NMF, but also permits a ready comparison
with the corresponding f(T) data for the parent
ferrocene. Before considering in detail the significance
of these data, a number of supportive comments appear
to be appropriate. First, the data obtained with a
natural abundance 57Fe sample of NMF (star data
points) and those obtained with the 57Fe-enriched
sample (full circles) are in good agreement with each
other as expected. In this context, however, it should be
noted that because of the large resonance effects
observed for the latter (especially at low temperatures)
the data were fitted using a transmission integral
program [12] to correct the area under the resonance
curve for saturation effects in the spectra. The use of
such a transmission integral fitting procedure has been
discussed earlier [3,13]. Second, the temperature depen-
dence of ln f is characterized by a much more gradual
decrease with increasing temperature than was noted for
A plausible interpretation of these results is as
follows: on warming a sample of NMF from 90 to Â
/
200 K, the normal thermal expansion factors of a
covalent solid readily account for the observed data
and are, in fact, in reasonably close agreement with the
results obtained for the parent ferrocene [3]. Above Â
/
200 K, there is a gradual onset of ring rotation. This ring
rotation, which involves an expansion of the ring-ring
distance due to mutual steric repulsion of the methyl
groups, causes an increase of the ringꢀmetal distance,
/
and hence in an increase of Žx2 of the iron atom,
leading to a significant decrease in f as observed.
It is now appropriate to consider why the vibrational
anomaly in NMF is less abrupt than it is in the
octamethyl ferrocenyl complexes referred to above. It
is clear that there is a competition between the effects of
the onset of ring rotation which lead to an increase of
Žx2 with increasing temperature, and the van der
Waals interactions between adjacent molecules of the
covalent matrix which tend to offset the former effect on
warming. The Fe-ring carbon distance in ferrocene has
been reported in numerous X-ray crystallographic
˚
˚
studies and is 2.030 A at 173 K [14], and 2.033 A at
298 K [15]. In the gas phase, electron diffraction
OMF, OMFÃ
less, the decrease in ln f is very pronounced at TRꢀ
0.45, although this is some 250 degrees below TMP
/
CÅ
/
CH, and OMFÃ
/
CHÄ
/
CH2. Nonethe-
/
Â
/
˚
measurements lead to a bond distance of 2.064 A [16].
!
The effect of methyl group interactions between the two
rings is reflected in the fact that in solid decamethyl
ferrocene, in which the methyl groups are bent 38 out of
Finally, the corresponding data for ferrocene are
included in Fig. 3 from which it can be seen that the
resonance effect here is still readily observable at TR
values well above 0.75. In this compound, it is clear that
the ‘‘gear wheel effect’’ attributed to the nonamethyl
compounds, which confers both librational as well as
rotational motion to the rings relative to each other, is
absent. A behaviour quite similar to that observed for
the Cp plane away from the iron atom, the Feꢀ
/ring
carbon distance determined by X-ray diffraction is 2.049
˚
˚
A [17], while in the gas phase this value is 2.064 A [18],
identical to that in ferrocene under the same conditions.
It is appropriate to note that a similar behaviour to that
reported here for NMF has been noted [19] in octa-
methylferrocenylmethanol
Cp(CH3)4CH2OH], (OMFÃ
for which TR for the onset of the Žx2 increase is Â
0.45, comparable to the value of TR 0.47 for NMF. In
comparison, the TR onset values for OMF, OMFÃCÅ
CH, and OMFÃCHÄCH2 are 0.81, 0.56, and 0.53,
[h5-(Cp(CH3)4]Fe[(h5-
CH2OH), (TMP 484 K),
/
ꢁ
/
/
Â
/
/
/
/
/
respectively. Thus, at the lower TR value, the inter-
molecular interactions, specifically hydrogen bonding,
play a more significant role in the temperature depen-
dence of Žx2 than at higher values of this parameter.
As noted earlier [9], the onset of ring rotation and
libration in NMF well below r.t. has precluded a
detailed single crystal X-ray determination of the
Fig. 3. Temperature dependence of the logarithm of the area under the
resonance curve, normalized to the 90 K data point, plotted as a
structure of this compound at
Â
/
300 K. A low
temperature (B220 K) X-ray structure determination
/
function of the reduced temperature TRꢁT/TMP. The comparison
/
would seem to be informative relative to the points
raised in this discussion.
with the ferrocene data in the high temperature regime illustrates the
effect on Žx2 due to the onset of ring rotation and libration in NMF.

166
R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ171
/
tetramethylcyclopentadienide. In order to subsequently
separate the resulting methylated ferrocenes, preparative
HPLC-methods were developed for neutral and oxidized
cationic mixtures as well [22]. However, besides the
additional problems of chromatographic recovery, if
only nonamethylferrocene is desired, two thirds of the
enriched iron isotope would be consumed by the
unwanted octa- and decamethylferrocene also formed
in this statistical approach.
Transition metal catalysed hydrogenations of octa-
methylferrocene carbaldehyde, [128925-12-2], have not
been reported so far. Since handling small quantities of
expensive progenitors in an autoclave is likely to involve
tedious workup, no attempts have been undertaken to
investigate such conversions of this most versatile
starting compound in substitution chemistry of octa-
methylferrocene [23]. Furthermore, the high steric
shielding of the aldehyde moiety by adjacent methyl
groups may require fine tuned reaction conditions for a
clean and complete conversion to NMF.
Therefore, it seemed encouraging to try a small scale
preparation by hydride addition to octamethylferroce-
nylmethylium tetrafluoroborate [128925-11-1], which
was reported to be fairly stable for limited periods at
low temperatures as an isolated salt [23].
However, the dissolution of this salt in any appro-
priate solvent to establish homogeneous reaction con-
ditions, which turned out to be necessary for the attack
of hydride transfer reagents, is accompanied by the
known dimerisation forming 1,2-bis(octamethylferroce-
nyl) ethane species [23], thus resulting in lower yields
and, unfortunately, causing problems of separation. For
example, according to nuclear magnetic resonance
measurements, a composition of 85% NMF besides
15% 1,2-bis[(octamethyl)ferrocenyl]ethane was found in
the apolar fractions when reacting octamethylferroce-
nylmethylium tetrafluoroborate with lithium triethyl-
borohydride in THF. The ethanediyl-bridged
biferrocenium salt, generated in the dimerisation step,
obviously undergoes reduction to the neutral apolar
species, which prevents its separation by means of
standard column chromatography on a preparative
scale. Upon addition of lithium or potassium hydride
in THF, no NMF could be obtained.
Fig. 4. Temperature dependence of the area ratio of the two
components of the quadrupole split doublet in NMF. The absence
of a significant departure from unity up to 273 K reflects the isotropy
of the metal atom motion in NMF.
A final parameter which should be considered in this
discussion is that of the Gol’danskiiꢀ/Karyagin effect
[20] in these solids. This effect, which arises from the
anisotropy of the Mossbauer active atom motion with
¨
respect to the two major symmetry axes running
through the metal atom, is reflected experimentally in
the intensity ratio of the two components of the
quadrupole doublet; that is, Rꢁ
the two areas are those at more positive (ꢃ
negative (ꢂ) velocities with respect to the spectrum
/
A(ꢃ
/
)/A(ꢂ
/
), where
/) and more
/
centroid. In a number of cyclopentadienyl iron com-
plexes this motional anisotropy has been shown to be a
major effect arising from the details of the local
molecular architecture, especially when there is a strong
anisotropy in the metal atom-ligand bonding present in
the structure. The temperature dependence of R for
NMF is summarized graphically in Fig. 4 from which it
can be seen that R is completely insensitive to tempera-
ture over the interval of the Mossbauer measurements
¨
(90B
/
T B273 K). This observation is completely con-
/
sistent with the damping effects on Žx2 due to the van
der Waals interactions at temperatures well below TMP
as referred to above.
,
3. Synthetic considerations
In Wittig reactions involving octamethylferrocenyl-
methyl triphenylphosphonium bromide [185543-98-0],
we regularly observed concomitant degradation of the
respective ylide intermediate, affording some trace
amounts of NMF after aqueous workup [24]. NMF
obtained by this side reaction was the primary source for
Nonamethylferrocene [41311-84-6] was first men-
tioned by Nesmeyanov and coworkers [21], and actually
was identified by mass spectroscopy as an inseparable
by-product obtained from exhaustive Friedel Crafts
alkylation: treating ferrocene in heptane with LiAlH4
and AlCl3 in equimolar proportions followed by passage
of CH3Cl gave decamethylferrocene and nonamethyl-
ferrocene in about 20% and 50% yields, respectively.
In a later synthetic approach, nonamethylferrocene
was obtained from ferrous chloride in tetrahydrofuran
by ligand combination of a mixture of pentamethyl- and
earlier Mossbauer investigations [9,11a]. The utilization
¨
of this reaction sequence for the intended formation of
[57Fe]NMF (4) also led us to identify a second elimina-
tion product, where the bond cleavage of the phospho-
nium ylide occurs between the phosphorus and a
phenylic carbon to form the novel ferrocene substituted

R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ
/171
167
Scheme 1. Small scale synthesis of nonamethylferrocene via phospho-
nium ylide.
Fig. 5. ORTEP plot of 5 showing 50% of thermal ellipsoids; methyl- and
Cp-protons are omitted for clarity.
4. Experimental
4.1. Instrumentation
Scheme 2. Protic formation of ether-type dimers from hydroxymethyl-
ferrocenes.
Bruker AC 300 (NMR), Nicolet 510 FTIR (IR);
Varian CH-7 (MS); Nonius Kappa CCD (X-ray).
Melting points were determined on a Kofler hot-plate
apparatus equipped with a thermocouple (Leica Galen
III; Fluke 50S; correction obsolete).
phosphine oxide 3. This derivative will be the subject of
future Mossbauer studies [25] (Scheme 1).
¨
The higher polarity of 3 compared to the extremely
apolar 4 allows a convenient separation of the product
mixture by simple chromatography on basic alumina
(see Section 4).
4.2. Synthetic procedures
It is worth noting that under acidic conditions,
especially in the absence of nucleophilic coreactants,
also the conversion of hydroxymethyl-octamethylferro-
cene, [128925-13-3], to a further by-product, not de-
scribed previously in the literature, namely
di[(octamethylferrocenyl)methyl] ether (5) takes place;
it is the expected predominant product under conditions
of water removal by acidic catalysis. This compound
represents the first ether-type product based on octa-
methylferrocene, its detailed characterisation including
The starting material for all ferrocenes described in
this work, iron(II) chloride enriched in 57Fe, was
prepared from ꢀ
by a modified literature procedure [26]. 57Fe-enriched
iron powder (ꢀ95%) was purchased from AMT,
/95% enriched metallic iron powder
/
Advanced Materials Technologies, 7 Mevo Dalya Str.,
Yehud 56478, Israel; sales@isotope-amt.com.
All organometallic reactions were carried out under
inert atmosphere conditions using standard Schlenk
techniques. Solvents were distilled, dried and deoxyge-
nated prior to use. Starting compounds not referenced
are commercially available. Purchased chemicals were
used without any further purification.
Mossbauer studies will also be published in a forth-
¨
coming paper (Scheme 2).
57Fe enriched octamethylferrocenyl methanol,
[4115698-09-8] [19,27], was prepared in analogy to
published procedures [23,24] for naturally abundant
isotope material, [128925-13-3], starting from the corre-
sponding carbaldehyde [25]. Small scale optimised
preparations were applied in analogy to published
procedures also for the preparation of 57Fe-enriched
octamethylferrocene [28]. All preparations requiring
57Fe-enriched chemicals have been previously checked
with unlabelled ferrous chloride and ferrocene starting
compounds (vide infra).
3.1. Refinement details of 5 (Fig. 5)
Reflected by the highly distorted thermal ellipsoids of
the two tetramethylꢀCp rings, the ring at Fe1 shows a
/
disordering, which could be caused by a free rotation of
the ring or by a normal 1:1 disorder in two positions
(rotation by 1808). The disorder was not solved and the
carbon atoms of the four methyl groups were refined
into five positions with an occupation factor of 0.8.

168
R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ171
/
Fig. 6. ORTEP plot of 1 showing 50% of thermal ellipsoids; methyl- and
Cp-protons are omitted for clarity.
4.2.1. 1-(Hydroxymethyl)-1?,2,2?,3,3?,4,4?,5-
Fig. 7. ORTEP plot of 2 showing 50% of thermal ellipsoids.
octamethyl[57Fe]ferrocene [415698-09-8] ([57Fe]1)
1?,2,2?,3,3?,4,4?,5-Octamethyl[57Fe]ferrocenylcarbalde-
hyde [25] (120 mg, 0.37 mmol) was dissolved in
anhydrous THF (15 ml) under argon to form a red
Hz, Ar-Cmeta), 117.50 (d, Jꢁ
Jꢁ5 Hz, fec-C), 82.50 (d, Jꢁ
C), 80.72 (d, Jꢁ2 Hz, fec-C), 72.62 (fec-C), 71.25 (fec-
CH), 25.94 (d, Jꢁ43 Hz, CH2), 11.12, 9.92, 9.29, 9.13
(CH3); 31P-NMR (121.49 MHz, CD2Cl2) d 12.93 (s,
P(Ph)3); MS (FAB pos.): [MꢃH]ꢃ calculated: 574.2,
/84 Hz, Ar-Cipso), 83.55 (d,
/
/5 Hz, fec-C), 82.27 (fec-
/
solution, and subsequently cooled to ꢂ
triethylborohydride (0.55 ml, 0.55 mmolꢁ
/
40 8C. Lithium
1.5 equiva-
/
/
lent) was then added dropwise via syringe. The resulting
yellow solution was warmed to room temperature (r.t.)
and stirred for another 1.5 h. Water (30 ml) was added
and most of the THF was removed azeotropically by
means of a vacuum line. The resulting suspension was
extracted with ether (40 ml) and the organic phase was
washed back with water (30 ml) and brine (20 ml).
Afterwards, the solvent was reduced in volume to about
2 ml. Subsequent column chromatography on basic
/
found: 574.0.
4.2.3. X-ray analysis
Single crystals were grown from CD2Cl2 (NMR
solution). The phosphonium species crystallized with a
dichloromethane solvate as a mixed halogenide salt,
statistically disordered with respect to the bromide and
chloride counterions, in a ratio of 2:1 (formed by partial
nucleophilic substitution of the solvent by bromide
anion, see Fig. 7).
alumina (Brockmann III, 9ꢄ3 cm) in ether yielded
/
[57Fe]1 (105 mg, 86.7%) as a yellow solid, which was
used without further purification in the following step
(Fig. 6).
4.2.4. Refinement details of [57Fe]2
2:1 disorder of the chlorine atoms Cl2:ClA and
Cl3:Cl3A of the solvent dichloromethane (for hydrogen
calculation the carbon atom was split into two parts
with equal coordinates and displacement parameters
4.2.2. (1?,2,2?,3,3?,4,4?,5-
Octamethyl[57Fe]ferrocenyl)methyl-
triphenylphosphonium bromide ([57Fe]2)
To a solution of 1 (105 mg, 0.32 mmol) in neat toluene
(30 ml), triphenylphosphine hydrobromide (109 mg,
C30Ã
/
C30A). Another 2:1 disorder (see above) appears
for the halide counterions; the same position is occupied
by 2/3 bromide Br1 and 1/3 chloride Cl.
0.32 mmolꢁ
was refluxed in order to azeotropically remove water by
means of a DeanꢀStark trap for 15 min; during this
/
1.0 equivalent) was added and the mixture
/
operation, a yellow solid gradually precipitated. After
cooling, the precipitate was filtered through a sintered
glass frit (G4). The product was washed twice with ether
(10 ml each) and dried to yield analytically pure [57Fe]2
4.2.5. (1?,2,2?,3,3?,4,4?,5-
Octamethyl[57Fe]ferrocenyl)methyl-diphenylphosphine
oxide ([57Fe]3) and 1,1?,2,2?,3,3?,4,4?,5-
Nonamethyl[57Fe]ferrocene ([57Fe]4)
(174 mg, 73% overall yield). m.p.:192ꢀ
/
193 8C; IR (KBr,
To a precooled (ꢂ
/
65 8C), well stirred yellow suspen-
cmꢂ1): 3052 w, 2960 m, 2939 m, 2898 s, 2854 m, 1585 w,
1481 m, 1436 vs, 1380 m, 1106 s, 1031 m, 997 m, 754 m,
sion
of
(1?,2,2?,3,3?,4,4?,5-octamethyl[57Fe]ferroce-
nyl)methyl-triphenylphosphonium bromide ([57Fe]2)
1
740 s, 692 s, 509 m, 497 s, 472 m, 449 m; H-NMR
t
(75 mg, 0.11 mmol) in anhydrous THF (20 ml), BuOK
(300.13 MHz, CD2Cl2) d 7.85 (t, 3H, Jꢁ
7.66 (t, 6H, Jꢁ7.3 Hz, Ar), 7.57ꢀ7.50 (m, 6H, Ar), 3.98
(s (br), 2H, CH2), 3.67 (s (br), 1H, H of fec), 1.45 (s,
12H, 4ꢄCH3), 1.37 (s, 6H, 2ꢄCH3), 0.81 (s, 6H, 2ꢄ
CH3); 13C-NMR (75.47 MHz, CD2Cl2) d 135.51 (Ar-
para), 134.74 (d, Jꢁ9 Hz, Ar-Cortho), 130.26 (d, Jꢁ12
/
7.3 Hz, Ar),
(42 mg, 0.37 mmolꢁ3.4 equivalents) was added all at
/
/
/
once under Ar. Afterwards, the cooling bath was
removed and the resulting deep red mixture was allowed
to reach r.t. over 1.5 h. After warming up to 50 8C for
1.5 h, the now yellow misty suspension was quenched
with water (30 ml) and most of the THF was removed
/
/
/
C
/
/

R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ
/171
169
azeotropically by means of a vacuum line. The resulting
suspension was extracted with ether (40 ml) and the
ethereal phase was washed back with water (2ꢄ30 ml)
/
and brine (20 ml). The yellow solution was dried over
Na2SO4 and the ether was removed to dryness. The
crude product was suspended in n-hexane and filtered
through cotton wool. The insoluble part was dissolved
in ether and chromatographed on basic alumina (Brock-
mann III, 8ꢄ
/
3 cm) in ether to yield [57Fe]3 (19 mg,
32.3%) as a pale yellow solid. The filtrate was chroma-
tographed on basic alumina (Brockmann III, 5ꢄ3 cm)
/
in n-hexane to yield [57Fe]4 (15 mg, 41.9%) as a yellow
solid.
4.2.6. Data for [57Fe]3
M.p.: 146ꢀ
148 8C; IR (KBr, cmꢂ1): 3052 m, 2962 s,
2940 s, 2900 vs, 2856 m, 1436 s, 1378 m, 1189 vs, 1176 s,
1118 m, 1027 m, 738 s, 721 m, 696 m, 530 m, 501 m, 474
/
Fig. 8. ORTEP plot of 3 showing 50% of thermal ellipsoids.
1
m; H-NMR (300.13 MHz, CD2Cl2) d 7.68ꢀ
/
7.61 (m,
7.40 (m, 6H, Ar), 3.20 (d, 2H, Jꢁ11.6
Hz, CH2), 3.17 (s, 1H, H of fec), 1.66, 1.65, 1.60, 1.34
(each: s, 6H, 2ꢄ
CH3); 13C-NMR (75.47 MHz, CD2Cl2)
d 134.40 (d, Jꢁ95 Hz, Ar-Cipso), 131.72 (d, Jꢁ3 Hz,
Ar-Cpara), 131.45 (d, Jꢁ9 Hz, Ar-Cortho), 128.55 (d,
Jꢁ10 Hz, Ar-Cmeta), 80.34, 80.01, 79.86, 74.10 (fec-C),
71.22 (fec-CH), 29.87 (d, Jꢁ68 Hz, CH2), 11.18, 10.29,
9.85, 9.25 (CH3); 31P-NMR (121.49 MHz, CD2Cl2) d
P(Ph)2); HR-EIMS: [M+ꢃ] calc.: 513.19363,
dryness. The crude product was suspended in n-hexane
and filtered through cotton wool. The insoluble part was
dissolved in ether and chromatographed on basic
4H, Ar), 7.54ꢀ
/
/
/
alumina (Brockmann III, 8ꢄ3 cm) in ether to yield 3
/
/
/
(33 mg, 20%) as a pale yellow solid. The filtrate was
chromatographed on basic alumina (Brockmann III,
3 cm) in n-hexane to yield 4 (40 mg, 39%) as a yellow
solid.
/
/
6ꢄ
/
/
Performing the reaction under the same conditions
t
but warming to r.t. over 3 h after addition of BuOK
26.37 (s, OÄ
/
found: 513.1936.
yielded 3 in 22.% and 4 in 38%. When heated to 55 8C
for 3 h after addition of BuOK the yield of 3 increased
to 29%, but the yield of 4 decreased slightly to 35%.
t
4.2.7. Data for [57Fe]4
M.p.: 180ꢀ
transition between 130ꢀ
/
182 8C (partial melting due to a phase
/
135 8C); IR (KBr, cmꢂ1): 2962
4.2.9. Data for 3
m, 2942 m, 2898 m, 2854 m, 1375 m, 1261 s, 1097 vs,
1025 vs, 802 vs, 455 m; 1H-NMR (300.13 MHz, CD2Cl2)
M.p.: 169ꢀ
1718C; IR (KBr, cmꢂ1): 3052 m, 2962 m,
2942 m, 2898 s, 2856 m, 1436 s, 1380 m, 1371 m, 1186 vs,
1118 m, 1027 m, 740 s, 719 s, 701 m, 526 s, 509 m, 476 m,
/
d 3.16 (s, 1H, H of fec), 1.70 (s, 15H, 5ꢄ
/
CH3), 1.69 (s,
CH3); 13C-NMR (75.47
6H, 2ꢄCH3), 1.64 (s, 6H, 2ꢄ
/
/
1
410 m; H-NMR (300.13 MHz, CD2Cl2) d 7.67ꢀ
/
7.60
7.38 (m, 6H, Ar), 3.22 (d, 2H, Jꢁ
11.6 Hz, CH2), 3.15 (s, 1H, H of fec), 1.67, 1.66, 1.60,
1.35 (each: s, 6H, 2ꢄ
CH3); 13C-NMR (75.47 MHz,
CD2Cl2) d 133.90 (d, Jꢁ95 Hz, Ar-Cipso), 131.71 (d,
Jꢁ2 Hz, Ar-Cpara), 131.45 (d, Jꢁ9 Hz, Ar-Cortho),
128.55 (d, Jꢁ10 Hz, Ar-Cmeta), 80.28, 80.24, 79.93,
79.78, 74.06 (fec-C), 71.18 (fec-CH), 29.89 (d, Jꢁ68 Hz,
CH2), 11.18, 10.29, 9.85, 9.25 (CH3); 31P-NMR (121.49
MHz, CD2Cl2) d 26.55 (s, OÄP(Ph)2); MS(EI pos.):
[M+ꢃ] calc.: 512.2, found: 512.2. X-ray analysis: single
crystals were grown from CH2Cl2ꢀn-hexaneꢁ1:1.
MHz, CD2Cl2) d 79.95, 79.63, 78.93 (fec-C), 70.96 (fec-
CH), 11.18, 9.88, 9.26 (CH3); HR-EIMS: [M+ꢃ] calc.:
312.1467, found: 312.15008.
(m, 4H, Ar), 7.54ꢀ
/
/
/
/
4.2.8. (1?,2,2?,3,3?,4,4?,5-Octamethylferrocenyl)methyl-
diphenylphosphine oxide (3), and 1,1?,2,2?,3,3?,4,4?,5-
Nonamethylferrocene, [41311-84-6] (4) (Fig. 8)
/
/
/
/
To a cooled (ꢂ65 8C) and well stirred yellow suspen-
/
sion of (1?,2,2?,3,3?,4,4?,5-octamethylferrocenyl)methyl-
triphenylphosphonium bromide (2), (215 mg, 0.33
mmol) in dry THF (40 ml), tBuOK (123 mg, 1.10
/
/
/
mmolꢁ3.3 equivalents) was added all at once under Ar.
/
The resulting deep red mixture was stirred for 2.5 h at
0 8C in an ice-water bath. Afterwards, water (40 ml) was
added. Most of the THF was removed azeotropically in
vacuo, the yellow suspension was extracted with ether
(60 ml) and the ethereal phase was washed back with
4.2.10. Data for 4
M.p.: 163 8C; IR (KBr, cmꢂ1): 3054 w, 2962 s, 2940 s,
2896 vs, 2854 s, 1475 m, 1448 m, 1425 m, 1375 s, 1261 m,
1
1068 m, 1027 s, 819 m, 804 m, 457 s; H-NMR (300.13
water (2ꢄ
/
30 ml) and brine (30 ml). The yellow solution
MHz, CD2Cl2) d 3.18 (s, 1H, H of fec), 1.65 (s, 21H,
CH3), 1.60 (s, 6H, 2ꢄ
CH3); 13C-NMR (75.47 MHz,
was dried over Na2SO4 and the ether was removed to
7ꢄ
/
/

170
R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ171
/
Table 2
Crystal data and structure refinement for 1, 2, 3 and 5
Octamethylferrocenylꢀmethyl OH
compound
/
1 OMFÃ
/
CH2Ã
/
[
57Fe]2 OMFÃ
P(C6H5)3Xꢂ
/
CH2Ã
/
[
57Fe]3 OMFÃ
PO(C6H5)2
/
CH2Ã
/
5 OMFÃ
/
CH2Ã
/
OÃ
/
CH2Ã
/
OMF
Molecular formula
C19H28FeO
C37H42FePꢃBr0ꢂ.67Cl0ꢂ.33
CH2Cl2
×
/
C31H37FeOP
C38H54Fe2O
Formula weight
Crystal system
Space group
a (pm)
328.26
723.54
512.43
Monoclinic
638.51
Monoclinic
C2/c (No.15)
3838.0(3)
871.15(4)
2835.7(2)
90
Monoclinic
P21/n (No.14)
957.35(2)
2201.02(4)
1696.54(3)
90
Monoclinic
P21/n (No.14)
1796.7(1)
857.25(6)
2237.2(1)
90
P21/c (No.14)
1160.01(3)
2330.02(7)
1123.21(2)
90
b (pm)
c (pm)
a (8)
b (8)
132.240(2)
90
91.306(2)
90
116.010(2)
90
90.366(3)
90
g (8)
V (nm3)
Z
7.0192(8)
16
3.57393(12)
4
233(2)
2.72838(12)
4
233(2)
3.4457(3)
4
233(2)
Temperature (K)
Dcalc (Mg mꢂ3
Absorption coefficient (mmꢂ1
F(000)
Colour, habit
2933(2)
1.243
0.857
)
1.345
1.412
1.247
0.632
1.231
0.869
)
2816
Yellow prism
1504
Yellow prism
1088
Yellow prism
1368
Yellow prism
Crystal size (mm)
Data collection u range (8)
Index ranges
0.4ꢄ
1.94ꢀ
05h 5
295l 5
8535
/
0.2ꢄ
20.00
40; ꢂ
22
/
0.1
0.3ꢄ
1.85ꢀ
9; 05h 5
205l 5
22 308
/
0.2ꢄ
25.00
11; ꢂ
20
/
0.15
0.6ꢄ
1.75ꢀ
135
26; ꢂ115
/
0.4ꢄ
24.00
h 5
l 5
/
0.06
0.25ꢄ
1.45ꢀ20.00
05h 517; ꢂ
215l 521
11 748
/
0.18ꢄ
/
0.14
/
/
/
/
/
/
/85
/
k 5
/
/
/
/265
/
k 5
/
26;
ꢂ/
/
/
0; ꢂ
/
265
/
k 5
/
/
/
/
85
/
k 58;
/
ꢂ/
/
/
ꢂ/
/
/
/
/
/12
ꢂ/
/
/
Reflections collected
14 139
Independent reflections
3162 [Rint
2639
ꢁ/0.0329]
6283 [Rint
5285
ꢁ/0.0307]
4275 [Rint
3505
ꢁ/0.0393]
3216 [Rint
2234
ꢁ/0.0726]
Number of reflections with I ꢀ
/
2s(I)
Abs correction
Refinement method
None
None
None
Full-matrix least-squares Full-matrix least-squares on Full-matrix least-squares
None
Full-matrix least-squares
on F2
on F2
F2
on F2
Data/restraints/parameters
Goodness-of-fit on F2
3037/2/399
1.058
6283/0/414
1.031
4275/0/316
1.030
3216/3/391
1.038
Final R indices [I ꢀ
/
2s(I)]
R1ꢁ
0.1544
R1ꢁ0.0804; wR2ꢁ
0.1709
Largest difference peak and hole 553 and ꢂ
(e nmꢂ3
/
0.0672; wR2ꢁ
/
R1ꢁ
R1ꢁ
243 and ꢂ
/
0.0306; wR2ꢁ
0.0417; wR2ꢁ
306
/
0.0713
0.0758
R1ꢁ
R1ꢁ
551 and ꢂ
/
0.0383; wR2ꢁ
/
0.0887 R1ꢁ
0.1750
0.0938 R1ꢁ0.0945; wR2ꢁ
0.1924
469 and ꢂ
/
0.0634; wR2ꢁ
/
R indices (all data)
/
/
/
/
/0.0520; wR2ꢁ
/
/
/
/
620
/
/274
/277
)
CD2Cl2) d 80.00, 79.68, 78.96 (fec-C), 70.99 (fec-CH),
11.18, 9.88, 9.26 (CH3); MS(EI pos.): M+ꢃ calc.: 312.15,
found: 312.1.
displacement parameters. All non-hydrogen atoms were
refined with anisotropic displacement parameters.
Crystallographic data are collected in Table 2.
4.2.11. X-ray measurement and structure determination
of 1, [57Fe]2, [57Fe]3 and 5
The data collection was performed on a Nonius
4.3. Mossbauer spectroscopic investigations
¨
The Mossbauer spectrometer, its method of calibra-
¨
tion, the method of data reduction and thermal control
have been described earlier [1,13,32].
Kappa CCD equipped with graphite-monochromatised
˚
Ka-radiation (lꢁ0.71073 A) and a nominal crystal
Moꢀ
/
/
to area detector distance of 36 mm. Intensities were
integrated using DENZO and scaled with SCALE-
PACK [29]. Several scans in f and v direction were
made to increase the number of redundant reflections,
which were averaged in the refinement cycles. The
structures were solved with direct methods SHELXS86
[30] and refined against F2 (SHELX97) [31]. Hydrogen
atoms at carbon atoms were added geometrically and
refined using a riding model, hydrogen atoms at the
hydroxyl groups of 1 was refined regular with isotropic
A 5.8 mg sample of solid nonamethylferrocene was
ground with quartz powder to reduce crystallite size,
mixed with BN to ensure random crystallite orientation
with respect to the optical axis of the experiment, and
packed into a plastic sample holder which was, in turn,
transferred to a variable temperature cryostat as de-
scribed earlier [13]. The details of the spectroscopic
measurements, as well as the data reduction and
transmission integral corrections have also been dis-
cussed previously. Spectrometer calibration was effected

R.H. Herber et al. / Journal of Organometallic Chemistry 682 (2003) 163ꢀ
/171
171
using an 18.81 mg cmꢂ2 natural abundance a-Fe
absorber at room temperature, and all isomer shifts
are reported with respect to the centroid of this
spectrum.
[9] R.H. Herber, in: M. Gielen, R. Willem, B. Wrackmeyer (Eds.),
Unusual Structures and Physical Properties in Organometallic
Chemistry, Wiley, West Sussex, UK, 2002, pp. 207ꢀ
/218.
[10] R.H. Herber, in: R.H. Herber (Ed.), Chemical Mossbauer
¨
216.
Spectroscopy, Plenum Press, New York, 1984, pp. 199ꢀ
/
[11] (a) R.H. Herber, Inorg. Chim. Acta 291 (1999) 74;
(b) A. Houlton, J.R. Miller, R.M.G. Roberts, J. Silver, J. Chem.
Soc. Dalton Trans. (1991) 467 (and references cited therein).
[12] G.K. Shenoy, J.M. Friedt, H.J. Maletta, S.L. Ruby, in: I.J.
5. Supplementary material
Gruverman, C.W. Seidel, D.K. Dieterly (Eds.), Mossbauer Effect
¨
Methodology, vol. 9, Plenum Press, New York, 1974, pp. 277ꢀ
305.
/
Tables of crystal data and structure refinement de-
tails, anisotropic thermal parameters, fractional atomic
coordinates and isotropic thermal parameters for the
non-hydrogen atoms, all bond lengths and angles, and
fractional atomic coordinates for the hydrogen atoms
have been deposited with the Cambridge Crystallo-
graphic Data Centre, CCDC nos. 212215 (1), 212216
([57Fe]2), 212217 ([57Fe]3) and 212218 (5). Copies of this
information may be obtained free of charge from The
Director, CCDC, 12 Union Road, Cambridge CB2 1EZ,
[13] R.H. Herber, I. Nowik, Solid State Sci. 4 (2000) 691.
[14] C.P. Brock, Y. Fu, Acta Crystallogr. B53 (1997) 928.
[15] P. Seiler, J.D. Dunitz, Acta Crystallogr. B35 (1979) 1068; see also
P. Seiler, J.D. Dunitz, Acta Crystallogr. B35 (1979) 2020.
[16] L. Phillips, A.R. Lacey, M.K. Cooper, J. Chem. Soc. Dalton
(1988) 1383.
[17] D.P. Freyberg, J.L. Robbins, K.N. Raymond, J.C. Smart, J. Am.
Chem. Soc. 101 (1979) 892.
[18] A. Almenningen, A. Haaland, S. Samdal, J. Brunvoll, J.C. Smart,
J. Organomet. Chem. 173 (1979) 293.
[19] R.H. Herber, I. Nowik, T. Asthalter, E. Reichel, H. Schottenber-
ger, J. Organomet. Chem. 642 (2002) 203.
[20] See Ref. [1] and references therein.
UK (Fax: ꢃ44-1223-336033; e-mail: deposit@ccdc.cam.
ac.uk or www: http://www.ccdc.cam.ac.uk).
/
[21] A.N. Nesmeyanov, N.S. Kochetkova, A.M. Vainberg, Yu.K.
Krynkina, Doklady Akademii Nauk SSSR 207 (1972) 868.
[22] (a) J. Maynollo, diploma thesis, Innsbruck, 1991;
(b) M. Andre, P. Jaitner, J. Maynollo, H. Schottenberger, XVth
International Conference on Organometallic Chemistry, Warsaw,
Acknowledgements
August 9ꢀ14, 1992, book of abstracts, P153.
/
The authors (R.H.H. and I.N.) are indebted to the
Israel Ministry of Science and the Racah Institute for
financial support. We thank Professor Karl-Hans On-
gania for recording the high resolution mass spectra.
[23] C.H. Zou, M.S. Wrighton, J. Am. Chem. Soc. 112 (1990) 7578.
[24] A. Hradsky, B. Bildstein, N. Schuler, H. Schottenberger, P.
Jaitner, K.-H. Ongania, K. Wurst, J.-P. Launay, Organometallics
16 (1997) 392.
[25] A.G. Mueller. E. Reichel, H. Kopacka, H. Schottenberger, K.
Wurst, K.H. Ongania, R.H. Herber, submitted to J. Organomet.
Chem.
References
[26] M. Aresta, C.F. Nobile, D. Petruzelli, Inorg. Chem. 16 (1977)
1817.
[27] R.H. Herber, I. Nowik, Hyperfine Interactions 141/2 (2002) 297.
[1] I. Nowik, R.H. Herber, J. Phys. Chem. Solids 64 (2003) 313.
[2] G. Laus, H. Schottenberger, N. Schuler, K. Wurst, R.H. Herber,
J. Chem. Soc. Perkin Trans. 2 (2002) 1445.
¨
[28] G. Schmitt, S. .Ozman, Chem. Ztg. 100 (1976) 143.
[29] Z. Otwinowski, W. Minor, Methods Enzymol. 276 (1997) 307.
[30] (a) G.M. Sheldrick, SHELXS-86: Program for Crystal Structure
Solutions, University of Gottingen, Gottingen, Germany, 1986;
[3] R.H. Herber, I. Nowik, Hyperfine Interactions 136/7 (2001) 699.
[4] I. Nowik, R.H. Herber, Inorg. Chim. Acta 310 (2000) 191.
[5] H. Schottenberger, M.R. Buchmeiser, R.H. Herber, J. Organo-
met. Chem. 612 (2000) 1.
¨
¨
(b) G.M. Sheldrick, SHELXL-93: Program for the Refinement of
Crystal Structures, University of Gottingen, Gottingen, Germany,
1993.
¨
¨
[6] H. Schottenberger, K. Wurst, R.H. Herber, J. Organomet. Chem.
625 (2001) 200.
[31] Sheldrick, G.M.: Program package SHELXTL V.5.1, Bruker
Analytical X-Ray Instruments Inc, Madison, USA, 1997
[32] (a) R.H. Herber, E. Bauminger, I. Felner, J. Chem. Phys. 104
(1996) 1;
[7] R.H. Herber, unpublished results. The author is indebted to Dr.
Ingrid Kohl and Professor Erwin Mayer for DSC data pertaining
to OMFÃ
/
CHÄCH2
/
[8] T. Asthalter, H. Franz, U. van Burck, K. Messel, E. Schreier, R.
(b) R.H. Herber, Hyperfine Interactions 108 (1997) 549 (and
references cited therein).
¨
Dinnebier, J. Phys. Chem. Solids 64 (2003) 677.