Chemistry of (pentabenzylcyclopentadienyl)iron compounds including 17-electron dithiocarbamate complexes

Article abstract of DOI:10.1021/om9606350

C₅(CH₂Ph)₅H, C₅Bz₅H, 1, the precursor of the useful C₅Bz₅ ligand that was introduced by Rausch, has been synthesized in 85% yield, a slight improvement over the 62% yield very recently reported by Rausch. [Fe(C₅Bz₅)(CO)₂], 2, also reported by Rausch, could be obtained in 72% yield as the only product of the reaction between 1 and [Fe(CO)₅] at 90°C for 4 days. Reaction of 1 with Br₂ gives [Fe(C₅Bz₅)(CO)₂Br], 3, in 84% yield. Photolysis of 3 in THF using visible light in the presence of Na⁺S₂CNMe₂^- (Na⁺dtc^-) gives a 73% yield of [Fe(C₅Bz₅)(CO)(η²-dtc)], 4, whose oxidation by [FeCp₂]⁺PF₆^- gives the 17-electron complex [Fe^III(C₅Bz₅)(η ²-dtc)(NCMe)]⁺PF₆^-, 5, the precursor of [Fe^IVCp*(η²-dtc)₂] ⁺-PF₆^-, 7. Cyclic voltammetry studies of the C₅Bz₅-iron complexes show that the electronic properties of C₅Bz₅ are intermediate between those of C₅H₅ (Cp) and C₅Me₅ (Cp*) and that C₅Bz₅ destabilizes paramagnetic iron complexes as opposed to the stabilizing property of Cp*.

Full text of DOI:10.1021/om9606350

5598
Organometallics 1996, 15, 5598-5604
Ch em istr y of (P en ta ben zylcyclop en ta d ien yl)ir on
Com p ou n d s In clu d in g 17-Electr on Dith ioca r ba m a te
Com p lexes
Marie-He´le`ne Delville-Desbois,* Stephan Mross, and Didier Astruc*,†
Laboratoire de Chimie Organique et Organome´tallique, URA CNRS No. 35,
Universite´ Bordeaux I, 351, Cours de la Libe´ration, 33405 Talence Ce´dex, France
Received J uly 30, 1996X
C5(CH2Ph)5H, C5Bz5H, 1, the precursor of the useful C5Bz5 ligand that was introduced
by Rausch, has been synthesized in 85% yield, a slight improvement over the 62% yield
very recently reported by Rausch. [Fe(C5Bz5)(CO)2], 2, also reported by Rausch, could
be obtained in 72% yield as the only product of the reaction between 1 and [Fe(CO)5]
at 90 °C for 4 days. Reaction of 1 with Br2 gives [Fe(C5Bz5)(CO)2Br], 3, in 84% yield.
Photolysis of 3 in THF using visible light in the presence of Na+S2CNMe2 (Na+dtc-) gives
-
a 73% yield of [Fe(C5Bz5)(CO)(η2-dtc)], 4, whose oxidation by [FeCp2]+PF6 gives the 17-
-
electron complex [FeIII(C5Bz5)(η2-dtc)(NCMe)]+PF6-, 5, the precursor of [FeIVCp*(η2-dtc)2]+-
PF6-, 7. Cyclic voltammetry studies of the C5Bz5-iron complexes show that the electronic
properties of C5Bz5 are intermediate between those of C5H5 (Cp) and C5Me5 (Cp*) and that
C5Bz5 destabilizes paramagnetic iron complexes as opposed to the stabilizing property of
Cp*.
In tr od u ction
cyclopentadienyl derivative [FeIV{C5(CH2Ph)5}(S2CN-
Me2)2]+PF6-, whose crystal structure was recently pub-
lished.9 In the course of this work, we improved upon
the known synthesis of pentabenzylcyclopentadiene and
obtained an 85% yield. The recent report by Rausch4
urges us to also report these improvements. In the
course of our work, we also needed a good-yield synthe-
sis of the starting material bis[(pentabenzylcyclopen-
tadienyl)dicarbonyliron] dimer. Indeed, we have devel-
oped a procedure leading to a 72% yield. We have also
synthesized the bromo derivative by cleavage of the
Fe-Fe bond, and from there, we have synthesized
several dithiocarbamate compounds, including para-
magnetic ones, which we are also reporting in the
present paper.
Ten years ago, Rausch et al. published the first
pentabenzylcyclopentadienyl complexes.1-3 A paper
recently appeared from Rausch’s group4 reporting an
improved procedure for the synthesis of pentabenzyl-
cyclopentadiene. In this paper, the reaction of this
ligand with iron pentacarbonyl is analyzed and reported
to give bis[(pentabenzylcyclopentadienyl)dicarbonyliron]
dimer and (η4-pentabenzylcyclopentadiene)tricarbonyl-
iron.
We have been interested for some time in cyclopen-
tadienyl-type ligands5-7 in order to stabilize and crys-
tallize new organometallic complexes,8 especially
paramagnetic ones.9 Our interest in the pentabenzyl-
cyclopentadienyl ligand started when we tried to
obtain crystals from the robust complex [FeIV(C5Me5)-
Resu lts a n d Discu ssion
-
(S2CNMe2)2]+PF6
. Our frustration was then com-
pensated by success in the case of the pentabenzyl-
P en t a b en zylcyclop en t a d ien e, 1. Pentabenzyl-
cyclopentadiene, first reported by Hirsch and Bailey in
1978,10 was originally obtained in 19% yield, and its
X-ray structure confirmed the proposed assignment of
this complex by these authors. Rausch’s group im-
proved the yield to 28%1 and finally to 62%.4 In our
work, at first we followed the original procedure given
in ref 1, but found that it could be improved considerably
by a slower addition of dicyclopentadiene to the reaction
mixture (benzylic alcohol, small pieces of sodium metal,
and diisopropyl benzene). We believe this slow rate of
addition is crucial to obtain a good yield. We note that
our rate of addition is about two times slower than that
reported by Rausch et al.,4 and this must account for
our improved yield (eq 1).
† Telefax: Int. code + (556) 84 66 46. E-mail: d.astruc@lcoo.u-
bordeaux.fr.
X Abstract published in Advance ACS Abstracts, November 15, 1996.
(1) Chambers, J .-W.; Baskar, A. J .; Bott, S. G.; Atwood, J . L.; Rausch,
M. D. Organometallics 1986, 5, 1635.
(2) Schumann, H.; J aniak, C.; Hahn, E.; Kolax, C.; Loebel, J .;
Rausch, M. D.; Zuckerman, J .; Heeg, M. Chem. Ber. 1986, 119, 2656.
(3) Schumann has also published several papers on pentabenzyl-
cyclopentadienyl complexes. See for instance ref 2 and: Schumann, H.;
Suhring, K.; Weimann, R. J . Organomet. Chem. 1995, 496, C5 and
references cited therein and in refs 4 and 7.
(4) Tsai, W.-M.; Rausch, M. D.; Rogers, R. D. Organometallics 1996,
15, 2591.
(5) Two excellent reviews on bulky cyclopentadienyl ligands have
appeared, see refs 6 and 7.
(6) Okuda, J . Top. Curr. Chem. 1992, 160, 97.
(7) J aniak, C.; Schumann, H. Adv. Organomet. Chem. 1991, 33, 291.
(8) (a) Gloaguen, B.; Astruc, D. J . Am. Chem. Soc. 1990, 112, 4607.
(b) Buchholz, D.; Gloaguen, B.; Fillaut, J .-L.; Cotrait, M.; Astruc, D.
Chem. Eur. J . 1995, 1, 374.
(9) (a) Desbois, M. H.; Astruc, D. Angew. Chem., Int. Ed. Engl. 1989,
28, 460. (b) Delville, M.-H.; Mross, S.; Astruc, D.; Linares, J .; Varret,
F.; Rabaa, H.; Le Beuze, A.; Saillard J .-Y.; Culp, R. D.; Atwood, D. A.;
Cowley, A. H. J . Am. Chem. Soc. 1996, 118, 4133.
(10) Hirsch, S. S.; Bailey, W. J . J . Org. Chem. 1978, 43, 4090.
(11) Schumann, H.; Gorlitz, F. H.; Esser, L. Acta Crystallogr., Sect.
C: Cryst. Struct. Commun. 1993, C49, 598. This X-ray structure was
also solved independently in Rausch’s group; see ref 4 for discussions.
S0276-7333(96)00635-8 CCC: $12.00 © 1996 American Chemical Society

(Pentabenzylcyclopentadienyl)iron Compounds
Organometallics, Vol. 15, No. 26, 1996 5599
Ver sa tile Rea ctivity of th e F e-Br Com p lex 3
a n d Syn th esis of [F e{C5(CH2P h )5}(CO)(S2CNMe2)],
4. We attempted some of the simple reactions known
in the previous series [Fe(C5R5)(CO)2Br], R ) H or CH3,
(1)
namely the replacement of the bromine ligand by CO
14,16
induced by Al2Cl6
and the substitution of Br by
dimethyldithiocarbamate (dtc-) using its sodium salt17
(Na+dtc-). Neither of these reactions work for 3,
although they had proved to be efficient for the C5H5
and C5Me5 analogues. The reaction of 3 with a 100-
bar pressure of CO and Al2Cl6 failed. Likewise, the
reaction of 3 with Na+dtc- in refluxing DME also failed.
Unreacted 3 was recovered almost quantitatively in
each case. Either of these reactions would have ulti-
mately led to the dithiocarbamate complexes. We found,
however, that photolysis of 3 in the presence of Na+dtc-
gives [Fe{C5(CH2Ph)5}(CO)(η2-dtc)] cleanly in 73% yield
of rhombic bright-brown crystals (eq 3).
Bis[(p en t a b en zylcyclop en t a d ien yl)d ica r b on yl-
ir on ] Dim er , 2. The report by Rausch of the stabiliza-
tion of (η4-pentabenzylcyclopentadiene)tricarbonyliron
is of interest since the parent Cp (C5H5) compound is
unstable. In their reaction between 1 and [Fe(CO)5],
Rausch’s group refluxed the reaction mixture for 1 day
in toluene (115 °C). Our conditions were different.
Since we noted that the conditions used with penta-
methylcyclopentadiene, refluxing xylene (145 °C),12-14
provoked a huge amount of decomplexation, we lowered
the reaction temperature to 90 °C and heated the
reaction mixture for 4 days at this temperature. Inter-
estingly, we did not obtain any side product (the
monometallic tricarbonyl complex shows three carbonyl
1
stretches in the 2000 cm-1 region), and the H and 13C
NMR spectra of the reaction product were clean. The
yield under these conditions is 72% of dark-violet
crystals (eq 2).
(2)
(3)
Br om od ica r b on yl(η5-p e n t a b e n zylcyclop e n t a -
d ien yl)ir on (II), 3. The iron-bromo derivatives
[FeII(C5R5)(CO)2Br] are very useful starting materials
to synthesize a series of sandwich15 and piano-stool16
Fe(C5R5) complexes. In 1981, we reported the synthesis
of [FeCp*(CO)2Br] (Cp* ) C5Me5) by clean cleavage of
the dimer in CH2Cl2 using Br2.15 By applying the same
procedure, we synthesized the pentabenzylcyclopenta-
dienyl analogue 3 in 84% yield as orange needles.
Syn th esis a n d Sta bility of 17-Electr on Dith io-
ca r b a m a t e
Com p lexes
[F e{C5(CH 2P h )5}(L)-
(S2CNMe2)]+P F 6-, L ) CO, NCMe. As in the Cp*
series, we attempted to oxidize 4 to the isostructural
17-electron (17e) complex using ferricinium hexafluo-
rophosphate. Although a color change is observed, it is
not possible to isolate this 17e product. The solution
quickly turns dark, indicating decomposition. In the
Cp* series, we had obtained a thermally stable but
extremely reactive, thus interesting, 17e complex.9 It
appears that this complex is not stable in the C5Bz5
series, probably because this ligand is not as electron
releasing as Cp*. The agitation of the phenyl substit-
uents (rapid rotation about the Cp-CH2 axes) also may
not be compatible with the use of weakly-bound ligands,
such as CO, as the legs of the piano-stool, especially in
a cationic 17e carbonyl complex.
(12) King reported the preparation of [FeCp*(CO)2]2, in 48% yield,
synthesized from the reaction of Cp*H with [Fe(CO)5] in refluxing 2,2,5-
trimethylhexane.13,14 King, R. B.; Bisnette, M. B. J . Organomet. Chem.
1967, 8, 287.
(13) For a seminal review on Cp* complexes, see: King, R. B. Coord.
Chem. Rev. 1976, 20, 155.
(14) The reaction was improved to a 65% yield using xylenes at
reflux (2 days): Catheline, D.; Astruc, D. Organometallics 1984, 3,
1094.
(15) Hamon, J . R.; Astruc, D.; Michaud, P. J . Am. Chem. Soc. 1981,
103, 758.
(16) Roma´n, E.; Astruc, D. Inorg. Chem. 1979, 18, 3284.
(17) (a) Catheline, D.; Roma´n, E.; Astruc, D. Inorg. Chem. 1984, 23,
4508. (b) Roma´n, E.; Catheline, D.; Astruc, D. J . Organomet. Chem.
1982, 236, 229.

5600 Organometallics, Vol. 15, No. 26, 1996
Delville-Desbois et al.
We carried out the same oxidation reaction in aceto-
nitrile with the hope that the electron-releasing aceto-
nitrile ligand would replace CO and stabilize the
electron-deficient 17-electron complex. Indeed, in the
Cp* series, the 17e carbonyl complex rapidly reacts with
CH3CN with an associative mechanism to give the 17e
acetonitrile complex.9a The oxidation of 4 in acetonitrile,
after a color change from brown to violet, gave a stable
paramagnetic dark-violet complex 5 that was character-
ized by elemental analysis and infrared and ESR
spectroscopies. This reaction also produced 1 equiv of
ferrocene (eq 4).
(5)
(4)
(dtc) complexes, and Figure 1 shows the ESR spectrum
of 6. The 17e form is also a necessary intermediate in
the thermal decomposition of this complex (see also the
electrochemical discussion).
The data on the C5Bz5 complex 6 taken alone are not
as informative as they are for the Cp* analogue since 6
is thermally less stable, and thus it is not possible to
observe both the 17e and 19e forms at the same time.
This unstability of 6 again seems logical because of the
mutual steric constraints that one encountered in the
rotation of the phenyl rings about the Cp-CH2 axis,
interfering with the rotation of the monodentate dithio-
carbamate about the Fe-S axis, especially in a 17e
complex. However, taken together with the background
knowledge that has been obtained for the Cp* analogue
of 6, the new trends are in accord with the previously
established equilibrium between the 17e and 19e forms,
decomposition proceeding via the 17e form. It was of
interest to verify experimentally that the C5Bz5 ligand
destabilizes paramagnetic complexes contrary to the
Cp* ligand.
Syn th esis a n d Sta bility of Dia - a n d P a r a m a g-
n et ic Bis(d it h ioca r b a m a t e) Com p lexes [F e{C5-
(CH2P h )5}(d tc)2]0/+, 6/7. The acetonitrile complex is
an ideal substrate for the introduction of a second
dithiocarbamate ligand using Na+dtc- in THF. In the
Cp* series, this led to neutral paramagnetic complexes
in which either one (17e complex) or two dtc ligands (19e
complex) acted as chelates.9b The two forms were
interconverting, which led to the first spectroscopic and
electrochemical characterization of a paramagnetic
complex which was observable under both the 17e and
19e forms. The reaction of 5 with Na+dtc- at room
temperature led to decomposition, the neutral bis(dtc)
complex again being less stable than the Cp* analogue.
However, when this reaction was carried out between
-35 and -20 °C, a color change from violet to bright
turquoise-blue was observed (eq 5).
The bis(dtc) complex could be characterized by ESR
spectroscopy. This spectrum shows that the acetonitrile
complex 5 has completely disappeared. The new ESR
spectrum contained three new g values indicating an
orthorhombic distortion that is attributed to the 17e
form. Table 1 contains the g values for the Cp* and
C5Bz5 derivatives of the 17e acetonitrile-dtc and bis-
Although 6 is thermally unstable above -20 °C, its
thermal stability up to -20 °C allowed the synthesis of
the oxidized FeIV cationic complex and, ultimately, the
resolution of the crystal structure of the latter (Figure
2). Indeed, oxidation of 6, as shown in eq 5, could be
achieved using 1 equiv of ferricinium hexafluorophos-
phate. Upon addition of solid [FeCp2)]+PF6- to the THF
solution of 6 at -50 °C, the turquoise-blue color pro-
gressively disappeared upon warming and 7 precipitated

(Pentabenzylcyclopentadienyl)iron Compounds
Organometallics, Vol. 15, No. 26, 1996 5601
as a green solid which was recrystallized (71.2% yield
from 5, eq 6).
F igu r e 1. ESR spectrum of 6 in frozen THF at 12 K.
Ta ble 1. ESR Sp ectr a
compound
R ) CH3
R ) CH2Ph
[Fe(C5R5)(dtc)(NCMe)]PF6 solid state (12 K) solid state (18 K)
g1 ) 2.6722
g2 ) 2.0073
g1 ) 2.5736
g2 ) 2.0362
[Fe(C5R5)(dtc)2]
solid state (18 K) frozen THF (12 K)
g1 ) 2.2680
g2 ) 2.1260
g3 ) 2.0350
g1 ) 2.2707
g2 ) 2.0611
g3 ) 2.0002
Complex 7 is robust; thus, the C5Bz5 ligand is
compatible with the coordination of two dtc ligands
because the metal now has 18 valence electrons and
both dtc ligands are bidentate. Oxidation of the 17e
form of 6 generates a 16e fragment to which the free
sulfur of the monodentate dtc ligand rapidly coordinates
to fill the coordination sphere of the metal. The
stabilization of such a high oxidation state results from
the strong inorganic nature (four Fe-S bonds) of this
complex. Nevertheless, it represents the first family of
stable (even robust) FeIV organometallic complexes.
Electr och em istr y of th e [F e{C5(CH2P h )5}(d tc)L]
Com p lexes. In order to investigate the stability and
behavior of odd-electron pentabenzylcyclopentadienyl
dithiocarbamate complexes, we performed cyclic voltam-
metry experiments for the 18-electron carbonyl complex
4, the 17-electron acetonitrile complex 5, and the 18-
electron bis(dtc) complex 7. The results are summarized
in Table 2 in which we compare the data of the C5Me5
and C5(CH2Ph)5 (C5Bz5) complexes.
redox system can be taken as the thermodynamic E°
value with a good degree of accuracy. The comparison
between the E° values of the Cp* and C5Bz5 complexes
can now be done and shows that C5Bz5 is not as electron
releasing as Cp*, the E° being 0.1-0.3 V more positive
for the C5Bz5 compounds than for the Cp* compounds.
This tendency also holds for 7, but at this time, the E°
values are not directly accessible from the CVs since
the waves are chemically irreversible for the C5Bz5
compounds even at -80 °C, contrary to the Cp* ana-
logue. For all the compounds examined, it is clear that
the chemical reversibility is greatly reduced in the C5Bz5
complexes as compared to the Cp* analogues. This is
related to the failure to isolate some odd-electron C5Bz5
complexes. It confirms that C5Bz5 is not a good ligand
to stabilize odd-electron complexes, contrary to Cp*.
This can be rationalized by the bulk and rotation of the
phenyl ring about the Cp-CH2 axis for the five benzyl
substituents. There may also be mutual steric con-
straints between the Cp* ligand and one or several legs
of the piano-stool. These constraints increase the
energy of the system, which is not often compatible with
the removal of one electron from a bonding or partly
bonding orbital of the complex. This is particularly true
for 6 in which one leg of the piano-stool, the dtc, is itself
Complexes 4 and 5 give two chemically quasirevers-
ible waves corresponding to the FeII/FeIII and FeIII/FeIV
redox systems at both negative and positive E° values,
SCE, respectively. Since these systems are quasi-
reversible, the average between Epa and Epc of each

5602 Organometallics, Vol. 15, No. 26, 1996
Delville-Desbois et al.
F igu r e 2. (a) ORTEP view of the X-ray structure of complex 6, [FeIV(C5Bz5)(dtc)2]. (b) ORTEP view of 6 showing the
clockwise orientation of the benzyl groups, from ref 9b.
Ta ble 2. Electr och em ica l Ch a r a cter iza tion s of th e
Dim eth yld ith ioca r ba m a to-Ir on Com p lexes w ith in
th e Cp * a n d C5Bz5 Ser ies in th e Solven t In d ica ted
in P a r en th eses
two waves are now found on the negative side of the E
scale. After the diffusion control is ensured for the first
wave by verifying that the peak current is linear to the
square root of the scan rate in CV and to 1/t1/2 in
chronoamperometry, the classical criteria are applied
for the choice between an electrochemical irreversible
wave (Ebslow) and a fast heterogeneous electron transfer
followed by a chemical reaction (EbfastCfast). The shift of
the potential for each 10-fold increase in scan rate is
36 mV. Since this is a one-electron system, we conclude
that this is the latter (EbfastCfast) system (theoretical value
at 25 °C, 29 mV). For the study of the Cp* series, we
know that the neutral 19-electron bis(dtc) complex 6
rapidly decoordinates to give a 17e monodentate com-
plex. Since this complex is not observed in the CV,
contrary to the Cp* case, this 17e complex rapidly
decomposes, due to the mutual bulk of the C5Bz5 and
η1-dtc ligand discussed above (eq 7).
L
C5Bz5
Cp*
CO (THF); 4
oxb
E (V/SCE) 0.682
0.57
∆E (mV)
ia/ic
65
0.63
70
1
D° (cm2/s) 1.56 × 10-6 2.7 × 10-4
redb E (V/SCE) -1.81
∆Ep (mV) 90
-1.945
100
ia/ic
0.67
0.92
D° (cm2/s) 3.81 × 10-6 2.4 × 10-4
NCCH3 (NCCH3); 5 oxc
E (V/SCE) 1.165
∆Ep (mV) 160
0.835
120
ia/ic
0.30
0.81
D° (cm2/s) 7.53 × 10-6 2.30 × 10-5
k° (cm/s)
3.40 × 10-3 1.30 × 10-2
redd E (V/SCE) -0.352
-0.570
60
0.62
∆E (mV)
ia/ic
60
0.89
D° (cm2/s) 9.16 × 10-6 2.30 × 10-5
dtc- (CH2Cl2); 7
red1 E (V/SCE) -0.305
∆Ep (mV) irr
-0.248
135
b
Con clu sion
The synthesis of pentabenzylcyclopentadiene has been
improved (85% yield) by slowing down the addition of
dicyclopentadiene to the reaction mixture.
ia/ic
irr
0.89
D° (cm2/s) 2.16 × 10-5 1.07 × 10-5
k° (cm/s)
red2 E (V/SCE) -0.995
irr
4.0 × 10-3
-0.765
210
b
The yield of bis[(pentabenzylcyclopentadienyl)di-
carbonyliron] dimer has been improved to a considerable
extent (72% yield) so that this compound can be the only
reaction product between iron pentacarbonyl and penta-
benzylcyclopentadiene if the reaction is carried out at
90 °C for 4 days.
∆E (mV)
ia/ic
irr
irr
0.88
D° (cm2/s) 3.60 × 10-6 1.46 × 10-5
k° (cm/s)
irr
2.3 × 10-3
a
n-Bu4NBF4 0.1 M; Pt electrode; E° are given vs SCE. L is the
ligand which is bonded to the FeCp(η2dtc)-moiety. E is the average
between Epa and Epc (∼E° if the wave is at least partly chemically
reversible). ∆Ep is equal to Epa - Epc. D° is the diffusion coefficient;
k°: heterogeneous electron-transfer rate constant. b Scan rate: 0.2
The new 18-electron complexes [FeII(C5Bz5)(CO)2Br],
-
3, [FeII(C5Bz5)(CO)(dtc)], 4, and [FeIV(C5Bz5)(dtc)2]+PF6
7, have been synthesized in high yields.
,
V/s. c Scan rate: 0.4 V/s. Scan rate: 0.8 V/s.
d
Pentabenzylcyclopentadiene destabilizes odd-electron
complexes, this effect being especially dramatic when
the metal bears another bulky ligand or an electron-
deficient ligand. The only stable 17-electron complex
which could be isolated was [Fe(C5Bz5)(η2-dtc)(NCMe)]+
PF6- in which the acetonitrile ligand is small, rigid, and
electron releasing.
nonrigid, but rotating about both the Fe-S and S-C
single bonds. Whereas the Cp* analogue of 7 shows
reversibility at room temperature and could be isolated
and characterized in its old-electron state, 7 is irrevers-
ible at -80 °C, which indicates that the odd-electron
state is totally unstable even for a fraction of second at
-80 °C. This chemical irreversibility of 7 for the redox
system FeIV/FeIII also exists for the FeIII/FeII system, and
Cyclic voltammetry of the complexes of the series
[Fe(C5Bz5)(η2-dtc)L]0/+, L ) CO, NCMe or η2-dtc, shows

(Pentabenzylcyclopentadienyl)iron Compounds
Organometallics, Vol. 15, No. 26, 1996 5603
13C NMR spectra were obtained in the pulsed FT mode at
50.327 MHz with a Bruker AC 200 spectrometer. All chemical
shifts are reported in parts per million (δ, ppm) with reference
to the solvent or Me4Si. Cyclic voltammetry data were
recorded with a PAR 273 potentiostat galvanostat. In the CV
experiments, care was taken to minimize the effects of the
solution resistance on the measurements of the peak potentials
(the use of positive feedback iR compensation and dilute
solutions (<10-4 mol/L) kept currents around 1 µA). Thermo-
dynamic potentials were recorded with reference to an aqueous
SCE in various solvents (0.1M n-Bu4NBF4). The diffusion
coefficients were obtained from analysis of the chronoampero-
metric i-t transient. Elemental analyses were performed by
the Center of Microanalytics of the Department of Inorganic
Chemistry of the Technische Universita¨t Mu¨nchen.
P r ep a r a tion . 1,2,3,4,5-P en ta ben zylcyclop en ta -1,3-d i-
en e, [C5(CH2P h )5H], 1. A 500 mL three-necked flask with a
magnetic stirrer was equipped with a Dean-Stark trap with
a reflux condenser, a gas inlet valve, and an addition funnel;
the system was evacuated and back-flushed with argon three
times. In the flask 157 g (1.45 mol) of benzylic alcohol was
reacted with 11.7 g (0.5 mol) of sodium metal that was cut
into small pieces. This procedure was carried out under
stirring and heating to about 120 °C under argon for 12 h.
After this time, 125 mL of degassed diisopropylic benzene was
added, and the mixture was heated to reflux. When the
Dean-Stark trap was filled, 0.8 mL of water was separated.
Then 3.45 g (0.025 mol) of dicyclopentadiene (95%) dissolved
in 25 mL of degassed diisopropylic benzene was added drop-
wise to the refluxing mixture for 2.5 h. The reaction mixture
continued to reflux for 30 h, and then the reaction solution
was cooled to room temperature under argon. The residue
consisted of a white solid which was extracted with 1 L of
diethyl ether and filtered from sodium benzylate. The organic
layer was washed to a neutral pH with 2 L of water, and the
organic phase was degassed and dried over Na2SO4. After
evaporation of the solvent, 75 mL of a yellowish residue
remained in the flask to which the same volume of degassed
n-pentane was added. This was cooled to -20 °C, which
induced crystallization of 21.7 g (85%, lit.: 28%, ref 1, and 62%,
ref 4 of C5(CH2Ph)5H (mp 61-70 °C). A further recrystalli-
zation gave 18.2 g (70%) of white crystals (mp 70-72 °C, lit.
69-75 °C). 1H NMR (CDCl3, 20 °C) δ (ppm): 7.44-7.21 (m,
C6H5, 20H), 6.92-6.88 (m, C6H5, 5H), 4.06, 4.00, 3.78, 3.71
(system AB, CH2C6H5, 4H), 3.65 (s, CH2C6H5, 4H), 3.40 (t,
CHCH2C6H5, 1H), 3.266 (d, CHCH2C6H5, 2H). 13C NMR δ
(ppm): 142.98, 141.43, 140.14, 138.66, (Cq, C6H5 + CdC),
129.18-125.78 (9CH, C6H5), 52.75 (CH-CH2C6H5), 34.31,
33.44, 31.84 (CH2C6H5) .
the two redox couples FeII/FeIII and FeIII/FeIV for these
three complexes. The C5Bz5 ligand is found, from the
compared E° values, to be less electron releasing than
the Cp* ligand. Destabilization of the odd-electron
C5Bz5 complexes is confirmed by the partial chemical
reversibility for 4 and 5 and by the total irreversibility
for 7. The destabilizing effect is found to be dramatic
for 7 (cathodic reduction irreversible even at -80 °C)
and for which the study of the potential variation as a
function of scan rate shows an EbfastCfast system.
Bis[(η5-p e n t a b e n zylc yc lop e n t a d ie n yl)d ic a r b on yl-
ir on (I)], 2. 1,2,3,4,5-Pentabenzylcyclopenta-1,3-diene (4.9 g,
0.01 mol) was dissolved in 30 mL of p-xylene under stirring.
Pentacarbonyliron (2 mL) was added through the reflux
condenser, and the mixture was heated to 90 °C for 4 days.
After consumption of [Fe(CO)5] (24 h), 2 mL more of this educt
was added. Slight traces of an iron mirror were observed.
After the solution was cooled to room temperature, the flask
was carefully opened in order to avoid the explosion of
pyrophoric iron; o-xylene was evaporated in vacuo, and the
solid residue was extracted with n-heptane while scratching
the surface of the flask. The solids were extracted with
Exp er im en ta l Section
Gen er a l Da ta . Reagent-grade tetrahydrofuran (THF),
diethyl ether, and pentane were predried on Na foil and
distilled from sodium benzophenone ketyl under argon just
before use. Acetonitrile (CH3CN) was stirred under argon
overnight with phosphorus pentoxide, distilled from sodium
carbonate, and stored under argon. Methylene chloride (CH2Cl2)
was distilled from calcium hydride under argon just before use.
All other chemicals were used as received. All manipulations
were done by Schlenk technique or in a nitrogen-filled Vacuum
Atmosphere drylab. Infrared spectra were recorded with a
Perkin-Elmer Model 1420 ratio recording infrared spectropho-
tometer which was calibrated with polystyrene. Samples were
examined in solution (0.1 mm cells with NaCl windows)
between NaCl disks in Nujol or in KBr pellet. 1H NMR spectra
were recorded with a Bruker AC 200 (200 MHz) spectrometer.
dichloromethane in
a Kumagawa apparatus in order to
separate the metallic iron from the product. The dark-purple
solution was concentrated, and absolute alcohol was added.
Further evaporation of CH2Cl2 induced crystallization of 4.3
g (72%) of the dark-violet product. 1H NMR (CDCl3, 20 °C) δ
(ppm): 7.050 (m, 20H) 6.795 (m, 20H) 3.680 (s, 10H, CH2C6H5).
13C NMR δ: 206.74 (C-O), 138.57 (Cq, C6H5), 129.10, 128.07,
126.10 (C6H5), 102.75 (C5Bz5), 30.45 (CH2). IR (Nujol, ν cm-1):
1930 (CO, terminal), 1740 (CO, bridging). Anal. Calcd for
C
84H70Fe2O4: C, 80.38; H, 5.62; Fe, 8.90; O, 5.10. Found: C,
80.03; H, 5.64; Fe, 8.93; O, 5.53.

5604 Organometallics, Vol. 15, No. 26, 1996
Delville-Desbois et al.
Br om od ica r b on yl(η5-p en t a b en zylcyclop en t a d ien yl)-
ir on (II), 3. 2 (1.0018 g, 0.8 mmol) was dissolved in 50 mL of
CH2Cl2 in a dry three-necked flask with a magnetic stirrer and
addition funnel under argon; 41.6 µL (0.8 mmol) of bromine
was added within 2 h under protection from the daylight. Thin
layer chromatography with a 3:2 mixture of n-pentane and
CH2Cl2 indicated that the reaction was finished after this time.
The mixture was concentrated and passed through an alumina
column with CH2Cl2 under protection from the daylight. To
the deep-red solution absolute ethanol was added. Evapora-
tion of the solvent precipitated 968 mg (84.4%) of bright-orange
needles. 1H NMR (CDCl3, 20 °C) δ (ppm): 7.105 (m, 15H),
6.629 (m, 10H), 3.661 (s, 10H). 13C NMR δ (ppm): 213.26 (C-
O), 137.65 (Cq, C6H5), 128.74, 128.44, 126.68 (CH, C6H5), 100.65
(C5Bz5), 31.23 (CH2). IR (Nujol, ν cm-1): 2020, 1980 (CO).
Anal. Calcd for C42H35FeO2Br: C, 71.30; H, 4.99; Fe, 7.89.
Found: C, 71.35; H, 5.02; Fe, 9.1.
(Nujol, ν cm-1): 1550 (CN), 1030 (CS). ESR (solid state
sample, 18 K): g1 ) 2.574, g2 ) 2.036. Cyclic voltammetry,
see text and Table 2. Anal. Calcd for C45H44FeN2PS2F6: C,
61.57; H, 5.05; N, 3.19. Found: C, 61.52; H, 5.18; N, 3.10.
In Situ P r ep a r a tion of Bis(d im eth yld ith ioca r ba m a to)-
(η5-p en ta ben zylcyclop en ta d ien yl)ir on (III) in THF Solu -
tion , 6. In a dry Schlenk flask, 228 mg (0.26 mmol) of 5 was
dissolved in 20 mL of dry THF at -35 °C. To the cooled
solution was added 46.6 mg of predried Na+dtc- in 10 mL of
THF, prepared under the same conditions, inducing a spon-
taneous color change from dark purple to blue-turquoise. After
15 min, an ESR sample of this air- and thermally-sensitive
complex was taken under argon from the Schlenk flask. ESR
(THF frozen solution, 12 K): g1 ) 2.2707, g2 ) 2.0611, g3
2.0002.
)
Bis(η2-d im e t h yld it h ioca r b a m a t o)(η5-p e n t a b e n zyl-
cyclop en ta d ien yl)ir on (IV) Hexa flu or op h osp h a te, 7. To
a solution of 6, prepared as described above, 86.1 mg (0.26
mmol) of ferricinium hexafluorophosphate was added under
a counterstream of argon at -50 °C. The solution was allowed
to warm to room temperature for at least 2 h. While the
solution was warmed, the blue-turquoise color of 6 disappeared
and a green solid precipitated. After evaporation of the THF,
the residue was extracted with ether giving a stoichiometric
amount of ferrocene, after filtration and removal of the solvent.
The solid which remained on the frit was extracted with 10
mL of dichloromethane. The dark-green solution was washed
three times with 10 mL of water and dried over sodium sulfate.
To the filtered CH2Cl2 phase was added 10 mL of absolute
alcohol, and after partial evaporation of the halogenated
solvent, 177.2 mg (71.2%) of crystalline complex 7 was
obtained. 1H NMR (CD3CN, 20 °C) δ (ppm): 7.019-6.538 (m,
25H, C6H5), 3.458 (s, 12H, N(CH3)2), 3.184 (s, 10H, CH2C6H5).
13C NMR δ (ppm): 198.67 (C-S), 134.62 (Cq, C6H5), 129.20,
128.32, 126.86 (C6H5), 107.43 (C5Bz5), 38.35 (CH3), 31.85 (CH2).
IR (Nujol, ν cm-1): 1545 (CN), 1020 (CS), 810 (PF6-). Anal.
Calcd for C45H47FeS4N2PF6: C, 57.74; H, 4.95; N, 2.93.
Found: C, 57.60; H, 5,25; N, 2.76. Cyclic voltammetry: see
text and Table 2. MS (FAB/NBA): m/e (811, M - PF6, 27),
691 (M - PF6 - dtc, 89), 296 (M - PF6 - C5Bz5, 100), 91, (M
- (C5Bz4)Fe(dtc)2 PF6, 49).
Ca r bon yl(η2-d im eth yld ith ioca r ba m a to)(η5-p en ta ben -
zylcyclop en ta d ien yl)ir on (II), 4. To 500 mg (0.71 mmol) of
3 in a Schlenk flask with a cooling mantle was added 126.6
mg of predried sodium dimethyldithiocarbamate in dry THF
under argon. The solution was irradiated with visible light
(Philips, 120 W) for 30 min. The color of the reaction mixture
changed from red to deep brown-red with precipitation of a
white solid. Completion of the reaction was verified by thin
layer chromatography on SiO2 with a 7:2 mixture of n-pentane
and CH2Cl2. The solution was filtered into a second Schlenk
flask over Al2O3 under argon and then evaporated. The orange
crude product was introduced into a glovebox and recrystal-
lized from diethyl ether and n-pentane yielding 370 mg (73%)
of crystallized substance. 1H NMR (CDCl3, 20 °C) δ (ppm):
7.020 (m, 15H, C6H5), 6.797 (m, 10H, C6H5), 3.553 (s, 10H,
CH2C6H5), 3.247 (s, 6H, N(CH3)2). 13C NMR δ (ppm): 218.11
(C-O), 208.38 (C-S), 139.29 (Cq, C6H5), 128.76, 127.99, 125.81
(CH, C6H5), 93.59 (C5Bz5), 38.40 (CH3), 31.64 (CH2). IR (Nujol,
ν cm-1): 1915 (CO). Anal. Calcd for C44H41FeS2NO: C, 73.42;
H, 5.74; N, 1.95. Found: C, 72.83; H, 5.90; N, 1.86. Cyclic
voltammetry: see text and Table 2.
Acet on it r ile(η2-d im et h yld it h ioca r b a m a t o)(η5-p en t a -
ben zylcyclop en ta d ien yl)ir on (III) Hexa flu or op h osp h a te,
5. In a dry Schlenk flask 242 mg (0.34 mmol) of 4 and 111.4
mg (0.34 mmol) of ferricinium hexafluorophosphate were
mixed in the solid state under argon; 20 mL of degassed
acetonitrile was transferred to these educts inducing the
immediate apparition of a dark-purple color. After this was
stirred for 15 min at room temperature, about half of the
solvent was removed and the concentrated solution was
extracted under argon with small amounts of dry n-pentane
until no more yellow color was observed in the pentane layer.
From the evaporated pentane solution, 62 mg (0.33 mmol) of
ferrocene, identified by its 1H NMR spectrum was recovered.
The reaction solution was evaporated and the solid residue
was introduced into a glovebox. Recrystallization from ether/
n-pentane gave 213 mg (71.4%) of dark-purple product. IR
Ack n ow led gm en t . We thank Dr. J . M. Dance
(ICMCB, Universite´ Bordeaux I) for ESR assistance and
helpful discussions, Dr. B. Barbe (CESAMO, Universite´
Bordeaux I) for performing the NMR experiments, the
Stiftung des Deutsches Volkes for a predoctoral grant
to S.M., and the Institut Universitaire de France, the
Universite´ Bordeaux I, the Centre National de la
Recherche Scientifique, and the Re´gion Aquitaine for
financial support.
OM9606350