Facile synthesis of lanthanidocenes by the borohydride/alkyl route and their application in isoprene polymerization

Article abstract of DOI:10.1002/ejic.201000184

Several borohydridolanthanidocenes were synthesized by means of a straightforward approach, starting from equimolar amounts of Ln(BH ₄)₃(thf)₃ (Ln = Nd, Sm) and n-butylethyl-magnesium (BEM) in the presence of 2 equiv. of a cyclopentadiene derivative. Depending on the nature of the cyclopentadiene, monometallic Cp*₂Nd(BH₄)(thf) (Cp* = C₅Me ₅) (1), heterodimetallic [Cp₂Nd.(BH₄)(μ- BH₄)]₂Mg(thf)₄ (Cp = C₅H ₅) (2), [(CMe₂C₅H₄)2Ln(BH ₄)(μ-BH₄)]₂Mg(thf)₃ (Ln = Nd, 3a; Ln = Sm, 3b), and monometallic anilidocyclopentadienyl compounds (C ₅Me₄CH₂SiMe₂NPh)Ln(BH ₄)(thf)₂ (Ln = Nd, 4a; Ln = Sm, 4b) were isolated. The complexes were characterized by ¹H NMR spectroscopy, elemental analysis, and X-ray structural analysis. Mg(BH₄)₂(thf) ₃ (S) was isolated as a byproduct and structurally characterized, revealing an unprecedented structural arrangement. In combination with BEM, complexes 2a, 3a, and 4a were found to be efficient in isoprene polymerization to afford highly stereoregular transpolyisoprene. Similar results were obtained by an all in situ approach starting from Nd(BH₄) ₃(thf)₃/BEM mixtures in the presence of 2 equiv. of cyclopentadiene derivative.

Full text of DOI:10.1002/ejic.201000184

FULL PAPER
DOI: 10.1002/ejic.201000184
Facile Synthesis of Lanthanidocenes by the “Borohydride/Alkyl Route” and
Their Application in Isoprene Polymerization
Marc Visseaux,*[a,b] Michael Terrier,[a,b] André Mortreux,[a,b] and Pascal Roussel[a,b]
Keywords: Lanthanides / Metallocenes / Structure elucidation / Polymerization / Homogeneous catalysis
Several borohydridolanthanidocenes were synthesized by
means of a straightforward approach starting from equimolar
amounts of Ln(BH4)3(thf)3 (Ln = Nd, Sm) and n-butylethyl-
= Sm, 4b) were isolated. The complexes were characterized
by 1H NMR spectroscopy, elemental analysis, and X-ray
structural analysis. Mg(BH4)2(thf)3 (5) was isolated as a by-
magnesium (BEM) in the presence of 2 equiv. of a cyclopen- product and structurally characterized, revealing an unprec-
tadiene derivative. Depending on the nature of the cyclopen- edented structural arrangement. In combination with BEM,
tadiene, monometallic Cp*2Nd(BH4)(thf) (Cp* = C5Me5) (1),
heterodimetallic [Cp2Nd(BH4)(µ-BH4)]2Mg(thf)4 (Cp = C5H5)
(2), [(CMe2C5H4)2Ln(BH4)(µ-BH4)]2Mg(thf)3 (Ln = Nd, 3a; Ln
= Sm, 3b), and monometallic anilidocyclopentadienyl com-
pounds (C5Me4CH2SiMe2NPh)Ln(BH4)(thf)2 (Ln = Nd, 4a; Ln
complexes 2a, 3a, and 4a were found to be efficient in iso-
prene polymerization to afford highly stereoregular trans-
polyisoprene. Similar results were obtained by an “all in
situ” approach starting from Nd(BH4)3(thf)3/BEM mixtures in
the presence of 2 equiv. of cyclopentadiene derivative.
stoichiometric amount of a dialkylmagnesium reagent.[5]
Using this strategy, we were able isolate in high yield a new
family of heterodimetallic trinuclear half-sandwiches of gene-
ral formula [CpRLn(BH4)3]2[Mg(thf)6] (CpR = C5H4, CpPh3,
C5Me5; Ln = Nd, La). We then wanted to validate this syn-
thetic strategy and to extend it to the preparation of lanthan-
idocenes. For this purpose, we chose four different cyclopen-
tadienyl-type ligands: the usual Cp*H (Cp* = C5Me5), the
nonsubstituted CpH (Cp = C5H5), an ansa ligand bearing a
tetra(methyl)ethylene bridge (CMe2C5H5)2, and an anilido-
cyclopentadienyl ligand HC5Me4CH2SiMe2NHPh recently
prepared by Okuda et al.[6] These cyclopentadiene reagents
were chosen in order to generalize our approach as an ef-
ficient synthetic method.
We describe in this contribution the synthesis and char-
acterization of several new borohydridolanthanidocenes
(For reasons of convenience, the term “lanthanidocenes”
refers to all mono(borohydrido) disubstituted complexes
described in this manuscript, including 4a and 4b.) obtained
by this B/A route. All complexes were structurally charac-
terized. Some results in isoprene polymerization involving
isolated and in situ synthesized complexes are also pre-
sented.
Introduction
Organometallic chemistry of the rare earths has been the
subject of increasing interest over the past two decades,
with high impact in the field of molecular catalysis[1] and
particularly polymerization.[2] Most organometallic synthe-
ses are carried out by ionic metathesis and with the very
common trihalide derivatives LnX3 (X = Cl, Br, I) as start-
ing materials, but we and others have shown that the tris-
(borohydride)s Ln(BH4)3(thf)n (n = 2,3)[3] could be a valu-
able alternative in this frame.[4] More soluble, the latter pre-
cursors allowed the isolation, by ionic metathesis, of a
growing assortment of organometallic structures, including
mono- or bis(cyclopentadienyl)-supported borohydrides.[3,4]
It was also observed that, depending on the substituents on
the Cp group, half-sandwich borohydrido complexes, which
were found to be more stable than their halo homologues
versus comproportionation reactions, could be easily iso-
lated.[4b,4c]
We just developed an advantageous one-pot synthetic
pathway for the preparation of organolanthanides, which
starts from the tris(borohydride)s, as an alternative to the
classical ionic metathesis. This approach, which we called the
“borohydride/alkyl route” (or B/A route), consists of the re-
action of a Ln(BH4)3(thf)3 compound, a cyclopentadienyl
reagent in its protonated form (i.e. cyclopentadiene), and a
Results and Discussion
[a] Université Lille Nord de France,
59000 Lille, France
Synthesis of Complexes 1–4
[b] Unité de Catalyse et de Chimie du Solide (UCCS, UMR 8181
CNRS), ENSCL,
The mono(borohydrido) disubstituted complexes,
(C5Me5)2Nd(BH4)(thf) (1), [(C5H5)2Nd(BH4)(µ-BH4)]2Mg-
(thf)4 (2), [(CMe2C5H4)2Ln(BH4)(µ-BH4)]2Mg(thf)3 (Ln =
Nd, 3a; Ln = Sm, 3b), and (C5Me4CH2SiMe2NPh)Ln-
Bât. C7, Cité Scientifique, B. P. 90108, 59652 Villeneuve d’Ascq
Cedex, France
Fax: +33-3-20336585
E-mail: marc.visseaux@ensc-lille.fr
Eur. J. Inorg. Chem. 2010, 2867–2876
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2867

M. Visseaux, M. Terrier, A. Mortreux, P. Roussel
FULL PAPER
(BH4)(thf)2 (Ln = Nd, 4a; Ln = Sm, 4b), were prepared freshly cracked cyclopentadiene alone, even at 70 °C for
starting from the protonated form HCpR (given for a gene- prolonged reaction times. On the bulk scale, we found that
ral substituted cyclopentadiene) by the straightforward re- it was necessary to perform the reaction in the presence of
action described in Scheme 1, that is, addition of one equiv- thf in order to improve the yield of isolated product. After
alent of BEM (n-butylethylmagnesium) to a solution of experimental work-up, blue microcrystalline solid 2 was iso-
Ln(BH4)3(thf)3 in the presence of 2 equiv. of HCpR.
lated in high yield (91%). According to an X-ray analysis
of suitable crystals obtained after recrystallization (thf/pen-
tane), the molecular formula was found as [(C5H5)2-
Nd(BH4)(µ-BH4)]2Mg(thf)4 (Scheme 1). Elemental analysis
was found to match a tris(thf) adduct, which denotes the
tendency to lose one thf molecule (see further). In the case
of the ansa-cyclopentadienyl ligand (CMe2C5H4)2, the reac-
tion afforded, after extraction of the crude product with thf,
crystalline complex [(CMe2C5H4)2Nd(BH4)(µ-BH4)]2Mg-
(thf)3 (3a, Scheme 1) (from X-ray data and elemental analy-
sis) in good yield (70%). Samarium analogue 3b (Scheme 1)
could also be isolated and characterized by X-ray crystal-
Scheme 1. Synthesis of complexes 1–4 by the B/A route.
1
lography and H NMR spectroscopy. However, it was not
possible to get a satisfactory elemental analysis result for
this compound, possibly because of a partial loss of a
Mg(BH4)2 unit. To confirm the molecular structure of 3b,
the latter was treated with 2 equiv. of allylLi(dioxane) in
the NMR tube (C6D6). As expected, a bis(allyl) derivative
(Scheme 3), a homologue of the dimethoxyethane adduct
previously described by some of us,[7] was obtained immedi-
ately and quantitatively.
1
As shown by preliminary H NMR spectroscopy-moni-
tored syntheses, the reaction was in every case immediate
and quantitative. Running the reaction in bulk is very sim-
ple, since no refluxing is necessary: the mixture is stirred at
room temperature for a couple of hours to make sure that
the reaction is complete (the reaction is immediate on the
NMR scale) and until the color change is obvious.
1
With HCp* (Cp* = C5Me5), the H NMR spectrum of
the resulting compound showed the presence of a com-
pound containing two Cp*, one BH4, and approximately
four thf ligands. This spectrum was identical to that of the
neodymocene Cp*2Nd(BH4)(thf) (1, Scheme 2), which had
been previously and alternatively prepared by ionic metath-
esis and structurally characterized.[4d] From the bulk syn-
thesis, crystals of this neutral monometallic complex were
isolated in low yield (24%), along with intractable blue oil,
which did not afford more crystals.
Scheme 3. Formation of a bis(allyl) derivative from 3b.
Finally, by using the B/A route, complexes 4a and 4b
(Scheme 1) were isolated from the reaction in thf of the
starting tris(borohydride) with the aniline-tethered cyclo-
pentadiene ligand (C5Me4)CH2SiMe2NHPh recently de-
scribed by the group of J. Okuda.[6] X-ray structure deter-
mination established that both complexes are mononuclear
and monomeric.
X-ray Structures of Complexes 2–4
Complex 2 is a very rare example of nonsubstituted cy-
clopentadienyl lanthanidocene in the early series.[1a] It is a
dimetallic trinuclear Nd2Mg compound that crystallizes in
the monoclinic P21/c space group (see Figure 1). The unit
cell comprises two nonsymmetrical Nd2Mg subunits, in
which each neodymium is linked to the magnesium atom
through a mono-BH4 bridge, with a quasi-linear arrange-
Scheme 2. Molecular structures of complexes 1–4.
ment of the three metal atoms [Nd3–Mg1–Nd1
=
176.29(7)°]. The Nd–Mg distances are between 5.518(4) and
The same strategy was undertaken with the nonsubsti- 6.008(4) Å, which accounts for a rather covalent trimetallic
tuted C5H5H. Prior to the synthesis of a bis(cyclopentadi- compound, in contrast to observations in the corresponding
enyl) compound, we first checked by 1H NMR spec- half-sandwich series, where discrete ionic moieties were
–
troscopy (C6D6) that Nd(BH4)3(thf)3 did not react with clearly identified [(C5H5)Nd(BH4)3 , Mg(thf)62+, (BH4)3-
2868
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2867–2876

Application of Lanthanidocenes in Polymerization
Nd(C5H5)–, Nd–Mg distance 7.22 Å].[5a] However, the two for 3a), and this Mg(BH4)2 moiety appears to be almost
Nd–B–Mg bridges are found to be clearly dissymmetric symmetrical [Mg–B distances: 2.467(12) and 2.511(12) Å in
[Mg1–B1 2.656(11) Å, Mg1–B7 3.253(10) Å], and if the 3a]. Since, on the other hand, the two Nd–B distances
Nd–B distances fall in the commonly observed range for (bridging BH4) are approximately equal to each other as
(µ-BH4) dimers,[4a,8] one can observe one particularly short well (mean value 2.80 Å), one can think that each BH4
Nd–B bridge length [Nd3–B7 2.738(10) Å]. Therefore, the group is shared equivalently between Mg and Nd atoms,
molecular arrangement is better described as: [(C5H5)2Nd- with shorter intermetallic distances [4.888(4) and
(BH4)2][Mg(BH4)(thf)4][(C5H5)2Nd(BH4)].
5.066(4) Å] relative to those in complex 2. Therefore, the
molecular arrangement in 3a (and 3b) can be seen as in-
cluding a more covalent Mg(BH4)2 unit than that in 2, that
is, (CMe2C5H4)2Nd(BH4)(µ-BH4)Mg(thf)3(µ-BH4)(BH4)-
Nd(C5H4CMe2)2. In both cases, the lanthanide (Nd or Sm)
bears a second BH4 group in a terminal η3 mode, allowing
a short mean Ln–B (terminal) distance of approximately
2.62 Å (2.59 Å for 3b). Mg–O distances are found to be
smaller in 3a (mean value 2.04 Å) than in complex 2. The
CP–Nd–CP (CP is the centroid of a cyclopentadienyl ring)
angles (115.4 and 116.0° in 3a, 116.1 and 117.1° in 3b) are
comparable to those observed in other complexes in this
ansa series[10] and are, as expected, lower than those in com-
plex 2 (mean value 121.4°), which clearly demonstrates the
constrained effect of the (CMe2–CMe2) link. Finally, it
should be noted that the two cyclopentadienyl cycles have
an almost eclipsed conformation when viewed perpendicu-
lar to the [Nd(BH4)2] plane, as already observed with such a
ligand,[10] with a mean torsion angle C(Cp)–CMe2–CMe2–
C(Cp) of 38.20° (37.59° for 3b).
Figure 1. One of the two entities of the asymmetric unit showing
the molecular structure of 2. Hydrogen atoms are omitted for clar-
ity. Selected distances [Å] and angles [°]: Mg1–Nd1 5.518(4), Mg1–
Nd3 5.918(4), Nd1–B1 2.905(10), Nd1–B2 2.657(12), Nd3–B3
2.652(12), Nd3–B7 2.738(10), Mg1–B1 2.656(11), Mg1–B7
3.253(10), Nd3–Mg1–Nd1 176.29(7).
Mg–B distances are in the range 2.65–3.33 Å [Mg1–B1
2.656(11), Mg1–B7 3.253(10), Mg2–B5 2.613(13), Mg2–B6
3.332(13) Å]. The consequence of one long Mg–B distance
in each subunit is the presence of four molecules of thf co-
ordinated to the magnesium atom in 2 (Mg–O mean value
2.06 Å), which denotes a partial ionic character of the mag-
nesium moiety (in contrast to complex 3a, see further),
while six thf molecules were necessary to complete the coor-
dination sphere in the true cationic Mg moiety of ionic half-
lanthanidocenes.[5] Despite the fact that the hydrogen atoms
could not be located due to the poor quality of the crystal,
Nd–B distances clearly indicate that the bridged BH4 here
are η2 rather than η3 coordinated, whereas each neodym-
ium center in 2 bears one additional terminal BH4, most
likely trihapto according to the Nd–B distance observed
(Nd–B is typically between 2.62 and 2.66 Å for an η3 coor-
dinated BH4).[4,9]
Figure 2. Asymmetric unit showing the molecular structure of 3a.
Non-BH4 hydrogen atoms are omitted for clarity. Selected distances
[Å] and angles [°]: Mg1–Nd1 5.066(4), Mg1–Nd2 4.888(4), Nd1–B2
2.834(14), Nd1–B1 2.629(10), Nd2–B3 2.613(8), Nd2–B4 2.773(13),
Mg1–B4 2.511(12), Mg1–B2 2.467(12), Nd1–Mg1–Nd2 88.29(5).
Complexes 3a and 3b {[(CMe2C5H4)2Ln(BH4)(µ-BH4)]2-
Mg(thf)3 Ln = Nd, Sm} are isostructural, and both crys-
¯
tallize in the P1 space group. Both can be described as di-
metallic trinuclear structures (Figure 2), as complex 2, but
the trinuclear arrangement in 3a (or 3b) is bent, with a Nd–
Mg–Nd angle of 88.29(5)° [88.68(6)° in 3b], in sharp con-
Complexes 4a and 4b [(C5Me4CH2SiMe2NPh)Ln(BH4)-
trast to that observed in 2. This is likely related to the (thf)2, Ln = Nd, 4a; Ln = Sm, 4b] crystallize in the Pnma
pentacoordinate trigonal bipyramidal geometry of the mag- space group and are isostructural. The molecular structure
nesium center, which bears only three thf molecules (one shows a distorted square pyramidal arrangement, similar
additional thf was found in 2). The magnesium is linked to to that observed with yttrium,[6] with a base around the
both lanthanide centers through a typical η2 BH4 bridge lanthanide atom comprising two thf, one BH4, and one ani-
(hydrogen atoms belonging to BH4 groups could be located lido ligand (Figure 3).
Eur. J. Inorg. Chem. 2010, 2867–2876
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2869

M. Visseaux, M. Terrier, A. Mortreux, P. Roussel
FULL PAPER
Scheme 4. Possible pathways leading to the mixed Ln/Mg lanthan-
idocenes.
Figure 3. Molecular structure of 4a (symmetrical subunit, top view,
left; side view, right). Non-BH4 hydrogen atoms are omitted for
clarity. Selected distances [Å] and angles [°]: Nd1–B1 2.706(8),
Nd1–O1 2.534(4), Nd–N 2.362(5), Nd–CP 2.42, N–Nd–B 142.5(2),
Nd–N–Cipso 118.9(4), C4–Si–N 107.4(3).
To probe which of the two possible pathways occurs, we
ran a reaction according to path A in an NMR tube start-
ing with bis(cyclopentadienyl)magnesium. One equivalent
of Nd(BH4)3(thf)3 was added to Cp2Mg in C6D6. The reac-
tion was found to be very slow (see Experimental Section)
and needed 48 h at 50 °C to finally lead to a mixture of
Cp3Nd (40%) and Cp2Nd(BH4) (60%). On the other hand,
we had already observed by NMR spectroscopic monitor-
ing that the reaction of Nd(BH4)3(thf)3 with BEM affords
a rapid darkening of the solution, along with formation of
α-olefins, probably resulting from β-H elimination.[14] Thus,
since it is well known that LnR species instantaneously re-
act with cyclopentadienes,[13a] it is likely that the lanthani-
docenes presented in this work are formed according to
path B.
The B/A Route
The one-pot strategy used to prepare the lanthanido-
cenes (and related compounds) described in this paper dis-
plays several advantages: (1) it involves the easily available
tris(borohydride)s as starting materials; (2) the tedious
preparation of the anionic CpRM reagent (M = alkaline
metal) used for ionic metathesis can be avoided; (3) it re-
quires short-time reactions at room temperature and affords
quantitative yields; (4) work-ups are simplified, as there is
no inorganic precipitate to eliminate.
Straightforward alternative synthetic routes to circum-
vent the drawbacks of ionic metathesis, that is, formation
of salts and disproportionation reactions as a result of the
use of thf, have been developed in the literature. The “sil-
ylamide route”,[11] described by Anwander, involves amido
starting materials, and this methodology requires heating
toluene at reflux for a long time. More recently, the same
author published another approach starting from tretraalk-
ylaluminolanthanide compounds, which also allows the
preparation of complexes at room temperature within short
reaction times.[12] However, it was recalled in this paper that
the volatile byproduct AlMe3 reacts violently when exposed
to air. Other methods use hydrocarbyl LnR3 derivatives,[13]
which are known to be thermally sensitive and hence not
easy to store and manipulate, in contrast to the very stable
tris(borohydride)s.
Heterodimetallic vs. Monometallic Complexes
The results presented above show that the molecular
structure of complexes 1–4 strongly depends on the nature
of the precursor ligand, affording trinuclear dimetallic or
monometallic compounds. Regarding the dimetallic com-
plexes, a Mg(BH4)2(thf)n moiety is found to be included in
the structure, under a rather ionic “Mg(BH4)(thf)4+” (in 2),
or neutral Mg(BH4)2(thf)3 (in 3a,b) form. It is noteworthy
that the half-sandwich compounds already prepared using
the same B/A route all display an ionic [CpRLn(BH4)2]–-
[Mg(thf)6]2+[(BH4)2LnCpR]– form, as mentioned previously.
Mixed lanthanide/magnesium complexes are rather
scarce in the organometallic chemistry of the lanthanides
and result mostly from syntheses involving Mg alkylating
reagents in stoichiometric amounts. Actually, a small
number of them have been structurally characterized.
Moreover, catalytic combinations in polymerization reac-
tions involving a lanthanide-based precatalyst associated to
a magnesium cocatalyst have been largely developed since
the 1980s, and heterodimetallic species have often been pro-
posed as acting as active species.[4d,14,15] As far as we know,
Mechanism of Formation by the B/A Route
In order to ensure the formation of the desired product, just one example resulting from a catalytic combination, [Nd-
the dialkylmagnesium reagent must be added to a solution (thf)(µ3-OtBu)2(µ2-OtBu)2(OtBu)Mg2(CH2SiMe3)2], could
of lanthanide tris(borohydride) in the presence of the cyclo- be isolated.[16] Regarding organometallic syntheses, mono-
pentadiene reagent. Two mechanistic pathways can be en- (cyclopentadienyl)lanthanide complexes have been cited as
visaged to rationalize the formation of the (dimetallic) com- incorporating a MgX2 molecule (X = halide), but the struc-
plexes, as represented in Scheme 4.
ture of these complexes remain unclear because crystallo-
2870
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2867–2876

Application of Lanthanidocenes in Polymerization
graphic studies were not carried out.[17] In the ansa series,
the reaction of [CMe2(C5H4MgCl)2(thf)4] (in 50% excess)
with SmCl3 in thf was reported to afford [(CMe2C5H4)2-
Sm(µ-Cl)2MgCl(thf)3(thf)]2.[10b] In this complex, the dimer
comprises two Mg–Sm subunits, linked together through
Sm(µ-Cl)2Sm bridges. The MgCl2 moiety is not shared be-
tween the two samarium atoms, as in complex 3a (or 3b),
but each magnesium atom is associated to its samarium
counterpart through a bis(chlorido) bridge and bears three
additional thf ligands. The Mg–Sm distance is 4.26(4) Å,
revealing a strong interaction through the Cl bridges. By
varying the experimental conditions (1:1 ansa ligand vs.
Sm), the same authors isolated a monometallic (i.e. without
Figure 4. Molecular structure of 5. Hydrogen atoms are omitted
for clarity. Selected distances [Å] and angles [°]: Mg–B1 2.438(7),
Mg–B2 2.496(7), Mg–O (average) 2.07, B1–Mg–B2 128.7(3).
any Mg included) samarocene.[10c] This emphasizes the oc- towards the lanthanide center. It is to be noted that, with
currence (and thus the difficulty of predicting) of the disso- the electron-rich Cp* and using the B/A route as synthetic
ciation of a MgX2(thf)n moiety from a dimetallic associa- strategy, monometallic Cp*2Nd(BH4)(thf) complex 1 was
tion in the (CMe2C5H4)2 series, as we observed during the isolated. The BH4 group seems to bring enough electronic
synthesis of 3a,b (vide supra). A few years ago, in the same density along with the two Cp* groups to ensure the forma-
ansa series, the presence of a MgCl2 unit was clearly estab- tion of neutral compound 1, in contrast to its chlorido
lished by XPS analysis in the molecular formula of allyl homologue, which is generally obtained in an “ate”
complexes bearing this bridging ligand.[18] Unfortunately, Cp*2NdCl2M(ether)2 form (M
=
alkaline metal)
none of them yielded suitable crystals for X-ray studies. [Cp*2NdCl(thf) has, however, been cited in the litera-
Other allyllanthanide complexes including MgCl2,[19] or ture].[1a] The molecular structure of 1, prepared by ionic
more recently Mg(allyl)2,[20] were also described. The mag- metathesis, was solved by us recently.[4d] The CP–Nd–CP
nesium moiety does not act as a bridging ligand between angle (133.18°) may prevent the formation of an ionic ate
two lanthanide atoms in the latter compound, in contrast complex, when compared with less substituted neodymo-
to observations in the present work. Moreover, depending cenes, though the “Cp*2NdCl2” unit is highly stable in this
on the experimental conditions, mixed Ln/Mg dimetallic or series. The isolation of neutral monometallic 1 by the B/A
monometallic complexes were isolated as well.
route thus most likely proceeds by loss of a [Mg(BH4)2]
In the course of our previous studies devoted to borohyd- unit, which compares well with the different behavior of the
ridolanthanide-based polymerization catalysts, it was pos- neodymium complex in the ansa series (see above, molecu-
tulated that Nd(µ-BH4)Mg species, coming from the binary lar structure of 3a,b).
Ln/MgR2 catalytic systems used,[4d,14,15c,15d,15e] could be re-
Complexes 4a and 4b exist as mononuclear bis(thf) ad-
lated to the high trans-selectivity observed for isoprene po- ducts. The coordination number (CN) of the lanthanide
lymerization. The unprecedented isolation of complexes 2, atom is 9 in complexes 1, 2, and 3a (3b). Complex 4a (4b)
3a, and 3b supports well the hypothesis that a borohydride displays the same CN value if one considers the linked anili-
bridge can indeed efficiently link Nd and Mg moieties. One docyclopentadienyl ligand as a L3X2 one.[23] The Nd–N dis-
can note that with the poor electron-donating C5H5 or tance in 4a is quite long [2.362(5) Å] and slightly higher
(C5H4CMe2)2, the magnesium moiety is retained within the than usual ones (ca. 2.20–2.30 Å with late lanthanoids) for
molecular structure. However, during attempted crystalli- linked amidocyclopentadienyl complexes.[6,23,24] Likewise,
zations of 3a, large colorless crystals of a byproduct were Nd–O and Nd–CP distances are rather long as well (average
1
also isolated. This compound was further identified, by H 2.53 and 2.42 Å, respectively) when compared to those
NMR spectroscopy and X-ray diffraction studies, as found for the yttrium analogue.[6] These structural features
Mg(BH4)2(thf)3 (5) (Figure 4). As the cell parameters were do not speak in favor of a poor electron density at the metal
different from those of the published data,[21] a full X-ray center, in contrast to what is generally observed in the Cp-
structure determination was carried out. We found that the amido “CGC” family. It is noteworthy that isolated com-
isolated compound corresponded to the same molecular plexes 4a and 4b are “Mg(BH4)2-free”: the anilido moiety
formula, Mg(BH4)2(thf)3, but in a different unprecedented obviously hampers the coordination of the latter. As a con-
crystalline form.
sequence, the Nd(BH4) moiety is trihapto, and it is to be
¯
Compound 5 crystallizes in the P1 group. BH4 hydrogen mentioned that, as observed during the preparation of 3a,
atoms could be located despite the low quality of the crystal Mg(BH4)2(thf)3 was obtained as a byproduct, under the
(5 was found to be highly reactive towards moisture), and same unprecedented crystalline form as before. Finally,
Mg–B distances fit well with a dihapto Mg(µ-H)2(BH2) ter- since there is obviously enough space to coordinate a
minal BH4 ligand.[22] It is therefore possible that, if the bis(borohydrido)magnesium moiety in complexes 4, the
Mg(BH4)2 moiety displays a covalent character, it may be preference of a thf molecule over a Mg(BH4)2(thf)3 in the
leached from the heterodimetallic trinuclear structure. This latter, as well as in 1, may be attributed to electronic factors
supports well the overall neutral character of complex 3a rather than to steric ones. To the best of our knowledge,
(and 3b), Mg(BH4)2(thf)3 acting as a simple donor ligand complexes 4a and 4b are the first borohydrido derivatives
Eur. J. Inorg. Chem. 2010, 2867–2876
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2871

M. Visseaux, M. Terrier, A. Mortreux, P. Roussel
FULL PAPER
supported by this linked anilidocyclopentadienyl ligand,
The spectrum of 2 showed the presence of two additional
and thus C5Me4CH2SiMe2NPh appears also well-suited for compounds in [D8]thf, which denotes a disproportionation
larger lanthanides.
(so-called “ligand scrambling”) process in this polar solvent
Finally, one can observe that in all the heterodimetallic (Scheme 5), as commonly reported for this chemistry.[23,25]
complexes, the thf molecules are in every case coordinated Cp3Nd[26] (17%), the monocyclopentadienyl derivative[5a]
to magnesium rather than to the lanthanide atom, which (17%), and the metallocene (66%) were identified. It is
denotes a higher oxophilicity of the former element.
noteworthy that, if structurally characterized Cp2LnX com-
pounds are very scarce with the larger lanthanides,[1a] this
is probably connected to the occurrence of such dispropor-
tionation reactions in solution.
1H NMR Spectroscopic Analysis
Despite the paramagnetism of the lanthanides, all 1H
NMR signals could be assigned (Table 1). The H NMR
Scheme 5. Disproportionation of 2 in thf solution (Mg moiety
omitted).
1
spectrum of complex 4a is given as an example in Figure 5.
It was interpreted on the basis of DEPT (distortionless en-
hancement by polarisation transfer) 135 and 1H-13C HSQC
(1H single quantum coherence) experiments.
Isoprene Polymerization Catalysis
Catalytic systems combining a lanthanide precatalyst
and a magnesium cocatalyst are well known to be highly
efficient
towards
olefins,[2e,4d,15a,15d,27]
conjugated
1
dienes,[2a,5,14,15c,28] styrene,[29] and copolymerizations.[15e,30]
Using such dual catalysts, it has been proposed that the
reaction proceeds through Ln/Mg dimetallic active species.
We and others demonstrated in this frame that borohydri-
dolanthanide precatalysts compare well with the more con-
ventional chloride ones regarding activities and selectivi-
ties.[4d,30d] When we used the B/A route to prepare hemilan-
thanidocenes, this approach also allowed the rapid screen-
ing of the impact of a given ligand towards the reactivity in
polymerization by an “all in situ” approach, consisting of
both preparing and alkylating the precatalyst in situ as rep-
resented in Scheme 6 (right).[5b,29c,31]
Table 1. H NMR spectroscopic data for complexes 1–4 (in C6D6
at 300 K; δ in ppm vs. TMS).
Comp.
1[a]
2[b]
3a
3b
4a
4b
Cp(H)
–
5.4
32.0 (4H) 14.9; 8.1
–20.0 (4H) (4 H, 4
H)
–
–
Cp(Me) 9.8 (30H)
–
–
–
–
14.8 (6H) –2.0 (6H)
–9.0 (6H)
1.1 (6H)
3.6 (2H)
1.8 (6H)
14.6 (2H)
CMe2
/
–
–1.5 (12H) 1.7 (12H) 14.4 (6H)
SiMe2
o-Ph
m-Ph
p-Ph
BH4
thf
–
–
–
–
–
–
–
–
–
–
–
2.9 (2H)
7.8 (2H)
6.5 (2H)
7.3 (2H)
–
5.4 (1 H) 7.1 (1 H)
[c]
35.1 (4H)
–11.7 (4H)
–34.9 (4H)
14 (8H)
–5.4 (8H) 35.0 (4H)[d] –1.3 (4H)
–
2.2 (12H) 2.7 (12H) –4.2 (8H)
0.9 (12H) 0.8 (12H) –6.6 (8H)
2.1 (8H)
0.4 (8H)
[a] Identical with data in ref.[4d] [b] In [D8]thf. [c] May be lost in
the baseline. [d] At 323 K.
Scheme 6. Isoprene polymerization with isolated (left) and in situ
generated (right) complexes as precatalysts.
We tested the metallocenes presented in this work in iso-
prene polymerization, in the presence of BEM as cocatalyst,
and the results were compared to the “all in situ” approach
(Scheme 6). In this frame, the ansa-(HC5H4CMe2)2 ligand
was of particular interest, because we had prepared some
years ago several complexes of samarium and neodymium
in this series, which revealed high efficiency for homo- and
copolymerizations.[7,18,32] Selected results are given in
Table 2.
Not surprisingly, these catalysts are not as active as the
half-sandwich derivative Cp*Nd(BH4)2(thf)2,[14] likely due
to steric limitations. The low yield observed with precata-
lysts 2 and 2Ј (runs 1, 2) may be related to their poor solu-
bility in toluene and to a possible ligand scrambling in solu-
1
Figure 5. H NMR spectrum of complex 4a in C6D6 at 300 K (the
BH4 signal appears in the range 30–40 ppm at higher temperatures;
* residual C6D5H, silicon grease).
2872
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2867–2876

Application of Lanthanidocenes in Polymerization
Table 2. Isoprene polymerization with isolated (2, 3a, 4a) and “in phism that exists in magnesium chemistry. The results pre-
situ” synthesized (2Ј, 3aЈ, 4aЈ) precatalysts.
sented in isoprene polymerization clearly validate the B/A
Run[a] Precatalyst
Yield trans-1,4 3,4
Mn(exp.)ϫ10–3 PDI[d]
route as an efficient method to assess a given ligand, di-
rectly used in its protonated form, in polymerization experi-
ments.
[%]
[%][b]
[%]
[gmol–1][c]
1
2
3
4
5
6
2
2Ј
3a
3aЈ
4a
4aЈ
18
27
72
83
83
82
90.0
81.2
92.5
93.1
85.2
84.9
9.0
8.0
6.5
5.9
14.5
14.3
21.3
15.0
9.9
1.80
1.40
2.12
1.76
1.73
1.57
12.1
8.4
Experimental Section
17.5
Materials and Methods: The solvents were dried with sodium/
benzophenone ketyl, deoxygenated by distillation, and stored on
molecular sieves (3 A) in a glovebox. They were condensed directly
in the reaction flask by trap-to-trap distillation. Isoprene (Aldrich)
was dried with calcium hydride, distilled twice, and stored over mo-
lecular sieves (3 A) in a glovebox. C5H6 was freshly cracked before
use. BEM (n-butylethylmagnesium, 20% solution in hexanes, Texas
Alkyl), (C5H5)2Mg (Aldrich), and Cp*H (Aldrich) were used as
[a] Experimental conditions: V(isoprene) = V(toluene) = 1 mL,
[isoprene]/[Nd] = 1000, [BEM]/[Nd] = 1, T = 50 °C, t = 72 h. [b]
Determined by 1H and 13C NMR spectroscopy, residual proportion
= cis-1,4 defects. [c] Determined by size exclusion chromatography
(SEC), with a correction factor of 0.5.[33] [d] Polydispersity index
Mw/Mn.
received. Nd(BH4)3(thf)3,[35] allylLi(dioxane),[36] C5Me4CH2Si-
tion, as noted above. All the catalysts are highly trans-selec-
tive, the most selective ones being observed in the ansa-
(C5H4CMe2)2 series, as expected.[7,18] The 3,4-content ob-
served with anilidocyclopentadienyl-supported complexes
could be explained by a single η2 coordination of the mono-
mer to the metal center rather than by an η4 one,[34] due to
the presence of the anilido group, which was already sus-
pected to hamper the inclusion of a Mg(BH4)2 moiety in
4a. Interestingly, there is a good correlation between the
results obtained with the isolated and the in situ generated
precatalysts, which clearly validates the in situ B/A route as
an efficient strategy to rapidly screen the efficiency of a
series of ligands towards a catalytic polymerization process.
Moreover, polymolecularities are significantly lower with
precatalysts 2Ј, 3aЈ, 4aЈ, which may be related to a faster
initiation stage with the latter ones, as compared to 2, 3,
4. This may be connected to the dissociation of isolated
complexes in solution to generate the active species.
[37]
Me2NHPh,[6] and (CMe2C5H5)2
were synthesized as previously
reported. H and 13C NMR spectra were recorded with a Bruker
Avance 300 instrument at 300 K, with specific DEPT and HSQC
Bruker sequences when necessary. The chemical shifts were cal-
ibrated by using the residual resonances of the solvent. Elemental
analyses were carried out by the Centre d’Analyses de l’Université
de Dijon. Size exclusion chromatography (SEC) was performed in
thf as eluent at 40 °C with a Waters SIS HPLC-pump, a Waters
410 refractometer and Waters styragel column (HR2, HR3, HR4,
and HR5E) calibrated with polystyrene standards. A correction
factor of 0.5 was applied for the determination of the true number-
average molecular weight of polyisoprenes.[33]
1
Crystallographic Data: Crystallographic and refinement data for
complexes 2–5 are gathered in Table 3. A parallelepiped-shaped
crystal with the dimensions reported in Table 3 was mounted on
top of a glass fiber and aligned on a Bruker SMART APEX CCD
diffractometer (platform with full three-circle goniometer). The dif-
fractometer was equipped with a 4 K CCD detector set 55.0 mm
from the crystal. The crystal was cooled to 100(1) K with an
OXFORD CRYOSTREAM 700 low-temperature device. Intensity
measurements were performed by using graphite monochromated
Mo-Kα radiation from a sealed ceramic diffraction tube (SIE-
MENS). Generator settings were 50 kV40 mA–1. APEX2[38] was
used for preliminary determination of the unit cell constants and
data collection control. Data integration and global cell refinement
was performed with the program SAINT.[39] The final unit cell was
obtained from the xyz centroids of the number of reflections re-
ported in Table 3 for each complex. Intensity data were corrected
for Lorentz and polarization effects, scale variation, for decay and
absorption: a multiscan absorption correction was applied, on the
basis of the intensities of symmetry-related reflections measured at
different angular settings (SADABS),[40] and reduced to Fo2. The
program suite JANA2006 was used for space group determination
Conclusions
The synthetic strategy described in this work allows the
one-pot preparation, in good yields, of lanthanidocene-like
complexes, by simply starting from the protonated form of
the ligand. Depending on the stoichiometry of the CpRH
ligand, it is now demonstrated that lanthanidocenes can be
easily prepared by the B/A route, as already shown for hemi-
lanthanidocenes. It is noteworthy that the nature of the
HCpR ligand has a strong impact on the molecular struc-
ture of the isolated complex. With the simple C5H5, or a
monosubstituted cyclopentadienyl ligand, mixed Ln/Mg di- and full-matrix least square refinements based on F.[41] They were
solved by direct methods with the program SIR97.[42] The posi-
metallic compounds are obtained, whereas monometallic
tional and anisotropic displacement parameters for the non-hydro-
neutral complexes are provided by a reaction with a per-
gen atoms were refined. Except hydrogen atoms pertaining to B,
alkyl-substituted ligand. The Ln/Mg mixed dimetallic com-
the positions of hydrogen atoms were generated by geometrical
plexes are the first examples of compounds comprising a
considerations, constrained to idealized geometries, and allowed to
BH4 group linking a lanthanide and a magnesium center,
ride on their carrier atoms with an isotropic displacement param-
eter related to the equivalent displacement parameter of their car-
rier atoms. Note that the crystal quality for complexes 2 and 3b
was not satisfactory enough to allow the location of BH4 hydrogen
atoms. The structures were refined to the final agreement factors
which supports well the hypothesis of formation of Ln(µ-
BH4)Mg active species involved in previously reported cata-
lytic processes. An unprecedented crystalline form of
Mg(BH4)2(thf)3 could also be isolated during the course of
our organometallic syntheses, as an example of polymor- given in Table 3. CCDC-764030, -764034, -764035, -764031,
Eur. J. Inorg. Chem. 2010, 2867–2876
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2873

M. Visseaux, M. Terrier, A. Mortreux, P. Roussel
Table 3. Summary of crystal and refinement data for complexes 2–5 [recorded at T = 100 K; with λ(Mo-Kα) = 0.71073 Å].
FULL PAPER
Compound
2[a]
3a
3b[a]
4a
4b
5
Color
light blue
green blue
yellow
light blue
light orange
colorless
Chemical formula
Formula weight
Crystal system
Space group
a [Å]
b [Å]
c [Å]
C36H68B4MgNd2O4 C44H80B4MgNd2O3 C44H80B4MgNd2O3 C26H45BNNdO2Si C26H45BNNdO2Si C12H32B2MgO3
920.96
1013.1
1009.3
586.8
orthorhombic
Pnma
17.2106(3)
44.5052(8)
10.9861(2)
90
90
90
8414.9(3)
12
1.389
592.90
orthorhombic
Pnma
17.2040(2)
44.3324(6)
10.9721(2)
90
90
90
8368.4(2)
12
1.413
270.3
monoclinic
P21/c
15.5438(4)
16.2479(4)
32.5322(8)
90
90.033(2)
90
8216.1(4)
8
triclinic
triclinic
triclinic[b]
¯
¯
¯
P1
P1
P1
13.414(6)
13.459(6)
16.499(7)
87.781(7)
66.608(6)
61.350(6)
2356.3(18)
2
13.392(2)
13.493(2)
16.545(3)
87.170(8)
66.370(9)
60.952(8)
2354.9(7)
2
8.495(2)
8.591(2)
12.339(3)
102.95(3)
104.028(3)
92.43(4)
847.0(4)
2
α [°]
β [°]
γ [°]
V [Å3]
Z
ρ [gcm–3]
1.489
1.428
1.423
1.059
θ range [°]
Collected reflections
Unique
Obsd. [IϾ3σ(I)]
Rint
1.25; 29.41
94804
22535
12723
0.0905
2.45; 28.61
12080
9865
6991
0.0602
1.37; 30.72
14294
14294
9986
0.0739
0.92; 30.63
13130
13130
10787
0.0422
1.84; 33.15
142417
16144
10016
0.0561
1.75; 31.33
11928
5559
3071
0.0562
Parameters
R(F), Rw(F) [IϾ3σ(I)]
R(F), Rw(F) (all data)
ρmin, ρmax [e Å–3]
341
536
246
470
451
196
0.0733, 0.0739
0.1324, 0.0783
–0.76; 5.36
0.0475, 0.0434
0.0734, 0.0465
–1.02; 1.60
0.0757, 0.0692
0.1186, 0.0706
–2.56; 3.57
0.0477, 0.0596
0.0626, 0.0613
–2.97; 2.38
0.0822, 0.1145
0.0984, 0.1174
–4.72; 3.78
0.1255, 0.1354
0.1795, 0.1420
–0.45; 1.23
[a] The crystal structure was refined from a twinned crystal. [b] Twin pseudomonoclinic.
-764032, and -764033 contain the supplementary crystallographic
data for 2, 3a, 3b, 4a, 4b, and 5, respectively. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
mixture turned from deep blue to blue-green and then to light blue,
while a blue solid deposited. After stirring for 1 h at room tempera-
ture, the solution was concentrated to dryness. Then, thf (50 mL)
was added by vacuum transfer into the flask, giving a limpid blue
solution. The solution was concentrated to dryness, and the residue
was extracted with toluene (50 mL). After filtration to eliminate
insoluble materials, the solution was concentrated (ca. 10 mL) and
allowed to stand at room temperature to afford light blue crystals,
which were collected for X-ray analysis. The mother liquor was
concentrated to dryness, and a crystalline solid was isolated after
washing with pentane (10 mL) and drying under vacuum. Yield:
412 mg (90%). 1H NMR (C4D8O): δ = 5.45 (s, 10 H, Cp), the
BH4 signal may be lost in the baseline. Additional peaks: δ = 2.80
{corresponding to Cp3Nd (ca. 17%)},[26] –2.50 {attributed to a
monoCp derivative (ca. 17%)}[5a] ppm. Elemental analysis was cor-
rect for a tris(thf) adduct. C36H68B4MgNd2O4 (920.96): calcd. C
45.28, H 7.12; found C 45.36, H 7.56.
Polymerizations: In a typical experiment conducted in a glovebox,
the precatalyst (10 µmol) was weighed and diluted in toluene. The
cocatalyst, BEM, was added with a microsyringe, and the monomer
immediately after it at once. The flask was taken out from the
glovebox and placed in a thermostatted bath at 50 °C for a given
time. The flask was then opened to air, and the viscous mixture
was diluted with undried toluene to quench the polymerization.
The resulting solution was poured into a large volume of a 50:50
methanol/isopropyl alcohol mixture containing tert-butylcatechol
as a stabilizing agent. An off-white polymer was filtered off and
dried in vacuo until it reached constant weight.
Organometallic Syntheses
(C5Me2)2Nd(BH4)(thf) (1): NMR scale: BEM (14.7 µL, 20.6 µmol)
solution was added dropwise, at room temperature, to a solution
of Nd(BH4)3(thf)3 (8.5 mg, 20.6 µmol) and HC5Me5 (6.4 µL,
41.2 µmol) in C6D6 (0.4 mL). After evaporation of the solvents and
dissolution in C6D6 only one product was observed. 1H NMR
(C6D6): δ = 44 (vbr, BH4, 4 H), 10.07 (s, Cp*, 30 H), 6.01 (s, thf,
16 H), 3.31 (s, thf, 16 H) ppm. Bulk synthesis: A hexanes solution
of BEM (0.8 mL, 1.00 mmol) was added dropwise to a solution of
Nd(BH4)3(thf)3 (406 mg, 1 mmol) and HC5Me5 (278 mg,
1.02 mmol) in toluene (40 mL). The mixture turned from light blue
to blue-green. After stirring for 2 h at room temperature, the solu-
tion was concentrated to afford a blue oily material. After treat-
ment with thf/pentane, a bluish crystalline powder was isolated.
Yield: 121 mg (24%). The residual oily mother liquor did not crys-
[(C5H4CMe2)2Nd(BH4)(µ-BH4)]2Mg(thf)3 (3a): Same procedure as
for 1, starting from BEM (0.8 mL, 1 mmol), Nd(BH4)3(thf)3
(400 mg, 1 mmol), and (HC5H4CMe2)2 (210 mg, 1 mmol) in tolu-
ene (50 mL). The mixture turned from green-blue to light orange
within 1 h at room temperature. A crystalline green-blue solid was
finally isolated after work-up. Yield: 0.355 g (70%). 1H NMR
(C6D6): δ = 32.0 (s, C5H4), 14.0 (br., BH4), 2.2 (br., thf), 0.9 (br.,
thf), –1.5 (s, CMe2), –20.0 (s, C5H4). C44H80B4MgNd2O3 (1013.1):
calcd. C 52.16, H 7.96; found C 52.34, H 8.30.
[(C5H4CMe2)2Sm(BH4)(µ-BH4)]2Mg(thf)3 (3b): Same procedure as
for 1, starting from BEM (0.8 mL, 1 mmol), Sm(BH4)3(thf)3
(415 mg, 1 mmol), and (HC5H4CMe2)2 (217 mg, 1 mmol) in tolu-
ene (50 mL). The mixture turned from pale yellow to deep orange
within 1 h at room temperature, without formation of any solid. A
crystalline yellow solid was finally isolated after work-up. Yield:
1
tallize properly. The H NMR spectrum of the crystals was found
to be identical to that of previously isolated (C5Me2)2Nd(BH4)(thf).
1
0.315 g (66%). H NMR (C6D6): δ = 14.9 (s, C5H4), 8.1 (s, C5H4),
[(C5H5)2Nd(BH4)(µ-BH4)]2Mg(thf)4 (2): BEM solution (0.8 mL,
1 mmol) diluted in toluene (10 mL) was added dropwise, at room
temperature, to a solution of Nd(BH4)3(thf)3 (405 mg, 1 mmol) and
an excess of HC5H5 (219 mg, 3.3 mmol) in toluene (50 mL). The
2.7 (s, thf), 1.7 (s, CMe2), 0.8 (s, thf), –5.4 (br., BH4) ppm.
Recrystallization from thf/pentane gave large yellow crystals
(105 mg), which were used for crystallographic analysis (H atoms
pertaining to BH4 could not be located). Reaction with allylLi-
2874
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2867–2876

Application of Lanthanidocenes in Polymerization
ama, H. Yasuda, J. Organomet. Chem. 2004, 689, 4489–4498;
e) J. Gromada, J. F. Carpentier, A. Mortreux, Coord. Chem.
Rev. 2004, 248, 397–410; f) F. Bonnet, D. Barbier-Baudry, A.
Dormond, M. Visseaux, Polym. Int. 2002, 51, 986–993.
M. Ephritikhine, Chem. Rev. 1997, 97, 2193–2242.
a) S. M. Cendrowski-Guillaume, M. Nierlich, M. Lance, M.
Ephritikhine, Organometallics 1998, 17, 786–788; b) D. Barb-
ier-Baudry, O. Blacque, A. Hafid, A. Nyassi, H. Sitzmann, M.
Visseaux, Eur. J. Inorg. Chem. 2000, 2333–2336; c) F. Bonnet,
M. Visseaux, D. Barbier-Baudry, A. Hafid, E. Vigier, M. M.
Kubicki, Inorg. Chem. 2004, 43, 3682–3690; d) M. Visseaux, T.
Chenal, P. Roussel, A. Mortreux, J. Organomet. Chem. 2006,
691, 86–92; e) N. Meyer, J. Jenter, P. W. Roesky, G. Eickerling,
W. Scherer, Chem. Commun. 2009, 4693–4695.
a) M. Visseaux, P. Zinck, M. Terrier, A. Mortreux, P. Roussel,
J. Alloys Compd. 2008, 451, 352–357; b) M. Visseaux, M. Ter-
rier, A. Mortreux, P. Roussel, C. R. Chim. 2007, 10, 1195–1199.
D. Robert, P. Voth, T. P. Spaniol, J. Okuda, Eur. J. Inorg. Chem.
2008, 2810–2819.
D. Baudry-Barbier, N. Andre, A. Dormond, C. Pardes, P. Rich-
ard, M. Visseaux, C. J. Zhu, Eur. J. Inorg. Chem. 1998, 1721–
1727.
(dioxane) (NMR scale experiment): allylLi(dioxane) (2 equiv.,
3.4 mg, 19 µmol) was added to 3b (4.9 mg, 4.9 µmol) in C6D6
(0.4 mL). The 1H NMR spectrum recorded immediately showed
the formation of a bis(allyl) derivative as the only product: 1H
NMR (C6D6): δ = 10.1 (s, 1 H, allyl), 8.9 (s, 2 H, allyl), 8.1 (s, 4
H, C5H4), 7.5 (s, 2 H, allyl), 5.9 (s, 4 H, C5H4), 3.3 (s, dioxane),
3.2 (s, thf), 1.9 (s, thf), 1.2 (s, 12 H, CMe2) ppm.
[3]
[4]
Mg(BH4)2(thf)3 (5): A small crop of highly air-sensitive large color-
less crystals (ca. 50 mg) was also isolated from the above mother
1
liquor after 12 h at –40 °C. H NMR (C4D8O): δ = –0.25 (q, J =
91 Hz, BH4). Elemental analysis was not possible due to the low
quantity of the isolated product.
[5]
[(C5Me4)CH2SiMe2NPh]Nd(BH4)(thf)2 (4a): Same procedure as for
1, starting from Nd(BH4)3(thf)3 (809 mg, 2 mmol), HC5Me4CH2Si-
Me2NHPh (570 mg, 2 mmol), and BEM (1.6 mL, 2 mmol) in thf
(100 mL). The mixture turned from purple to light green within 1 h
at room temperature. The thf was concentrated to ca. 10 mL, and
pentane (10 mL) was added, affording large deep blue crystals after
one night at –30 °C. Total yield of crystalline material: 288 mg
[6]
[7]
1
(33%). H NMR (C6D6):δ = 35 (vbr, BH4), 14.8 (s, C5Me4, 6 H),
[8]
a) G. G. Skvortsov, M. V. Yakovenko, G. K. Fukin, E. V. Bar-
anov, Y. A. Kurskii, A. A. Trifonov, Russ. Chem. Bull. Int. Ed.
2007, 56, 456–460; b) E. B. Lobkovsky, Y. K. GunЈko, B. M.
Bulychev, G. L. Soloveichik, M. Y. Antipin, J. Organomet.
Chem. 1991, 406, 343–352.
14.6 (s, CH2, 2 H), 14.4 (s, SiMe2, 6 H), 7.8 (s, m-Ph, 2 H), 5.4 (s,
p-Ph, 1 H), 2.9 (s, o-Ph, 2 H), –4.2 (s, thf, 8 H), –6.6 (s, thf, 8 H),
–9.0 (s, C5Me4, 6 H) ppm. C26H45BNNdO2Si (586.8): calcd. C
53.22, H 7.73, N 2.39; found C 52.37, H 7.75, N 2.26.
[9]
W. Nie, C. Qian, Y. Chen, S. Jie, J. Organomet. Chem. 2002,
[(C5Me4)CH2SiMe2NPh]Sm(BH4)(thf)2 (4b): Same procedure as
for 1, starting from Sm(BH4)3(thf)3 (432 mg, 1.05 mmol),
HC5Me4CH2SiMe2NHPh (300 mg, 1.05 mmol), and BEM
(0.8 mL, 1 mmol) in thf (80 mL). The mixture turned from yellow
to deep red-purple. The thf was concentrated to ca. 5 mL, and pen-
tane (10 mL) was added, affording after one night at –30 °C a crop
of orange crystals. Yield: 165 mg (28%). 1H NMR (C6D6): δ = 7.29
(s, m-Ph, 2 H), 7.10 (s, p-Ph, 1 H), 6.49 (s, o-Ph, 2 H), 3.65 (s, CH2,
2 H), 2.15 (s, thf, 8 H), 2.00 (s, C5Me4, 6 H), 1.82 (s, SiMe2, 6 H),
1.08 (s, C5Me4, 6 H), 0.37 (s, thf, 8 H), –1.35 (vbr, BH4) ppm.
647, 114–122.
[10]
a) P. Yan, N. Hu, Z. Jin, W. Chen, J. Organomet. Chem. 1990,
391, 313–320; b) C. Sun, G. Wei, Z. Jin, W. Shen, J. Organomet.
Chem. 1993, 453, 61–64; c) C. Sun, G. Wei, Z. Jin, W. Chen,
Polyhedron 1994, 13, 1483–1487.
a) R. Anwander, Top. Organomet. Chem. 1999, 2, 1–61; b) J.
Eppinger, M. Spiegler, W. Hieringer, W. A. Herrmann, R. An-
wander, J. Am. Chem. Soc. 2000, 122, 3080–3096.
H. M. Dietrich, C. Zapilko, E. Herdtweck, R. Anwander, Orga-
nometallics 2005, 24, 5767–5771.
a) H. Schumann, D. M. Freckmann, S. Dechert, Z. Anorg. Allg.
Chem. 2002, 628, 2422–2426; b) S. Bambirra, M. W. Bouwk-
amp, A. Meetsma, B. Hessen, J. Am. Chem. Soc. 2004, 126,
9182–9183; c) S. Bambirra, A. Meetsma, B. Hessen, Organome-
tallics 2006, 25, 3454–3462.
[11]
[12]
[13]
Mechanistic Study of the Formation of Lanthanidocenes: An equi-
molar amount of Nd(BH4)3(thf)3 was added to a C6D6 solution of
Cp2Mg (3.2 mg, 20 µmol) in an NMR tube equipped with a Teflon
valve. The 1H NMR spectrum recorded immediately showed the
absence of any reaction. After 24 h at room temperature, a red-blue
precipitate appeared, with just some residual neodymium tris-
(borohydride) in solution. After 48 h at 50 °C, the precipitate re-
mained insoluble. The solvents were then evaporated off, and [D8]-
[14]
[15]
F. Bonnet, M. Visseaux, A. Pereira, D. Barbier-Baudry, Macro-
molecules 2005, 38, 3162–3169.
Selected references: a) X. Olonde, A. Mortreux, F. Petit, J. Mol.
Catal. 1993, 82, 75–82; b) S. Bogaert, T. Chenal, A. Mortreux,
G. Nowogrocki, C. W. Lehmann, J. F. Carpentier, Organome-
tallics 2001, 20, 199–205; c) M. Terrier, M. Visseaux, T. Chenal,
A. Mortreux, J. Pol. Sci. Pol. Chem. A 2007, 45, 2400–2409; d)
R. Godoy Lopez, C. Boisson, F. DЈAgosto, R. Spitz, F. Bois-
son, D. Gigmes, D. Bertin, Macromol. Rapid Commun. 2006,
27, 173–181; e) A. Valente, P. Zinck, A. Mortreux, M. Vis-
seaux, Macromol. Rapid Commun. 2009, 30, 528–531.
J. Gromada, A. Mortreux, G. Nowogrocki, F. Leising, T. Ma-
thivet, J. F. Carpentier, Eur. J. Inorg. Chem. 2004, 3247–3253.
W. Kretschmer, K. H. Z. Thiele, Z. Anorg. Allg. Chem. 1995,
621, 1093–1097.
D. Barbier-Baudry, F. Bonnet, B. Domenichini, A. Dormond,
M. Visseaux, J. Organomet. Chem. 2002, 647, 167–179.
a) J. Lin, Z. Wang, J. Organomet. Chem. 1999, 589, 127–132;
b) W. Wu, M. Chen, P. Zhou, Organometallics 1991, 10, 98–
104.
L. F. Sanchez-Barba, D. L. Hughes, S. M. Humphrey, M.
Bochmann, Organometallics 2005, 24, 5329–5334.
1
thf was added. The H NMR spectrum showed the presence of a
mixture of Cp3Nd (40%) and Cp2Nd(BH4) (60%).
Acknowledgments
The Région Nord-Pas-de-Calais is gratefully acknowledged
(ARCIR Nanocat project). We thank Aurélie Malfait and Anne-
Marie Cazé for SEC measurements.
[16]
[17]
[18]
[19]
[1] a) H. Schumann, J. A. Meese-Marktscheffel, L. Esser, Chem.
Rev. 1995, 95, 865–986; b) F. T. Edelmann, D. M. Freckmann,
H. Schumann, Chem. Rev. 2002, 102, 1851–1896; c) P. M. Zei-
mentz, S. Arndt, B. R. Elvidge, J. Okuda, Chem. Rev. 2006,
106, 2404–2433; d) E. Kirillov, J. Y. Saillard, J. F. Carpentier,
Coord. Chem. Rev. 2005, 249, 1221–1248; e) P. L. Arnold, I. J.
Casely, Chem. Rev. 2009, 109, 3599–3611.
[2] a) L. Friebe, O. Nuyken, W. Obrecht, Adv. Polym. Sci. 2006,
204, 1–154; b) A. Fischbach, R. Anwander, Adv. Polym. Sci.
2006, 204, 155–281; c) X. Li, M. Nishiura, L. Hu, K. Mori, Z.
Hou, J. Am. Chem. Soc. 2009, 131, 13870–13882; d) Y. Nakay-
[20]
[21]
Nöth et al. determined the following crystal data for Mg(BH4)
2(thf)3: C2/c group, a = 12.181(4) Å, b = 12.642(5) Å, c =
12.460(5) Å, α = 90°, β = 111.53(3)°, γ = 90°, see H. Nöth, Z.
Naturforsch., Teil B 1982, 37, 1499–1503.
Eur. J. Inorg. Chem. 2010, 2867–2876
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2875

M. Visseaux, M. Terrier, A. Mortreux, P. Roussel
FULL PAPER
[22] M. Bremer, H. Nöth, M. Warchhold, Eur. J. Inorg. Chem. 2003,
111–119.
[23] J. Okuda, Dalton Trans. 2003, 2367–2378.
[24] a) W. X. Zhang, M. Nishiura, Z. Hou, Chem. Eur. J. 2007, 13,
4037–4051; b) K. C. Hultzsch, P. Voth, K. Beckerle, T. P.
Spaniol, J. Okuda, Organometallics 2000, 19, 228–243; c) S.
Tian, V. M. Arredondo, C. L. Stern, T. J. Marks, Organometal-
lics 1999, 18, 2568–2570.
[25] J. Gräper, R. D. Fischer, G. Paolucci, J. Organomet. Chem.
1994, 471, 87–95.
Chem. Phys. 2007, 208, 973–978; d) J. Thuilliez, R. Spitz, C.
Boisson, Macromol. Chem. Phys. 2006, 207, 1727–1731.
[31] F. Bonnet, C. Da Costa Violante, P. Roussel, A. Mortreux, M.
Visseaux, Chem. Commun. 2009, 3380–3382.
[32] a) M. Visseaux, D. Barbier-Baudry, F. Bonnet, A. Dormond,
Macromol. Chem. Phys. 2001, 202, 2485–2488; b) F. Bonnet,
M. Visseaux, D. Barbier-Baudry, A. Dormond, Macromole-
cules 2002, 35, 1143–1145; c) D. Barbier-Baudry, F. Bonnet, A.
Dormond, E. Finot, M. Visseaux, Macromol. Chem. Phys.
2002, 203, 1194–1200.
[26] The chemical shift was not available in [D8]thf: R. von Am-
mon, B. Kanellakopulos, R. D. Fischer, P. Laubereau, Inorg.
Nucl. Chem. Lett. 1969, 5, 315–319. We found δCp = 2.8 ppm
in [D8]thf for a pure Cp3Nd sample.
[27] Selected references: a) J. F. Pelletier, A. Mortreux, X. Olonde,
K. Bujadoux, Angew. Chem. Int. Ed. Engl. 1996, 35, 1854–
1856; b) S. Bogaert, T. Chenal, A. Mortreux, J. F. Carpentier,
J. Mol. Catal. A 2002, 190, 207–214; c) J. Gromada, T. Chenal,
A. Mortreux, J. W. Ziller, F. Leising, J. F. Carpentier, Chem.
Commun. 2000, 2183–2184.
[33] P. Zinck, M. Terrier, A. Mortreux, M. Visseaux, Polym. Testing
2009, 28, 106–108.
[34] L. Friebe, O. Nuyken, H. Windisch, W. Obrecht, Macromol.
Chem. Phys. 2002, 203, 1055–1064.
[35] U. Mirsaidov, I. B. Shaimuradov, M. Khikmatov, Russ. J. In-
org. Chem. 1986, 31, 753–754.
[36] R. Taube, S. Maiwald, J. Sieler, J. Organomet. Chem. 1996, 513,
37–47.
[37] H. Schwemlein, H. H. Brintzinger, J. Organomet. Chem. 1983,
254, 69–73.
[28] Selected references: a) D. Jenkins, Polymer 1985, 26, 152–158;
b) V. N. Urazbaev, V. P. Efimov, Z. M. Sabirov, Yu. B.
Monakov, J. Appl. Polym. Sci. 2003, 89, 601–603; c) F. Bonnet,
M. Visseaux, A. Pereira, F. Bouyer, D. Barbier-Baudry, Macro-
mol. Rapid Commun. 2004, 25, 873–877.
[29] Selected references: a) Y. Sarazin, T. Chenal, A. Mortreux, H.
Vezin, J. F. Carpentier, J. Mol. Catal. A 2005, 238, 207–214; b)
P. Zinck, M. Visseaux, A. Mortreux, Z. Anorg. Allg. Chem.
2006, 632, 1943–1944; c) P. Zinck, A. Valente, M. Terrier, A.
Mortreux, M. Visseaux, C. R. Chim. 2008, 11, 595–602.
[30] Selected references: a) J. Thuilliez, L. Ricard, F. Nief, F. Bois-
son, C. Boisson, Macromolecules 2009, 42, 3774–3779; b) J.
Gromada, A. Mortreux, T. Chenal, J. W. Ziller, F. Leising, J. F.
Carpentier, Chem. Eur. J. 2002, 8, 3773–3788; c) P. Zinck, M.
Terrier, A. Mortreux, A. Valente, M. Visseaux, Macromol.
[38] Bruker APEX2, Bruker AXS Inc., Madison, Wisconsin, USA,
2007.
[39] Bruker SAINT, Bruker AXS Inc., Madison, Wisconsin, USA,
2007.
[40] G. M. Sheldrick, SADABS, University of Göttingen, Germany,
2007.
[41] V. Petricek, M. Dusek, L. Palatinus, JANA2006 Structure De-
termination Software Programs, Institute of Physics, Praha,
Czech Republic, 2008.
[42] A. Altomare, M. C. Burla, M. Camalli, G. Cascarano, C. Gia-
covazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R.
Spagna, SIR97, A Package for Crystal Structure Solution by
Direct Methods and Refinement, Bari, Rome, Italy, 1997.
Received: February 15, 2010
Published Online: April 28, 2010
2876
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2867–2876