Bonding of H2, N2, Ethylene, and Acetylene to Bivalent Lanthanide Metallocenes: Trends from DFT Calculations on Cp 2M and Cp*2 M (M = Sm, Eu, Yb) and Experiments with Cp*2Yb

Article abstract of DOI:10.1021/om034206v

The results of DFT calculations have been used to define the trends in the interactions of H₂, N₂, C₂H₄, C ₂H₂, and C₂Me₃ with the bivalent lanthanide metallocenes Cp₂M (Cp = η⁵C ₅H₅) and Cp*₂M (Cp* = η ⁵-C₂Me₅), where M = Sm, Eu, Yb. These results, together with those previously published for the bonding of CO to Cp ₂M (M = Ca, Eu, Yb), suggest that the interaction of these ligands with the lanthanide metallocenes results from a subtle balance between attractive (dipole-dipole or dipole-induced dipole) and repulsive (electron-electron repulsion within the f shell) forces. The balance between the attractive and repulsive forces, and therefore the net bond energy, depends on the f-electron count in these bivalent lanthanide metallocenes. The computational results are compared with experimental observations on paramagnetic Cp*₂Eu and diamagnetic ytterbocene, Cp* ₂Yb.

Full text of DOI:10.1021/om034206v

Organometallics 2003, 22, 5447-5453
5447
Bon d in g of H2, N2, Eth ylen e, a n d Acetylen e to Biva len t
La n th a n id e Meta llocen es: Tr en d s fr om DF T
Ca lcu la tion s on Cp 2M a n d Cp *2 M (M ) Sm , Eu , Yb) a n d
Exp er im en ts w ith Cp *2Yb
L. Perrin,† L. Maron,*,‡ O. Eisenstein,*,† D. J . Schwartz,§ C. J . Burns,§ and
R. A. Andersen*,§
LSDSMS (UMR 5636), Universite´ Montpellier 2, 34095 Montpellier Cedex 05,
France, Laboratoire de Physique Quantique (UMR 5626), IRSAMC, Universite´ Paul Sabatier,
118 route de Narbonne, 31064 Toulouse Cedex 04, France, and Chemistry Department and
Chemical Sciences Division of Lawrence Berkeley National Laboratory,
University of California, Berkeley, California 94720
Received September 29, 2003
The results of DFT calculations have been used to define the trends in the interactions of
H2, N2, C2H4, C2H2, and C2Me2 with the bivalent lanthanide metallocenes Cp2M (Cp ) η5-
C5H5) and Cp*2M (Cp* ) η5-C5Me5), where M ) Sm, Eu, Yb. These results, together with
those previously published for the bonding of CO to Cp2M (M ) Ca, Eu, Yb), suggest that
the interaction of these ligands with the lanthanide metallocenes results from a subtle
balance between attractive (dipole-dipole or dipole-induced dipole) and repulsive (electron-
electron repulsion within the f shell) forces. The balance between the attractive and repulsive
forces, and therefore the net bond energy, depends on the f-electron count in these bivalent
lanthanide metallocenes. The computational results are compared with experimental
observations on paramagnetic Cp*2Eu and diamagnetic ytterbocene, Cp*2Yb.
In tr od u ction
responsible for electron-electron (e-e) repulsion with
the electrons in the f shell. In the case of Cp2Eu (4f7)
the increase in the e-e repulsion with the half-filled 4f
shell is less than the attractive forces, resulting in a
net bond between Cp2Eu and CO. In the case of Cp2Yb
(4f14) the increase in e-e is larger with the full 4f shell,
which is minimized when the electronegative end (viz.,
the oxygen atom) of CO binds to the metal center. This
results in a lower νCO value relative to νCO for free CO,
consistent with the experimental result. Thus, the Cp2M
fragment handles the bonding to dipolar CO in different
ways, depending upon whether the 4f shell is filled or
half-filled. This begets the question of how this bond
model can accommodate bonding when the diatomic
ligand does not have a permanent dipole moment, the
subject of this paper. It is well-known that DFT calcula-
tions underestimate dispersion forces.3 However, this
method has been used successfully for metal complexes
of rare gas.4
It has been shown recently that the CO pressure and
temperature dependence for the equilibrium reaction
between CO and Cp*2M (Cp* ) η5-C5Me5; M ) Ca, Eu)
in hydrocarbon solution yields a monocarbonyl adduct
with a change in enthalpy of about -4 kcal mol-1.1 For
Ca and Eu, the CO stretching frequency of the 1:1
adduct is greater than that of free CO. In contrast, for
Yb, an equilibrium exists between 1:1 and 1:2 adducts,
both of which have CO stretching frequencies lower that
that of free CO.1 DFT calculations model properly the
experimental variation (∆νCO) of CO stretching frequen-
cies relative to that of free CO for Cp2M(CO) (Cp ) η5-
C5H5), when M is Ca or Eu and when the CO ligand is
C-bound. In the case of Cp2Yb, the experimental lower-
ing of the CO stretching frequencies is properly repro-
duced only for an O-bonded carbon monoxide in both
adducts.2 The principle that emerges from the DFT
calculations is that π-back-bonding from the 4f core
electrons is not involved in the bond to CO. The bonding
between Cp2M and CO is the result of dipole-dipole
interactions along with a small amount of charge
transfer from CO to Cp2M; the charge transfer is
Resu lts a n d Discu ssion
(I) Molecu la r Hyd r ogen a n d Nitr ogen . (a ) Ex-
p er im en ta l Obser va tion s. Dihydrogen has been sug-
gested to interact reversibly with Cp*2Eu, since the H
NMR spectrum shows that the chemical shift of dihy-
† Universite´ Montpellier 2.
1
‡ Universite´ Paul Sabatier.
§ University of California.
* Corresponding author.
(1) (a) Schultz, M.; Burns, C. J .; Schwartz, D. J .; Andersen, R. A.
Organometallics 2001, 20, 5690. (b) Selg, P.; Brintzinger, H. H.;
Schultz, M.; Andersen, R. A. Organometallics 2002, 21, 3100.
(2) Perrin. L.; Maron, L.; Eisenstein, O.; Andersen, R. A. J . Am.
Chem. Soc. 2002, 124, 5614.
(3) Wesolowski, T. A.; Parisel, O.; Ellinger, Y.; Weber, J . J . Chem.
Phys. A 1997, 101, 7818.
(4) Andrews, L.; Liang, B. Y.; Li, J .; Bursten, B. E. J . Am. Chem.
Soc. 2003, 125, 3126.
10.1021/om034206v CCC: $25.00 © 2003 American Chemical Society
Publication on Web 11/20/2003

5448 Organometallics, Vol. 22, No. 26, 2003
Perrin et al.
Ta ble 1. Ca lcu la ted En er gy Va r ia tion s a n d
In fr a r ed Str etch in g F r equ en cy Ch a n ges for th e
Equ ilibr iu m Cp 2M(ga s) + L(ga s) ) Cp 2M‚L(ga s)a
The calculations were then performed on the equilib-
rium reaction between Cp2M and dinitrogen. As shown
in Table 1, the energy change for all three metallocenes
is favorable by about -4 kcal mol-1 in each case, but
the perturbation in N-N stretching frequency is small.
From consideration of the calculated stretching frequen-
cies, we conclude that the electron density on N2 is not
changed in the presence of the Cp2M or Cp*2M frag-
ments so that a weak adduct, if it exists, should not be
spectroscopically detectable. This is consistent with the
experimental evidence mentioned above for Cp*2Eu and
Cp*2Yb but contrasts with the published result for
Cp*2Sm, since the latter gives a 2:1 adduct in which
the dinitrogen molecule forms a symmetrical bridge
between the two Cp*2Sm fragments.9a This apparent
contradiction between the calculation and experiment
can be resolved by postulating that the N2 compound
of Cp*2Sm is not an adduct between Cp*2SmII and N2
but is the result of electron transfer in which two [Cp*2-
SmIII] fragments interact with a disordered [N2]2-
fragment. This interpretation is consistent with the
recent observation on the 1:2 N2 adducts of LnX2 (Ln )
∆E (kcal mol-1
)
∆ν (cm-1
)
M
L
Cp
Cp*
Cp
Cp*
Sm
Eu
Ybb
Sm
Eu
Yb
Sm
Eu
Yb
Sm
Eu
Yb
Eu
Yb
H2
H2
H2
N2
N2
N2
C2H4
C2H4
C2H4
C2H2
C2H2
C2H2
C2Me2
C2Me2
-0.7
-0.8
-0.5
-4.1
-4.2
-4.1
-6.2
-6.2
-5.7
-8.1
-8.2
-7.8
-7.4
-7.0
-89
-86
-113
1
1
-5
-27
-26
-29
-29
-29
-32
-31
-30
0.0
0.0
-61
-4
-3.1
-2.9
-4
-15
-3.9
-2.9
-21
-24
-4.8
-3.7
-30
-32
aStretching frequencies for the free ligands (cm-1): H2, νHH
4455.6; N2, νNN 2.472.7; C2H4, νCC 1717.2; C2H2, νCC 2087.0; C2Me2,
b
νCC 2384.1. See ref 8.
drogen is deshielded in the presence of Cp*2Eu.5 The
change in chemical shift of H2 in the presence of the
paramagnetic Cp*2Eu relative to that of free H2 implies
that the electron density in the dihydrogen molecule is
perturbed on bonding. In contrast, exposure of a solution
of Cp*2Yb in C6D12 to an atmosphere of either N2 or H2
does not perturb the ring resonance of diamagnetic
Cp*2Yb or that of dissolved H2 relative to free H2. In
addition, the T1 value of H2 is not changed in the
presence of Cp*2Yb (see the Experimental Section for
details). This result might be interpreted to mean that
the dihydrogen molecule is perturbed by Cp*2Eu but not
by Cp*2Yb. However, the chemical shift of H2 induced
by the paramagnetic Cp2*Eu will be larger than that
by the diamagnetic Cp*2Yb, and no conclusion can be
drawn about the relative strength of the interaction.6
(b) Com p u ta tion a l Stu d ies. The qualitative experi-
mental observations are numerically verified by DFT
calculations. As shown in Table 1, the equilibrium
reaction of Cp2M (Cp ) η5-C5H5) with H2 in the gas
phase lies in favor of the H2 complex by about -0.5 kcal
mol-1, for each of the three metallocenes. The calculated
changes in the H-H stretching frequency mirror the
energy changes.7 The most important deduction is that
Cp2Eu binds slightly less strongly to H2 than does
Cp2Yb. The calculations were repeated for the Cp*2M
fragment in order to examine the change induced by this
substitution; these values are listed in Table 1. The
donor-acceptor interaction is weaker by about 0.5 kcal
mol-1 relative to the Cp2M case, so that ∆E is ap-
proximately zero for each of these decamethylmetal-
locenes.8
Tm, Dy, Nd, X ) NSiMe3, OC6H3 Bu2-2,6).9b
t
(II) Alk en e a n d Alk yn e. (a ) Exp er im en ta l Ob-
ser va tion s. The experimental observation of adduct
formation or the lack thereof between Cp*2M and
nonpolar diatomic molecules (H2 and N2) can be ex-
tended to simple nonpolar organic molecules such as an
alkene or alkyne. An equilibrium reaction between
ethylene and Cp*2Eu was implied by the change in the
1H NMR chemical shift of the C2H4 resonance in the
presence of Cp*2Eu in C6D12.5 In contrast, exposure of
Cp*2Sm10a or Cp*2Yb to C2H4 in aliphatic hydrocarbon
solvent yields polyethylene; see the Experimental Sec-
tion for details for Cp*2Yb. Although mechanistic details
are incomplete, adduct formation presumably precedes
the electron-transfer event that initiates the polymer-
ization reaction, which has been suggested for bivalent
samarium.10b-e The adduct between Cp*2Yb and MeCt
CMe has been isolated and characterized by X-ray
crystallography.11 In C6D6 solution, an equilibrium
exists between the adduct and free MeCtCMe, since
the change of the chemical shift in the NMR spectra
(1H, 13C) show that electron density in MeCtCMe is
perturbed by Cp*2Yb addition. The Experimental Sec-
tion gives 1H and 13C{1H} NMR chemical shifts of
isolated adducts between Cp*2Yb and the alkynes PhCt
CPh, MeCtCPh, and MeCtCCMe3 in C6D6. In each
case, the average chemical shifts of the alkyne reson-
ances are perturbed by coordination, but the slow
exchange spectra are not obtained by cooling. No
experimental studies have been reported for Sm and Eu,
but it seems reasonable to assume these two metal-
locenes would form similar 1:1 adducts.
(5) Nolan S. P.; Marks, T. J . J . Am. Chem. Soc. 1989, 111, 8538.
Eu(II) is 4f7, and therefore the resonance due to the Cp* hydrogens is
not observed in the NMR experiment.
(6) Bertini, I.; Luchinat, C.; Parigi, G. Solution NMR of Paramag-
netic Molecules; Current Methods in Inorganic Chemistry 2; Elsevier:
Amsterdam, 2001.
(7) The calculated changes in stretching frequency upon coordination
to the metallocene fragment are more accurate and therefore more
trustworthy for the small changes in the values relative to the
uncoordinated values. For instance, see ref 4.
(8) In the case of Yb, two minima were located on the potential
energy surface, which is correlated with the different conformations
(eclipsed and staggered) of the C5R5 (R ) H, Me) ligand. The energy
difference between these two rotamers is small, and the results for
the rotamer that is lower in energy (staggered) is listed in Table 1.
(9) (a) Evans, W. J .; Ulibarri, T. A.; Ziller, J . W. J . Am. Chem. Soc.
1988, 110, 6877. (b) Evans, W. J .; Zucchi, G.; Ziller, J . W. J . Am. Chem.
Soc. 2003, 125, 10.
(10) (a) Evans, W. J .; Decoster, D. M.; Greaves, J . Macromolecules
1995, 28, 7929. (b) Boffa, L. S.; Novak, B. M. Tetrahedron 1997, 53,
15367. (c) Ihara, E.; Nodono, M.; Katsura, K.; Adachi, Y.; Yasuda, H.;
Yamagashira, M.; Hashimoto, H.; Kanehisa, N.; Kai, Y Organometallics
1998, 17, 3945. (d) Ihara, E.; Nodono, M.; Katsura, K.; Yasuda, H.;
Kanehisa, N.; Kai, Y. Macromol. Chem. Phys. 1996, 197, 1909. (e) Hou,
Z.; Zhang, Y.; Tezuka, H.; Xie, P.; Tardif, O.; Koizumi, T.; Yamazabi,
H.; Wakatsuki, Y. J . Am. Chem. Soc. 2000, 122, 10533.
(11) Burns, C. J .; Andersen, R. A. J . Am. Chem. Soc. 1987, 109, 941.

Bonding to Divalent Lanthanide Metallocenes
Organometallics, Vol. 22, No. 26, 2003 5449
(b) Com p u ta tion a l Stu d ies. Equilibrium adduct
formation between Cp2M and C2H4 in the gas phase was
explored by calculations, the results of which are shown
in Table 1. All three bivalent Cp2M fragments form
adducts with an energy change of about -6 kcal mol-1
and a lowering of the CdC stretching frequency by
about 30 cm-1 for each metallocene. The ethylene
adducts are thus energetically favorable, and the changes
in νCdC parallel the energy change. These trends are
maintained for the decamethylmetallocenes. The ir-
reversible polymerization that results in the case of
Cp*2Sm and Cp*2Yb presumably is due to electron
transfer after adduct formation. The feasibility of this
pathway is related to the reduction potential of the
metal fragments (Sm(II) > Yb(II) > Eu(II))12 and the
electron affinity of ethylene; the energy changes in these
redox processes are being studied computationally and
will be reported at a later date.
The calculated energy change upon reversible binding
of Cp2M or Cp*2M and C2H2 or C2Me2 also is shown in
Table 1. In each case, the binding is about 2 kcal mol-1
more favorable than for ethylene and the changes in νCC
upon coordination are similar for the ethylene and
acetylene adducts. The binding energy for C2H2 suggests
that an acetylene complex might be isolable, since the
calculated free energy changes at 25 °C are -0.8 kcal
mol-1 for Eu, -0.2 kcal mol-1 for Sm, and +2 kcal mol-1
for Yb: this is indeed the case for Cp*2Yb with C2Me2.8
(III) Bon d in g Tr en d s. The experimental deductions
on the relative binding abilities of the Cp*2M fragments
toward H2, N2, C2H4, C2H2, and CO2 are reproduced by
calculations on Cp2M and Cp*2M. These calculations
support the notion that adduct formation precedes
further reactions in the case of Cp*2Sm with C2H4 and
N2 and of Cp*2Yb with C2H4. Several patterns emerge
from inspection of the values in Table 1. (i) All of the
values for the change in energy on adduct formation are
small. (ii) The values of ∆E parallel the change in X-X
stretching frequency upon binding; a larger change in
∆E correlates with a larger change in ∆ν. In the cases
of binding N2, C2H4, and C2H2 (C2Me2), ∆E and ∆ν follow
the order N2 < C2H4 < C2H2 (C2Me2), for a given
metallocene. Furthermore, there is no large difference
between Sm, Eu, or Yb; however, the trend is for Yb to
bond more weakly than the two other metals. Dihydro-
gen behaves differently with respect to the same set of
Cp2M fragments,8 since Sm and Eu bind slightly more
strongly than Yb and the H2 stretching frequency is
changed more than the stretching frequency of N2, C2H4,
and C2H2 upon interaction with these metallocenes and
decamethylmetallocenes. The correspondence between
calculated binding energies and stretching frequencies
and the currently available qualitative experimental
results shows that the calculations account for the small
changes involved when a ligand bonds to these metal-
locene fragments. The next step in the theoretical
treatment is to compare the calculated molecular struc-
tures with the experimental results, where the latter
are available.
and Cp*2M fragments are described. The Cp(centroid)-
metal-Cp(centroid) angle for Cp*2M is always larger
by 1-2° than for Cp2M. The metal-Cp(centroid) dis-
tance is always shorter for Cp*2M by 0.02 Å; the
staggered conformation in all cases is the lower mini-
mum.
(a ) Dih yd r ogen . The calculated molecular structures
that result from the interaction between Cp2M and
dihydrogen are shown in Figure 1. As dihydrogen
approaches the Cp2MII fragment, the metallocene frag-
ment bends, allowing bonding to develop. The Eu‚‚‚H
distances are longer than the Yb‚‚‚H distances, and this
clearly parallels the change in H2 stretching frequencies
but not the energy,7 as shown in Table 1. The geometry
of adducts in the Cp*2M fragment is different from that
found in the Cp2M fragment.8 The ring(centroid)-M-
ring(centroid) angles in the adducts of Cp*2M increase
by 1-2° relative to those with Cp2M, and the ring-
(centroid)-metal distance decreases by 0.02 Å from
Cp*2M to Cp2M. The M‚‚‚H2 distance increases in the
Cp*2M adduct relative to the unsubstituted species, and
this lengthening parallels the change in νH and binding
2
energy. The M‚‚‚H2 distance depends on the metal, as
the changes are larger for Yb than for Eu.
The H2 complex can be used to identify the interac-
tions at work between the ligands and metal fragments
studied here. In the d transition metal dihydrogen
complexes, the stretching frequency of H2 is determined
by delocalization of the electrons in σH toward the
2
empty metal d orbital and by the back-bonding of the
electrons in an occupied metal d orbital toward σ*H . In
2
the present adducts, the back-bonding is not present.2
In addition, delocalization of the electron density from
the ligand toward the metal is drastically decreased,
which appears only in the form of dynamic polariza-
tion.13 Dynamic polarization corresponds to the defor-
mation of the electron cloud in the dihydrogen bond by
the electric field created by the dipole moment of Cp2M
and the dihydrogen molecule. Dynamic polarization is
expressed as an excitation of the occupied σH toward
2
the empty metal orbitals. This differs from classical
ligand to metal charge transfer, which is larger in
magnitude than the latter because polarization is
exclusively associated with an instantaneous interac-
tion, whereas the traditional charge transfer is associ-
ated with an explicit electron transfer. Since only the
metal fragment has a permanent dipole, only polariza-
tion of the ligand plays a role. Thus, the amount of
dynamic polarization increases on raising the energy of
the occupied orbitals of the ligand and on increasing the
overlap between the orbitals of the two fragments. Static
polarization is associated with the occurrence of an
induced dipole on the ligand by the electric field of the
metallocene. It is expressed in terms of excitation from
the occupied orbitals toward the empty orbitals of the
ligand itself. This interaction relocalizes the charges
within the ligand and thus influences the stretching
frequency. It increases with decreasing distances be-
tween partners. These properties expressed for H2
complexes can be generalized to the other ligand sys-
tems. Dihydrogen is η2 bonded with a slight distortion
toward an η1 bonding mode, which increases the elec-
(IV) Ad d u ct Geom etr y a n d Bon d in g. Before dis-
cussing the adduct formation between Cp2M and Cp*2M,
the geometrical differences between the base-free Cp2M
(12) Finke, R. G.; Keenan, S. R.; Schiraldi, D. A.; Watson, P. L.
Organometallics 1986, 5, 598.
(13) McDowell, S. A. C.; Amos, R. A.; Handy, N. C. Chem. Phys. Lett.
1995, 235, 1.

5450 Organometallics, Vol. 22, No. 26, 2003
Perrin et al.
8
F igu r e 1. Optimized geometries of adducts of Cp*2Ln and Cp2Ln (Ln ) Eu, Yb) with (a) H2 and (b) N2. Distances are
given in Å. Values in parentheses and italics are for Cp. Calculated distances for free ligands: H-H ) 0.743 Å; N-N )
1.104 Å.
trostatic interaction with Cp2M and Cp*2M through
(c) Eth ylen e a n d Acetylen e. The calculated geom-
etry of the complexes shows that the HOMO of ethylene
and acetylene approaches the open wedge of the met-
allocene fragment (Figure 2). This maximizes the dy-
namic polarization of the ligand, due to the relatively
high energy of their π-orbitals. The nature of the HOMO
in N2, C2H4, and C2H2 clearly controls the different
orientations and the nature of the interaction: static
polarization for N2 and dynamic polarization for C2H4
and C2H2. From these results, it seems that variation
in the stretching frequencies is mostly determined by
the dynamic polarization, which involves the π-systems.
In the case of acetylene, the two π-systems are involved,
which may be the reason for its greater binding energy
relative to C2H4. This is consistent also with the shorter
M-L distance for C2H2 relative to C2H4.
It should be noted that ethylene and acetylene are
found to be perpendicular to the mirror plane of the
Cp2M fragment; however, they are in the plane of the
Cp*2M fragment, as found in Cp*2Yb(η2-MeCtCMe).11
The latter orientation is preferred sterically, as shown
in the calculations on Cp*2M. The calculated Yb-
C(alkyne) distance is longer (3.14 Å with C-C ) 1.215
Å) than in Cp*2Yb(η2-MeCtCMe) (2.85 ( 0.01 Å).
static polarization of H2, illustrated as Mδ+‚‚‚Hδ-Hδ+
The binding energies are related to the extent of bending
and the resulting dipole moment in the Cp2M fragment.
The calculated dipole moment of Cp2Eu (2.2 D for a Cp-
Eu-Cp angle of 168° and 1.7 D for Cp*2Eu for an angle
of 159°) is substantially larger than for Cp2Yb (0.14 D)
and Cp*2Yb (0.0 D) for angles of 179 and 180°, respec-
tively.
.
(b) Din itr ogen . Dinitrogen binds weakly to these
metallocenes, and the geometry is end-on (Figure 1). The
metallocene interacts with the electrons in the HOMO
of N2, which results in the occurrence of an induced
dipole moment (static polarization) illustrated as
Mδ+‚‚‚Nδ-Nδ+. No significant dynamic polarization can
occur, since the low-lying π system overlaps poorly with
the empty orbitals of the metal. Dihydrogen and dini-
trogen both interact with Cp2M or Cp*2M via static
polarization, but dynamic polarization only exists with
H2. The lack of variation on the NtN stretching
frequency, in contrast with the significant lowering of
the H-H stretching frequency (Sm, Eu), points toward
the dominant influence of dynamic polarization in the
dihydrogen ligand.

Bonding to Divalent Lanthanide Metallocenes
Organometallics, Vol. 22, No. 26, 2003 5451
F igu r e 2. Optimized geometries of adducts of Cp*2Ln and Cp2Ln (Ln ) Eu, Yb) with (a) C2H4 and (b) C2Me2. Distances
are given in Å. Values in parentheses and italics are for Cp. Calculated C-C distances for free ligands: C2H4, 1.330 Å;
C2Me2, 1.210 Å.
(V) Gen er a l P r in cip les a n d Resu ltin g Bon d in g
Mod el. The general trend within the lanthanide met-
allocenes is that Cp2Sm and Cp2Eu tend to bind more
strongly than the Yb analogue in all of the adducts
considered. The calculations also document the effects
of changing from Cp to Cp*, with the general result that
the ligands bind slightly less well to Cp*M. This is
perhaps due to the neutralization of the positive charge
on M that increases by 0.1e, which weakens the donor-
acceptor interaction. A question that naturally arises
is why an interaction forms at all. DFT calculations
underestimate dispersion forces,3 and these are the
dominant forces in these molecules. This is apparent
in the comparison between the Yb-C(alkyne) distance
in the calculated MeCtCMe adduct of 3.12 Å and the
experimental value of 2.85 ( 0.01 Å for the MeCtCMe
complex.
associated with the permanent dipole of the metal
fragment (dynamic polarization). The bent-sandwich
metallocenes have a permanent dipole moment in which
the positive end is orientated along the z axis away from
the bending. The ease of bending a metallocene frag-
ment is related to the size of central metal atoms; those
with larger radii are easier to bend, which in turn
results in a larger permanent dipole moment. The
calculated dipole moment is larger for Cp2Eu or Cp*2-
Eu and Cp2Eu than for Cp*2Yb or Cp*2Yb, accounting
for the slightly stronger interaction calculated for Eu
relative to Yb.
As with CO bonding, these attractive interactions due
to static and dynamic polarization must overcome the
electron-electron repulsion of the electrons in the 4f
shell, which is greater in YbII (4f14) than in SmII (4f6)
and EuII (4f7). Thus, the net bonding is a compromise
between small attractive and repulsive forces, the net
outcome of which is ∆E < 0. Thus, the calculated energy
changes that give rise to small bonding energies de-
scribed here for molecular organometallic compounds
The following qualitative bond model may be ad-
vanced on the basis of the DFT calculations. The weak
binding interaction results from a dipole-induced dipole
interaction (static polarization) and from electron cloud
deformation within the ligand by the electric field

5452 Organometallics, Vol. 22, No. 26, 2003
Perrin et al.
within minutes. On a synthetic scale, exposure of Cp*2Yb in
pentane in a thick-walled pressure bottle to ethylene (12 atm)
produced an immediate color change from orange to green and
within seconds a precipitate of polyethylene formed. In a
separate experiment, exposure of Cp*2Yb in pentane to CO (7
atm) resulted in a color change from orange to green. When
this solution was exposed to ethylene (to a total pressure of
14 atm), no polyethylene was formed over a period of 1 h. A
similar inhibition was observed when methane or xenon was
used, but in these cases, no color change was observed and
polyethylene was not formed within 1 h. In contrast, neither
dihydrogen nor dinitrogen inhibits polyethylene formation
under similar conditions. Exposure of Cp*2Yb in pentane to
propylene (5 atm) did not result in a color change or polymer
formation; when the pressure was released, the orange color
reappeared and Cp*2Yb was recovered by crystallization.
Exposure of Cp*2Yb to styrene (1 atm) resulted in a green
solution, but no polystyrene formed.
are, in a sense, like the “van der Waals forces” between
atoms and/or nonpolar molecules in the gas phase.14 If
this is a good analogy, then the decamethylmetallo-
cenes of the 4f block metals are good models for learning
about the nature of weak metal-ligand bonds, in
general.
Exp er im en ta l Section
Gen er a l P r oced u r es. The experimental techniques used
in this work were similar to those previously described.15 All
NMR solvents were dried and deoxygenated by distillation
from sodium or potassium under nitrogen before use. The NMR
experiments were done by dissolving (Me5C5)2Yb in the desired
solvent, benzene-d6, toluene-d8, cyclohexane-d12, or methylcy-
clohexane-d14, in an NMR tube that was fitted with a J . Young
valve which was used to connect the NMR tube to a small-
volume vacuum line. The J . Young NMR tube was cooled in
ice, or to a lower temperature, the contents of the NMR tube
were exposed briefly to dynamic vacuum, and then the
contents of the NMR tube were allowed to melt, while under
static vacuum. This freeze-thaw procedure was repeated two
more times. The NMR tube, while still in the cooling liquid,
was isolated from the vacuum line, the small-volume vacuum
manifold was filled to 1 atm with the desired gas, which
expanded into the evacuated NMR tube, and the J . Young
valve was closed. In this way, the pressure of gas in the NMR
tube was slightly greater than 1 atm at room temperature.
When the added ligand was a liquid, a similar transfer
technique was used, except that the purified liquid was
contained in a flask with a greaseless joint in order to connect
the flask to the small-volume vacuum manifold. The vacuum
manifold and the NMR tube were evacuated, and the contents
of the flask containing the ligand were expanded into the
manifold and attached NMR tube. Proton and carbon NMR
spectra were recorded on a Bruker AMX spectrometer operat-
ing at 400 and 125.7 MHz, respectively, at the specified
temperature.
Rea ction of Cp *2Yb w ith Acetylen es. (a ) With 2-Bu -
tyn e. Decamethylytterbocene (0.22 g, 0.50 mol) dissolved in
pentane (15 mL) was added to a degassed solution of 2-butyne
(0.5 mL, 0.35 g, 6.5 mmol) in pentane (5 mL). The solu-
tion color changed immediately from orange to deep red. The
volume of the solution was reduced to 5 mL, and the solu-
tion was cooled to -78 °C, resulting in formation of dark
purple-red needles. When isolated and exposed to reduced
pressure, the needles seemed to lose solvent, but they did not
crumble or change color. The yield was 0.18 g (73%). Mp: 170-
173 °C. Anal. Calcd for C24H36Yb: C, 57.9; H, 7.31. Found: C,
1
54.6; H, 7.33. H NMR (C6D6, 30 °C): δ 1.99 (s, 30H), 1.27 (s,
6H). 13C{1H} NMR (C6D6, 30 °C): δ 113.4 (ring-C), 76.86
(CCH3), 10.88 (ring-Me), 3.73 (CCH3). 1H NMR of 2-butyne
1
(C6D6, 30 °C): δ 1.52. 13C NMR: δ 74.60 (s) and 3.08 (q, J CH
) 125 Hz).
(b) With Dip h en yla cetylen e. This reaction was carried
out in a manner similar to that described above. The adduct
was isolated as black blocks from hexane at -25 °C in 91%
yield. Mp: 121-123 °C. Anal. Calcd for C34H40Yb: C, 65.7; H,
1
6.50. Found: C, 64.9: H, 6.62. H NMR (C6D6, 30 °C): δ 1.95
In ter a ction of Cp *2Yb w ith H2, Sila n es, Meth a n e,
Olefin s, a n d Xen on . Several experiments were done in which
the perturbations of the 1H and 13C NMR chemical shifts of
Cp*2Yb in the presence of various gases were monitored in
cyclohexane-d12 at 25 °C. Placing an atmosphere of xenon over
(s, 30H), 6.95 (m), 7.47 (m, the last two resonances give a
1
combined integral of 10H). H NMR of diphenylacetylene: δ
7.47 (m) and 6.98 (m).
(c) With 4,4-Dim eth yl-2-p en tyn e. This reaction was
performed in a manner similar to that described above. The
adduct was isolated from a hexane solution at -25 °C as dark
green-black feathers in 83% yield. Mp: 147-150 °C. Anal.
1
Cp*2Yb did not perturb the H (1.871 ppm) and 13C chemical
shifts (114.32 and 10.60 ppm for the ring-C and the methyl-C
resonances, respectively) relative to the values for the sample
under dinitrogen; the chemical shift values were identical to
within 1 and 10 ppb in the 1H and 13C NMR spectra,
respectively. A similar lack of perturbation in the 1H NMR
spectra of Cp*2Yb was observed when the sample was exposed
to dihydrogen, methane, and trifluoromethane; in each case
the resonances due to the free ligands were observed at their
unperturbed chemical shift values. In addition, the relaxation
time, T1, for free dihydrogen of 220 ms and the chemical shift
of 4.54 ppm were identical with those of dihydrogen in the
presence of Cp*2Yb. A sample of Cp*2Yb in benzene-d6 was
exposed to 1 equiv of 1,1,2,2-tetramesityldisilane in an NMR
tube; no perturbation of the chemical shift of Cp*Yb or that of
the free ligand and its J SiH coupling constant was observed. A
similar lack of perturbation on the chemical shift of Cp*2Yb
in the presence of 1,1,1-tris(trimethylsilyl)dimethyldisilane
was observed.
C
27H42Yb: C, 60.1: H, 7.86. Found: C, 59.1; H, 7.90. 1H NMR
(C6D6, 20 °C): δ 2.00 (s, 30H), 1,41 (s, 3H), 1.13 (s, 9H).
1H NMR of 4,4-dimethyl-2-pentyne: δ 1.54 (s, 3H), 1.14 (s,
9H).
(d ) With 1-P h en ylp r op yn e. In a manner similar to that
above, the adduct was isolated as dark brown-black needles
in 85% yield. Mp: 126-129 °C. Anal. Calcd for C29H38Yb: C,
62.2; H, 6.86. Found: C, 61.8; H, 6.98. 1H NMR (C6D6, 20 °C):
δ 7.27 (m), 6.93 (m, the combined integral is 5H), 1.99 (30H),
1.48 (s, 3H). 1H NMR of 1-phenylpropyne: δ 7.42 (m) 7.01 (m,
the combined integral is 5H), 1.67 (s, 3H).
Com p u ta tion a l Deta ils. The Stuttgart-Bonn large core
relativistic effective core potentials (RECPs) were chosen for
the lanthanide centers.16 For Eu and Yb, 10e large-core
RECPs, adapted to the +II oxidation state of lanthanide, were
used. Basis sets adapted to the RECPs were used. For Sm,
Eu, and Yb, basis sets (7s6p5d1f) contracted in [5s4p3d1f] were
used. C, N, and H were represented with a 6-31G(d,p) (double-ú
quality) basis set.17 The C and H atoms of the Cp* methyl
Exposure of Cp*2Yb in cyclohexane-d12 in an NMR tube at
25 °C to an atmosphere of ethylene produced polyethylene
(14) (a) Buckingham, A. D.; Fowler, P. W.; Hutson, J . M. Chem. Rev.
1988, 88, 963. (b) Bellert, D.; Breckenridge, W. H. Chem. Rev. 2002,
102, 1595.
(15) Schultz, M.; Burns, C. J .; Schwartz, D. J .; Andersen, R. A.
Organometallics 2000, 19, 781.
(16) (a) Dolg, M.; Stoll, H.; Savin, A.; Preuss, H. Theor. Chim. Acta
1989, 75, 173. (b) Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta
1993, 85, 441.
(17) Hariharan, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28, 213.

Bonding to Divalent Lanthanide Metallocenes
Organometallics, Vol. 22, No. 26, 2003 5453
groups were represented with a 3-21G(d,p) basis set. Calcula-
tions were carried out at the DFT-B3PW9118 level with
Gaussian 98.19 Geometry optimizations were carried out
without any symmetry restrictions. The nature of the minima
was verified with analytical frequency calculations.
a fellowship (C.J .B.). We are grateful to the French
computing centers CINES and CALMIP for a generous
donation of computational time.
OM034206V
(19) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.9; Gaussian, Inc.: Pittsburgh, PA,
1998.
Ack n ow led gm en t. This work was partially sup-
ported by the Director, Office of Energy Research, Office
of Basic Energy Sciences, Chemical Sciences Division,
of the U.S. Department of Energy, under Contract No.
DE-AC03-76F00098. We thank the NSF for a fellowship
(D.J .S.) and the Fanny and J ohn Hertz Foundation for
(18) (a) Becke, A. D. J . Chem. Phys. 1993, 98, 5648. (b) Burke, K.;
Perdew, J . P.; Yang, W. In Electronic Density Functional Theory:
Recent Progress and New Directions; Dobson, J . F., Vignale, G., Das,
M. P., Eds.; Plenum: New York, 1998.