Preparation, solution behavior, and solid-state structures of (1,3-R2C5H3)2UX2, where R Is CMe3 or SiMe3 and X is a one-electron ligand

Article abstract of DOI:10.1021/om9805988

High-yield preparations of the uranium metallocenes [1,3-(Me₃Si)₂C₅H₃] ₂UCl₂ (Cp″₂UCl₂) and [1,3-(Me₃C)₂C₅H₃] ₂UCl₂ (Cp?₂UCl₂) have been developed from the reaction of UCl₄ and the corresponding magnesocenes Cp″₂Mg and Cp?₂Mg in diethyl ether. The chloride ligands can be exchanged with either Me₃SiBr or Me₃Sil to give the uranium metallocene bromides or iodides. The fluorides were prepared by the reaction of BF₃·OEt₂ with Cp″₂U(NMe)₂, Cp?₂U-(OMe)₂, or Cp?₂UMe₂. The crystal structures of Cp″₂UCl₂, Cp?₂UCl₂, Cp″₂UMe₂, Cp?₂UF₂, and dimeric (Cp″₂UF₂)₂ are reported. The idealized symmetry of the monomers Cp″₂UX₂ and Cp?₂UCl₂ is C_2v when X is F, Cl, or Br and C₂ when X is I or Me; in the dimer, the idealized symmetry is C_i. This preference is rationalized by intramolecular and intermolecular steric effects. The solution ring conformations and intramolecular exchange processes have been studied by variable-temperature ¹H NMR spectroscopy. In all cases, the low-temperature limiting spectra are consistent with the idealized symmetry observed in the solid state.

Full text of DOI:10.1021/om9805988

Organometallics 1999, 18, 1235-1246
1235
P r ep a r a tion , Solu tion Beh a vior , a n d Solid -Sta te
Str u ctu r es of (1,3-R2C5H3)2UX2, Wh er e R Is CMe3 or SiMe3
a n d X Is a On e-Electr on Liga n d
Wayne W. Lukens, J r., Sharon M. Beshouri, Laura L. Blosch,
Anthony L. Stuart, and Richard A. Andersen*
Chemistry Department and Chemical Sciences Division of Lawrence Berkeley Laboratory,
University of California, Berkeley, California 94720
Received J uly 14, 1998
High-yield preparations of the uranium metallocenes [1,3-(Me3Si)2C5H3]2UCl2 (Cp′′2UCl2)
and [1,3-(Me3C)2C5H3]2UCl2 (Cpq UCl2) have been developed from the reaction of UCl4 and
2
the corresponding magnesocenes Cp′′2Mg and Cpq Mg in diethyl ether. The chloride ligands
2
can be exchanged with either Me3SiBr or Me3SiI to give the uranium metallocene bromides
or iodides. The fluorides were prepared by the reaction of BF3‚OEt2 with Cp′′2U(NMe)2, Cpq U-
2
(OMe)2, or Cpq UMe2. The crystal structures of Cp′′2UCl2, Cpq UCl2, Cp′′2UMe2, Cpq UF2,
2
2
2
and dimeric (Cp′′2UF2)2 are reported. The idealized symmetry of the monomers Cp′′2UX2
and Cpq UCl2 is C2v when X is F, Cl, or Br and C2 when X is I or Me; in the dimer, the
2
idealized symmetry is Ci. This preference is rationalized by intramolecular and intermolecular
steric effects. The solution ring conformations and intramolecular exchange processes have
been studied by variable-temperature 1H NMR spectroscopy. In all cases, the low-temperature
limiting spectra are consistent with the idealized symmetry observed in the solid state.
In tr od u ction
tendency to undergo ligand redistribution to form the
unknown Cp*3UCl.
The metallocene chemistry of tetravalent uranium
has been dominated by two starting materials: Cp3-
UCl1-4 and Cp*2UCl2,5-8 where Cp* is Me5C5. The bulky
pentamethylcyclopentadienyl ligand was developed to
prevent ligand redistribution reactions, a feature that
dominates the solution chemistry of unsubstituted f-
block metallocenes. For example, neither Cp2UCl2 nor
Cp*(Cp)UCl2 can be isolated due to the redistribution
of their ligands in solution.9-11 Redistribution also
dominates the solution behavior of monosubstituted
cyclopentadienyl ligands since (RC5H4)2UCl2, where R
) SiMe3 or CMe3, cannot be isolated. In contrast, the
crystal structure of Cp*2UCl2 shows that it is a normal,
monomeric bent metallocene.12 This complex shows no
A few disubstituted cyclopentadienylmetallocenes of
the type Cp′′2UX2, where Cp′′ is 1,3-(Me3Si)2C5H3, have
been prepared and shown to be typical bent metal-
locenes by crystallography.13-15 These complexes and
the related Cpq UX2 complexes, where Cpq is 1,3-
2
(Me3C)2C5H3, should be ideal for studying ring confor-
mations in the solid state and ring rotation and inter-
molecular ligand exchange in solution, since the R
groups in (R2C5H3)2U(X)(Y) are diastereotopic. In addi-
tion, Lappert has shown that these U(IV) metallocenes
are ideal synthons for dimeric trivalent uranium met-
allocenes in which similar dynamic processes can be
studied. In this paper we describe the synthesis and
properties of Cp′′2UX2 and Cpq UX2, and in an ac-
2
* To whom correspondence should be addressed at the Chemistry
Department, University of California.
companying paper we describe the solution behavior of
their U(III) reduction products.
(1) Reynolds, L. T.; Wilkinson, G. J . Inorg. Nucl. Chem. 1956, 2,
246-253.
(2) Kanellakopolous, B. In Organometallics of the f-Elements;
Marks, T. J ., Fischer, R. D., Eds.; D. Reidel: Dordrecht, Holland, 1978;
pp 1-36.
Resu lts a n d Discu ssion
(3) Fischer, R. D. In Organometallics of the f-Elements; Marks, T.
J ., Fischer, R. D., Eds.; D. Reidel: Dordrecht, Holland, 1978; p 337.
(4) Marks, T. J . Prog. Inorg. Chem. 1979, 25, 223.
(5) Manriquez, J . M.; Fagan, P. J .; Marks, T. J . J . Am. Chem. Soc.
1978, 100, 3939.
(6) Fagan, P. J .; Manriquez, J . M.; Maatta, E. A.; Seyam, A. M.;
Marks, T. J . J . Am. Chem. Soc. 1981, 103, 6650-6667.
(7) Marks, T. J .; Day, V. W. In Fundamental and Technological
Aspects of Organo-f-Element Chemistry; Marks, T. J ., Fragala, I. L.,
Eds.; D. Reidel: Dordrecht, Holland, 1985; pp 115-157.
(8) Fischer, R. D. In Fundamental and Technological Aspects of
Organo-f-Element Chemistry; Marks, T. J ., Fragala, I. L., Eds.; D.
Reidel: Dordrecht, Holland, 1985; p 277.
Syn th esis a n d P h ysica l P r op er ties. The starting
metallocenes, Cp′′2UCl2 and Cpq UCl2, were prepared
2
by the reaction of UCl4 with Cp′′2Mg16 or Cpq Mg in
2
diethyl ether, followed by crystallization from hexane
(12) Spirlet, M. R.; Rebizant, J .; Apostolidis, C.; Kanellakopolus, B.
Acta Crystallogr. 1992, C48, 2135.
(13) Hitchcock, P. B.; Lappert, M. F.; Singh, A.; Taylor, R. G.; Brown,
D. J . Chem. Soc., Chem. Commun. 1983, 561.
(14) Blake, P. C.; Lappert, M. F.; Taylor, R. G.; Atwood, J . L.;
Hunter, W. E.; Zhang, H. J . Chem. Soc., Chem. Commun. 1986, 1394-
1395.
(15) Blake, P. C.; Lappert, M. F.; Taylor, R. G. Inorg. Synth. 1990,
27, 172.
(16) Duff, A. W.; Hitchcock, P. B.; Lappert, M. F.; Taylor, R. G.;
Segal, J . A. J . Organomet. Chem. 1985, 293, 271-283.
(9) Ernst, R. D.; Kennelly, W. J .; Day, C. S.; Day, V. W.; Marks, T.
J . J . Am. Chem. Soc. 1979, 101, 2656.
(10) Dormond, A. J . Organomet. Chem. 1981, 256, 47.
(11) Dormond, A.; Duval-Huet, C.; Tirouflet J . Organomet. Chem.
1981, 209, 341.
10.1021/om9805988 CCC: $18.00 © 1999 American Chemical Society
Publication on Web 03/04/1999

1236 Organometallics, Vol. 18, No. 7, 1999
Ta ble 1. 1H NMR Ch em ica l Sh ifts of th e U(IV) Meta llocen e Com p lexesa
Lukens et al.
δ(Me3X) (X ) C, Si)
δ(a2)
δ(b)
δ(other protons)
δb + δa
2
Cpq2UF2
-1.38 (24)
0.31 (5)
-16.55 (36)
-40.91 (14)
-43.85 (45)
-46.20 (43)
-14.61 (9)
-34.24 (15)
-36.23 (24)
-37.80 (60)
-39.02 (9)
-28.42 (6)
-29.52 (12)
-41.78 (10)
-25.62 (8)
-32.00 (8)
-25.89 (8)
-25.09 (4)
-10.82 (11)
-9.61 (30)
97.36 (15)
105.56 (38)
108.31 (64)
-16.14 (13)
90.46 (24)
98.29 (30)
101.30 (60)
18.18 (12)
7.61 (6)
-26.2
56.5
61.7
62.1
-30.7
56.2
62.1
63.5
-20.8
-20.8
20.0
Cpq2UCl2
Cpq2UBr2
Cpq2UI2
1.54 (15)
3.69 (13)
-0.76 (2)
-2.60 (5)
-1.60 (8)
-0.12 (10)
-0.64 (6)
-1.19 (3)
3.46 (4)
-3.82 (4)
0.14 (3)
-3.77 (3)
-1.56 (5)
-1.24 (3)
0.07 (3)
Cp”2UF2
Cp”2UCl2
Cp”2UBr2
Cp”2UI2
Cpq2UMe2
Cp”2UMe2
Cpq2U(Me)Cl
-35.43 (14) [U-Me]
-21.03 (7) [U-Me]
-52.42 (20) [U-Me]
55.6 (12)
Cp”2U(Me)Cl
46.82 (9)
-49.90 (20) [U-Me]
18.0
Cpq2U(OMe)2
Cp”2U(OMe)2
Cp”2U(NMe2)2
-9.02 (5)
-10.75 (4)
-5.70 (6)
37.12 (5) [U-OMe]
46.32 (4) [U-OMe]
9.82 (5) [U-NMe2]
-34.9
-35.8
-16.5
a
The chemical shifts are expressed in δ units with δ >0 to high field in C6D6 or C7D8 at 30 °C. The values in parentheses are the full
widths at half-maximum (Hz).
as gold or dark red needles, respectively, in yields from
50 to 80%. Cp′′2UCl2 was prepared previously by the
reaction of Cp′′Li and UCl4 in tetrahydrofuran (thf).15
Surprisingly, the reaction of Cp′′2Mg with UCl4 works
very poorly in tetrahydrofuran, presumably due to the
formation of Cp′′UCl3(thf)2. The magnesocenes were
used rather than the alkali-metal cyclopentadienyl
reagents since the magnesocenes are easily purified by
crystallization from hexane or by sublimation and since
the magnesium reagents can be isolated as pure materi-
als, enabling accurate control of stoichiometry. The
the methyl chloride complexes give molecular ions
resulting from the redistribution of the methyl and
chloride ligands. The bis(dimethylamide) complex Cp′′2U-
(NMe2)2 was obtained by metathesis using LiNMe2 in
tetrahydrofuran followed by crystallization from hex-
ane.19 The methoxides were obtained similarly using
KOMe and are very soluble in hexane. Both methoxides
give monomeric molecular ions in their mass spectra.
Curiously, like the difluorides, the colors of the meth-
oxides differ from the dihalides; Cp′′2U(OMe)2 is purple,
and Cpq U(OMe)2 is green.
2
chloride ligands in either Cp′′2UCl2 or Cpq UCl2 can be
The proton NMR chemical shifts at 30 °C of the U(IV)
bent metallocenes are listed in Table 1. The resonances
for all of the chemically inequivalent protons are
observed, and the line widths of the Me3C or Me3Si
hydrogens are between 5 and 15 Hz, generally less than
those of the a2- and b-methine protons on the cyclopen-
tadienyl ring, the widths of which range from 5 to 60
Hz. The chemical shifts of the resonances strongly
depend on the identity of X. In the dihalide compounds
of the Me3C and Me3Si series, the chemical shifts of the
a2-protons move upfield (δ <0) by 20-30 ppm as X
varies from iodide to fluoride. However, the chemical
shifts of the b-protons move downfield (δ >0) as X
changes from iodide to fluoride with a much larger
change of about 100 ppm. The average in these two
shifts may be conveniently expressed in the parameter
∆ ) δ(b) + δ(a2), which reflects the net shift of the
averaged ring C-H protons in the Me3C- and Me3Si-
substituted metallocenes. The net shift decreases in the
order I > Br > Cl > F, and the ring protons experience
a net shielding (upfield shift) of 88 ppm (Me3C) and 94
ppm (Me3Si) upon changing from X ) I to X ) F; this is
not a trivial shift. The change in the shift of the Me3C
or Me3Si resonance is small (5 ppm or less) and does
not vary systematically with the identity of X.
2
metathetically exchanged using commercially available
Me3SiX to give Cp′′2UX2 or Cpq UX2 in good yield, where
2
X is Br or I.17 This method is more convenient than the
previous use of BX3 to prepare Cp′′2UX2 (X ) Br, I),
since Me3SiX is easier to handle and use.15 All of the
dihalide complexes are soluble in hexane, from which
they were crystallized. Each gives a monomeric molec-
ular ion in its mass spectrum. All of the Cp′′2UX2
complexes are orange, while all of the Cpq UX2 com-
2
plexes are red. All of the complexes melt cleanly; the
Cpq derivatives melt ca. 20 °C higher than those of Cp′′.
The fluorides were prepared by the method used to
prepare Cp2ZrF2, viz., exchange with BF3‚OEt2.18 The
Cp′′ derivative was prepared by the reaction between
Cp′′2U(NMe2)2 and 2 equiv of BF3‚OEt2, and Cpq UF2
2
was prepared from Cpq UMe2 or Cpq U(OMe)2 and BF3‚
2
2
OEt2. Both fluorides are hexane-soluble, but Cpq UF2
2
is significantly less soluble. Curiously, the colors of the
fluorides are different from those of the other halides;
Cp′′2UF2 is dark green while Cpq UF2 is mustard yellow.
2
The metallocene fluorides give monomeric molecular
ions in their mass spectra.
The chloride ligands can also be sequentially replaced
with methyl groups, as observed previously in Cp*2UCl2
chemistry.6 One equivalent of MeLi gives the methyl
chloride complexes, isolated as orange-red crystals from
hexane. Two equivalents of MeLi gives the highly
soluble dimethyl complexes, which again may be iso-
lated by crystallization from hexane. The dimethyl
complexes yield M - Me ions in their mass spectra, but
A related, though less extensive, correlation has been
noted by Fisher in the Cp3UX series.3 In these com-
pounds, the averaged Cp proton chemical shift moves
upfield by ca. 7 ppm as X is changed from Cl to F. Fisher
also observed that δ(Cp) moves upfield when X is Et2N
and EtO in the order Cl > F > NEt2 > OEt. This order
was suggested to result from ligand-metal π-bonding
(17) Andersen, R. A. Inorg. Nucl. Chem. Lett. 1980, 16, 31.
(18) Druce, P. M.; Kingston, B. M.; Lappert, M. F.; Spalding, T. R.;
Srivastava, R. C. J . Chem. Soc. A 1969, 2106-2110.
(19) Blake, P. C.; Lappert, M. F.; Taylor, R. G.; Atwood, J . L.; Zhang,
H. Inorg. Chim. Acta 1987, 129, 12-20.

Uranium Metallocenes
Organometallics, Vol. 18, No. 7, 1999 1237
F igu r e 1. ORTEP drawing of Cp′′2UCl2 with 50% thermal
ellipsoids.
with the better π-donors causing greater shielding of
the Cp hydrogen atoms. It is tempting to offer a similar
rationalization for the dihalides reported here, since the
value of ∆ moves upfield with increasing π-donor ability
of X. To allow a closer comparison of the bent metal-
locene values with those of Cp3UX, the bent metal-
locenes with X ) NMe2 and MeO were prepared. As
shown in Table 1, for these bent metallocenes, the ∆
values move upfield in the order Cl . Me2N > F > OMe.
These data are consistent with Fisher’s π-bonding
hypothesis, although the ordering is slightly different
in the two classes of complexes.
F igu r e 2. ORTEP drawing of Cpq UCl2 with 50% thermal
ellipsoids.
2
The ∆ values for the dimethyl compounds lie between
those of the difluorides and the dichlorides, as expected
if chloride is reasonably assumed to be a poor π-donor.
Further, in the mixed methyl-chloro species, the ∆
values are additive; the observed ∆ values are nearly
the geometric mean of the dimethyl and dichloro values.
Table 1 clearly shows that the average room-temper-
1
ature H NMR chemical shifts are very sensitive to the
nature of the X ligand, presumably due to changes in
the bonding of X to the metallocene fragment. A
postulate of π-bonding is offered without proof as a
reasonable explanation for these trends. The origin of
the isotropic chemical shifts in uranium compounds is
determined by contact and dipolar contributions, and
separating these contributions is very difficult, as
previously noted by Fischer. However, the chemical
shifts are nonetheless dependent upon the ligands and
therefore diagnostic of differences in U-X bonding. The
variable-temperature data will be discussed after the
solid-state crystal structures are described, since the
variable-temperature data cannot be understood with-
out knowing the idealized molecular symmetry.
F igu r e 3. ORTEP drawing of Cp′′2UMe2 with 50% thermal
ellipsoids.
Cr ysta l Str u ctu r es. To determine whether the mo-
lecular structures of the complexes containing the
postulated strong π-donor ligands are different from
those with weak π-donor ligands, a representative series
was crystallographically characterized. The molecular
structures of Cp′′2UCl2, Cpq UCl2, Cp′′2UMe2, and Cpq -
F igu r e 4. ORTEP drawing of Cpq UF2 with 50% thermal
ellipsoids.
2
2
2
UF2 are shown in Figures 1-4. The important distances
and angles in these complexes are given in Table 2. In
the solid state, Cp′′2UCl2 has crystallographic C2 sym-
metry with staggered cyclopentadienyl ligands. The
bulky SiMe3 groups are nearly eclipsed, and the Cp′′
ligands point toward the front of the metallocene wedge
(the front of the Cp′′ ligand is defined by the unique
b-hydrogen atom located between the R groups of the
Cp′′ ring). The ring conformations in Cpq UCl2 and Cpq -
U-Cl distance in Cpq UCl2, in agreement with the
2
difference between the sizes of fluoride and chloride.20
The structure of Cp′′2UMe2 is quite different from that
of the other metallocenes. Although the idealized sym-
metry is C2, the conformation of the rings is different
from that in the dihalides. The Cp′′ rings are staggered,
with one SiMe3 group on each cyclopentadienyl ring
pointing between the methyl groups bonded to the
uranium center. This conformation results in an opening
2
2
UF2 are similar. All three compounds have similar
X-U-X angles, and both of the dichlorides have the
same U-Cl distance as that in Cp*2UCl2, 2.583(6) Å.
(20) Pauling, L. The Nature of the Chemical Bond; Cornell Univer-
sity Press: Ithaca, NY, 1960.
The U-F distance in Cpq UF2 is 0.5 Å shorter than the
2

1238 Organometallics, Vol. 18, No. 7, 1999
Lukens et al.
Ta ble 2. Str u ctu r a l P a r a m eter s for Mon om er ic
U(IV) Meta llocen esa
Cp′′2UCl2 Cpq2UCl2 Cp′′2UMe2 Cpq2UF2
U-X, Å
2.573(1)
2.71(2)
2.42
124.7
95.2(2)
2.577(4)
2.77(5)
2.49
128.1
91.0(2)
2.42(2)
2.72(3)
2.44
130.8
105.0(7)
2.086(2)
2.74(3)
2.46
125.3
95.4
U- Cring , Å
U-Cp, Å
Cp-U-Cp′, deg
X-U-X′, deg
a
Cp is the centroid of the cyclopentadienyl ring.
of the Me-U-Me angle by almost 10° relative to that
found in the dihalides.
The conformations of the Cp′′ or Cpq ligands are
directly related to the radii of the X ligands in these
Cp2UX2 complexes, which have an idealized C2v struc-
ture when X is F, Cl, or Br and a C2 structure when X
is I or Me.19 The van der Waals radii of F, Cl, Br, I, and
Me are 1.35, 1.8, 1.95, 2.15, and 2.0 Å, respectively.20
The role of the radius of X is shown by the trend in the
close contacts between the X ligand and the methyl
carbons of the SiMe3 or CMe3 groups; the smaller the
ligand, the shorter the closest contact distance. For Cpq -
2
UF2, the closest F‚‚‚CH3C distance is 3.28 Å. The
analogous Cl‚‚‚CH3C contact distance is larger in Cpq -
2
UCl2 at 3.60 Å, while the shortest Cl‚‚‚CH3Si distance
in Cp′′2UCl2 is 3.71 Å. When the C2 structure is adopted
rather than the C2v structure, the H3C‚‚‚CH3Si distance
in Cp′′2UMe2 is increased to 3.82 Å in accord with the
larger radius of CH3.
F igu r e 5. Packing diagram of Cp′′2UCl2 showing the “head
to tail” packing of the Cp′′2UCl2 molecules.
in opposite directions presumably to maximize the
interactions between their molecular dipoles. The spac-
ing between the metallocenes is the b-axis length. Note
A similar explanation can account for the observed
geometry of [1,2-(Me3C)2C5H3]2MX2, which adopts a C2
geometry.21 Since the Cl ligands of [1,2-(Me3C)2C5H3]2-
UCl2 would lie directly under the Me3C groups if this
that in Cpq UF2 the uranium atoms are 6.30 Å apart,
2
versus 6.89 Å in Cpq UCl2. Presumably the greater
2
dipole-dipole interaction in solid Cpq UF2 is responsible
2
molecule adopted the C2v geometry of Cp′′2UCl2 or Cpq -
2
for its greatly decreased solubility in aliphatic hydro-
carbon solvents. Therefore, crystal packing forces rather
than intramolecular steric considerations appear to be
responsible for the preference for the C2v geometry,
particularly when X is small. When X is large, the
complexes adopt the C2 geometry to minimize steric
interactions between X and the SiMe3 or CMe3 groups.
In solution, the intermolecular forces are presumably
very small and therefore ligand size and electronic
effects dominate the molecular properties.
UCl2, [1,2-(Me3C)2C5H3]2UCl2 adopts the C2 geometry
of Cp′′2UMe2 in which the SiMe3 groups of each Cp′′
ligand straddle the CH3 ligands. It is worth noting that
monomeric [1,2-(Me3C)2C5H3]2MX complexes should
adopt the eclipsed C2v geometry, since the loss of one of
the Cl ligands would remove the unfavorable interac-
tions with the CMe3 groups which force [1,2-(Me3C)2-
C5H3]2MX2 to adopt the C2 geometry.
While it appears that the C2v geometry is preferred
due to the minimization of the steric interactions
between the Cp ring and the SiMe3 or CMe3 group, this
cannot be the only reason that this geometry is pre-
While the previous complexes all have the typical bent
metallocene structure, the crystal structure of [Cp′′2-
UF2]2 is different, since it is a centrosymmetric dimer,
as shown in Figure 6. Distances and angles in [Cp′′2-
UF2]2 are given in Table 3. The structure of [Cp′′2UF2]2
is strongly reminiscent of the structures of [(MeCp)2ZrH-
(µ-H)]222 and the (Cp2WH)2(µ-H) cation.23 The Cp′′rings
and SiMe3 groups on each metallocene fragment are
staggered such that one Cp′′ group points to the front
of the metallocene wedge and one points to the rear,
giving the complex idealized Ci symmetry. The terminal
ferred. In Cpq UF2, the closest Cp‚‚‚CH3C contact is 4.30
2
Å. In Cpq UCl2 and Cp′′2UCl2, the closest analogous
2
distances are 3.69 and 3.87 Å, respectively. In Cp′′2-
UMe2, the closest Cp‚‚‚Me3Si contact between Cp” rings
is 3.83 Å. In comparing the structures of Cp′′2UCl2 and
Cp′′2UMe2, the Cp‚‚‚CH3Si distances are similar al-
though the Cl‚‚‚CH3Si distance is shorter than the
H3C‚‚‚CH3Si distance. If only intramolecular steric
interactions were important, both complexes should
have the same C2 geometry, but they do not.
In addition to intramolecular steric interactions,
crystal packing, and therefore intermolecular effects,
appears to be important in determining the geometry
of the ligands in these complexes. As shown in Figure
5, Cp′′2UCl2 packs such that the metallocenes are “head
to tail” with adjacent chains of metallocenes running
U-F distance is similar to that found in Cpq UF2, and
2
the bridging U-F distances are longer at 2.30 and 2.34
Å. The U‚‚‚U distance is 3.95 Å, shorter than that in
[Cp′′2UX]2, where X is Cl or Br.14 Cp′′2UF2 is able to
dimerize due to the small radius of F. Since F is small,
three fluoride ligands can coordinate to the uranium
center while three chloride, bromide, or iodide ligands
(22) J ones, S. B.; Petersen, J . L. Inorg. Chem. 1981, 20, 2889.
(23) Klingler, R. J .; Huffman, J . C.; Kochi, J . K. J . Am. Chem. Soc.
1980, 102, 208.
(21) Hughes, R. P.; Lomprey, J . R.; Rheingold, A. L.; Haggerty, B.
S.; Yap, G. P. A. J . Organomet. Chem., 1996, 517, 89.

Uranium Metallocenes
Organometallics, Vol. 18, No. 7, 1999 1239
F igu r e 6. ORTEP drawing of [Cp′′2UF2]2 with 50%
thermal ellipsoids.
Ta ble 3. Dista n ces (Å) a n d An gles (d eg) in
F igu r e 7. Plots of δ versus 1/T for [Cp′′2UF2]2 (a, top) and
a
[Cp ′′2UF 2]2
Cpq UF2 (b, bottom).
2
U-F1
U-F2
U-F2′
U- Cring
2.073(5)
2.297(5)
2.343(5)
2.76(4)
U-Cp1
U-Cp2
U-U
2.48
2.47
3.950(1)
the type of disubstituted cyclopentadienyl compounds
described in this paper. They all have C2 or higher
symmetry in the solid state, and the predominance of
an isomer of lower symmetry in solution is unlikely.
This deduction is supported by the observation that the
averaged chemical shifts of the Me3Si groups below the
coalescence temperature is different from the observed
chemical shift above the coalescence temperature (see
Figure 7a). This disagreement shows that the Curie-
Weiss law is not obeyed and that some temperature-
dependent process (or processes) is operating in solution.
The temperature dependence of the NMR chemical
F1-U-F2
F1-U-F2′
F2-U-F2′
75.6(2)
138.9(2)
63.3(2)
Cp1-U-Cp2
U-F2-U′
126.3
116.7(2)
a
Cp1 and Cp2 are the centroids of the two crystallographically
independent cyclopentadienyl rings.
would presumably occupy too great a volume. Since the
dimeric nature of Cp′′2UF2 may play a role in the
difference in the NMR spectrum of this complex relative
to that of the other Cp′′2UX2 complexes, the variable-
temperature NMR study of these molecules was initi-
ated to study the dynamics of these complexes in
solution.
shifts of monomeric Cpq UF2 is shown in Figure 7b.
2
Since the solid-state structure of this molecule has C2
symmetry, two Me3C resonances are expected at low
temperature, and two broad resonances are observed
at -100 °C. These resonances coalesce by -90 °C, but,
as in [Cp′′2UF2]2, the δ vs 1/T plots are again nonlinear.
One explanation of these observations is that, in solu-
tion, a rapid equilibrium between monomers and dimers
exists for both complexes. Since the ratio of monomer
to dimer changes with temperature, the chemical shifts
of the protons will be nonlinear in 1/T. Similar behavior
observed in the variable-temperature NMR spectra of
Cp3UF was also attributed to the presence of a mono-
mer-dimer equilibrium.24 Further, this type of metal-
locene dimer is likely to be a model for the transition
state that is involved in the halide exchange between
metallocenes Cp2UX2 and Cp2UY2 to give two Cp2U(X)-
(Y) species.
Var iable-Tem per atu r e NMR Spectr oscopy. While
[Cp′′2UF2]2 is dimeric in the solid state, its degree of
association in solution is unknown. To explore the
solution constitution of these metallocenes, their vari-
able-temperature NMR behavior was examined. Plots
of δ versus 1/T for [Cp′′2UF2]2 and Cpq UF2 are shown
2
in Figure 7 and reveal that the variable-temperature
behaviors of these two difluoride metallocenes are very
different. For [Cp′′2UF2]2, the ring a2- and b-protons
broaden and disappear at ca. -30 °C and do not
reappear. The Me3Si resonances, however, decoalesce
at ca. -40 °C to give four equal-area resonances. This
is the expected slow exchange pattern for a dimeric
molecule with inversion symmetry as found in the solid-
state structure (Figure 6). In order for the monomer to
have all of its Me3Si groups inequivalent, it must be
asymmetric, a condition we have not observed among
(24) Fischer, R. D.; von Ammon, R.; Kanellakopolos, B. J . Orga-
nomet. Chem. 1970, 25, 123.

1240 Organometallics, Vol. 18, No. 7, 1999
Lukens et al.
F igu r e 9. Plots of ø versus 1/T for Cpq UCl2 (a, top) and
2
F igu r e 8. Plots of δ versus 1/T for Cp′′2U(OMe)2 (a, top)
Cpq UF2 (b, bottom).
and Cpq U(OMe)2 (b, bottom).
2
2
The variable-temperature behavior of Cpq U(OMe)2
ing in different NMR spectra for these complexes
seemed small, the variable-temperature magnetic sus-
ceptibility of these complexes was examined to deter-
mine whether the complexes had similar electronic
structures in the solid state. Plots of 1/ø versus T for
Cpq UCl2 and Cpq UF2 are shown in Figure 9. The
2
was also studied. The plots of δ vs 1/T are linear to -100
°C, where the lines begin to broaden. Thus, above -100
°C, all of the resonances follow Curie-Weiss behavior,
and no temperature-dependent equilibria need to be
postulated. Presumably, the broadening is due to a
decoalescence of the ring Me3C resonances into two
resonances in a molecule of C2 symmetry. The behavior
of Cp′′2U(OMe)2 is similar, except that the Me3Si
resonances follow Curie-Weiss behavior to -70 °C,
where they decoalesce into two equal-area resonances
(Figure 8). At temperatures greater that Tc, all of the
resonances follow Curie-Weiss law. Unlike the difluo-
rides, the chemical shifts of the protons of the dimethox-
ides display Curie-Weiss behavior, which implies that
no monomer-dimer equilibrium is present in these
complexes. At low temperature, the resonances of the
a2-protons and the SiMe3 groups decoalesce, each form-
ing two new resonances; however, the resonances of the
b-proton and OMe group remain sharp. Since the
molecule cannot have inversion symmetry and since no
single mirror plane can make both of the b-protons and
the OMe groups equivalent, the complex presumably
has the same C2 structure as Cp′′2UMe2. The barriers
to site exchange calculated using the SiMe3 and a2-
protons are 8.5 and 8.6 kcal/mol, respectively. Since the
difluorides and dimethoxides have similar NMR spectra
at room temperature but have very different geometric
structures, it is unlikely that the molecular structures
of the complexes are responsible for the change in NMR
spectra versus the complexes of the heavier dihalides.
Solid -Sta te Ma gn etic Su scep tibility. Since the
likelihood of a difference in molecular structure result-
2
2
electronic ground states of the complexes are different.
In Cpq UCl2, the susceptibility is temperature-indepen-
2
dent at low temperature, typical of a U(IV) complex with
low symmetry, since no degeneracy remains among the
electronic states. Surprisingly, the low-temperature
susceptibility of Cpq UF2 is not temperature indepen-
2
dent, which suggests that the two lowest lying states
are coincidentally degenerate or very nearly so. At
higher temperatures, Cpq UF2 has a slightly lower
2
magnetic moment than Cpq UCl2, again supporting the
2
premise that the electronic structures of the complexes
are not the same. Since the electronic structures of the
complexes are different, the dipolar contributions to the
chemical shifts of the protons of the complexes are also
different, and the complexes will have different NMR
spectra. Presumably, all of the complexes possessing
ligands that are strong π-donors have electronic struc-
tures similar to Cpq UF2 and this similarity is respon-
2
sible for their similar solution chemical shift behavior.
Due to the low symmetry of the complexes, their
electronic structures were not further examined.
Con clu sion . The syntheses of several U(IV) metal-
locene complexes have been described. Complexes with
ligands that are strong π-donors have different NMR
spectra from those with poor π-donor ligands. Some of
these complexes have been crystallographically char-
acterized. The conformations of the Cp rings in the

Uranium Metallocenes
Organometallics, Vol. 18, No. 7, 1999 1241
and filtered. The volume of the filtrate was reduced to ca. 20
mL, and the mixture was heated to redissolve the solid.
Cooling to -20 °C produced large, blood red prisms (7.5 g,
93%). Mp: 165-167 °C. IR: 3070 (w), 3060 (w), 1270 (w), 1258
(w), 1248 (w), 1200 (w), 1165 (w), 1162 (w), 1050 (w), 1020
(w), 922 (w), 833 (w), 792 (w), 775 (w), 670 (w), 664 (w), 470
(w), 355 (w), 280 (w) cm-1. MS (M+) m/z (calcd, found): 662
(100, 100), 663 (30, 61), 664 (68, 66), 665 (19, 21), 666 (13,
11). Anal. Calcd for C26H42Cl2U: C, 47.1; H, 6.33; Cl, 10.7.
Found: C, 47.1; H, 6.45; Cl, 10.7.
complexes differ, and the source of the differences is
discussed. One of the complexes, [Cp′′2UF]2, is a dimer.
Variable-temperature NMR spectroscopy suggests that
a monomer-dimer equilibrium is present in the difluo-
rides. Variable-temperature magnetic susceptibility
shows that the electronic structures of Cpq UCl2 and
2
Cpq UF2 are different, which is thought to be responsible
2
for the difference observed in the NMR spectra of these
complexes.
(b) A mixture of Cpq2Mg (5.00 g, 13.2 mmol) and UCl4 (6.26
g, 16.5 mmol) was suspended in 200 mL of diethyl ether. The
solution slowly turned cherry red. After 2 days of stirring, the
solvent was removed under reduced pressure. The red solid
residue was suspended in 200 mL of hexane, and the mixture
was heated to 70 °C and then allowed to settle. The red
solution was filtered, and the volume of the filtrate was
reduced to 40 mL. The solution was heated to 70 °C to
redissolve the solid. Cooling to -20 °C produced bright red
needles (4.77 g, 50%, based upon Cpq2Mg). The 1H NMR
spectrum was identical with that of the product of route a.
Cp ′′2UCl2.15 A mixture of UCl4 (2.54 g, 6.68 mmol) and Cp′′2-
Mg16 (2.96 g, 6.68 mmol) was suspended in 100 mL of diethyl
ether. The reaction mixture slowly turned orange. After the
mixture was stirred for 3 days, the diethyl ether was removed
under reduced pressure, and the orange solid residue was
suspended in 150 mL of hexane. The orange solution was
filtered, and the volume of the solution was reduced to ca. 30
mL. The solution was heated to redissolve the solid, and slow
cooling to -80 °C gave orange needles (3.3 g, 68%). The 1H
NMR spectrum agrees with that reported previously.15 (Note:
this reaction does not work in tetrahydrofuran presumably due
to the formation of Cp”UCl3(thf)2.)
Exp er im en ta l Deta ils
All reactions and manipulations were carried out in an inert
atmosphere using standard Schlenk and drybox techniques.
Hexane, diethyl ether, and tetrahydrofuran were dried over
sodium benzophenone ketyl, distilled, and degassed immedi-
ately prior to use. Toluene, methylcyclohexane, and deuterated
NMR solvents were dried over and distilled from potassium
or sodium. Me3SiBr and Me3SiI were distilled under argon and
stored over copper powder before using.
Infrared spectra were recorded on a Perkin-Elmer 283
spectrometer as Nujol mulls between CsI plates. 1H NMR
spectra were measured on a J EOL FX-90Q FT NMR spec-
trometer operating at 89.56 MHz. Chemical shifts were
referenced to tetramethylsilane (δ 0), with positive values at
lower field. Unless otherwise noted, all spectra were acquired
at 30 °C in C6D6, and the chemical shifts are listed in Table 1.
Melting points were measured on a Thomas-Hoover melting
point apparatus in sealed capillaries and are uncorrected.
Susceptibility measurements were carried out on a SHE Model
500 SQUID susceptometer. Electron impact mass spectra were
recorded by the mass spectroscopy laboratory, and elemental
analyses were performed by the analytical laboratories, both
at the University of California, Berkeley. Numerical modeling
of susceptibilities was done using the program Horizon.
The compounds Cp′′2UCl2, Cp′′2UBr2, and Cp′′2UI2 have been
reported previously.15 Our syntheses are somewhat different
and so are reported here. Cp′′2U(NMe2)2 has been reported,
but with no experimental details or characterization.19
Cp qK. Potassium (4.6 g, 120 mmol) was washed with
hexane, cut into little pieces, and dried under vacuum in a
500 mL Schlenk flask. Tetrahydrofuran (150 mL) was added,
and CpqH25 (20 mL, 16.6 g, 93 mmol) was added by syringe.
The reaction is slow. The mixture was stirred, and the flask
was periodically vented to the Schlenk line. After 7 days, gas
was no longer evolved, and the tetrahydrofuran was removed
under reduced pressure. The remaining light pink solid was
dried under vacuum. The flask was taken into the drybox, and
the remaining potassium was physically removed from the
powder (18.9 g, 94%). The compound was used without further
characterization. The compound is insoluble in all common
solvents except for hot tetrahydrofuran, in which it partially
decomposes.
Cp q2UBr 2. Cpq2UCl2 (0.50 g, 0.75 mmol) was dissolved in
50 mL of hexane, and Me3SiBr (0.49 g, 3.0 mmol) was added
using a syringe. The solution was stirred for 12 h, and then
the volume was reduced to ca. 5 mL. Cooling to -80 °C
produced red blocks (0.52 g, 91%). Mp: 185-187 °C. IR: 3110
(w), 3081 (w), 2721 (w), 1290 (w), 1249 (w), 1205 (m), 1195
(m), 1167 (s), 1058 (m), 1029 (w), 1024 (w), 938 (m), 930 (w),
864 (m), 844 (s), 780 (s), 734 (w), 722 (w) cm-1. MS (M+) m/z
(calcd, found): 750 (50, 87), 751 (15, 49), 752 (100, 100), 753
(29, 72), 754 (52, 89), 755 (14, 43). Anal. Calcd for C26H42
-
Br2U: C, 41.5; H, 5.63. Found: C, 41.9; H: 5.69. (Note: the
product often needed a second treatment with Me3SiBr to
completely metathesize all of the chlorides.) The NMR spec-
trum of the impurity, Cpq2U(Br)(Cl), is distinct from the
spectrum of Cpq2UBr2 and Cpq2UCl2. In all cases, the mixed
halide complex is clearly visible in the NMR spectrum when
the metathesis is incomplete.
Cp ”2UBr 2.15 Cp′′2UCl2 (2.00 g, 2.75 mmol) was dissolved
15
in 50 mL of diethyl ether, giving an orange-green solution. Me3-
SiBr (1.1 mL, 1.3 g, 8.4 mmol) was added using a syringe. The
solution slowly darkened. After the mixture was stirred for
10 h, the volatile components were removed under reduced
pressure. The orange solid residue was dissolved in 50 mL of
hexane, giving a purple-red solution, which was then filtered.
The volume of the filtrate was reduced to ca. 25 mL, and the
solution was heated to dissolve the solid. Cooling to -20 °C
produced bright orange needles (1.85 g, 82%). The 1H NMR
spectrum agrees with that reported previously.15 (Note: the
product often needed a second treatment with Me3SiBr to
completely metathesize all of the chlorides. This reaction does
not work in hexane.)
Cp q2UI2. Cpq2UCl2 (1.5 g, 2.3 mmol) was dissolved in 50 mL
of diethyl ether, and Me3SiI (0.92 mL, 1.4 g, 6.8 mmol) was
added using a syringe. The initially red solution became
purple. After the mixture was stirred for 3 days, the ether was
removed under reduced pressure, and the dark solid residue
was dissolved in 75 mL of hexane. The solution was filtered,
and the volume of the filtrate was reduced to ca. 45 mL.
Cp q2Mg. CpqH25 (18.8 mL, 15.6 g, 87.3 mmol) was added by
syringe to a stirred solution of Bu2Mg (60 mL, 0.64 M in
heptane, 38.4 mmol). The solution became slightly cloudy and
was heated to reflux for 3 days. The solution was cooled and
then was filtered. Cooling the filtrate to -80 °C for 7 days
1
produced colorless crystals (11.8 g, 81%). H NMR (C6D6): δ
5.90 (m, 3H, ring protons), 1.31(s, 18H, CMe3). The compound
was used without further characterization.
Cp q2UCl2. (a) A mixture of KCpq (5.25 g, 23.1 mmol) and
UCl4 (4.60 g, 11.6 mmol) was suspended in 100 mL of
tetrahydrofuran. The reaction mixture immediately turned
deep red and became hot. After the mixture was stirred for 8
h, the tetrahydrofuran was removed under reduced pressure.
The solid residue was suspended in 150 mL of diethyl ether
(25) Venier, C. G.; Casserly, E. W. J . Am. Chem. Soc. 1990, 112,
2808-2809.

1242 Organometallics, Vol. 18, No. 7, 1999
Lukens et al.
Cooling to -20 °C yielded purple-orange needles (1.52 g, 80%).
Mp: 180-186 °C. IR: 3062 (w), 2722 (w), 1247 (s), 1198 (w),
(M)+ m/z (calcd, found): 694 (100, 100), 695 (45, 68), 696 (23,
33), 697 (7, 12). Anal. Calcd for C22H42F2Si4U: C, 38.0; H, 6.09.
Found: C, 37.5; H, 6.26.
1167 (m), 1076 (s), 1057 (w), 1025 (m), 847 (s), 788 (s) cm-1
.
MS (M)+ m/z (calcd, found): 846 (100, 100), 847 (30, 31). Anal.
Calcd for C26H42I2U: C, 36.9; H, 5.00. Found: C, 36.9; H, 5.14.
(Note: the product often needed a second treatment with Me3-
SiI to completely metathesize all of the chlorides.)
Cp ′′2U(NMe2)2. Cp′′2UCl215 (4.00 g, 5.50 mmol) and LiNMe2
(0.57 g, 11 mmol) were suspended in 100 mL of tetrahydro-
furan. The solution immediately became warm and turned
dark yellow-green. After the mixture was stirred for 12 h, the
tetrahydrofuran was removed under reduced pressure. The
green solid residue was suspended in 90 mL of hexane, and
the solution was filtered. The volume of the filtrate was
reduced to ca. 10 mL. Cooling to -20 °C produced orange
blocks (2.82 g, 69%). Mp: 125-127 °C. IR: 3050 (w), 2767 (s),
1317 (w), 1245 (s), 1208 (m), 1141 (s), 1122 (w), 1079 (s), 1058
(m), 919 (s), 833 (s), 779 (s), 754 (s), 689 (m), 636 (m), 616 (w)
cm-1. MS (M)+ m/z (calcd, found): 744 (100, 100), 745 (50, 52),
746 (25, 26). Anal. Calcd for C26H54N2Si4U: C, 41.9; H, 7.30;
N, 3.76. Found: C, 41.9; H, 7.33; N, 3.29.
Cp ′′2UI2.15 Cp′′2UCl2 (1.5 g, 2.1 mmol) was dissolved in
15
50 mL of hexane, and Me3SiI (0.84 mL, 1.2 g, 6.2 mmol) was
added by syringe. The initially orange solution quickly turned
red. After the mixture was stirred for 10 h, the volatile
components were removed under reduced pressure, and the
dark red solid residue was dissolved in 75 mL of hexane. The
solution was filtered, and the volume of the filtrate was
reduced to ca. 45 mL. Cooling to -20 °C gave purple-brown
1
blocks (1.52 g, 81%). The H NMR spectrum agrees with that
reported previously.15 (Note: the product often needed a second
treatment with Me3SiI to completely metathesize all of the
chlorides.)
Cp q2U(OMe)2. Cpq2UCl2 (3.00 g, 4.52 mmol) and KOMe
(0.65 g, 9.3 mmol) were suspended in 50 mL of tetrahydrofu-
ran. After the mixture was stirred at 70 °C for 12 h, the
tetrahydrofuran was removed under reduced pressure, yielding
a viscous green oil, which slowly solidified under vacuum. The
green solid residue was suspended in 50 mL of hexane, and
the solution was filtered. The volume of the filtrate was
reduced to ca. 5 mL, and additional white solid precipitated.
After 12 h the solution was filtered and the volume was
reduced to ca. 2.5 mL. Cooling to -20 °C produced green blocks
(2.05 g, 69%). Mp: 86-88 °C. IR: 3067 (w), 2726 (w), 1297
(w), 1251 (s), 1201 (m), 1166 (m), 1115 (s), 1093 (s), 1055 (m),
1025 (m), 975 (w), 936 (m), 822 (s), 753 (s), 723 (w), 679 (m),
657 (m) cm-1. MS (M)+ m/z: 654. Anal. Calcd for C28H48O2U:
C, 51.4; H, 7.39. Found: C, 50.3; H, 7.25.
Cp q2UF 2. (a) Cpq2UMe2 (1.23 g, 2.00 mmol) was dissolved
in 30 mL of diethyl ether, and BF3‚OEt2 (0.51 mL, 0.59 g, 4.1
mmol) was added using a syringe. The solution immediately
became warm and turned green. After 12 h, the diethyl ether
was removed under reduced pressure, and the tube was heated
to 80 °C under vacuum for 1 h to remove MeBF2. The orange
solid residue was dissolved in 50 mL of hexane, and the
solution was filtered. The volume of the filtrate was reduced
to ca. 25 mL. Cooling to -20 °C produced orange-yellow
needles (1.00 g, 80%). Mp: 173-176 °C. IR: 3100 (w), 2730
(w), 1290 (w), 1255 (s), 1205 (m), 1165 (m), 1060 (m), 1035
(w), 925 (m), 825 (m), 820 (s), 785 (m), 725 (w), 685 (m), 665
(m), 510 (s), 480 (s), 430 (w), 355 (m), 245 (m) cm-1. MS (M)+
m/z: 630. Anal. Calcd for C26H42F2U: C, 49.5; H, 6.71. Found:
C, 49.5; H, 6.86.
15
Cp ′′2U(OMe)2. Cp′′2UCl2 (2.00 g, 2.57 mmol) and KOMe
(0.40 g, 5.6 mmol) were suspended in 50 mL of tetrahydrofu-
ran. The mixture was heated to 70 °C and was stirred for 12
h. The mixture was filtered, the tetrahydrofuran was removed
under reduced pressure, and the solid was heated to 60 °C
under vacuum for 1 h. The dark solid residue was suspended
in 50 mL of hexane and the suspension filtered. The volume
of the solution was reduced to ca. 10 mL, and a white solid
precipitated. After 12 h, the solution was filtered, and the
volume of the solution was reduced to ca. 2 mL. Cooling to
-20 °C produced dark purple blocks (1.65 g, 84%). Mp: 95-
99 °C. IR: 3043 (w), 2812 (m), 1316 (w), 1247 (s), 1212 (w),
1113 (s), 1086 (s), 1055 (m), 920 (s), 835 (s), 781 (m), 754 (s),
689 (w), 634 (s), 620 (w) cm-1. MS (M)+ m/z: 718. Anal. Calcd
for C24H48O2Si4U: C, 40.1; H, 6.73. Found: C, 39.7; H, 6.70.
(b) Cpq2U(OMe)2 (1.53 g, 2.34 mmol) was dissolved in 25 mL
of diethyl ether, and BF3‚OEt2 (0.60 mL, 0.70 g, 4.9 mmol) was
added using a syringe. The color of the solution immediately
changed from green to red. After the mixture was stirred for
6 h, the ether was removed under reduced pressure, and the
tube was heated to 70 °C under vacuum for 1 h. The red solid
residue was suspended in 50 mL of hexane, and the solution
was filtered. The volume of the filtrate was reduced to ca. 30
mL. Cooling to -20 °C gave yellow needles (1.1 g, 74%). The
1H NMR spectrum was identical with that of Cpq UF2 produced
2
by route a.
Cp q2UMe2. Cpq2UCl2 (1.00 g, 1.51 mmol) was dissolved in
50 mL of ether, and MeLi (4.2 mL, 0.72 M in diethyl ether,
3.0 mmol) was added by syringe. The solution immediately
turned orange and cloudy. After the mixture was stirred for 1
h, the ether was removed under reduced pressure. The dark
orange solid residue was suspended in 25 mL of hexane, and
the solution was filtered, giving a deep orange solution. The
volume of the solution was reduced to ca. 2 mL. Cooling to
-20 °C gave orange-brown blocks (0.73 g, 78%). Mp: 120-
125 °C. IR: 1250 (s), 1200 (m), 1165 (m), 1110 (s), 1055 (m),
1025 (m), 935 (m), 845 (m), 825 (s), 810 (m), 765 (s), 675 (m),
655 (m), 405 (m) cm-1. MS (M - CH3)+ m/z: 659. Anal. Calcd
for C28H48U: C, 54.0, H, 7.71. Found C, 53.7; H, 7.83.
[Cp ′′2UF 2]2. Cp′′2U(NMe2)2 (0.50 g, 0.67 mmol) was dis-
solved in 25 mL of diethyl ether, and BF3‚OEt2 (0.17 mL, 0.19
g, 1.3 mmol) was added using a syringe. The orange solution
turned bright green after 5 min. After the mixture was stirred
for 12 h, the volatile components were removed under reduced
pressure, and the reaction mixture was heated to 70 °C under
vacuum to remove Me2NBF2. The green solid residue was
dissolved in 40 mL of hexane, and the solution was filtered.
The volume of the solution was reduced to ca. 2 mL. Cooling
to -20 °C produced green blocks (0.34 g, 72%). Mp: 114-116
°C. IR: 3041 (m), 1318 (w), 1247 (s), 1204 (m), 1079 (s), 917
(s), 839 (s), 793 (s), 753 (s), 692 (m), 637 (s), 619 (w), 515 (s),
475 (s), 370 (s), 355 (s), 310 (s), 270 (m), 245 (w) cm-1. MS
15
Cp ′′2UMe2. Cp′′2UCl2 (2.00 g, 2.75 mmol) was dissolved
in 50 mL of diethyl ether and MeLi (10.6 mL, 0.52 M in diethyl
ether, 5.5 mmol) was added by syringe. The solution im-
mediately turned red-orange and cloudy. After the mixture was
stirred for 10 h, the solvent was removed under reduced
pressure. The orange solid residue was suspended in 50 mL
of hexane, and the solution was filtered. The volume of the
filtrate was reduced to ca. 3 mL. Cooling to -20 °C produced
red-orange crystals (1.45 g, 77%). Mp: 120-125 °C. IR: 1250
(s), 1200 (m), 1165 (m), 1110 (s), 1055 (m), 1025 (m), 935 (m),
845 (m), 825 (s), 810 (m), 765 (s), 675 (m), 655 (m), 405 (m)
cm-1. MS (M - CH3)+ m/z: 659. Anal. Calcd for C24H48Si4U:
C, 54.0; H, 7.71. Found: C, 53.7; H, 7.83.
Cp q2U(Me)Cl. Cpq2UCl2 (2.00 g, 3.01 mmol) was dissolved
in 50 mL of hexane. MeLi (5.80 mL, 0.52 M in ether, 3.0 mmol)
was added using a syringe. The solution immediately turned
a cloudy red-orange. After the mixture was stirred for 8 h, the
solution was filtered, and the volume of the solvent was
reduced to ca. 7 mL. Cooling to -20 °C produced red-orange
blocks (1.35 g, 70%). Mp: 128-132 °C. IR: 3091 (w), 2740 (w),
2720 (w), 1293 (w), 1248 (s), 1199 (m), 1166 (m), 1114 (m),
1056 (m), 1023 (m), 935 (m), 927 (m), 832 (s), 811 (w), 775 (s),
659 (w) cm-1. MS peaks at m/z 607, 627, and 662 are seen,
corresponding to Cpq2UMe, Cpq2UCl, and Cpq2UCl2, respec-

Uranium Metallocenes
Organometallics, Vol. 18, No. 7, 1999 1243
Ta ble 4. Cr ysta l Da ta
Cp′′2UCl2
Cpq2UCl2
Cp′′2UMe2
Cpq2UF2
[Cp′′2UF2]2
space group
a, Å
b, Å
c, Å
I2/a
C2/c
P212121
10.806(2)
16.185(8)
18.034(6)
90
90
90
3154(2)
4
P2/n
P21/n
22.28(1)
7.069(4)
20.558(8)
90
102.03(3)
90
25.4903(6)
6.8922(2)
18.7124(3)
90
124.077(1)
90
14.155(4)
6.302(1)
14.250(5)
90
92.46(3)
90
11.519(3)
21.892(7)
12.846(3)
90
116.07(2)
90
R, deg
â, deg
γ, deg
V, Å3
3166.87
4
2723.0(1)
4
1270
2
2910
2
Z
fw
680.98
1.428
50.154
5822
2818
0.030
2179
663.55
1.618
61.69
5448
2341
0.129
1880
687.02
1.448
50.438
2366
630.65
1.65
60.83
1930
1389.90
1.586
54.708
5868
d(calcd), g/cm3
µ(calcd), cm-1
no. of rflns collected
no. of unique rflns
Rint
2343
1684
4007
no. of rflns, Fo2 > 3σ(Fo
R,%
)
2075
4.2
1415
1.8
3109
3.4
2
2.3
10.4
Rw,%
2.8
6.8
5.4
2.2
3.4
GOF
0.83
6.85
1.39
1.18
1.97
tively. Anal. Calcd for C27H45ClU: C, 50.4; H, 7.05. Found: C,
50.0; H, 6.84.
(Fo). Azimuthal scans showed fairly flat absorption curves. No
absorption correction was applied. The systematic absences
indicated that the space group was I2/a or I2/c. I2/a was
chosen.
Cp ′′2U(Me)Cl. Cp′′2UCl215 (2.00 g, 2.75 mmol) was dissolved
in 50 mL of hexane, and MeLi (2.75 mL, 1.0 M in ether, 2.8
mmol) was added by syringe. The solution immediately turned
red-orange and cloudy. After the mixture was stirred for 12
h, it was filtered, and the volume of the filtrate was reduced
to ca. 7 mL. The solution was heated to 70 °C to redissolve
the solid. Cooling to -20 °C produced red-orange, flower-
shaped crystals (1.42 g, 72%). Mp: 122-124 °C. IR: 1403 (w),
1317 (m), 1247 (s), 1202 (m), 1112 (w), 1075 (s), 1053 (w), 912
(s), 840 (s), 793 (s), 751 (s), 689 (m), 627 (m), 469 (m), 411 (w),
362 (m), 312 (m), 278 (m), 245 (w) cm-1. MS peaks at m/z 671,
691, and 726 are seen corresponding to Cp′′2UMe, Cp′′2UCl,
and Cp′′2UCl2, respectively.
The cell volume indicated that four molecules were present
in the unit cell. The uranium atom had to be on a special
position. Since the molecule could have C2 symmetry, the first
special position was chosen. The uranium atom position and
the positions of the chlorine and silicon atoms were obtained
by solving the Patterson map. Successive Fourier searches
yielded the rest of the heavy-atom positions. The heavy-atom
structure was refined by standard least-squares and Fourier
techniques. The heavy atoms were refined anisotropically. The
hydrogen atoms were located in a difference Fourier, and the
hydrogen positions were then calculated on the basis of
idealized bonding geometry and assigned thermal parameters
equal to 5 Å2 for the cyclopentadienyl hydrogen atoms and 10
Å2 for the trimethylsilyl hydrogen atoms. After one least-
squares cycle, the hydrogen atoms were refined isotropically.
A final difference Fourier map showed no additional atoms in
the asymmetric unit. Examination of intermolecular close
contacts (<3.5 Å) showed that the molecule was a monomer.
Su scep tibility Mea su r em en ts. In an argon-filled drybox,
a ca. 100 mg sample was weighed into a cylindrical holder
consisting of two tightly fitting Kel-F cups, which were sealed
with silicone grease. The sample was quickly inserted into the
instrument, a SHE model 500 SQUID susceptometer. The
sample was centered, and data were obtained at 0.5 and 4.0
T from 5 to 300 K.
The data were treated as follows. First, the contribution
from the container was subtracted from the raw data using
data from susceptibility measurements of the container alone.
The magnetization was then converted to øgram and then ømol
using the mass of the sample and the molecular weight of the
complex. The diamagnetic contribution of the ligands was
calculated and was added to ømol to yield øcorr. Effective
magnetic moments were calculated by linear regression from
plots of 1/øcorr versus T in Kelvin.
Cr ysta llogr a p h y. Cp ′′2UCl2. Orange needles of the com-
pound were grown by cooling a saturated hexane solution to
-30 °C. The crystals were dried under reduced pressure and
loaded into quartz capillaries in an inert-atmosphere box. A
crystal measuring 0.08 × 0.20 × 0.60 mm was transferred to
an automated Picker FACS-1 diffractometer. The crystal was
centered in the beam. Automatic peak search and indexing
The final residuals for 216 variables refined against the
2179 unique data with Fo > 3σ(Fo) were R ) 2.3%, Rw ) 2.8%,
and GOF ) 0.83. The R value for all data (including unob-
served reflections) was 4.2%. The quantity minimized by the
least-squares refinements was w(|Fo| - |Fc|)2, where w is the
weight given to a particular reflection. The p factor, used to
reduce the weight of intense reflections, was set to 0.05.27 The
analytical forms of the scattering factor tables for neutral
atoms were used, and all non-hydrogen scattering factors were
corrected for both the real and imaginary components of
anomalous dispersion.28 The largest positive and negative
(26) The data reduction formulas are
ω
Lp
2
Fo
)
(C - 2B)
ω
Lp
σo(Fo2) )
(C + 4B)1/2
procedures indicated that the crystal possessed
a body-
centered monoclinic cell and yielded the cell parameters. The
cell parameters and data collection parameters are given in
Table 4.
The 5882 raw intensity data were converted to structure
factor amplitudes and their esds by correction for scan speed,
background, and Lorentz-polarization effects.26 Inspection of
the intensity standards showed no great change in intensity.
An intensity correction varying from 1.073 to 0.982 was applied
to the data. The systematic absences (h,0,l), h odd, and (h,0,l),
2
1/2
Fo ) (Fo
)
2
ωo(F) ) [Fo + σo(Fo2)]1/2 - Fo
where C is the total count of the scan, B is the sum of the two
background counts, ω is the scan speed used in deg/min, and
(sin 2θ)(1 + cos2 2θm)
1 + cos2 2θm - sin2 2θ
1
Lp
)
is the correction for Lorentz and polarization effects for a reflection
with scattering angle 2θ and radiation monochromatized with a 50%
perfect single-crystal monochromator with scattering angle 2θm.
l odd, were then rejected, and the data were averaged (Rint
0.030), yielding 2818 unique data, of which 2179 had Fo > 3σ-
)

1244 Organometallics, Vol. 18, No. 7, 1999
Lukens et al.
peaks in the final difference Fourier map have electron
forms of the scattering factor tables for neutral atoms were
used, and all non-hydrogen scattering factors were corrected
for both the real and imaginary components of anomalous
dispersion.28
densities of 0.51 and -0.49 e/Å2, respectively.
Cp q2UCl2. Red crystals of the compound were grown by
cooling a hexane solution to -20 °C. The crystals were then
placed in a Petri dish along with Paratone N in a glovebox. A
plate-shaped crystal measuring 0.12 × 0.26 × 0.23 mm was
mounted on the end of a 0.2 mm thin-walled quartz capillary.
The crystal was transferred to an Siemens SMART diffracto-
meter and cooled to -95 °C under a cold stream previously
calibrated by a thermocouple placed in the sample position.
The crystal was centered in the beam. Automatic peak search
and indexing procedures indicated that the crystal possessed
a C-centered monoclinic cell and yielded the unit cell param-
eters. The cell parameters and data collection parameters are
given in Table 4. On the basis of a statistical analysis of
intensity distribution and the successful solution and refine-
ment of the crystal structure, the space group was found to
be C2/c.
An arbitrary hemisphere of data was collected using the
default parameters for the diffractometer. The data were
collected as 30 s images with an area detector. Two images
were averaged to give the net image data. The image data were
integrated to give intensity data using the program SAINT.
The 5448 raw intensity data were converted to structure factor
amplitudes and their esds by correction for scan speed,
background, and Lorentz-polarization effects.26 Inspection of
the intensity standards showed no decrease in intensity over
the duration of data collection. An empirical absorption
correction using an ellipsoidal model for the crystal was
applied to the intensity data on the basis of the intensities of
all intense equivalent reflections (Tmax ) 0.979, Tmin ) 0.227).
Averaging equivalent reflections gave 2341 unique data (Rint
) 0.129).
The uranium atom positions were obtained by direct meth-
ods. Refinement on the uranium positions followed by a
difference Fourier search yielded the other heavy-atom posi-
tions. The heavy-atom structure was refined by standard least-
squares and Fourier techniques. The hydrogen positions were
calculated on the basis of idealized bonding geometry and
assigned thermal parameters equal to 1.15 Å2 larger than the
carbon atom to which they were connected. The hydrogen
positions were included in the structure factor calculations but
not refined by least squares. A final difference Fourier map
showed no additional atoms in the asymmetric unit. No close
(<3.5 Å) intermolecular contacts were found.
The final residuals for 132 variables refined against the
1880 unique data with I > 3σ(I) were R ) 10.4%, Rw ) 6.8%,
and GOF ) 6.85. The quantity minimized by the least-squares
refinements was w(|Fo| - |Fc|)2, where w is the weight given
to a particular reflection. The p factor, used to reduce the
weight of intense reflections, was set to 0.03.27 The analytical
Inspection of the residuals ordered in the ranges of (sin θ)/
λ, |Fo|, and parity and values of the individual indexes showed
no trends other than the previously mentioned secondary
extinction. No reflections had anomalously high values of w∆2.
The largest positive and negative peaks in the final difference
Fourier map have electron densities of 6.42 and -6.01 e/Å2,
respectively.
Cp ′′2UMe2. Orange crystals of the compound were grown
by cooling a saturated ether solution to -30 °C. The crystals
were then placed in a small Petri dish and covered with
Paratone N, a high-molecular-weight hydrocarbon oil. A block-
shaped single crystal measuring 0.45 × 0.40 × 0.40 mm was
mounted on the end of a 0.4 mm diameter quartz capillary.
The crystal was transferred to an Enraf-Nonius CAD-4 dif-
fractometer and cooled to -98 °C under a cold stream of
nitrogen gas previously calibrated by a thermocouple placed
in the sample position. The crystal was centered in the beam.
Automatic peak search and indexing procedures indicated that
the crystal possessed a primitive orthorhombic cell and yielded
the cell parameters. The cell parameters and data collection
parameters are given in Table 4.
The 2366 raw intensity data were converted to structure
factor amplitudes and their esds by correction for scan speed,
background, and Lorentz-polarization effects.26 Inspection of
the intensity standards showed no loss of intensity during the
data collection. The 23 systematic absences (h,0,0), h odd,
(0,k,0), k odd, and (0,0,l), l odd, were then rejected, yielding
2343 unique data, of which 2075 possessed Fo > 3σ(Fo).
Azimuthal scan data showed a difference of Imin/Imax ) 0.69
for the average curve. An empirical absorption correction was
applied on the basis of the average curve. The systematic
absences indicated that the space group was P212121.
The cell volume indicated that four molecules were present
in the unit cell. The uranium atom position was obtained by
solving the Patterson map. Successive Fourier searches yielded
the rest of the heavy-atom positions. The heavy-atom structure
was refined by standard least-squares and Fourier techniques.
The heavy atoms were refined anisotropically. Hydrogen atoms
were placed at calculated positions with thermal parameters
1.3 times the thermal parameters of the carbon atoms to which
they were bound. At the end of the refinement, the enantiomer
was changed and the structure refined. The refinement was
very slightly worse, so the enantiomer was changed back to
the original one, and the structure was re-refined. A final
difference Fourier map showed no additional atoms in the
asymmetric unit. Examination of intermolecular close contacts
(<3.5 Å) showed that the molecule was a monomer.
The final residuals for 263 variables refined against the
(27)
||Fo| - |Fc||
∑
2075 unique data with Fo > 3σ(Fo) were R ) 4.22%, Rw
)
R )
5.38%, and GOF ) 1.39. The R value for all data (including
unobserved reflections) was 5.01%. The quantity minimized
by the least-squares refinements was w(|Fo| - |Fc|)2, where w
is the weight given to a particular reflection. The p factor, used
to reduce the weight of intense reflections, was set to 0.03
initially, but later changed to 0.06.27 The analytical forms of
the scattering factor tables for neutral atoms were used, and
all non-hydrogen scattering factors were corrected for both the
real and imaginary components of anomalous dispersion.28
|Fo|
∑
1/2
(|Fo| - |Fc|)2
∑
Rw
)
2
[
]
wFo
∑
1/2
(|Fo| - |Fc|)2
(no - nv)
∑
GOF )
[
]
Inspection of the residuals ordered in the ranges of (sin θ)/
λ, |Fo|, and parity and values of the individual indexes showed
that secondary extinction was occurring. A secondary extinc-
tion coefficient was included in the refinement; its final value
was 1.2 × 10-8. One reflection had anomalously high values
where no is the number of observations and nv is the number of variable
parameters, and the weights were given by
1
w )
σ2(Fo)
σ(Fo2) ) [σo2(Fo2) + (pF2)2]1/2
where σ2(Fo) is calculated as above from σ(Fo2) and where p is the factor
used to lower the weight of intense reflections.
(28) Cromer, D. T.; Waber, J . T. In International Tables for X-ray
Crystallography; Kynoch Press: Birmingham, England, 1974; Vol. IV.

Uranium Metallocenes
Organometallics, Vol. 18, No. 7, 1999 1245
of w∆2 and was weighted to zero toward the end of the
refinement. The largest positive and negative peaks in the final
difference Fourier map have electron densities of 2.44 and
-0.29 e/Å2, respectively, and are associated with the uranium
atom.
cooling slowly to room temperature. The supernatant was
removed using a cannula, and degassed Paratone N, a high-
molecular-weight, aliphatic hydrocarbon oil was poured into
the Schlenk tube. A suitable, roughly cubic crystal measuring
0.28 × 0.30 × 0.30 mm was mounted on the end of a 0.2 mm
thin-walled quartz capillary. The crystal was transferred to
an Enraf-Nonius CAD-4 diffractometer and cooled to -111 °C
under a cold stream previously calibrated by a thermocouple
placed in the sample position. The crystal was centered in the
beam. Automatic peak search and indexing procedures indi-
cated that the crystal possessed a primitive monoclinic cell and
yielded the unit cell parameters. The cell parameters and data
collection parameters are given in Table 4.
Cp q2UF 2. Light orange crystals of the compound were grown
by cooling a hexane solution to -20 °C. The supernatant was
removed using a cannula, and degassed Paratone N, a high-
molecular-weight, aliphatic hydrocarbon oil was poured into
the Schlenk tube. A suitable square crystal measuring 0.28
mm × 0.28 mm × 0.08 mm was mounted on the end of a 0.2
mm thin-walled glass capillary. The crystal was transferred
to an Enraf-Nonius CAD-4 diffractometer and cooled to -88
°C under a cold stream previously calibrated by a thermocouple
placed in the sample position. The crystal was centered in the
beam. Automatic peak search and indexing procedures indi-
cated that the crystal possessed a primitive monoclinic cell and
yielded the unit cell parameters. The cell parameters and data
collection parameters are given in Table 4.
The data were collected in two blocks (+h,(k,+l) and
(+h,+k,-l). The 5868 raw intensity data were converted to
structure factor amplitudes and their esds by correction for
scan speed, background, and Lorentz-polarization effects.26
Inspection of the intensity standards showed no decrease in
intensity over the duration of data collection for the first block
of data and a 5% decrease in intensity for the second block of
data. The second block was then corrected for a decay of 5%.
The 157 systematic absences (0,k,0), k odd, and (h,0,l), h + l
odd, and the 1499 redundant data (+h,-k,+l) and the 195
duplicated (h,k,0) data were then rejected, yielding 4007
unique data. No empirical absorption correction could be
applied because a large piece of ice had grown on the glass
capillary, and when the goniometer head was raised to 90° in
ø, the capillary snapped off. Inspection of the collected data
showed systematic absences for (0,k,0), k odd, and (h,0,l), h +
l odd, indicating that the space group was P21/n.
Inspection of the raw data indicated that the space group
contained an n glide plane, so the centric space group P2/n
was chosen. The 1930 raw intensity data were converted to
structure factor amplitudes and their esds by correction for
scan speed, background, and Lorentz-polarization effects.26
Inspection of the intensity standards showed no decrease in
intensity over the duration of data collection. Inspection of the
azimuthal scan data showed a variation of Imin/Imax ) 0.52 for
the averaged curve. An empirical absorption correction was
applied to the intensity data on the basis of the average curve.
Removal of the 181 systematic absences (h,0,l), h + l odd, and
the 68 redundant data (+h,+k,-l) left 1684 unique data.
The molecule was initially believed to be a monomer with
Z ) 4. The uranium atom positions were obtained by solving
the Patterson map. Refinement on the uranium positions
followed by a difference Fourier search yielded the other heavy-
atom positions. The heavy-atom structure was refined by
standard least-squares and Fourier techniques. The heavy
atoms were refined isotropically, and the hydrogen atom
positions were calculated. A numerical absorption correction
(DIFABS) was applied, and the heavy-atom positions were
refined anisotropically with the hydrogen atoms in calculated
positions. The hydrogen positions were calculated on the basis
of idealized bonding geometry and assigned thermal param-
eters equal to 1.3 Å2 larger than the carbon atom to which
they were connected. The hydrogen positions were included
in the structure factor calculations but not refined by least
squares. A final difference Fourier map showed no additional
atoms in the asymmetric unit. Examination of intermolecular
close contacts(<3.5 Å) showed that the molecule was actually
a dimer with two bridging fluoride ligands.
The uranium atom positions were obtained by solving the
Patterson map. The uranium atom was on a special position
(0.25, 0.13, 0.25) with 2-fold rotational symmetry. Refinement
on the uranium positions followed by a difference Fourier
search yielded the other heavy-atom positions. The heavy-atom
structure was refined by standard least-squares and Fourier
techniques. The heavy atoms were refined with isotropic and
then anisotropic thermal parameters. The hydrogen positions
were calculated on the basis of idealized bonding geometry and
assigned thermal parameters equal to 1.15 Å2 larger than the
carbon atom to which they were connected. The hydrogen
positions were included in the structure factor calculations but
not refined by least squares. Toward the end of the refinement,
examination of the extinction test listing indicated that
secondary extinction was occurring. The secondary extinction
coefficient was initially set to 1.7 × 10-7 and was refined to a
final value of 2.67 × 10-7. A final difference Fourier map
showed no additional atoms in the asymmetric unit. No close
(<3.5 Å) intermolecular contacts were found.
The final residuals for 133 variables refined against the
The final residuals for 263 variables refined against the
1415 unique data with Fo > 3σ(Fo) were R ) 1.78%, Rw
)
4007 unique data with Fo > 3σ(Fo) were R ) 3.35%, Rw
)
2.20%, and GOF ) 1.19. The R value for all data (including
unobserved reflections) was 2.02%. The quantity minimized
by the least-squares refinements was w(|Fo| - |Fc|)2, where w
is the weight given to a particular reflection. The p factor, used
to reduce the weight of intense reflections, was set to 0.03.27
The analytical forms of the scattering factor tables for neutral
atoms were used, and all non-hydrogen scattering factors were
corrected for both the real and imaginary components of
anomalous dispersion.28
3.58%, and GOF ) 1.97. The R value for all data (including
unobserved reflections) was 5.42%. The quantity minimized
by the least-squares refinements was w(|Fo| - |Fc|)2, where w
is the weight given to a particular reflection. The p factor, used
to reduce the weight of intense reflections, was set to 0.03
initially but was later changed to 0.02.27 The analytical forms
of the scattering factor tables for neutral atoms were used,
and all non-hydrogen scattering factors were corrected for both
the real and imaginary components of anomalous dispersion.28
Inspection of the residuals ordered in the ranges of (sin θ)/
λ, |Fo|, and parity and values of the individual indexes showed
no trends other than the previously mentioned secondary
extinction. No reflections had anomalously high values of w∆2.
The largest positive and negative peaks in the final difference
Fourier map have electron densities of 1.02 and -0.54 e/Å2,
respectively, and are associated with the uranium atom.
Inspection of the residuals ordered in the ranges of (sin θ)/
λ, |Fo|, and parity and values of the individual indexes showed
no trends. Fourteen reflections had anomalously high values
of w∆2 and were weighted to zero toward the end of the
refinement. The largest positive and negative peaks in the final
difference Fourier map have electron densities of 0.93 and
-0.30 e/Å2, respectively, and are associated with the uranium
atom.
[Cp ′′2UF 2]2. Dark green crystals of the compound were
grown by heating a hexane solution to about 80 °C and then

1246 Organometallics, Vol. 18, No. 7, 1999
Lukens et al.
Su ppor tin g In for m ation Available: Tables giving atomic
positions and anisotropic thermal parameters for Cp′′2UCl2,
Cpq2UCl2, Cp′′2UMe2, Cpq2UF2, and [Cp′′2UF2]2. This material
is available free of charge via the Internet at http://pubs.ac-
s.org. Tables of structure factors are available from the authors
upon request.
Ack n ow led gm en t. We thank F. J . Hollander and
A. Zalkin for help with crystallography. W.W.L. thanks
the National Science Foundation for a graduate fellow-
ship. This work was partially supported by the Director,
Office of Energy Research, Office of Basic Energy
Sciences, Chemical Sciences Division, of the U.S. De-
partment of Energy under Contract No. DE-AC03-
76SF00098.
OM9805988