Synthesis and properties of di-π-[8]annulenecerium(IV), cerocene

Article abstract of DOI:10.1021/om049743+

An efficient synthesis of di-π-([8]annulene)cerium(IV) (cerocene) is reported by the reaction of K[Ce(C₈H₈)₂] with allyl bromide. Cerocenes can also be prepared by oxidation of the corrresponding cerates with silver iodide in THF, but this reaction is more variable. Cerocene reacts only slowly with UCI₄ to give uranocene. K[Ce(C ₈H₈)₂] and cerocene undergo electron exchange on the NMR time scale. Electrochemical study of cerocene shows redox with the cerate ion, Ce(C₈H₈)₂^-, to be reversible with a relatively low reduction potential (-0.6 V vs NHE). The corresponding praeseodymium salt, K[Pr(C₈H₈)₂], does not undergo comparable reversible oxidation; its cyclic voltammetry is compared with that of K₂[Yb(C₈H₈)₂]. The chemistry of cerocene resembles that of di-π-([8]annulene)actinide(IV) compounds.

Full text of DOI:10.1021/om049743+

Organometallics 2004, 23, 5169-5175
5169
Syn th esis a n d P r op er ties of Di-π-[8]a n n u len ecer iu m (IV),
Cer ocen e
Andrew Streitwieser,* Steven A. Kinsley, Chris Henry J enson, and
J ohn T. Rigsbee
Department of Chemistry, University of California, Berkeley, California 94720-1460
Received April 8, 2004
An efficient synthesis of di-π-([8]annulene)cerium(IV) (cerocene) is reported by the reaction
of K[Ce(C
corrresponding cerates with silver iodide in THF, but this reaction is more variable. Cerocene
reacts only slowly with UCl to give uranocene. K[Ce(C ] and cerocene undergo electron
exchange on the NMR time scale. Electrochemical study of cerocene shows redox with the
8 8 2
H ) ] with allyl bromide. Cerocenes can also be prepared by oxidation of the
4
8 8 2
H )
-
cerate ion, Ce(C
NHE). The corresponding praeseodymium salt, K[Pr(C
reversible oxidation; its cyclic voltammetry is compared with that of K
chemistry of cerocene resembles that of di-π-([8]annulene)actinide(IV) compounds.
8
H
8
)
2
, to be reversible with a relatively low reduction potential (-0.6 V vs
], does not undergo comparable
[Yb(C ]. The
8
8 2
H )
2
8 8 2
H )
In tr od u ction
lational studies suggest the possibility of oxidation of
other lanthanide(III) COT salts, K[Ln(C8H8)2], to their
neutral counterparts, Ln(C8H8)2.
Di-π-[8]annulenecerium(IV), cerocene, Ce(C8H8)2, is
an especially interesting metallocene that is formally a
lanthanide(IV) analogue of uranocene, U(C8H8)2, but
combines a strong oxidizing agent, Ce(IV), and a strong
reducing agent, cyclooctatetraene (COT) dianion, in the
same molecule. The first successful synthesis of cerocene
resulted surprisingly from the use of a reducing agent,
triethylaluminum, in a reaction using cerium(IV) iso-
propoxide with neutral cyclooctatetraene as a reactant
9
Although we confirmed the reported preparation of
cerocene, its use of COT as a solvent does not permit a
practical extension to substituted derivatives. In a 1985
communication we described a better approach based
on the mild oxidation of potassium di-π-[8]annulenecer-
1
3
ate(III) with solid silver iodide in THF. In the present
paper we describe more details of this procedure, a
4
method that has also now been used by others. We also
2
3
and solvent. Subsequent work from this group and
others has provided further characterization of cerocene
discuss problems with this approach and describe a new
and simpler method. We have also further studied the
electrochemistry of cerocene and report a new value for
the Ce(IV)/Ce(III) couple based on more recent informa-
4
and various ring-substituted analogues. Crystal struc-
5
6
tures of Ce(C8H7Me)2 and Ce[C8H5(SiMe3)3]2 show that
cerocenes have a uranocene-type sandwich structure.
10
tion on electrochemical standards in tetrahydrofuran.
Cerocene has received further recent attention from
theoretical studies, which suggest that it is better
To study the possibility of the existence of additional
neutral compounds of the type Ln(C H ) , we have also
8
8 2
described as a neutral compound of Ce(III) involving two
explored the electrochemistry and oxidation of the
praeseodynium compound K[Pr(C H ) ].
7
[
8]annulene rings with a total charge of -3. X-ray
8
8 2
absorption near-edge structure measurements (XANES)
are consistent with these calculations. Further calcu-
8
Resu lts a n d Discu ssion
*
To whom correspondence should be addressed. E-mail: astreit@
socrates.berkeley.edu.
1) Streitwieser, A.; Muller-Westerhoff, U. J . Am. Chem. Soc. 1968,
0, 7364.
Oxid a tion of Cer a tes. The COT derivatives of
trivalent lanthanides have long been known and are
readily prepared from the reaction of COT dianions with
(
9
1
1
1
(2) Greco, A.; Cesca, S.; Bertolini, G. J . Organomet. Chem. 1976,
the lanthanide halides. The Ce(III) compounds are
13, 321-330. It is surprising that this method is successful; the
analogous reaction was used by Wilke to make Ti(C H ) , which does
8 8 2
12
now conveniently accessible by reaction of CeI3(THF)4
with COT dianions. Thus, we considered that a conve-
nient preparation of cerocene would be based on oxida-
tion of potassium di-π-([8]annulene)cerate(III). A num-
ber of oxidizing agents were tried including O2, NOBF4,
Br2, (NH4)2Ce(NO3)6, and K2S2O8. The first two pro-
duced cerocene in yields of about 5%; the rest gave
not have both COT rings planar. See: Breil, H.; Wilke, G. Angew.
Chem., Int. Ed. Engl. 1966, 5, 898-899.
(
3) Streitwieser, A.; Kinsley, S. A.; Rigsbee, J . T.; Fragal a` , I. L.;
Ciliberto, E.; R o¨ sch, N. J . Am. Chem. Soc. 1985, 107, 7786-7788.
4) For recent review of Ce(IV) [8]annulene complexes see:
Edelmann, F. T.; Freckmann, D. M. M.; Schumann, H. Chem. Rev.
002, 102, 1851-1896.
5) Boussie, T. R.; Eisenberg, D. C.; Rigsbee, J .; Streitwieser, A.;
Zalkin, A. Organometallics 1991, 10, 1922-1928.
6) Kilimann, U.; Herbst-Irmer, R.; Stalke, D.; Edelmann, F. T.
Angew. Chem., Int. Ed. Engl. 1994, 33, 1618-1621.
7) (a) Dolg, M.; Fulde, P.; K u¨ chle, W.; Neumann, C. S.; Stoll, H. J .
(
a
2
(
3
+
overoxidation to COT and Ce . A suspension of silver
iodide in THF, however, gave the desired result. The
(
(
Chem. Phys. 1991, 94, 3011-3017. (b) Dolg, M.; Fulde, P.; Stoll, H.;
(9) Liu, W.; Dolg, M.; Fulde, P. Inorg. Chem. 1998, 37, 1067-1072.
(10) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877-910.
(11) Hodgson, K.; Mares, F.; Starks, D.; Streitwieser, A. J . Am.
Chem. Soc. 1973, 95, 8650-8658.
Preuss, H.; Pitzer, R. M. Chang, A. Chem. Phys. 1995, 195, 71-82.
(8) Edelstein, N. M.; Allen, P. G.; Bucher, J . J .; Shuh, D. K.; Sofield,
C. D. J . Am. Chem. Soc. 1996, 118, 13115-13116.
1
0.1021/om049743+ CCC: $27.50 © 2004 American Chemical Society
Publication on Web 09/18/2004

5
170 Organometallics, Vol. 23, No. 22, 2004
Streitwieser et al.
Ta ble 1. Visible Sp ectr a of (RC
8
H
7
)
2
Ce An ion a n d
In an NMR tube, allyl bromide showed reaction with
cerates within 2 min and no decomposition reactions
even after several days. A typical and general procedure
consists of dissolving completely a sample of K[Ce-
(C8H8)2] or other substituted COT derivative in THF,
then adding a 3-fold excess of allyl bromide and allowing
the solution to stir for 12 h. The product is recovered
by removing solvent, organic reaction byproducts, and
unreacted allyl bromide under vacuum and extracting
the remaining material with the appropriate hydrocar-
bon solvent. The method was applied to the synthesis
of cerocene, 1,1′-di-tert-butylcerocene, and 1,1′,4,4′-tet-
rakis(trimethylsilyl)cerocene. The last compound was
not obtained analytically pure but was characterized by
MS and NMR and appears to be identical to the product
Neu tr a l in THF
+
-
R
K (RC8H7)2Ce
(RC8H7)2Ce
472 (570)b
477 (570)b
692a
H
methyl
n-butyl
tert-butyl
tert-butoxy
706
708
707
724
b
478 (∼585)
483 (567)b
492
a
Lit.13 691 nm. b Shoulder; λmax by Gaussian deconvolution.
use of a 10-fold excess of suspended silver iodide gave
up to 78% yields of cerocene. This procedure was also
used to prepare several substituted cerocenes: 1,1′-
dimethyl (whose crystal structure has been deter-
5
mined), 1,1′-di-n-butyl, 1,1′-di-tert-butyl, and 1,1′-di-
tert-butoxy. These compounds were characterized primar-
ily by spectroscopic measurements (Table 1). Cerocene
itself has a pronounced λmax at 469 nm (ꢀ ) 8000) with
a shoulder (Figure S1, Supporting Information). This
spectrum was deconvoluted using two Gaussian func-
tions, also shown in Figure S1 (Supporting Information),
to give a least-squares fit.
6
prepared using AgI by Kilimann et al.
Som e P r op er ties a n d Rea ction s. Although Ln(C8-
-
2-
H8)2 salts behave as ionic salts of COT , cerocene does
not. For example, K[Ce(C8H8)2] reacts rapidly with UCl4
11
to give uranocene in high yield, whereas a THF
solution of cerocene with UCl4 in an NMR tube shows
a small peak for uranocene at -36 ppm only after
several days. In this respect, cerocene behaves much like
a thorocene, which also reacts slowly with UCl4 to give
The method has also been applied by others; for
3
+
example, the Ce salts of 1,4-bis(trimethylsilyl)cyclooc-
tatetraene and 1,3,6-tris(trimethylsilyl)cyclooctatraene
dianions were oxidized with AgI to the corresponding
cerocenes. The latter cerocene, Ce(1,3,6-TMS3-C8H5)2,
16
the corresponding uranocene. Similarly, cerocene does
not react with trimethylchlorosilane.
Electron exchange has been shown to occur on the
NMR time scale (second-order rate constants on the
6
was characterized by its X-ray crystal structure.
Nevertheless, we have found the AgI oxidation to be
inconsistent. Yields were frequently variable, and some
samples gave overoxidation. We therefore considered
approaches involving one-electron oxidation by alkyl
halides based on the work of Kagan et al.1 in the
oxidation of Sm(II) to Sm(III). We also note that we have
previously used benzyl bromide to oxidize the Yb(II) salt
of a COT to the Yb(III) salt.15 Benzyl chloride did not
react with K[Ce(C8H8)2] over several days. Benzyl
bromide did give cerocene plus some decomposition
products. 1,2-Diphenylethane (bibenzyl) was identified
in the reaction mixture by NMR. This is the expected
byproduct of a one-electron transfer process between the
cerium(III) salt and benzyl bromide to form benzyl
radicals that couple in solution. The solubility of both
cerocene and bibenzyl in hydrocarbon solvents makes
the separation and purification of cerocene difficult. It
was desirable, therefore, to find a reagent that would
provide byproducts that would be more easily removed.
Ethyl iodide produced low yields of cerocene with
considerable decomposition. Allyl bromide gave by far
the best results, producing cerocene in 80% yields after
isolation, with no apparent decomposition products.
These are the highest consistent yields seen using any
of the previous oxidizing agents. The byproduct formed
was determined by NMR to be 1,5-hexadiene. Again,
this is undoubtedly formed via the free radical process
discussed earlier. This byproduct has a low boiling point
of 60 °C and is easily removed along with any unreacted
allyl bromide under vacuum at the end of the reaction.
7
-1 -1
order of 10 M
s
) between potassium bis(tert-butyl-
[
[
8]annulene)ytterbate(II) and potassium bis(tert-butyl-
8]annulene)ytterbate(III). Similar rapid electron ex-
1
5
change on the NMR time scale has been shown between
the III and IV oxidation states of actinocenes of ura-
nium, neptunium, and plutonium. The corresponding
4
1
7
1
behavior has now been found for cerium. The H NMR
spectrum of a nearly saturated solution of the diamag-
netic bis(tert-butyl[8]annulene)cerium(IV) in THF-d8 is
shown in Figure 1A. It shows a clean singlet at δ 2.2
ppm for the tert-butyl protons. The ring protons show
up as a muliplet aound δ 6 ppm but will not be relevant
in this study. Other peaks are due to solvent (δ 1.73
and 3.58 ppm) and residual stopcock grease (δ 0.05
1
ppm). The H NMR spectrum of the paramagnetic
potassium bis(tert-butyl[8]annulene)cerate(III) at 20 °C
is shown in Figure 1B. The ring protons are now too
broad to be seen and the tert-butyl protons occur as a
broad resonace at δ 4.4 ppm. The position of this peak
is temperature dependent, as is usual for such para-
magnetic species. An approximately 50:50 mixture of
these two compounds is shown in Figure 1C. Resonances
corresponding to the solvent and grease are unchanged,
but the metallocene resonances are gone. Close exami-
nation shows a small broad peak in the baseline
overlapping with the downfield solvent peak. If this is
the exchanging peak, it should be affected by changes
in temperature.
The NMR spectrum at several temperatures is shown
in Figure 2. As the temperature is raised, the broad
peak sharpens with the high-temperature fast exchange
limit about 65 °C. At this temperature the shift of the
(
12) Hazin, P. N.; Huffman, J . C.; Bruno, J . W. Organometallics
987, 6, 23-27.
13) Mares, F.; Hodgson, K. O.; Streitwieser, A., J r. J . Organomet.
Chem. 1971, 28, C24-C26.
14) Kagan, H. B.; Namy, J . L.; Girard, P. Tetrahedron Supp. 1981,
, 175.
15) Eisenberg, D. C.; Kinsley, S. A.; Streitwieser, A. J . Am. Chem.
Soc. 1989, 111, 5769-5772.
1
(
(
(16) LeVanda, C.; Streitwieser, A., J r. Inorg. Chem. 1981, 20, 656-
9
659.
(
(17) Eisenberg, D. C.; Streitwieser, A.; Kot, W. K. Inorg. Chem. 1990,
29, 10-14.

Di-π-[8]annulenecerium(IV), Cerocene
Organometallics, Vol. 23, No. 22, 2004 5171
1
1
F igu r e 1. (A) H NMR of Ce(tert-butylC
°
8
H
7
)
2
in THF-d
8
at 20 °C. (B) H NMR of K[Ce(tert-butylC
8 7 2 8
H ) ] in THF-d at 20
1
C. (C) H NMR of a 50:50 mixture of Ce(tert-butylC
8
H
7
)
2
and K[Ce(tert-butylC ] in THF-d at 20 °C.
8
H
7
)
2
8
exchange peak is about 3 ppm, about that expected for
the average of the two species. The coalescence temper-
ature is about room temperature, and the difference in
chemical shifts is too small to freeze out the exchange.
The phenomenon is clearly an electron exchange process
analogous to that observed with other actinide and
lanthanide metallocenes. In particular, the similarity
of temperature ranges indicates that the cerocene-

5
172 Organometallics, Vol. 23, No. 22, 2004
Streitwieser et al.
1
F igu r e 2. Effect of temperature on the H NMR of a 50:50 mixture of Ce(tert-butylC
8
H
7
)
2
and K[Ce(tert-butylC
8
H
7
)
2
] in
8
THF-d .
cerate exchange, like that of the actinocenes, has rate
constants of comparable orders of magnitude to the
Yb(II)-Yb(III) exchange cited above.
Electr och em istr y of K[Ce(C8H8)2], K2[Yb(C8H8)2],
a n d K[P r (C8H8)2]. The (bis-[8]annulene)cerate anion
exhibits a fully reversible oxidation-reduction wave at
-
1.4 V relative to either internal or external ferrocene.
+
1/0
Using a value of the ferrocene
couple of 0.56 V vs
SCE (in THF with [Bu4N][PF6]),10 the cerocene
couple is -0.8 V vs SCE or -0.6 V vs NHE. This value
is less negative than originally reported from this
0/-1
3
group, but the major difference is the more accurate
+
1/0
18
value for the ferrocene
couple in the exact media.
F igu r e 3. Cyclic voltammogram of K[Ce(C
in THF. The electrochemically interrelated sandwich com-
plexes are shown as Ce(C H ) and [Ce(C H ) ] . The scan
8 8 2 8 8 2
8
H
8
)
2
], ∼1 mM
The less negative value of the cerocene0/-1 couple is more
consistent with the -0.4 V vs NHE value for the cerium-
-
1
9
(
IV/III) catecholate couple found by Sofen et al.
rate is 0.100 V/s and referenced to external ferrocene.
Although not as negative, a value of -0.6 V vs NHE is
still abnormally low for a Ce(IV/III) couple. Figure 3
reversible, the peak-to-peak separation of the cerocene
and ferrocene is identical, indicating a one-electron
process, and the same value of Eo is obtained from
measurement of the difference between the oxidation
waves or the reduction waves.
shows the one-electron oxidation-reduction wave for
cerocene0
/-1
.
One of the difficulties in obtaining the value of E° for
the cerocene0
/-1
couple is due to an apparent reaction
of one of the cerocene species with ferrocene or ferro-
cenium ion. Shown in Figure S3 (Supporting Informa-
tion) is the cyclic voltammogram of cerocene with an
internal standard of ferrocene. Although not completely
The ability to reversibly remove an electron from the
bis([8]annulene)cerate anion does not prove the exist-
ence of Ce(IV) in the neutral sandwich compound. If the
electron comes from the central metal, then the oxida-
0
4
tion state of cerium would be formally Ce(IV) (4f π ); if
the electron comes from the π-system of the rings, then
(
18) At the time of the original study, a value of 0.4 V vs NHE was
used for ferrocene. See ref 10.
19) Sofen, S. R.; Cooper, S. R.; Raymond, K. N. Inorg. Chem. 1979,
8, 1611-1616.
1
3
the oxidation state of cerium would be Ce(III) (4f π ).
(
1
But an analogous situation is that of the oxidation of

Di-π-[8]annulenecerium(IV), Cerocene
Organometallics, Vol. 23, No. 22, 2004 5173
F igu r e 4. Cyclic voltammogram of K
2
[Yb(C
8
H
8
)
2
], ∼1 mM in THF. The scan rate is 0.100 V/s.
F igu r e 5. Cyclic voltammogram of K[Pr(C
8
H
8
)
2
], ∼1 mM in THF. The scan rate is 0.100 V/s.
K2Yb(C8H8)2 to KYb(C8H8)2. The kinetics of this process
was cited above,15 and there is little doubt that the
oxidation state of the central metal changes from Yb-
sandwich complex. It frequently disappeared on repeat
scans, which suggests an electrochemical event followed
by a chemical reaction or rearrangement into something
else having a different potential. A possible candidate
is the peroxide initially formed when oxygen reacts with
COT. Assuming that the first, lower oxidation (-0.22
V vs NHE) is the removal of one electron from
(
II) to Yb(III). The cyclic voltammogram is shown in
2
-
Figure 4. Wave A is the reversible redox of Yb(C8H8)2
Yb(C8H8)2 (at -0.3 V vs NHE); wave B is some further
irreversible oxidation. Waves C and D resemble closely
the successive one-electron oxidations of C8H8
C8H8 and to C8H8 observed for COT in THF by
/
-
2
-
-
to
[Pr(C H ) ] to give neutral Pr(C H ) , then Figures S4-
8
8 2
8
8 2
-
S7 (Supporting Information) show that this oxidation
is not reversible. When ferrocene is added, the solution
becomes unstable, and the waves for both the ferrocene
2
0
Lehmkuhl et al. The Yb compounds behave essentially
as ionic salts of COT dianion, and such oxidations to
COT are to be expected. Moreover, any inadvertent
oxidation during preparation of the sample will result
in COT; thus, detection of COT is not meaningful.
The electrochemisty of bis-[8]annulenepraeseodymate
ion differs significantly from its two analogues. The
voltammogram in Figure 5 shows two nonreversible
oxidations. The oxidation wave at E ) -0.8 V is
sometimes present in the oxidation of sandwich com-
plexes in THF; however, this wave is not present in all
voltammograms and probably is not that of a C8H8
-
and [Pr(C8H8)2] become irreversible and unstable. To
overcome this difficulty, cerocene was used as an
internal standard (Figure S8, Supporting Information).
In this system, both the reversible cerocene standard
-
and the irreversible [Pr(C8H8)2] are observed, the
-
oxidation of [Pr(C8H8)2] occurring 0.38 V higher than
-
that of [Ce(C8H8)2] . Waves C and D again show the
2-
-
normal oxidation of C8H8 to C8H8 and to C8H8.
The initial oxidation reaction that does occur is
suggested by the results of chemical oxidation. On
treatment of K[Pr(C8H8)2] with allyl bromide, most of
the K[Pr(C8H8)2] was recovered unchanged. A small
(
20) Lehmkuhl, H.; Kintoff, S.; J anssen, E. J . Organomet. Chem.
1
973, 556, 41-52.

5
174 Organometallics, Vol. 23, No. 22, 2004
Streitwieser et al.
yield was obtained, however, of the known triple-
sandwich compound Pr2(C8H8)3,21 which was identified
by the molecular ion in its mass spectrum. We suggest
-
that removal of one electron from the Pr(C8H8)2 gives
a praeseodymium complex that disproportionates to Pr2-
(C8H8)3 plus cyclooctatetraene. The chemical reaction
suggests that the initial voltammetric oxidation of
-
Pr(C8H8)2 is coupled to the disproportionation that
F igu r e 6. Schematic diagram of ring-metal bonding in
2
-
results only in the net oxidation of one C8H8 to COT.
The product Pr2(C8H8)3 is still a salt of C8H82 that can
undergo additional further oxidation to COT in waves
C and D.
metallocenes.
-
the metal, the result is transfer of electron density from
ligand to metal. If the metal was originally +4, some
measures of the effective oxidation state of the metal
would be expected to give values less than 4. In the case
of cerocene this transfer of electron density is apparently
substantial, but it is still chemically reasonable and
consistent to regard the central metal in this case as
formally +4. Accordingly, the simple model in Figure 6
is an effective model at the one-electron MO level that
rationalizes the observed chemistry.
It is interesting to note that the class of compounds
Ln2(C8H8)3 is the only product observed from the reac-
tion of cyclooctatetraene with lanthanide metal vapor,
even for cerium. The reductive elimination of neutral
cyclooctatetraene is observed upon the heating of any
lanthanide sandwich complex where the central metal
atom is in the higher of two available oxidation states.
The trivalent complex K[Yb(C8H8)2] is converted to the
divalent K2[Yb(C8H8)2] by reductive elimination of
neutral cyclooctatetraene when heated in vacuo at 200
Exp er im en ta l Section
2
2
°
C. We note further that the reaction of Ce2(C8H8)3
with cyclooctatetraene is apparently reversible. When
Ce(C8H8)2 is heated in vacuo, the brown Ce(C8H8)2
initially sublimes, but upon continued heating, the
green compound Ce2(C8H8)3 sublimes and neutral cy-
Gen er a l P r oced u r es. All air-sensitive materials were
handled under argon either in a Vacuum Atmospheres Model
HE-243 glovebox with a Model HE-93A MOD drytrain or using
standard Schlenk line techniques. Unless otherwise noted,
materials were obtained from commercial suppliers and used
2
clooctatetraene is observed. The reverse reaction is in
1
without further purification. H NMR spectra were obtained
2
the original report of cerocene; reaction of Ce2(C8H8)3
on UC Berkeley 200 and 250 MHz superconducting FT
spectrometers equipped with Cryomagnets Inc. magnets and
Nicolet Model 1280 data collection systems, a Bruker 500 MHz
AM 500 superconducting FT spectrometer with an Aspect 3000
computer system, and a 90 MHz J EOL FX90Q FT spectrom-
eter. Temperatures were measured with a thermocouple
mounted next to the NMR tube. Cyclooctatetraene (COT)
(obtained from BASF) was vacuum distilled or vacuum trans-
ferred prior to use and stored over 3 Å molecular sieves. Allyl
bromide was washed two times with a saturated aqueous
with neutral cyclooctatetraene in the presence of a
catalytic amount of triethylaluminum gave Ce(C8H8)2.
At room temperature, the equilibrium favors cerocene;
at higher temperatures, in vacuo, the equilibrium shifts
to the right.
On the basis of NMR studies of the electron exchange
1
5
between K2[Yb(C8H8)2] and K[Yb(C8H8)2], and between
K [U(C8H8)2] and U(C8H8)2,17 the first step of oxidation
is probably an outer sphere mechanism, removing an
electron from the ring. This is followed by a rapid
transfer of an electron from the central metal in the case
of ytterbium and cerium, but not for praeseodymium.
Whether viewed as a true tetravalent central metal
sandwich compound or as a trivalent central metal in a
3 4
NaHCO solution and then with distilled water. MgSO was
added and the mixture was fractionally distilled. The highest
boiling fraction was degassed using three freeze-pump-thaw
cycles, taken into the glovebox, and stored over 4 Å Linde
molecular seives. All other reagents, unless otherwise noted,
were obtained from commercial suppliers and used without
further purification. Elemental analyses were performed by
the Microanalytical Laboratory operated by the College of
Chemistry, University of California at Berkeley.
9
Kondo type ion, the energy required is simply too high
for praeseodymium, and other pathways result.
P ota ssiu m Bis-π-(8-a n n u len e)cer a te(III).13 The earlier
Con clu sion s
12
preparation was improved by the use of CeI
3
(THF)
4
.
To a
stirring mixture solution of 2.00 g (19.2 mmol) of freshly
distilled cyclooctatetraene (BASF) in 40 mL of THF was added
In its chemistry, cerocene behaves remarkably like
actinide(IV) metallocenes with [8]annulene ligands. Like
thorocene and uranocene, in particular, it does not
1
.62 g (41.4 mmol) of freshly cut potassium metal. After
stirring for 4 h, most of the potassium had reacted. The
mixture was filtered and added slowly to a stirring mixture
of 4.09 g (5.10 mmol) of CeI (THF) and 80 mL of THF. The
3 4
resulting green solution was allowed to stand for several hours
without stirring to allow the precipitate to settle to the bottom.
The mixture was filtered, and the volume of the green filtrate
was reduced to 25 mL. The green precipitate was removed from
the mother liquor and washed three times with 5 mL of THF
to yield after drying in vacuo 1.56 g (79%) of product. This
product can be crystallized from diglyme to give [K(diglyme)]-
2
-
behave as an ionic salt of C8H8 . Like the higher An-
(
C8H8)2 metallocenes it undergoes facile electron ex-
-
change with the corresponding An(C8H8)2 ions on the
NMR time scale. The conventional MO picture of the
actinide sandwiches is given schematically in Figure 6.
Instead of a simple electron transfer redox reaction
between an electron-rich ligand orbital and a lower-lying
empty metal orbital, orbital overlap gives a doubly
occupied bonding orbital that provides ring-metal
bonding. Since this bonding MO has a contribution from
[
7
Ce(C
.5. Found: C, 48.0; H 6.0; K, 9.2.
P r ep a r a tion of Ce(C . F r om AgI. To 0.200 g (0.52
mmol) of K[Ce(C ] in an argon atmosphere glovebox was
8 8 2 30 3
H ) ]. Anal. Calcd for C22H O KCe: C, 50.7; H, 5.8; K,
8
8 2
H )
(
21) DeKock, C. W.; Ely, S. R.; Hopkins, T. E.; Brault, M. A. Inorg.
Chem. 1978, 17, 625-631.
22) Kinsley, S. A., Streitwieser, A., Zalkin, A. Organometallics 1985,
, 52-57.
8
8 2
H )
added 30 mL of THF, with stirring until all dissolved. To this
pale green solution under red light was then added 1.21 g (5.2
(
4

Di-π-[8]annulenecerium(IV), Cerocene
Organometallics, Vol. 23, No. 22, 2004 5175
mmol) of AgI, which had been freshly ground in a mortar. The
deep brown mixture was stirred for 2 min and filtered with
fine filter paper to give a clear dark brown solution. The filter
paper was rinsed several times with small portions of THF,
leaving a fine black material in the filter paper. The solvent
was removed from the filtrate in vacuo, to leave a brown-black
d
8
was added, and the cerocene was allowed to dissolve into a
brown solution. To this tube was added an approximately
comparable amount of UCl , and the tube was removed from
the box and sealed under vacuum. After several days, a small
peak in the H NMR appeared at -36 ppm corresponding to
4
1
uranocene.
powder, which was rinsed several times with small portions
1
Electr och em istr y. Gen er a l. Electrochemical measure-
ments were conducted with a Bioanalytical Systems CV-50
potentiostat with platinum disk working, platinum wire
auxiliary, and silver wire reference electrode. Samples were
of hexane, yielding 0.142 g (0.41 mmol, 78%) of Ce(C
NMR (THF-d ): δ 5.9 (s); (toluene-d ): δ 5.75 (s).
8
H
8
)
2
. H
8
8
F r om Allyl Br om id e. To a 500 mL round-bottom flask
equipped with a magnetic stirbar was added 0.65 g (1.8 mmol)
of potassium bis-(8-annulene)cerate(III) and 200 mL of dry
THF. To this light green solution was added 0.70 g (5.9 mmol)
of allyl bromide. This solution was allowed to stir overnight,
the solvent was removed in vacuo, and the remaining solid
was extracted with 3 × 250 mL portions of toluene. The
extracts were filtered and combined, and the toluene was
removed in vacuo. The resulting finely crystalline black powder
4 4
measured with 0.1 M Bu NBF as the supporting electrolye,
and ferrocene was added as either an internal or external
reference after measurements involving cerocene. Due to an
apparent reaction with ferrocene, cerocene was used as an
internal standard for electrochemical measurements of K[Pr-
(
(
C
C
8
H
H
8
)
)
2
8 8 2 2
]. The lanthanide complexes K[Ce(C H ) ], K [Yb-
8
8
2
], and K[Pr(C
8
H
8
)
2
] for the electrochemical measure-
was washed three times with hexane and dried in vacuo, giving
11,22
ments were prepared from previously reported methods
and purified by crystallization.
1
0
d
.56 g (89% yield); m/z ) 348, λmax ) 470 nm; H NMR (THF-
8
) δ 5.90 (s). Anal. Calcd for C16H16Ce: C, 55.1; H, 4.6.
Tetrahydrofuran (THF) for the electrochemical measure-
ments was distilled from sodium benzophenone, stirred with
lithium aluminum hydride, vacuum transferred and degassed
via freeze-pump-thaw, and stored over activated 3 Å molec-
ular sieves. Tetra-n-butylammonium tetrafluoroborate was
obtained from Fluka (puriss) and recrystallized from either
ethyl acetate/hexane or acetone/ether and dried in vacuo or
prepared as given below. Tetra-n-butylammonium hexafluo-
rophosphate was obtained from Fluka (puriss) and dried in
vacuo prior to use. Ferrocene was obtained commercially
Found: C, 54.4, H, 4.5. The UV-vis spectrum and mass
spectrum are shown in Figures S1 and S2 (Supporting
+
Information). The molecular ion is correct for Ce(C
8
H
8
)
8
2
, and
+
+
important fragment peaks occur for Ce(C
8
H
8
)
and C
H
8
.
Bis-π-[b is(1,4-t r im e t h ylsilyl)-8-a n n u le n e )]ce r iu m -
6
6
(
IV). The preparation differs from that of Kilimann et al.,
, and allyl bromide. To a solution of
.0 g (20 mmol) of 1,4-bis(trimethylsilyl)cyclooctatriene23 in 100
by the use of NaH, CeI
3
5
mL of THF was added with stirring a large excess (4.79 g, 200
mmol) of freshly ground NaH over 4 h. The dark green mixture
was allowed to stir overnight and filtered, and the THF was
removed in vacuo from the filtrate. The resulting material was
washed with hexane to remove unreacted cylcooctatriene,
leaving 0.84 g (2.9 mmol) of a cream-colored disodium 5,8-bis-
(Aldrich) and sublimed prior to use.
Cyclic Volta m m etr y Meth od s. Approximately 10 mL of
a stock solution of the electrolyte in THF (in the range of 0.1
M) was added to the electrochemical cell in an argon glovebox.
The electrodes were then set up, and the background volta-
mmogram was recorded. If this was clear, the sample was
(
trimethylsilyl)cyclooctateraenediide. To this material dis-
solved in 100 mL of THF was added 1.17 g (1.4 mmole) of
CeI (THF) in 50 mL of THF. The dark green mixture was
3
4
-
3
filtered, and the THF was removed from the filtrate in vacuo,
leaving a dark green solid. This solid was washed with
pentane, dried under vacuum, and dissolved in THF. To the
green solution was added 0.73 g (6.0 mmol) of allyl bromide.
This solution became dark brown over several minutes and
was allowed to stir overnight. The THF, excess allyl bromide,
and reaction byproducts were removed under vacuum. The
product was extracted from the remaining material with
pentane. Pentane was removed from the combined washes in
added (in the general range of 1 × 10 M, although this is
not critical and was usually not weighed out), and the
appropriate sweeps were carried out. Scan rates were in the
range 100-400 mV/s, with the majority of the measurements
being made at 100 mV/s. A small amount of ferrocene was
added until the ferrocene peaks were clearly visible. Electrodes
were wiped with Kimwipes and rinsed with THF between
samples, unless they were dirty enough to require repolishing
or cleaning, in which case they had to be removed from the
glovebox. The bis[8]annulene lanthanates are extremely air-
sensitive. Noticeable color changes in solutions were apparent
after 30 min exposure to the glovebox atmosphere, even though
standard tests (e.g., lack of fuming from diethylzinc) indicated
an oxygen “free” atmosphere.
1
vacuo, leaving 0.96 g of a dark brown semisolid material. H
NMR (250 MHz, toluene-d
m/z 636; λmax ) 470 nm.
8
): δ 0.06 (s, 3H), 5.6-6.2 (m, 1H);
Oxid a tion of K[P r (C
8 8 2
H ) ] w ith Allyl Br om id e. In an
argon atmosphere glovebox, 0.16 g (0.41 mmol) of K[Pr(C
8
8 2
H ) ]
was added to a 100 mL flask equipped with a stir bar. Addition
of 20 mL of degassed, anhydrous THF dissolved the bright
yellow powder to give a golden yellow solution. Allyl bromide
Ack n ow led gm en t. This research was supported in
part by the Office of Basic Energy Sciences, Chemical
Scieces Division of the U.S. Department of Energy, and
by grants from the National Science Foundation. Special
thanks to Profs. R. A. Anderson for his donation of
anhydrous PrCl3 and J . R. Long for use of electrochemi-
cal equipment and glovebox. S.K. also thanks Laurance
Beauvais, Miriam Bennett, and Matthew Shore for their
assistance in obtaining the cyclic voltamagrams.
(
0.06 g, 0.5 mmol), degassed by four freeze-pump-thaw cycles
on a Schlenk line, was added to the flask, via syringe. After
4 h, removal of the THF in vacuo from the now pale yellow
2
solution left a yellow powder. The powder was extracted two
times with 30 mL of toluene. The toluene-insoluble powder
was redissolved in THF; removal of the THF in vacuo left 0.095
1
3
g of a bright yellow powder, λmax ) 504 nm (lit. for K[Pr-
], λmax ) 502 nm.). The toluene was removed from the
8 8 2
(C H )
faint yellow extract, leaving a small amount (<20 mg) of pale
yellow powder. UV-vis: (in THF) λmax ) 426, 400, 376, 254
nm. MS: m/z 594 (Pr
Liga n d Exch a n ge Rea ction . A sample of Ce(C
placed in an NMR tube equipped with a stopcock joint. THF-
2 8 8 3
(C H ) ) 594).
8
H
8
)
2
was
Su p p or tin g In for m a tion Ava ila ble: Eight figures. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(23) Bellema, J . M.; Davison, J . B. J . Organomet. Chem. 1975, 86,
6
9-74.
OM049743+