Simple preparation of cobaloxime dienyl complexes and their exo selective Diels-Alder cycloadducts. Progress toward transition metal-mediated Diels-Alder reactions which are catalytic in metal dienyl complex

Article abstract of DOI:10.1016/S0022-328X(03)00587-4

A new method for the synthesis of cobaloxime dienyl complexes (cobaloxime=pyridine(dimethylglyoxime)₂Co) which involves a zinc-mediated hydrocobaltation of enynes is reported. Use of this dienyl complex synthesis method allows subsequent Diels-Alder reactions of these complexes to be accomplished without isolation of the dienyl complexes. A new silane-based cobaloxime-substituted cycloadduct demetallation protocol is then reported which provides Diels-Alder cycloadducts and cobalt complexes which have been recycled into the zinc-mediated dienyl complex synthesis.

Full text of DOI:10.1016/S0022-328X(03)00587-4

Journal of Organometallic Chemistry 681 (2003) 120Á133
/
www.elsevier.com/locate/jorganchem
Simple preparation of cobaloxime dienyl complexes and their exo
selective DielsÁAlder cycloadducts
Progress toward transition metal-mediated DielsÁ
/
/
Alder reactions
which are catalytic in metal dienyl complex
Kerry A. Pickin, Jennifer M. Kindy, Cynthia S. Day, Mark E. Welker *
Department of Chemistry, Wake Forest University, Winston-Salem, NC 27109-7486, USA
Received 9 April 2003; received in revised form 11 June 2003; accepted 16 June 2003
Abstract
A new method for the synthesis of cobaloxime dienyl complexes (cobaloximeꢀ
zinc-mediated hydrocobaltation of enynes is reported. Use of this dienyl complex synthesis method allows subsequent DielsÁ
reactions of these complexes to be accomplished without isolation of the dienyl complexes. A new silane-based cobaloxime-
substituted cycloadduct demetallation protocol is then reported which provides DielsÁAlder cycloadducts and cobalt complexes
which have been recycled into the zinc-mediated dienyl complex synthesis.
2003 Elsevier B.V. All rights reserved.
/
pyridine(dimethylglyoxime)
2
Co) which involves a
/
Alder
/
#
Keywords: DielsÁ
/
Alder; Dienyl complexes; [4ꢁ2] Cycloaddition; Demetallation
/
1
. Introduction
The use of cobalt complexes to mediate or catalyze
(1) for use in 4ꢁ2 cycloadditions [2]. We initially
/
prepared a number of cobaloxime- and cosalen-substi-
tuted 1,3-dienes (1) and found them to be air-stable
solids of high thermal stability that could be prepared
on a multi-gram scale from inexpensive starting materi-
als [3]. The cobalt dienyl complexes (1) reacted with a
variety of dienophiles (2) both regioselectively and
stereoselectively (with high exo diastereoselectivity and
enantioselectivity [4]) to produce air-stable cobalt-sub-
stituted cyclohexenes (3) in high yield. We also devel-
organic reactions has grown tremendously in the last
few years and there are now many reported examples of
cobalt-mediated cycloaddition or cyclization reactions
[
1]. Of these cycloaddition reaction types, the DielsÁ
/
Alder reaction is the most thoroughly studied in all of
organic chemistry. ISI’s citation database contains
almost 1000 publications per year for each of the last
oped a number of methods for cleavage of the cobaltÃ
/
4
chemistry.
years that deal with some aspects of DielsÁ
/
Alder
carbon bonds in the cycloadducts which would yield
organic products (4) as well as a cobalt complex (5)
which could be recycled into the synthesis of the starting
dienyl complexes (1). The exo diastereoselectivity of
these cycloaddition reactions provides access to cyclo-
hexene stereochemistries which are complimentary to
About 10 years ago, we first became interested in
preparing 2-transition metal-substituted 1,3-butadienes
*
Corresponding author. Tel.: ꢁ
321.
E-mail address: welker@wfu.edu (M.E. Welker).
/
1-336-758-5758; fax: ꢁ1-336-758-
/
those obtained from traditional endo selective DielsÁ
/
4
0
Alder reactions.
022-328X/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00587-4

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
121
This chemistry would be more useful to the organic
community if the stoichiometric reaction sequence
described above could be affected without isolation
and purification of 1 and 3 or, even better, through a
series of reactions which were catalytic rather than
stoichiometric in metal complex. In theory, a catalytic
cycle for the reaction sequence can be drawn (Fig. 1) and
all the individual steps involved in the cycle (cobalt
hydride (6) generation, enyne (7) hydrometallation,
to address them. We wondered if we used boron hydride
for catalyst generation and reductant whether our
enynes (such as 7) would be stable in the presence of
these reagents. We thought that even if enynes were
hydroborated [5] faster than hydrocobaltation occurred
those species must be transmetallating to cobalt or else
we would never have been able to isolate good yields of
cobaloxime dienes from our hydrometallation experi-
ments [6]. However, since the proposed cycle would
require enynes to be unreactive, relative to Co(II)
toward mild reducing agents, and the order of addition
of reagents was different from our earlier work [6], we
ran the control experiment. trans-Pentenyne (7) in
DielsÁAlder reactions of cobalt dienyl complexes (8)
/
with dienophiles (9) to generate metal-substituted cy-
cloadducts (10), and reductive demetallation of cycload-
ducts (10) to produce compounds like 11) have been
reported previously by our group [3]. In practice, there
are numerous challenges to be overcome to effect the
cycle shown and our progress toward that goal is
presented here.
methanol was treated with NaBH first, followed by
4
addition of (pyr)(dmg) CoCl and then dropwise addi-
2
tion of acetic acid to bring the pH back to 8. Quenching
of this reaction with ice water containing a few drops of
pyridine produced dienyl complex (8) in the same purity
1
(
by H-NMR comparison) and in the same isolated yield
as that observed from our typical order of addition [6].
Even if boron-substituted dienes were made as transient
intermediates, they are unlikely to compete with transi-
2
. Results and discussion
2.1. Control experiments to test the viability of the cycle
in Fig. 1
tion metal-substituted dienes in DielsÁ/Alder reactions.
(This observation may prove useful in that we may be
able to make transition metal dienyl complexes by
transmetallation rather than hydrometallation in the
future.) 1,3-Dienes substituted with main group ele-
ments in the 2 position are much less reactive than
cobalt-substituted dienes [7]. For instance, 2-stannyl or
We will first discuss some of the potential problems in
the proposed cycle and what control experiments we ran
seleno-substituted dienes require 2Á
react with maleic anhydride whereas similar cobalt-
substituted diene reactions are complete in 15Á30 min
8].
The catalytic cycle proposed in Fig. 1 also requires
/
3 days at 25 8C to
/
[
that dienophiles such as dimethyl fumarate (9) be stable
in the presence of cobalt anions and hydrides (6) and we
were concerned that these species might react faster with
dienophiles than the enynes [9]. The cobaloxime anion/
hydride mixture generated by the addition of NaBH to
4
(pyr)(dmg) CoCl was treated with a 1:1 mixture of
2
trans-2-pentenyne (7) and dimethyl fumarate (9) at 0 8C.
We isolated only cobaloxime diene (8, L ꢀ(dmg) pyr)
/
n
2
and cobaloxime-substituted cycloadduct (10) (60% mass
balance, 1:2 diene:cycloadduct). Cobaloximes have rea-
sonable water solubility and so when we have to do an
extraction to isolate complexes our mass balances for
cobalt are usually in the 60Á/70% range. No product of
Fig. 1. A possible catalytic cycle for effecting transition metal-
mediated DielsÁAlder reactions.
/
addition of cobaloxime to dimethyl fumarate was

122
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
isolated. This competition experiment was also encoura-
ging in that two of the three cycle steps, enyne
these individual steps led us to the conclusion that, while
the cycloaddition and demetallation steps could be
hydrometallation and DielsÁ
curred in the desired sequence under very mild condi-
/
Alder reaction, have oc-
performed in ꢀ
by the hydrometallation used since chemical yields for
this step are routinely in the 50Á60% range. This
limitation is presumably due to the thermal sensitivity
of the cobalt hydride (6) [9] and we viewed the
temperature cycling that this reaction sequence required
as a significant disadvantage. A metal dienyl complex
generation procedure which could be performed at
elevated temperatures in high yield was first needed.
/
90% yield, we were going to be limited
tions (0 8C). We also ran the control of treating dimethyl
fumarate with NaBH in methanol to make sure the
dienophile was not reduced by our reaction conditions.
/
4
2
.2. Demetallation questions/controls
An absolutely critical chemical requirement for the
2
cycle in Fig. 1 is that the metal sp Ã
/
carbon bond in the
2
diene (8) be more stable than the cobalt sp Ã
/carbon
2.3. Development of a new method for cobaloxime dienyl
complex preparation
bond in the cycloadduct (10). We had noticed this to be
true on many occasions when we ran demetallations [8],
i.e. cobalt-substituted cycloadducts (such as 10) contain-
ing small amounts of unreacted cobalt diene (8) could be
treated with reductants and the demetallated cycload-
duct (11) isolated along with unreacted cobalt-substi-
tuted diene (8) but we had never run a formal
competition experiment. A 1:1 mixture of dienyl com-
plex (8) to cycloadduct (10) was treated with an
We first screened a number of possible reductants for
the enyne hydrocobaltation step. In general, we found
that aluminum hydrides were too reactive, whereas
silanes were generally unreactive as possible reductants.
A cobaloxime reduction/electrophile trapping method
which used zinc as a reductant had been developed
previously by Roussi and Widdowson [10]. We had used
2
equimolar amount of NaBH in methanol to produce
4
a modification of this method previously to prepare sp
demetallated cycloadduct (11) and recovered unreacted
diene (8). We also found that we could treat cobalt-
substituted cycloadduct (10) with 1:1 mixture of NaBH4
and the enyne (7) and produce a 1:1 mixture of
demetallated cycloadduct (11) (27% isolated yield) to
cobaloxime dienyl complex (8) (29%). The isolated
yields here were not great but it did show that the
demetallation and hydrometallation steps of the cycle
could also be performed in the same pot.
carbonÃ
/
cobalt bonds from vinyl triflates [11] and
alkynones [12].
The successful metallation of the alkyne of alkynones
led us to consider the enyne as a substrate for the
modified Widdowson protocol. By substituting the
alkynone with (E)- and (Z)-3-pentene-1-yne (12), we
were able to produce the 2-substituted dienyl complexes,
13 and 7, in excellent yield. The zinc-assisted metallation
of the enyne provides a replacement method for the
synthesis of the initial cobalt dienyl complex in the cycle
and can be performed at elevated temperatures where
cycloaddition reactions should also be rapid. The ratio
of E to Z isomers produced by this method is similar to
the ratio we had noted previously when the (Z)-
pentenyne was hydrometallated using a borohydride-
based procedure [6]. We had previously demonstrated
that (E)-cobaloxime dienyl complex (7) participated in
Once the control experiments were completed, we
were ready to test the cycle outlined in Fig. 1 as a one-
pot reaction. Cobaloxime chloride [pyr(dmg) CoCl] was
2
treated with NaBH4 at 0 8C to generate cobaloxime
hydride (6). This solution was then warmed to 25 8C
after enyne (7) was added. Dimethyl fumarate (9) was
then added and the solution was refluxed for 1 h.
Additional NaBH was added to effect demetallation
4
1
and the crude product was analyzed by H-NMR.
DielsÁAlder reactions much faster than (Z)-dienyl
/
Compounds 11, 10 and 8 were present in a 1.6:1:1 ratio.
Considerable effort to optimize conditions for each of
complex (13) and that Z:E isomerization was much
faster than cycloaddition [13]. These observations per-

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
123
mitted us to use the more readily available E:Z mixture
of enynes (12) [14] in the complex preparation.
2.5. Tandem dienyl complex synthesis/DielsÁ
reactions
/
Alder
Mindful of the fact that we were trying to simplify
preparation of organic cycloadducts from cobalt-
2
reactions
.4. One-pot dienyl complex synthesis/DielsÁ
/Alder
mediated DielsÁAlder reactions, we decided that re-
/
moval of zinc dust, prior to the cycloaddition, could
probably be accomplished without chromatography but
Once this initial dienyl complex preparation step
proved successful, the reaction was taken to the next
step of the cycle in the same reaction flask. Upon
completion of the initial 1-h reflux, dienophile dimethyl
fumarate was added to the reaction and the mixture was
refluxed for an additional 1 or 24 h.
removal of Zn(OAc)
(which has reasonable THF
2
solubility) would be more problematic. Removal of
most zinc dust from the reaction pot was accomplished
using a specially adapted double-ended needle where
The results for the one-pot reaction showed that
synthesis of the dienyl complex and cycloaddition could
occur in the same reaction vessel. However, the presence
of a significant amount of unreacted (Z)-dienyl complex
one side of the needle was fitted with a glass frit. The
reaction solution was then transferred through the filter
fitted needle into a clean reaction vessel.
Once the zinc removal system had been successfully
developed, the scope of the dienyl complex synthesis/
cycloaddition sequence was determined using a number
of dienophiles. When dimethyl fumarate was used again
as the dienophile, the amount of unreacted dienyl
complex (13) dropped to 17% and the isolated yield of
the one diastereomer produced (14) improved to 65%.
N-Phenyl maleimide cyclized to produce a 9:1 mixture
of diastereomers. The major diastereomer is assigned as
the anti diastereomer (15) and the minor diastereomer
was assigned the syn stereochemistry (16). We had
(
13) in the 24-h reaction suggested possible interference
from one of the other reactants during the cycloaddition
reaction. When the purified E/Z mixture of dienyl
complexes (13 and 7) was heated to reflux in THF
with five equivalents of dimethyl fumarate overnight,
cycloadduct 14 was isolated in 90Á98% yield. Likewise,
/
additional control cycloadditions run using 13 and 7
with dimethyl fumarate plus 20 mol% pyr(dmg) CoCl or
2
2
0 mol% pyrHCl (compounds likely to be in the one-pot
cycloaddition) produced 14 only again in near-quanti-
tative yield. However, a control cycloaddition run with
1
earlier noted that H-NMR anti cycloadduct methyl
resonances in maleic anhydride cycloadducts (d 0.99
2
0 mol% zinc and 20 mol% Zn(OAc) present produced
2
ppm, CDCl ) were considerably upfield of the corre-
3
cycloadduct 14 in only 55% isolated yield. Competing
zinc metal reduction of a significant portion of the
dienophile [15] prior to cycloaddition appeared to be
one possible cause for the reduced yields of cycloadduct
seen in the one-pot reaction. Another possibility might
sponding syn adducts (d 1.37 ppm, CDCl ) [8,16]. We
3
see the same trend for the major and minor diaster-
eomers (15 and 16) here (major diastereomer methyl
doublet: d 1.04 ppm, CDCl ; minor diastereomer methyl
3
doublet: d 1.35 ppm, CDCl ). Maleic anhydride cyclized
3
be that Zn(OAc) exchanges with the bridging OH in the
2
to produce a 3:1 mixture of anti:syn diastereomers
(17:18) as we had seen in our earlier work [8]. A
dmg ligands and that complex is significantly less
reactive in DielsÁ/Alder reactions [13]. Such a dienyl
DielsÁAlder reaction which proceeds through predomi-
/
complex would be hydrolyzed back to the protonated
ligand form upon work up or chromatography [13].
nantly an exo rather than endo transition state rationa-
lizes 15 and 17 as the major products and the larger size

124
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
of N-phenyl maleimide would account for the larger
proportion of anti adduct for that dienophile. The small
dienophile, methyl vinyl ketone, which also contains
only one electron-withdrawing group, cyclizes in lower
yield and has a preference for the product predicted
from an endo (20) rather than exo (19) transition state
mediated reactions typically favor products which have
arisen from exo rather than endo transition states when
di-, tri- or tetra-substituted dienophiles are used [3].
Dimethyl fumarate is unusual among the dienophiles in
this series because it does not have different possible
regiochemical outcomes and endo or exo selectivity
requires clarification due to the opposing, symmetrical
substitution pattern of the methyl esters. In any
approach pattern of the dienophile to the diene, one of
the methyl esters will be endo and one will be exo (21
[
8]. It should be noted that the yields reported for
cycloadduct below are the cumulative isolated yields for
two chemical steps, dienyl complex synthesis and DielsÁ
/
Alder reaction.
The stereochemical outcome of the cycloaddition with
dimethyl fumarate deserves additional comment. The
stereochemical outcome of cobalt-mediated cycloaddi-
tion reactions is many times unique. These cobalt-
and 24). However, the reaction did produce a single,
unique stereochemical outcome in the cycloadduct.
Steric arguments would predict that the dienophile
would favor an approach that places the methyl ester

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
125
that is endo away from the cobaloxime moiety (24). The
relative stereochemistry of 14, which would arise from
the diene-dienophile approach shown in 24, was con-
moiety must be cleaved from the cycloadduct and a
viable metal species needed to be regenerated for reuse
in the cycle.
firmed by X-ray crystallography (see Fig. 2). Table 1
contains details of the structure determination. A partial
list of bond lengths and angles for 14 appears in Table 2.
Chlorosilane-mediated demetallations met the re-
quirements outlined above in that they could be
performed in THF at 25 8C or above. The results of
treatment of the cobaloxime cycloadduct (14) with
various silanes under both thermodynamic and photo-
chemical conditions are presented in Table 3. The best
cleavage reaction combined six equivalents of dichlor-
omethyl silane with reflux for 1 h to produce the desired
cycloadduct (26) in high yield. During this reaction, the
solution changed color from yellow to bright green and
a fine teal solid precipitated out of solution. Analysis of
the precipitate by MS showed it to be a mixture of
2
.6. Demetallation reactions
Many of the demetallation reactions we reported in
our earlier work were performed in protic solvents
where transesterification or hydrolysis of some sub-
(
strates could be a problem) or at low temperatures [3].
Since the diene complex synthesis and cycloadditions
reported for the first two steps of the cycle were now
being performed in THF at 25 8C or above, ideally we
would develop a demetallation which could be effected
under these same conditions. In order to complete the
reaction cycle initially proposed (Fig. 1), the metal
ꢂ
ꢁ
cobalt salts including [pyrCoCl
(pyr) CoCl . The exact composition of this salt mixture
] (pyrH)
and
3
2
2
was not determined since what we were really interested
in was a demetallation which could result in the

126
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
Table 1
Crystal data and structure refinement for (1,2-dicarboxylic acid dimethyl ester-3 methyl-4-cyclohexen-5-yl)pyridinebis(dimethylglyoximato)cobalt
Empirical formula
Formula weight
Temperature (K)
Co(NC
727.92
228(2)
5
H
5
2
)(N C
2
3 2 2 2 6 3 6 2 3 2 2 2 4
(CH ) O H) (C (CH )H (CO CH ) )×/1.5Cl C H
˚
Wavelength (A)
0.71073
monoclinic, C2/c-
6
h
Crystal system, space group
/
C2 (no. 15)
Unit cell dimensions
˚
a (A)
˚
35.207(5)
9.163(1)
23.743(3)
114.19(1)
6987(2)
8, 1.384
0.771
b (A)
˚
c (A)
b (8)
3
˚
V (A )
ꢂ3
Z, calculated density (g cm
Absorption coefficient (mm
F(0 0 0)
)
ꢂ1
)
3032
3
Crystal size (mm )
Theta range for data collection (8)
Limiting indices
0.09ꢃ
2.31Á24.15
15h 540, ꢂ
/0.44ꢃ0.50
/
/
ꢂ/
/
/
/
15
/
k 5
/
10, ꢂ
/
275
/
l 525
/
Reflections collected/unique
Absorption correction
6061/5111 [Rint
ꢀ/0.048]
empirical
Max. and min. transmission factors
Absolute structure parameter
Refinement method
0.978 and 0.643
0.09(2)
Full-matrix least-squares on F
2
Data/parameters
Goodness-of-fit on F
Final R indices
5111/451
1.020
2
[
2896F
all 5111 data]
o
ꢀ
/
4s(F ) data]
o
R
R
1
1
ꢀ
ꢀ/
/
0.063, wR
2
2
ꢀ
ꢀ/
/
0.128
[
0.050, wR
0.106
Largest diff. peak and hole (e Aꢂ3)
0.463 and ꢂ0.304
/
˚
regeneration of the cobaloxime dienyl complexes (13
and 7) needed for cycloaddition. The cobalt salt mixture
was suspended in a solution of THF and followed by the
dienyl complexes (13 and 7) followed by filtration of the
solution to remove excess of Zn dust. Dienophile was
added and the reaction mixture was refluxed for 17 h.
addition of dmg, pyridine, enyne and Zn dust in that
order. The reaction mixture was refluxed for 1 h to
regenerate the dienyl cobaloxime complexes (13 and 7)
in 99% yield.
Silane was then added to the reaction pot to cleave the
cobalt moiety from the cycloadduct. The solution was
filtered to remove the green precipitate and the cleaved
cycloadduct (26) was isolated by pentane extractions
and radial chromatography. The green precipitate was
regenerated into cobaloxime dienyl complexes (13 and
2
.7. Completion of Fig. 1 reaction cycle
7
). The complete cycle was attempted with dimethyl
fumarate producing 23% (62% average yield for each of
the three steps with no intermediate isolation) of the
cleaved cycloadduct (26) and 28% of the regenerated
dienyl complexes (13 and 7). (The original cobalt used
Once the individual steps of the alternative reaction
cycle had been optimized, a complete cycle with no
intermediate purifications or isolations was attempted.
The cycle began with the synthesis of the cobaloxime

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
127
3. Summary
In conclusion, we report here a new method for
cobaloxime dienyl complex generation and DielsÁAlder
cycloadditions which does not require the isolation and
/
purification of the cobalt dienyl complexes. We also
2
report a new silane-mediated cobalt sp Ã
/
carbon bond
cleavage which yields DielsÁAlder cycloadducts and
/
cobalt salts which have been converted back into dienyl
complexes in high yield. Both these observations are
significant advances over our previously reported pro-
cedures [3]. While all three steps of the cycle proposed in
Fig. 1 can be performed without intermediate character-
ization and purification, in practice, the chemical yields
are higher if just the first two steps of the cycle are run in
this manner. Using our optimized protocol, cobaloxime-
substituted DielsÁAlder cycloadducts are now isolated
/
and purified prior to performing the silane-mediated
demetallation reaction. We are directing current efforts
toward development of a cycle which uses transmetalla-
tion rather than hydrometallation for initial dienyl
complex synthesis.
Fig. 2. Perspective drawing of the solid-state structure for Co(N-
)(N (CH H) (C (CH )H (CO CH ) (14). All non-hydro-
C
5
H
5
2
C
2
3
)
2
O
2
2
6
3
6
2
3 2
)
gen atoms are represented by thermal vibration ellipsoids drawn to
encompass 50% of their electron density. Hydrogen atoms are
represented by arbitrarily small spheres which are in no way
representative of their true thermal motion. Hydrogen-bonding inter-
actions are represented with dashed open bonds.
4. Experimental
4.1. General
Proton nuclear magnetic resonance (NMR) spectra
were obtained using a Bruker Avance 300 MHz spectro-
has gone through four reaction steps: initial dienyl
meter operating at 300.1 MHz or a Bruker 500 MHz
1
3
complex synthesis, DielsÁ
/
Alder reaction, demetallation
and dienyl complex regeneration to get to this point.)
spectrometer operating at 500.1 MHz. C-NMR spec-
tra were obtained using a Bruker 300 MHz spectrometer

128
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
Atlantic Microlabs, Inc., Norcross, GA. High-resolu-
tion mass spectrometry was performed at the Duke
Mass Spectrometry Facility, Duke University in Dur-
ham, NC. Melting points were recorded using a Mel-
temp apparatus and are reported uncorrected. Thin-
layer chromatography was carried out on 0.25 mm pre-
coated Polygram SIL G/UV plates purchased from
Bodman. Silica gel 60 purchased from Natland Interna-
tional Corporation was used for standard column
chromatography. Silica gel 60/UV 254 purchased from
EM Science was used for radial chromatography.
All reactions were carried out under a dry nitrogen
atmosphere in flame-dried glassware unless otherwise
specified. Tetrahydrofuran and diethyl ether were dis-
tilled under nitrogen from sodium/benzophenone im-
mediately prior to use. Dichloromethane, pentane,
toluene and benzene were distilled under nitrogen from
calcium hydride immediately prior to use. Dimethyl
fumarate, maleic anhydride, methyl vinyl ketone, N-
phenyl maleimide, dichloromethylsilane, dimethylchlor-
osilane, diphenylsilane and triethylsilane were purchased
from Aldrich and used as received. Dimethylglyoxime
Table 2
Selected bond lengths and angles in crystalline (1,2-dicarboxylic acid
dimethyl ester-3-methyl-4-cyclohexen-5-yl)pyridinebis(dimethylglyoxi-
a
mato)cobalt
˚
Bond length (A)
Co
Co
Co
1
1
1
Ã
/
N
1
2
1.894(5)
1.894(5)
2.001(6)
1.344(7)
1.453(7)
1.193(7)
1.323(8)
1.519(8)
1.514(7)
1.531(7)
1.540(8)
1.895(4)
1.896(5)
2.079(5)
1.347(7)
1.433(7)
1.194(7)
0.99(4)
Ã
/N
Ã
/
C
10
15
O
5
O
5
O
6
Ã
/
C
C
C
Ã
/
16
15
10
14
Ã
/
C
C
C
C
C
9
9
Ã
Ã
10
11
13
/C
/C
Ã
/
C
11
12
14
Ã
/
C
C
Ã
/
Co
Co
Co
1
1
1
Ã
Ã
Ã
/
N
3
4
5
/
N
/
N
O
7
O
7
O
8
Ã
/
C
C
C
17
18
17
Ã
/
Ã
/
C
C
C
C
C
9 9
ÃH
/
12
13
12
14
Ã
/
C
15
17
13
19
1.494(8)
1.502(8)
1.525(8)
1.523(9)
Ã
/
C
C
C
Ã
/
was purchased from Aldrich and recrystallized from
ꢂ1
Ã
/
95% EtOH (12 ml g ) prior to use. Pyridine was
Bond angle (8)
N
N
N
N
N
N
N
N
2
2
1
3
2
1
3
4
Ã
/
Co
Co
Co
Co
Co
Co
Co
Co
1
1
1
1
1
1
1
1
Ã
/
N
N
N
N
1
3
4
4
81.6(2)
98.5(2)
97.7(2)
82.2(2)
88.3(2)
90.0(2)
89.1(2)
90.3(2)
115.8(5)
121.6(5)
120.9(4)
117.4(4)
111.9(5)
90.9(2)
92.03(19)
88.95(19)
90.4(2)
179.0(2)
178.5(2)
177.8(2)
116.2(5)
124.8(5)
124(3)
purchased from either Aldrich or Acros and use as
received. Zinc dust and cobalt acetate hexahydrate were
purchased from Strem and used as received. 3-Penten-1-
yne was prepared according to a known literature
Ã
/
Ã
/
Ã
/
Ã
/
Ã
/
Ã
/
Ã
/
Ã
/
C
C
C
C
10
10
10
10
procedure [14]. Control experiments where NaBH was
4
Ã
/
Ã
/
Ã
/
Ã
/
used for reduction or demetallation were performed
Ã
/
Ã
/
using procedures we have previously reported [8].
C
C
C
C
C
17
Ã
/
O
7
Ã
/
C
18
9
Ã
Ã
/
C
10
10
ÃC
/
11
9
/C
Ã
/
Co
1
11
17
Ã
/
C
10
Ã
/
Co
1
4.2. Synthesis of (3E)- and (3Z)-1,3-pentadien-2-
Ã
/
C
13
ÃC
14
/
ylpyridinebis(dimethylglyoximato)cobalt(III) (7 and
13) using Zn as reductant
N
2
N
1
N
3
N
4
N
1
N
2
Ã
/
Co
1
1
1
1
1
1
Ã
/
N
N
N
N
N
N
5
5
5
5
3
4
Ã
/
Co
Co
Co
Co
Co
Ã
/
Ã
/
Ã
/
Ã
/
Ã
/
Cobalt(II) acetate hexahydrate (0.489 g, 2.0 mmol)
and dimethylgloxime (0.464 g, 4.0 mmol) were dissolved
in degassed THF (50 ml). Pyridine (0.475 g, 6.0 mmol)
was added followed by Zn dust (0.654 g, 10.0 mmol).
This mixture was refluxed for 15 min. The solution was
cooled briefly to stop the reflux and 3-penten-1-yne
Ã
/
Ã
/
Ã
/
Ã
/
C
C
C
C
C
C
C
10
15
10
10
14
10
15
/
1 5
ÃCo Ã/N
Ã
Ã
Ã
Ã
Ã
Ã
/
O
5
Ã
/
C
16
/
C
C
C
C
C
9
9
9
ÃC
/
14
/
Ã
/
H
9
9
/
Ã
/
H
111(3)
113.1(5)
112.0(5)
(0.200 g, 3.0 mmol) was added. Reflux was resumed for
/
11
12
Ã
Ã
/
C
12
13
/
/
C
1 h. The reaction was cooled to 25 8C and the solution
was filtered through a pad of celite (10 mm) to remove
the excess Zn dust. The celite was washed with THF
Atoms are labeled in agreement with Fig. 1.
a
The numbers in parentheses are the estimated standard deviations
in the last significant digit.
(3ꢃ15 ml). THF was removed from the combined
/
extracts by rotary evaporation. The crude product was
recrystallized from a minimum amount of hot methanol
to yield 13 and 7 (0.754 g, 1.73 mmol, 87%, 1.8:1 Z to E
ratio). These compounds were determined to be iden-
operating at 75.48 MHz. All spectra were referenced to
the residual proton or carbon signals of the respective
deuterated solvents. Infrared (IR) spectra were obtained
1
on a Mattson Genesis II FTIR or PerkinÁ/Elmer 1600
tical to previously reported material by H-NMR
comparison [8,13].
Series FTIR. All elemental analyses were performed by

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
129
Table 3
Results from silane cleavage reactions
Silane
Cl MeSiH
Equivalents of silane
Heat source
Yield (cycloadduct, %)
2
Four equivalents
Four equivalents
Six equivalents
Six equivalents
Six equivalents
Six equivalents
Six equivalents
hn
D
hn
D
D
D
D
19
44
62
85
ClMe
Ph SiH
Et SiH
2
SiH
22
14
2
2
3
No Cleavage
4
adducts
.3. General procedure for synthesis of DielsÁ
/
Alder
the transferring end) to remove excess Zn dust. The Zn
residue was washed with additional THF (15 ml) and the
washings were transferred through the cannula. Five
equivalents of the dienophile were added and the
solution was refluxed for the specified amount of time.
The solution was cooled to 25 8C and the solvent was
removed by rotary evaporation. The residue was chro-
matographed on silica using EtOAc. In every case
except one, the cobalt diene and the cobalt cycloadduct
came off as a single band. All attempts to separate the
two compounds are failed. All reported yields where
residual dienyl complex (13) was present are percentages
of the isolated products based on the NMR ratio of the
Cobalt(II) acetate hexahydrate (0.489 g, 2.0 mmol)
and dimethylgloxime (0.464 g, 4.0 mmol) were dissolved
in degassed THF (50 ml). Pyridine (0.475 g, 6.0 mmol)
was added followed by Zn dust (0.654 g, 10.0 mmol).
This mixture was refluxed for 15 min. The solution was
cooled briefly to stop the reflux and 3-penten-1-yne
(
0.200 g, 3.0 mmol) was added. Reflux was resumed for
h. The reaction pot was cooled to stop the reflux and
the warm solution was transferred to another flask via
1
cannula fitted with a glass/fiberglass filter apparatus (on

130
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
methyl doublets for the cobalt diene and the cobalt
cycloadduct.
ended needle filtration prior to the addition of dimethyl
fumarate (1.441 g, 10.0 mmol) and 17-h reflux produced
1
3 (0.147 g, 0.339 mmol, 17%) and 14 (0.752 g, 1.30
1
4
3
.3.1. Synthesis of (1,2-dicarboxylic acid dimethyl ester-
-methyl-4-cyclohexen-5-
mmol, 65%). The product (14) was identical by H-
NMR spectroscopic comparison to the material isolated
in method A.
yl)pyridinebis(dimethylglyoximato)cobalt(III) (14)
This complex was synthesized using three different
methods.
4
3
.3.2. Synthesis of trans- and cis-4-methyl-2-phenyl-
a,4,7,7a-tetrahydro-1,3-dioxoisoindol-4-
4
.3.1.1. Method A (Zn was removed by chromatographic
purification of 13 and 7 prior to cycloaddition). Co diene
mixture, 13 and 7 (1.010 g, 2.32 mmol) and dimethyl
fumarate (0.670 g, 4.645 mmol) were dissolved in
degassed THF (50 ml) and the solution was refluxed
for 17 h. The solvent was removed by rotary evapora-
tion. The residue was chromatographed on silica gel
ylpyridinebis(dimethylglyoximato)cobalt(III) (15 and
6)
This complex, like 14, was synthesized with and
without dienyl complex purification prior to the
DielsÁAlder.
1
/
(
EtOAc). The product was collected as a yellow solid
14) (1.195 g, 2.064 mmol, 90%). m.p.: 195Á197 8C dec.
5.6, 1.3 Hz, 2H), 7.69
7.0, 1.2 Hz, 1H), 7.28 (d, Jꢀ6.5 Hz, 2H), 5.14
dd, Jꢀ5.0, 1.9 Hz, 1H), 3.64 (s, 3H), 3.59 (s, 3H), 2.83
dd, Jꢀ12.1, 6.0 Hz, 1H), 2.70 (dd, Jꢀ11.5, 4.9 Hz,
H), 2.64Á2.58 (m, 1H), 2.48 (dd, Jꢀ17.2, 4.9 Hz, 1H),
.08 (s, 6H), 2.05 (s, 6H), 1.87Á1.78 (m, 1H), 0.64 (d,
6.9 Hz, 3H) ppm. C-NMR (CDCl ): 176.2, 174.7,
(
1
/
H-NMR (CDCl ): 8.58 (dd, Jꢀ
/
4.3.2.1. Method A (Zn removal prior to DielsÁAlder by
/
3
(
(
(
tt, Jꢀ
/
/
chromatographic purification of 13 and 7). Purified
cobalt dienes, 13 and 7 (0.211 g, 0.485 mmol) and N-
phenyl maleimide (0.125 g, 0.725 mmol) were dissolved
in degassed THF (15 ml) in a sealed tube. The tube was
/
/
/
1
2
Jꢀ
/
/
/
sealed under N and heated to 100 8C for 24 h. The tube
2
1
3
/
3
was cooled to 25 8C and the solvent was removed by
rotary evaporation. The residue was purified by column
chromatography on silica gel (1.5 column volumes of 1:1
1
3
1
50.1, 150.0, 149.6, 137.5, 129.4, 125.1, 51.7, 51.3, 46.0,
9.4, 35.4, 32.8, 17.8, 12.1 ppm. IR (NaCl): 2946, 1733,
559, 1442, 1232, 1166, 1086 cm . Anal. Calcd. for
ꢂ1
pent: Et O followed by 1.5 column volumes of EtOAc).
2
C H CoN O : C, 49.66; H, 6.08. Found: C, 49.57; H,
5
2
4
35
5
The products (15 and 16, 9:1) were collected as a yellow
powder (0.248 g, 0.407 mmol, 89%) after solvent
6
.05%.
removal under vacuum. m.p.: 185Á187 8C dec. NMR
/
4
.3.1.2. Method B (no chromatographic purification of
3 and 7 and no Zn removal prior to cycloaddition).
1
data for major diastereomer (15): H-NMR (CDCl ):
3
1
8
1
7
.60 (dd, Jꢀ
H), 7.41 (t, Jꢀ
.28 (t, Jꢀ7.0 Hz, 2H), 7.20 (d, Jꢀ
6.5, 2.4 Hz, 1H), 3.04Á3.02 (m, 1H), 2.80 (dt,
8.9, 2.5 Hz, 1H), 2.73 (dd, Jꢀ8.9, 2.5 Hz, 1H), 2.65
dd, Jꢀ16.7, 3.3 Hz, 1H), 2.34 (ddd, Jꢀ16.7, 6.8, 2.4
Hz, 1H), 1.98 (s, 12H), 1.04 (d, Jꢀ
/
6.8, 1.4 Hz, 2H), 7.69 (tt, Jꢀ
7.6 Hz, 2H), 7.33 (t, Jꢀ7.4 Hz, 1H),
7.4 Hz, 2H), 5.49
/
7.6, 1.4 Hz,
Initial optimization of the general procedure outlined
above was completed using dimethyl fumarate. The
variations were as follows: the above procedure was
used without removal of the Zn from the reaction flask
prior to the addition of the dimethyl fumarate (0.432 g,
/
/
/
/
(dd, Jꢀ
/
/
Jꢀ
(
/
/
/
/
3
2
.0 mmol). This reaction pot was then refluxed for 1 or
4 h. The reaction was cooled to 25 8C and the solution
/7.1 Hz, 3H) ppm.
13
C-NMR (CDCl ): 179.7, 178.9, 150.3, 150.1, 149.8,
37.6, 132.4, 131.7, 128.8, 128.4, 126.4, 125.2, 46.4, 42.2,
3.8, 29.7, 20.5, 12.0 ppm. The minor diastereomer (16)
3
was filtered through a pad of celite (10 mm) to remove
the excess Zn dust. The celite was washed with THF
(3ꢃ15 ml). The THF washings were combined and the
solvent was removed by rotary evaporation. The residue
1
3
/
exhibited a methyl doublet, Jꢀ7.1 Hz, at d 1.35 ppm.
/
ꢂ1
IR (CDCl ): 3152, 2260, 1705, 947 cm . Anal. Calcd.
for C H CoN O : C, 55.26; H, 5.47. Found: C, 54.79;
3
was purified by column chromatography on silica gel
EtOAc) to produce 13 (0.357 g, 0.820 mmol, 41%) and
2
8
33
6
6
(
H, 5.59%.
1
1
(
4 (0.336 g, 0.581 mmol, 29%) as a yellow solid from the
-h reaction and 13 (0.311 g, 0.715 mmol, 36%) and 14
0.717 g, 1.24 mmol, 62%) as a yellow solid from the 24-
1
h reaction. The product (14) was identical by H-NMR
spectroscopic comparison to the material isolated in
method A.
4.3.2.2. Method B (Zn removed using a filter fitted
double-ended needle). The general procedure using Zn
removal, N-phenyl maleimide (1.732 g, 10.0 mmol) and
17-h reflux produced 13 (0.0992 g, 0.228 mmol, 11%)
and 15:16 (9:1) (0.880 g, 1.45 mmol, 72%). The product
1
was identical by H-NMR spectroscopic comparison to
the material isolated and characterized in method A.
4
.3.1.3. Method C (Zn removal by double-ended needle
filtration prior to cycloaddition). The general procedure
outlined above using Zn removal by glass frit double-

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
131
4.3.3. Synthesis of tran- and cis-1,3,3a,4,7,7a-
hexahydro-7-methyl-1,3-dioxo-isobenzofuran-5-
experiment, the cleavage reaction described in part (iv)
was coupled with the cobalt regeneration procedure
described below using 0.195 g (0.336 mmol) of 14 as
starting material. The desired cycloadduct, 26, was
isolated in 75% yield (53.6 mg, 0.253 mmol) and 99%
of cobalt was regenerated as 13 and 7 (144 mg, 0.331
mmol).
ylpyridinebis(dimethylglyoximato)cobalt(III) (17 and
1
8)
The general procedure using zinc filtration, maleic
anhydride (0.981 g, 10.0 mmol) and 17-h reflux pro-
duced only 17 and 18 (3:1) (0.656 g, 1.23 mmol, 62%)
1
identical to previously reported material by H-NMR
comparison [8].
4.4.2. Regeneration of cobalt salt mixture to 13 and 7
The bluish-green solid (0.226 g) obtained from the
4.3.4. Synthesis of cis- and trans-1-acetyl-2-methyl-3-
cyclohexene-4-
optimized Cl MeSiH cleavage of 14 (0.336 mmol)
2
described above was suspended in degassed THF (15
ml). Dimethylglyoxime (0.080 g, 0.69 mmol), pyridine
(0.082 g, 1.03 mmol), 3-penten-1-yne (0.046 g, 0.69
mmol) and Zn dust (0.113 g, 1.73 mmol) were added in
this order. The order of addition for these reactants is
important. The suspension was refluxed for 1 h. During
this time, the reaction pot turned from a bluish-green
suspension to a yellow solution. The solution was cooled
to 25 8C and filtered through a pad of celite (10 mm) to
remove the excess Zn dust. The celite was washed with
ylpyridinebis(dimethylglyoximato)cobalt(III) (20 and
1
9)
The general procedure using zinc filtration, methyl
vinyl ketone (810 ml, 0.700 g, 10.0 mmol) and 17-h reflux
produced 13 (0.279 g, 0.640 mmol, 32%) and 20 and 19
(
3.3:1) (0.462 g, 0.913 mmol, 46%). Compounds 20 and
9 were determined to be identical to previously
1
reported material by H-NMR comparison [8].
1
4
bonds with silanes
.4. General procedure for the cleavage of cobaltÃ
/
carbon
THF (3ꢃ15 ml). The THF washings were combined
with the filtered reaction mixture and the solvent was
removed by rotary evaporation. The residue was dis-
/
A typical procedure for the cleavage of 14 with a
silane is as follows: complex 14 (0.150 g, 0.258 mmol)
was dissolved in degassed THF (12 ml) and six
equivalents of silane were added. The solution was
solved in CHCl and washed with 2ꢃ
/
10 ml distilled
3
water, 2ꢃ
2ꢃ
combined and back-extracted with 2ꢃ
/
10 ml 1.2-M HCl, 1ꢃ
/
10 ml NaHCO and
3
/10 ml 1-M NaOH. The aqueous extracts were
/
20 ml CHCl3.
refluxed under N for 1 h. During this time, the solution
2
All CHCl3 extracts were combined and dried with
MgSO . MgSO was removed by filtration and the
turned from yellow to bluish-green due to the formation
of a bluish-green precipitate. The reaction pot was
cooled to 25 8C and the solvent was evaporated under
4
4
solvent was removed by rotary evaporation. The pro-
duct was dried in vacuo to produce 13 and 7 as a yellow
powder (1:1 E to Z ratio) (0.144 g, 0.331 mmol, 99%).
The product was determined to be identical to pre-
a stream of N . The residue was extracted with 3ꢃ3 ml
/
2
of pentane to isolate the cleaved cycloadduct (26) as a
clear, pale yellow oil. The isolated product was deter-
mined to be identical to previously reported material by
1
viously reported 13 and 7 by H-NMR comparison [8].
1
H-NMR comparison [17]. The remaining solid was
washed with CHCl leaving a bluish-green precipitate.
4.4.3. Cleavage of 14 with ClMe SiH
2
The cleavage procedure outlined above but using
ClMe SiH (180 ml, 153.4 mg, 1.62 mmol) produced 26
(12.0 mg, 0.057 mmol, 22%).
3
The solid was analyzed by MS and determined to be a
mixture
2
of
salts
ꢁ
consisting
(pyrCoCl )] (pyrH ) and (pyr) CoCl . The exact
primarily
of
ꢂ
[
composition of the salt was not able to be determined.
3
2
2
4.4.4. Cleavage of 14 with Ph SiH
2
The cleavage procedure outlined above but using
Ph SiH (300 ml, 297.9 mg, 1.62 mmol) produced 26 (7.8
mg, 0.037 mmol, 14%).
4
.4.1. Cleavage of 14 with Cl MeSiH and regeneration of
2
2
2
1
3 and 7
Initial optimization of the general procedure above
was performed using Cl MeSiH. Variations were as
2
4.5. Reaction cycle procedure
follows: (i) Cl MeSiH (165 ml, 121.6 mg, 1.58 mmol) and
2
1
-h irradiation with light (flood lamp) produced 26 (10.4
Cobalt(II) acetate hexahydrate (0.489 g, 2.0 mmol)
and dimethylgloxime (0.464 g, 4.0 mmol) were dissolved
in degassed THF (50 ml). Pyridine (0.475 g, 6.0 mmol)
was added followed by Zn dust (0.654 g, 10.0 mmol).
This mixture was refluxed for 15 min. The solution was
cooled briefly to stop the reflux and 3-penten-1-yne
(0.200 g, 3.0 mmol) was added. Reflux was resumed for
1 h. The reaction pot was cooled to stop reflux and the
mg, 0.049 mmol, 19%); (ii) Cl MeSiH (110 ml, 121.6 mg,
2
1
.06 mmol) and 1-h reflux produced 26 (24.0 mg, 0.113
mmol, 44%); (iii) Cl MeSiH (165 ml, 182.3 mg, 1.58
2
mmol) and 1-h irradiation with light (flood lamp)
produced 26 (33.9 mg, 0.160 mmol, 62%); (iv) Cl MeSiH
2
(
2
165 ml, 182.3 mg, 1.58 mmol) and 1-h reflux produced
6 (46.8 mg, 0.221 mmol, 86%); (v) in a separate

132
K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á133
/
warm solution was transferred via cannula fitted with a
glass/fiberglass filter apparatus (on the transferring end)
to remove excess Zn dust. The Zn residue was washed
with additional THF (15 ml) and transferred through
the cannula. Five equivalents of dienophile were added
and refluxed for 17 h. The solution was cooled to 25 8C
F)ꢀ
2T(MoÁ
based on F)ꢀ
0.157 for all 5111 reflections} using counter-weighted
full-matrix least-squares techniques and a structural
model which incorporated anisotropic thermal para-
meters for all non-hydrogen atoms (except disordered
/
0.063 for 2896 independent reflections having
2
2
/
K )B
/
48.38 and F ꢀ
/
2s(F )} {R (unweighted,
1
a
2
/
0.131 and wR (weighted, based on F )ꢀ
/
2
and Cl MeSiH (1.38 g, 12.0 mmol) was added. The
2
reaction pot was refluxed for 1 h. During this time, the
solution turned from yellow to green due to the
formation of a bluish-green precipitate. The reaction
pot was cooled to 25 8C and the solid precipitate was
filtered out and washed with a minimum amount of
pentane (10 ml). The precipitate was set aside for dienyl
complex (13 and 7) regeneration. The reaction pot
solution and the pentane wash were combined and the
solvent was removed by rotary evaporation. The residue
solvent atoms C Ã
/
C37 which were refined as isotropic
and H25B
25A
2
7
atoms). Hydrogen atoms H3O, H4O, H , H
9
were located from a difference Fourier map and refined
as independent isotropic atoms. The methyl groups (C5,
C , C , C , C , C , C and their hydrogen atoms) were
6
7
8
16
18
19
3
refined as rigid rotors with idealized sp -hybridized
˚
H bond length of 0.97 A. The
geometry and a CÃ
/
remaining hydrogen atoms were included in the struc-
ture factor calculations as idealized atoms (assuming
2
3
was extracted with 3ꢃ
/
5 ml of pentane. The residue is
sp - or sp -hybridization of the carbon atoms and CÃ
/H
˚
extracted to remove the cleaved cycloadduct from the
large excess of dimethyl fumarate which has a limited
solubility in pentane. The pentane extracts were filtered
through a cotton plug and evaporated under a stream of
bond lengths of 0.94Á0.98 A) ‘‘riding’’ on their respec-
/
tive carbon atoms. The isotropic thermal parameters for
H3O, H4O and H refined to final U values of 0.10(3),
9
iso
2
˚
0.10(3) and 0.01(1) A , respectively. The isotropic
thermal parameters of the remaining hydrogen atoms
were fixed at values 1.2 (non-methyl) or 1.5 (methyl)
times the equivalent isotropic thermal parameters of the
carbon atoms to which they are covalently bonded.
The crystal lattice contains three regions occupied by
dichloroethane or pentane solvent molecules of crystal-
lization. The first (half) dichloroethane molecule is
ordered and sits on a crystallographic inversion center
N . The residue from the pentane washes was chroma-
2
tographed using radial chromatography on a silica gel
plate eluted with 8:1 pentane:Et O. The solvent was
2
removed under a stream of N2 and the cleaved
cycloadduct was isolated. The bluish-green precipitate
was washed with 2ꢃ20 ml of CHCl to remove any
/
3
excess of dienophile and then regenerated into 13 and 7
using the method previously discussed. The optimized
reaction cycle using dimethyl fumarate (1.44 g, 10.0
mmol) produced the cleaved cycloadduct 26 (96 mg,
at (0, ꢂ
Cl , C , H
25A
/
1/2, 1/2) in the unit cell; it is specified by atoms
and H25B. The second dichloroethane
1
25
0
.453 mmol, 23%) and recovered the cobalt as the diene
3 and 7 (1:1 E to Z ratio) (0.246 g, 0.565 mmol, 28%)
molecule is disordered about the twofold axis at (0, y, 1/
4) in the unit cell and specified by half-occupancy
chlorine and carbon atoms designated as Cl , Cl , C
26
1
identical by H-NMR spectroscopic comparison to
material reported above.
1
2
3
and C27 (C27 lies on the twofold axis). Hydrogen atoms
were not included in the structural model for this or the
third solvent molecule. The identity of the third
disordered solvent molecule could not be unambigu-
ously established as dichloroethane or pentane. A
difference Fourier map of this area contained several
peaks of electron density above the background level
without meaningful chemical connectivity. Half-occu-
pancy carbon atoms were placed at 10 of these sites and
refined as isotropic atoms.
4.6. X-ray crystallographic characterization of
Co(NC H )(N C (CH ) O H) (C (CH )H (CO
2-
5
5
2
2
3 2
2
2
6
3
6
CH ) )×
/
1.5Cl C H (14)
2 2 4
3
2
Single crystals of Co(NC H )(N C (CH ) O H) -
2
5
5
2
2
3 2
2
(C (CH )H (CO CH ) )×
monoclinic, space group C2/c-
/
1.5 Cl C H are, at 218(2) K,
6
3
6
2
3 2 2 2 4
6
/
C2 (No. 15) with aꢀ
/
h
˚
˚
˚
3
5.207(5) A, bꢀ
/
9.163(1) A, cꢀ
/
23.743(3) A, bꢀ
8 {dcalcd 1.384 g
0.771 mm }. A total of 6061
independent absorption-corrected reflections having
T(MoÁK )B48.38 (the equivalent of 0.7 limiting
CuÁ spheres) were collected on a computer-con-
trolled Bruker-AXS P4 autodiffractometer using v
scans and graphite-monochromated MoÁK radiation.
Of these, 5111 were unique and gave Rint
/
3
˚
1
14.19(1)8, Vꢀ6987(2) A and Zꢀ
/
/
ꢀ
/
All calculations were performed using the SHELXTL-PC
(Version 6.12) interactive software package (G. Shel-
drick, Bruker AXS, Madison, WI).
ꢂ3
ꢂ1
cm , m (MoÁ
/
K )ꢀ
/
a
a
2
/
/
a
/
K
a
5. Supplementary material
/
a
ꢀ/0.048. The
Tables giving details of the X-ray structure determi-
nations, atomic coordinates and isotropic thermal
parameters, bond lengths and angles and anisotropic
displacement parameters for 14 have been deposited at
the Cambridge Crystallographic Data Centre (CCDC
structure was solved using ‘‘Direct Methods’’ techniques
with the Bruker AXS SHELXTL-PC (Version 6.12) soft-
ware package. The resulting structural parameters have
been refined to convergence {R (unweighted, based on
1

K.A. Pickin et al. / Journal of Organometallic Chemistry 681 (2003) 120Á
/
133
133
A.L. Rheingold, Adv. Cycloaddition 4 (1997) 149
For a review of our earlier work in this area see:
No. 212517). Copies of this information may be
obtained free of charge from The Director, CCDC, 12
[
[
4] J.J. Chapman, C.S. Day, M.E. Welker, Eur. J. Org. Chem. (2001)
273.
Union Road, Cambridge CB2 1EZ, UK (Fax: ꢁ
/
44-
2
1223-336033; e-mail: deposit@ccdc.cam.ac.uk or www:
http://www.ccdc.cam.ac.uk).
5] P.Y. Renard, J.Y. Lallemand, Bull. Soc. Chim. Fr. 133 (1996)
143.
[
[
[
6] H.L. Stokes, M.E. Welker, Organometallics 15 (1996) 2624.
7] G.S. Bates, M.D. Fryzuk, C. Stone, Can. J. Chem. 65 (1987) 2612.
8] M.W. Wright, T.L. Smalley, Jr., M.E. Welker, A.L. Rheingold, J.
Am. Chem. Soc. 116 (1994) 6777.
Acknowledgements
[9] D. Dodd, M.D. Johnson, J. Organomet. Chem. 52 (1973) 1.
We thank the National Science Foundation for their
support of this work (CHE-0104083) and the NMR and
X-ray instrumentation used to characterize the com-
pounds reported here.
[10] P.F. Roussi, D.A. Widdowson, J. Chem. Soc. Perkin I (1982)
1025.
[
[
11] K.A. Pickin, M.E. Welker, Organometallics 19 (2000) 3455.
12] B.L. Hayes, T.A. Adams, K.A. Pickin, C.S. Day, M.E. Welker,
Organometallics 19 (2000) 2730.
[
[
13] M.W. Wright, M.E. Welker, J. Org. Chem. 61 (1996) 133.
14] G. Eglinton, M.C. Whiting, J. Chem. Soc. (1950) 3650.
References
[15] S. Torii, T. Inokuchi, K. Kawai, Bull. Chem. Soc. Jpn. 52 (1979)
61.
8
1
[
16] There is a typographical error in the H-NMR experimental
section for compounds 57 and 58 of Ref. [8]. The methyl doublet
at d 1.37 ppm corresponds to the minor diastereomer and the
[
[
[
1] M.E. Welker, Curr. Org. Chem. 5 (2001) 785
For a review see:
2] T.L. Smalley, Jr., M.W. Wright, S.A. Garmon, M.E. Welker,
A.L. Rheingold, Organometallics 12 (1993) 998.
3] M.E. Welker, M.W. Wright, H.L. Stokes, B.M. Richardson, T.A.
Adams, T.L. Smalley, S.P. Vaughn, G.J. Lohr, L. Liable-Sands,
methyl doublet at
diastereomer.
[17] M. Kakushima, Can. J. Chem. 57 (1979) 2564.
d 0.99 ppm corresponds to the major