Synthesis of Dicarbonyl(η4-tricarbonylcobaltacyclopentadiene)cobalt Complexes from Co2(CO)8. A General Route to Intermediates in Cobalt Carbonyl Mediated Alkyne Trimerization

Article abstract of DOI:10.1021/om980544y

We have shown that cobaltacyclopentadiene complexes, derived from cobalt carbonyl and two alkyne molecules and long believed to be intermediates in cyclotrimerization as well as in the formation of the (three-alkyne-derived) flyover complexes, can be isolated in selected cases if the mono(alkyne)-Co₂(CO)₆ complexes are treated at room temperature with just 1 equiv of alkyne in the presence of trimethylamine N-oxide. The complex (PhC=CHCMe= CMe)Co₂(CO)₅ has been characterized by X-ray crystallography.

Full text of DOI:10.1021/om980544y

Organometallics 1999, 18, 197-205
197
Syn th esis of
Dica r bon yl(η4-tr ica r bon ylcoba lta cyclop en ta d ien e)coba lt
Com p lexes fr om Co2(CO)8. A Gen er a l Rou te to
In ter m ed ia tes in Coba lt Ca r bon yl Med ia ted Alk yn e
Tr im er iza tion
R. J ohn Baxter, Graham R. Knox,* Peter L. Pauson, and Mark D. Spicer*
Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street,
Glasgow G1 1XL, Scotland
Received J une 29, 1998
We have shown that cobaltacyclopentadiene complexes, derived from cobalt carbonyl and
two alkyne molecules and long believed to be intermediates in cyclotrimerization as well as
in the formation of the (three-alkyne-derived) “flyover” complexes, can be isolated in selected
cases if the mono(alkyne)-Co2(CO)6 complexes are treated at room temperature with just 1
equiv of alkyne in the presence of trimethylamine N-oxide. The complex (PhCdCHCMed
CMe)Co2(CO)5 has been characterized by X-ray crystallography.
In tr od u ction
major (66%) and stable product [together with 1; R1, R2
) (CH2)6] and shown by X-ray crystallography to have
the structure 2a , reminiscent of a class of alkyne-derived
ferracyclopentadienes, it was not so much viewed as
filling a gap but rather as a very special case.
Two more recent examples, the bicyclic complex 2b,
obtained from a substituted N,N-dipropargylmethyl-
amine,4 and the bis(pentafluorothio) derivative 2c from
HC2SF5,5 albeit alkynes of very different reactivity
compared to cyclooctyne, could not escape being labeled
special cases.
The reaction of octacarbonyldicobalt with almost any
alkyne (R1C2R2) yields 1:1 complexes, the well-known
dicobaltatetrahedranes 1 (see Scheme 1). They are now
widely used in a variety of ways as reagents in organic
synthesis and have also long been recognized as inter-
mediates in and catalysts for alkyne cyclotrimerization
(although some reactions of complexes 1 with additional
alkyne are limited by steric and possibly electronic
effects).
The most thorough and comprehensive study of these
compounds and their reactions appeared in an early
paper (1961) by Kru¨erke and Hu¨bel.1 Inter alia they
described the isolation of the complexes Co2(CO)4-
(C6R1R2R3 R4 ) from reaction of the mono(alkyne) com-
Three further complexes of a similar type have been
reported, but all result from other reaction paths:
Yamazaki and co-workers6 obtained the triphenylphos-
phine-substituted analogue 6 as a product of the reac-
tion of octacarbonyldicobalt with the mononuclear co-
baltacyclopentadiene 5a , a process which must involve
complex ligand transfers and which did not occur with
analogues including the symmetrical isomer 5b. Moldes
2
2
pounds 1 with two further molecules of an alkyne
(R3C2R4), later shown2 to have the “flyover” structure 3
(see following paper). They provided evidence however,
that these were stable end products of a competing path,
not intermediates in the cyclotrimerizations to arenes
[4a and/or (rarely) 4b]. Rejecting the unlikely possibility
of simultaneous incorporation of two alkynes into
complex 1, Kru¨erke and Hu¨bel searched for intermedi-
ates containing two alkyne units but had to report that
“All attempts, starting from these complexes [Co2(CO)6-
(RC2R′) and Co4(CO)10(RC2R′)], to isolate intermediate
compounds with a system [derived] from two reactive
alkyne ligands were without success.” It seemed most
likely, therefore, that whatever 1:2 complexes might be
involved were too short-lived for ready detection or
isolation.
and co-workers7 have obtained the simple diphenyl-
substituted complex 2d (10% yield) from a mixture of
ethynylbenzene, Fe2(CO)9, [Co(CO)4]-, and trifluoroace-
When, some 17 years later,3 the highly strained
cyclooctyne was shown to give a 1:2 complex as the
(4) Predieri, G.; Tiripicchio, A.; Tiripicchio Camellini, M.; Costa, M.;
Sappa, E. J . Organomet. Chem. 1992, 423, 129.
(5) (a) Henkel, T.; Klauck, A.; Seppelt, K. J . Organomet. Chem. 1995,
501, 1. (b) Wessel, J .; Hartl, H.; Seppelt, K. Chem. Ber. 1986, 119, 453.
(6) Yamazaki, H.; Yasufuku, K.; Wakatsuki, Y. Organometallics
1983, 2, 726.
(7) Moldes, I.; Papworth, T.; Ros, J .; Alvarez-Larena, A.; Piniella,
J . F. J . Organomet. Chem. 1995, 489, C65.
* Corresponding
m.d.spicer@strath.ac.uk.
(1) Kru¨erke, U.; Hu¨bel, W. Chem. Ber. 1961, 94, 2829.
(2) Mills, O. S.; Robinson, G. Proc. Chem. Soc. 1964, 137.
(3) Bennett, M. A.; Donaldson, P. B. Inorg. Chem. 1978, 17, 1995.
authors.
E-mail:
g.r.knox@strath.ac.uk;
10.1021/om980544y CCC: $18.00 © 1999 American Chemical Society
Publication on Web 12/23/1998

198 Organometallics, Vol. 18, No. 2, 1999
Baxter et al.
Sch em e 1
4b
2
4a
tic acid, apparently by a process involving the anionic
iron complex 7 as a source of at least one of the “alkyne”
(alkyne) complexes, 1, react further with the same or
different alkynes to give flyover complexes, 3, or arenes,
4, could be expected if the third alkyne molecule reacts
much faster than the second under the chosen reaction
conditions. The isolation of complexes 2 in the unusual
cases mentioned in the Introduction led us to assume
that similar structures are probably involved as tran-
sient intermediates in the above reactions. Hence,
addition of the second alkyne would be accompanied by
loss of just one molecule of carbon monoxide. Control of
such a process would be aided by making available no
more than 1 mole of alkyne/mol of complex and by using
the lowest possible reaction temperature in order to slow
the next step. Moreover, removal of one CO ligand by
means other than alkyne displacement might well assist
the first step. Our experience, particularly with a wide
range of reactions of the type discovered by Khand et
al.,9 shows that amine N-oxides are not only well suited
to this purpose but vary substantially in reactivity so
that the reaction temperature and rate can be adjusted
by a suitable choice of these oxidants.
units and hence a metal exchange. These authors
conclude “This shows that the formation of [2] is possible
only in unusual circumstances (reaction conditions or
particular alkynes).” Very recently8 Giordano et al.
reported a cobaltacyclopentadiene complex of this type,
derived from two molecules of HC2CMe2OH, as having
been formed from the corresponding flyover by a novel
decomposition, apparently promoted by the silica of a
TLC plate. This interesting result requires confirmation.
We now describe a more directed and more widely
applicable, though still limited, route to such complexes.
We have therefore treated a variety of (alkyne)-
hexacarbonyldicobalt complexes with either the same
Resu lts a n d Discu ssion
(9) Khand, I. U.; Knox, G. R.; Pauson, P. L.; Watts, W. E. Chem.
Commun. 1971, 36. This reaction is widely referred to as the Pauson-
Khand reaction. For reviews see: (a) Schore, N. E. Org. React. 1991,
40, 1. (b) Schore, N. E. In Comprehensive Organometallic Chemistry;
Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon: London,
1995; Vol. 12, p 703.
The failure, in most cases, to detect an intermediate
(e.g., 2), derived from two alkynes, when the mono-
(8) Giordano, R.; Sappa, E.; Predieri, G. Inorg. Chim. Acta 1995,
228, 139. These authors give the formula and analysis as Co2C15H10O7.

(Cobaltacyclopentadiene)cobalt Complexes
Organometallics, Vol. 18, No. 2, 1999 199
Ta ble 1. Rea ction s of (R1C2R2)Co2(CO)6, 1, w ith Alk yn es R3C2R4
starting materials
products
complex 1
alkyne R3C2R4
R4
mg mmol
215 2.11
2570 24.6
complex 1
complex 2
complex 3a
mg %
R1
R2
mg mmol
818 2.11
9550 24.6
R3
mgb
%
no.
2d
2d
2e
2f
2g
2h
2i
mg
%
c
Ph
Ph
COOMe
COOMe
COOMe
COOEt
H
H
H
H
Ph
Ph
458
56
302 31 (70)d
3520 31 (70)
5350
373
377
1350
983
56
63e
45
70
76
77
77
COOMe
H
593
839
1.38 COOMe COOMe
197
190
490
364
440
120
1.38
2.27
5.0
3.25
4.67 1370
1.68 478
373 24e (65) 117
2.27
5.0
H
COOMe
COOMe Me
3.25 COOEt Me
164 17 (31)
31f
Me
Me
H
1920
1290
1780
621
70
125
3 (10)
8 (33)
62 + 14g
55 + 18 3 + 1h
n-C5H11
CH2OMe
4.67
1.68
H
H
n-C5H11
CH2OMe
357 17 (74)
H
2j
+2k
2o
53
13
44
23
8 (35)
2 (8)
2 (8)
3
COOMe
COOMe
Ph
COOMe
1560
721
3.49 COOMe Ph
560
172
3.49 1150
1.68
74
223
9i
6
1.68
H
Ph
122 (33)
17 (5) 2d
+2l
345 41 (49)
181 14 (37)
390 18 (40)
COOMe
Ph
Ph
COOMe
Me
COOMe 1220
2.84
H
COOMe
Me
4.32 COOMe COOMe
240
300
610
240
640
2.84
5.62 1680 (76)
4.32 1020 (111) 61 (6) 2l
4.42 1260 (30) 67 (2) 2e
6.20 1610 (120) 76 (5) 2n
756 (42)
62 (4) 2m
54 (4) 2n
H
H
2180
1670
5.62 Me
43
72
2 (5)
3 (9)
160
COOMe 1890
Me 2120
4.42 Me
6.20
Me
Ph
H
373 12 (50)
a
b
For details see the following paper.19 Yield recovered and yield (in parentheses) from alkyne exchange. c Yield based on unrecovered
complex 1 in parentheses. Arene also isolated (978 mg, 13%).a e Use of NMO gives 1 (74%) + 2b (19%). f Impure; contains arene. Two
d
g
isomers + arene (4 mg).a Two isomers. +arenes (167 mg, 10%).a
h
i
or a different alkyne together with an N-oxide. By this
approach, we have indeed obtained all the cobaltacy-
clopentadienes, 2, listed in Table 1. Of these, complex
2d has previously been reported.7 The yields, which we
have not attempted to optimize, are low to modest, and
in many cases (listed in the Experimental Section) the
expected cobaltacyclopentadienes were detected either
in only trace amounts or not at all. Our concern,
however, in the present study was to show that the
straightforward route does yield such complexes in
“nonexceptional” cases and that they do react further
to yield flyover complexes (3) and behave as catalysts
of alkyne cyclotrimerization to arenes. The latter aspects
are discussed in the following paper.19
Our initial study was devoted to the case of hexacar-
bonyl(dimethyl acetylenedicarboxylate)dicobalt reacting
with the same alkyne to yield complex 2e. This mixture
failed to react at room temperature when trimethyl-
amine N-oxide was added as its dihydrate, in contrast
to the successful use of the much more active anhydrous
form. Table 1 also shows that the latter could be
replaced by anhydrous N-methylmorpholine N-oxide,
(“NMO”), but with a slight lowering of yield. Use of
tetrahydrofuran as sole solvent, in place of a 1:1 mixture
with dichloromethane, gave a more complex mixture of
products, while, in the synthesis of complex 2d , use of
acetone as solvent lowered the yield to 17%. The use of
Me3NO/THF/CH2Cl2 was therefore adopted for all other
examples studied. As shown in Table 1, the long-
known10 alkyne exchange reaction plays a significant
but never a dominant part. In some cases, recovery of
the alkyne-exchanged hexacarbonyldicobalt complex is
observed, while, in two cases, these react further to yield
cobaltacyclopentadienes in which the initially coordi-
nated alkyne is not present. Nevertheless, the fact that
it occurs at all under the mild conditions used may
suggest that it too is promoted by the removal of a CO
ligand.
No simple pattern of behavior emerges from compari-
son of the successful and unsuccessful reactions. Al-
though a majority of the former involve alkynes with
ester and/or phenyl groups, they also include, e.g.,
1-heptyne and if the 77% recovery of starting complex
is considered, the 17% conversion becomes a 74% yield
based on complex actually consumed. Moreover, among
internal alkynes, whereas the diester complex and that
of methyl 2-butynoate each reacted with the corre-
sponding alkyne (giving 2e and 2g, respectively), methyl
phenylpropiolate gave low yields and both diphenyl-
ethyne (tolan) and 2-butyne, long known to give stable
mononuclear cyclopentadienylcobaltacyclopentadienes,11
reacted with their hexacarbonyldicobalt complexes giv-
ing the corresponding arenes, without allowing detection
of intermediates under our conditions. An attempt to
make the di-tert-butyl-substituted complex (2; R1 ) R4
) But; R2 ) R3 ) H) gave 2,5-di-tert-butylcyclopenta-
dienone, a product previously obtained1 from mercury-
cobalt carbonyl with 3,3-dimethyl-1-butyne to yield
“Co2Hg(CO)6(Me3CC2H)4”, followed by pyrolysis or bro-
mine oxidation, and probably formed in the present
work by degradation of the expected compound of type
2. Despite the modest yield, this route to 2,5-di-tert-
butylcyclopentadienone is more convenient than the
known alternatives.1,12
Two of the “mixed” complexes included in Table 1
have been approached by treating the hexacarbonyldi-
cobalt complex of one alkyne (i) with free alkyne (ii) and
vice versa. Whereas, in the formation of the cobaltacy-
clopentadiene 2n , the formation from (i) ) PhC2H and
(ii) ) C2Me2 compares reasonably with its alternative
formation from (i) ) C2Me2 and (ii) ) PhC2H (18% or
40% based on unrecovered starting complex vs 12% or
50% based on unreacted starting complex), the forma-
tion of complex 2l from (i) ) C2(COOMe)2 and (ii) )
PhC2H in 41% yield [with only 17% recovery of (i)] is in
rather sharp contrast to its formation in only 2% yield
(10) (a) Cetini, G.; Gambino, R.; Rossetti, R.; Sappa, E. J . Orga-
nomet. Chem. 1967, 8, 149. (b) Cetini, G.; Gambino, R.; Stanghellini,
P. L.; Rossetti, R. Rend.sAccad. Naz. 40 [Quaranta] 1968, [4] 18/ 19,
45; Chem. Abstr. 1968, 72, 71104.
(11) (a) Yamazaki, H.; Wakatsuki, Y. J . Organomet. Chem. 1977,
139, 157. (b) McAlister, D. R.; Bercaw, J . E.; Bergman, R. G. J . Am.
Chem. Soc. 1977, 99, 1666.
(12) Maier, G.; Bosslet, F. Tetrahedron Lett. 1972, 1025.

200 Organometallics, Vol. 18, No. 2, 1999
Baxter et al.
Ta ble 2. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for 2n
Molecule 1
Bond Lengths
Co(1)-Co(2)
Co(1)-C(1)
Co(1)-C(2)
Co(1)-C(3)
Co(2)-C(4)
Co(2)-C(5)
Co(1)-C(7)
Co(1)-C(11)
Co(2)-C(7)
Co(2)-C(8)
Co(2)-C(10)
Co(2)-C(11)
2.498(1)
1.813(9)
1.785(9)
1.825(9)
1.741(9)
1.747(9)
1.992(7)
1.962(6)
2.011(7)
2.120(7)
2.108(7)
2.018(7)
C(1)-O(1)
C(2)-O(2)
C(3)-O(3)
C(4)-O(4)
C(5)-O(5)
C(6)-C(7)
C(7)-C(8)
C(8)-C(9)
C(8)-C(10)
C(10)-C(11)
C(11)-C(12)
1.12(1)
1.14(1)
1.12(1)
1.14(1)
1.14(1)
1.51(1)
1.40(1)
1.52(1)
1.39(1)
1.43(1)
1.47(1)
Bond Angles
Co(2)-Co(1)-C(1)
Co(2)-Co(1)-C(2)
Co(2)-Co(1)-C(3)
Co(2)-Co(1)-C(7)
Co(2)-Co(1)-C(11)
C(1)-Co(1)-C(2)
C(1)-Co(1)-C(3)
C(1)-Co(1)-C(7)
C(1)-Co(1)-C(11)
C(2)-Co(1)-C(3)
C(2)-Co(1)-C(7)
C(2)-Co(1)-C(11)
C(3)-Co(1)-C(11)
132.1(3) C(7)-Co(1)-C(11)
109.0(3) Co(1)-Co(2)-C(4)
108.8(3) Co(1)-Co(2)-C(5)
51.7(2) C(4)-Co(2)-C(5)
52.1(2) Co(1)-C(7)-C(6)
104.6(4) Co(1)-C(7)-C(8)
80.4(3)
114.4(3)
128.9(3)
92.9(4)
123.8(5)
114.0(5)
104.9(4) Co(1)-C(11)-C(12) 124.8(5)
95.2(3) Co(1)-C(11)-C(10) 112.9(5)
93.9(3) C(7)-C(8)-C(9)
88.8(4) C(7)-C(8)-C(10)
91.4(3) C(8)-C(10)-C(11)
160.3(4) C(10)-C(11)-C(12) 119.0(6)
92.7(3)
124.4(6)
113.6(6)
115.1(6)
Molecule 2
Bond Lengths
F igu r e 1. ORTEP plot of the two crystallographically
unique molecules of 2n . Thermal ellipsoids are drawn at
the 40% level, and the atom numbering schemes are shown.
Co(3)-Co(4)
Co(3)-C(18)
Co(3)-C(19)
Co(3)-C(20)
Co(4)-C(21)
Co(4)-C(22)
Co(3)-C(24)
Co(3)-C(28)
Co(4)-C(24)
Co(4)-C(25)
Co(4)-C(27)
Co(4)-C(28)
2.503(1)
1.794(8)
1.808(9)
1.802(8)
1.728(9)
1.738(9)
1.967(7)
1.965(7)
2.010(7)
2.116(8)
2.106(7)
2.011(7)
C(18)-O(6)
C(19)-O(7)
C(20)-O(8)
C(21)-O(9)
C(22)-O(10)
C(23)-C(24)
C(24)-C(25)
C(25)-C(26)
C(25)-C(27)
C(27)-C(28)
C(28)-C(29)
1.13(1)
1.13(1)
1.13(1)
1.15(1)
1.14(1)
1.53(1)
1.41(1)
1.52(1)
1.41(1)
1.42(1)
1.49(1)
further to give flyovers and arenes are undoubtedly a
major factor, and fast reactions must account both for
low yields and for those cases where we failed to isolate
such complexes although most of the mono(alkyne)
complex had reacted.
In the metal-carbonyl stretching region of the infra-
red spectra, the cobaltacyclopentadienes show a rather
characteristic pattern of three bands: a very strong one
at 2022-2067 cm-1 flanked by strong bands of ap-
proximately equal intensity at 2074-2101 and at 1968-
1999 cm-1. All the complexes 2 are orange or brown
solids, of which 2d and 2n crystallized most readily.
They were isolated after chromatography on silica,
except for 2j, 2k , and 2n , which required the use of
neutral alumina or silica pretreated with triethylamine.
For each series of derivatives the order of elution was
1, 2, 3, then 4.
X-r a y Cr ysta l Str u ctu r e of 2n . As the known
structures relate to symmetrical compounds,3-7 we
selected the unsymmetrical complex 2n for X-ray crys-
tallographic study. Selected bond lengths and angles are
shown in Table 2, and the structure of 2n is shown in
Figure 1. The complex crystallizes with two unique
molecules in the asymmetric unit, and since the complex
is chiral (vide infra) and the space group is centric, both
enantiomers coexist in the crystal. In addition to the
known direct analogues (2a ,3 2b,4 2c,5 2d 7) a range of
related metallacyclopentadienes (5a , 5b, 8a ,13 8b,14 9,15
Bond Angles
Co(4)-Co(3)-C(18) 130.7(2) C(24)-Co(3)-C(28)
80.2(3)
Co(4)-Co(3)-C(19) 103.3(3) Co(3)-Co(4)-C(21) 120.1(3)
Co(4)-Co(3)-C(20) 114.1(2) Co(3)-Co(4)-C(22) 127.1(3)
Co(4)-Co(3)-C(24)
Co(4)-Co(3)-C(28)
51.8(2) C(21)-Co(4)-C(22)
51.8(2) Co(3)-C(24)-C(23) 123.0(5)
92.4(4)
C(18)-Co(3)-C(19) 108.2(4) Co(3)-C(24)-C(25) 119.7(6)
C(18)-Co(3)-C(20) 102.8(4) Co(3)-C(28)-C(29) 125.4(5)
C(18)-Co(3)-C(24)
C(18)-Co(3)-C(28)
C(19)-Co(3)-C(20)
C(19)-Co(3)-C(24)
90.5(3) Co(3)-C(28)-C(27) 114.1(5)
96.8(3) C(24)-C(25)-C(26) 125.1(7)
90.4(4) C(24)-C(25)-C(27) 112.5(7)
90.8(3) C(25)-C(27)-C(28) 114.2(6)
C(19)-Co(3)-C(28) 153.5(3) C(27)-C(28)-C(29) 117.8(6)
C(20)-Co(3)-C(28) 92.5(3)
from (i) ) PhC2H and (ii) ) C2(COOMe)2 [with 61%
recovery of (i)]. More facile exchange of alkyne in the
latter case partly accounts for the difference, but varia-
tion of the mono(alkyne) complexes in reactivity toward
addition must also be important, as shown by the
different extents of their recovery. Products of ligand
exchange, including 2e (3% yield), were the only results
from the two approaches involving C2Me2 and C2-
(COOMe)2, but 2m was isolated from (i) ) C2(COOMe)2
and (ii) ) HC2COOMe. Similarly, cobaltacyclopenta-
dienes were formed in the reaction of (i) ) PhC2H and
(ii) ) HC2COOMe, but extensive ligand exchange and
instability precluded satisfactory separation. The dif-
fering rates at which the cobaltacyclopentadienes react
(13) Rosenblum, M.; North, B. J . Am. Chem. Soc. 1968, 90, 1060.
(14) Lee, W.-S.; Brintzinger, H. H. J . Organomet. Chem. 1977, 127,
93.
(15) Tyler, S. J .; Burlitch, J . M. J . Organomet. Chem. 1989, 361,
231.

(Cobaltacyclopentadiene)cobalt Complexes
Organometallics, Vol. 18, No. 2, 1999 201
Ch a r t 1
tures. This degree of distortion may be expected to
weaken any π-overlap with metal-based orbitals.
By analogy with the long-known Fe2(CO)6 analogue18
and the predominant formation of 1,2,4-triphenylben-
zene in most cyclotrimerization reactions of ethynyl-
benzene, complex 2d was expected to have its phenyl
substituents in the 2- and 5-positions of the cobaltacy-
clopentadiene ring. This was confirmed when the report
of its independent synthesis and crystal structure
analysis7 became available. It is also consistent with the
position of the phenyl group in complex 2n (see above).
1016) have also been structurally characterized. The key
NMR Sp ectr a a n d Regioch em istr y of Coba lta -
cyclop en ta d ien es 2. In principle, an unsymmetrical
alkyne, XC2Y, could yield three isomeric cobaltacyclo-
pentadienes, 2, the symmetrical complexes A and B
(Chart 2) and the unsymmetrical isomer C. No simple
pattern of regiochemistry emerges from the present
work, and examples of A (Y ) H), C, and possibly B (Y
) H) have been found. There is compelling evidence
from flyover formation19 that methyl 2-butynoate yields
at least one transient symmetrical cobaltacycle (A or
B) as well as C; only one alkyne yielded two isolable
isomers, A and C. The reasons for such diverse behavior
are not obvious.
feature of the structure is the central η5-cobaltacyclo-
pentadiene unit and its relationship to the second cobalt
atom (see Chart 1). Interestingly, this grouping is
remarkably invariant between the structures, appar-
ently not being influenced to any significant degree by
a range of different substituents either on the ring or
on the cobalt atoms. This is illustrated in Table 3, where
the key parameters for 2a -d and 2n are compared. In
2n , the C4 unit of the cobaltacyclopentadiene group is
planar with a maximum deviation from the least-
squares plane of 0.001(8) Å, while the C-C bond lengths
are approximately equal (for molecule 1, 1.397(9), 1.389-
(9), and 1.428(9) Å; for molecule 2, 1.412(9), 1.412(9),
and 1.415(9) Å), indicative of a delocalized system rather
than a discrete diene structure. This is in keeping with
the first structurally characterized complex of this
type,16 the [HFe2(CO)8]-/2-butyne derived complex (10)
which, with C-C bond distances of 1.423, 1.430, and
1.424 Å, also showed no bond alternation. It is inviting
to formulate these complexes as 2′, i.e. as analogues of
CpCo(CO)2, with the Cp ligand replaced by an aromatic
Co-containing heterocyclopentadienyl ligand. It is not
clear whether this is genuinely the case. For instance,
the Co(1)-C(7), Co(1)-C(11), Co(3)-C(24), and Co(3)-
C(28) bond lengths at an average of 1.972(11) Å are
perhaps slightly shorter than typical Co-C σ-bonds
(e.g., 1.998 and 1.996 Å in [(η5-1,2,3-C5Me3H2)CoMe2-
(PMe3)]17a and 2.016 Å in [(η4-di-(p-tolylfulvalene)CoMe-
(PMe3)2]17b) but are not strongly suggestive of the
multiple-bond character which might be expected for an
“aromatic” structure. In addition, the Co atoms in the
heterocyclopentadienyl rings are disposed some 0.45 Å
above the C4 plane, resulting in an “envelope” confor-
mation. The dihedral angles between the C4 planes
described by C(7), C(8), C(10), C(11) and by C(24), C(25),
C(27), C(28) and the CoC2 planes described by Co(1),
C(7), C(11) and by Co(3), C(24), C(28) are 17.6 and 17.1°
respectively. This parallels closely other observed struc-
The substitution patterns for all the complexes 2,
except 2f and 2j, can be assigned with confidence by
1
comparison of H and 13C NMR spectra (Tables 4 and
5) with those of the crystallographically characterized
materials 2a , 2b, 2d , and 2n .20 The results parallel
assignments reported21 for related iron complexes, e.g.
10. The unsymmetrical 2,4-disubstituted structures
(Chart 2; C), 2g, 2h , 2k , and 2o, are easily recognized
because the 1H NMR spectra exhibit two different
signals for each substituent, X or Y, while the sym-
metrical complexes (Chart 2; A, B) exhibit a single
signal for each substituent. Thus the product from
1-heptyne, 2i, is assigned a structure (A, Y ) H)
analogous to 2d on the basis of the similarity of the
chemical shifts of the â-protons. This parameter is also
used to assign structures 2l and 2m by comparison with
2n . In these cases, the chemical shift of the â-proton
falls in the range δ 6.0-6.3. Methyl propargyl ether is
the only alkyne we have found to give both a sym-
metrical (2j, major product) and an unsymmetrical
cobaltacyclopentadiene, 2k . In the latter case, the R-H
(18) (a) Hu¨bel, W.; Braye, E. H. J . Inorg. Nucl. Chem. 1959, 10, 250.
(b) Aime, S.; Milone, L.; Sappa, E. J . Chem. Soc., Dalton Trans. 1976,
838. (c) Todd, L. J .; Hickey, J . P.; Wilkinson, J . R.; Huffman, J . C.;
Folting, K. J . Organomet. Chem. 1976, 112, 167.
(19) Baxter, R. J .; Knox, G. R.; Moir, J . H.; Pauson, P. L.; Spicer,
M. D. Organometallics 1999, 18, 206.
(20) Structure 2c has also been confirmed by X-ray analysis,5 but
1H and 19F NMR spectra each show two groups of signals suggesting
the possible presence of an additional isomer.
(16) Hock, A. A.; Mills, O. S. Acta Crystallogr. 1961, 14, 139.
(17) (a) Wakatsuki, Y.; Yamazaki, H. J . Chem. Soc. J pn. 1985, 586.
(b) Klein, H.-F.; Auer, E.; J ung, T.; Rohr, C. Organometallics 1995,
14, 2725.
(21) Mann, B. E.; Taylor, B. F. 13C NMR Data for Organometallic
Compounds; Academic Press: London, 1981; p 214.

202 Organometallics, Vol. 18, No. 2, 1999
Ta ble 3. Com p a r ison of th e Str u ctu r a l Da ta for Coba lta cyclop en ta d ien e Com p lexesa
Baxter et al.
complex
r1, Å
r2,
Å
r3, Å
r4, r5, Å
r6, Å
r7, Å
φ, deg
r8, Å
2n , molecule 1
2n , molecule 2
2a
2b
2c
2d
1.397(9)
1.412(9)
1.422(3)
1.403(7)
1.41(2)
1.389(9)
1.412(9)
1.428(3)
1.402(7)
1.45(2)
1.428(9)
1.415(9)
1.419(3)
1.400(8)
1.39(2)
1.992(7), 1.962(6)
1.967(7), 1.965(7)
1.973(2), 1.968(2)
1.971(6), 1.968(5)
1.98(1), 1.97(1)
2.498(1)
2.503(1)
2.4738(7)
2.474(1)
2.476(3)
2.503(1)
1.5577
1.5609
1.5659(3)
1.552(1)
1.551
17.6
17.1
14.2
17.7
15.8
16.6
0.4553
0.4423
0.3722(3)
0.457(1)
0.420
1.417(4)
1.399(5)
1.414(5)
1.982(3), 1.972(3)
1.562
0.431
a
Chart 1.
Ch a r t 2
an upfield shift of the whole spectrum and clear splitting
of the signals due to the ring hydrogen and the methyl
group of â-COOMe but not of the ester group closer to
the cobalt atom. We surmise that coordination occurs
between Pr3+ and the former group, whereas the
R-COOMe group may be too hindered.
Rea ction s of Coba lta cyclop en ta d ien es. The reac-
tions of cobaltacyclopentadienes with alkynes, which
lead to arenes and flyover complexes, are dealt with in
the following paper.19 From some of them, but also from
other reactions, we obtained evidence that the formation
of cobaltacyclopentadienes is reversible. One example
is provided by the reactions of methyl propiolate, which,
surprisingly, had not been among the many alkynes
investigated by Hu¨bel et al.1,24 The preparation of
cobaltacyclopentadiene 2f (vide supra) yields an unusu-
ally complex mixture of byproducts, including a 1,3,5-
trisubstituted flyover complex, only the symmetrical
(1,3,5-) trimethyl benzenetricarboxylate, and pyrone 11f.
signal appears at δ 6.30 and the â-H at δ 7.03.22 The
chemical shift of the ring protons (δ 6.43) for the
symmetrical isomer, 2j, falls between the previously
1
cited values for R-H and â-H, so that H NMR alone does
not permit an unambiguous decision in favor of orienta-
tion A or B (Chart 2; Y ) H). A similar situation obtains
for the symmetrical cobaltacyclopentadiene 2f from
1
methyl propiolate, which exhibits a H NMR signal at
δ 6.69. Strongly electron-attracting and deshielding
groups, in, e.g., 2c or 2m , appear to exert little influence
on δ(â-H), and these protons in 2f might be expected
1
also to resonate in the range δ 6.0-6.3. The H NMR
chemical shifts of substituents other than hydrogen
exhibit the expected trends. R-Methyl groups (δ 2.15)
are slightly more shielded than â-methyl groups (δ 2.44),
and there are small but significant differences between
R- and â-methoxycarbonyl groups. However, these do
not assist in assignment of 2f or 2j and the orientations
have been deduced from 13C NMR data.
However, complex 2f reacts thermally25 with a further
molecule of methyl propiolate to give a 99% yield of a
mixture of trimethyl benzene-1,2,4- and -1,3,5-tricar-
boxylates in a ratio similar to that of the Co2(CO)8-
catalyzed thermal trimerization. But neither the 1,3,5
substituted flyover complex nor the arene can arise via
a symmetrical structure, but only via the unsymmetrical
2,4-disubstituted cobaltacyclopentadiene (2, R1 ) R3 )
COOMe, R2 ) R4 ) H), which was not detected. We
conclude that this last complex reacts too rapidly with
free alkyne to accumulate but must be formed reversibly
from complex 2f via fission to free alkyne and a mono-
(alkyne) complex, thus establishing the equilibrium
shown in Scheme 2.
More direct evidence for such cleavage of the cobal-
tacyclopentadienes comes from degradative reactions of
the 2,5-diphenyl-1-cobaltacyclopentadiene 2d . When
this complex is heated alone [in hexane or in acetoni-
trile/toluene (1:1)] or heated with sulfur (in benzene) or
with carbon monoxide, it gives 1,2,4-triphenylbenzene
as the principal product, in the last case in as high as
83% yield. The rapid room-temperature oxidation of this
complex, 2d , with ceric ammonium nitrate also gives
the same arene as the main product, accompanied by
2,5-diphenylfuran, 12d , and 3,6-diphenyl-2H-pyran-2-
one, 11d . Oxidation with trimethylamine N-oxide (6
13C NMR data for the complexes 2 are more consistent
1
and reinforce the H-based structure assignments. The
values for the ring carbons (Table 5) are found to fall
into two quite distinct groups: All the R-carbons (i.e.,
those adjacent to Co) give signals in the range δ 138-
176.5, whereas the â-carbon signals fall in the narrower
range δ 103-122.23 13C NMR signals of methyl substit-
uents fall in regions characteristic of R (δ 29.4-30.7)
and â (δ 16.7-17.45). The values for ester carbonyls also
fall into two regions, R (δ 174-175) and â (δ 167).
Collectively, the 13C NMR data for the complexes 2 from
methyl propiolate and from methyl propargyl ether are
in favor of structures 2f and 2j, respectively. It was not
possible to obtain suitable crystals for X-ray analysis,
so the apparent ambiguity of 1H assignments for these
complexes remains unresolved.
Those complexes that have unsymmetrical cobalta-
cyclopentadiene groups are chiral, a fact which we have
verified for complex 2l by examining its 1H NMR
spectrum in the presence of (-)-Pr(hfc)3. This caused
(22) The 1H NMR spectrum of 2k is unusual and suggests a mixture
of conformers or rotamers. The single R- and â-protons each appear
as pairs of singlets which are not cross-ring coupled. Each MeO group
is a singlet, but the corresponding (diastereotopic) CH2O groups appear
as temperature-dependent AB multiplets with the less shielded â-CH2
more strongly influenced than the R-CH2: â-CH2, ∆δ(AB) ) 0.09 ppm
at 313 K, ∆δ(AB) ) 0.17 ppm at 233 K; R-CH2, ∆δ(AB) ) 0.07 ppm at
313 K, ∆δ(AB) ) 0.11 ppm at 233 K.
(24) Hu¨bel, W.; Hoogzand, C. Chem. Ber. 1960, 93, 103.
(25) A room-temperature reaction of 2f, methyl propiolate, and
N-oxide gives <5% yield of the isomeric arenes.
(23) The â-carbon of 2a is reported3 to give a signal at δ 138; for
2b, the corresponding value is δ 140.4

(Cobaltacyclopentadiene)cobalt Complexes
Organometallics, Vol. 18, No. 2, 1999 203
Ta ble 4. 1H NMR Da ta for Com p lexes 2 (CDCl3; δ)
complex
R1
R2
R3
R4
2d
2e
2f
2g
2h
2i
7.72-7.36 (m, 5H)
3.81 (s, 3H)
3.83 (s, 3H)
6.01 (s, 1H)
3.85 (s, 3H)
6.69 (s, 1H)
3.90 (s, 3H)
2.15 (s, 3H)
2.44 (s, 3H)
2.45 (s, 3H)
6.11 (s, 1H)
3.81 (s, 3H)
1.29 (t, 3H, J ) 7.1 Hz) 4.25-4.39
2.16 (s, 3H), 4.25-4.39 (m, 2H)
0.91 (t, 3H, J ) 7.0 Hz), 1.36 (m, 4H),
1.63 (m, 2H), 2.51 (t, 1H, J ) 2.8 Hz),
2.89 (t, 1H, J ) 2.8 Hz)
3.45 (s, 3H), 4.30 (s, 2H)
6.303, 6.313 (1H)
7.22-7.41 (m, 5H)
3.850 (s, 3H)
7.28-7.39 (m, 5H)
3.27 (s, 3H)
1.34 (t, 3H, J ) 7.1 Hz)
2j
6.43 (s, 1H)
2k a
2l
3.965 (dd2H), 3.057 (s3H) 7.026, 7.039 (1H) 3.66 (dd, 2H), 3.03 (s, 3H)
6.23 (s, 1H)
6.23 (s, 1H)
6.29 (s, 1H)
7.28-7.46 (m, 5H)
3.75 (s, 1H)
3.86 (s, 3H)
3.856 (s, 3H)
2.45 (s, 3H)
3.57 (s, 3H)
3.84 (s, 3H)
3.845 (s, 3H)
2.17 (s, 3H)
2m
2n
2o
2p
7.28-7.46 (m, 5H)
3.50 (s, 1H)
a
See text; toluene-d8.
Ta ble 5. 13C NMR Da ta for Com p lexes 2 (CDCl3, δ)a
complex
C1(R)
C2(â)
C3(â′)
C4(R′)
othera
2d
2e
2f
2g
2h
2i
2j
2k
2l
148.62
138.77
139.26
135.57
138.12
168.85
176.47
168.86
148.81
140.17
150.2
114.66
103.64
116.75
117.52
119.17
117.22
114.22
122.02
114.61
117.53
114.44
122.11
114.66
103.64
116.75
114.20
114.41
117.22
114.22
114.45
109.88
109.91
121.07
102.12
148.62
138.77
139.26
166.33
165.91
168.85
176.47
174.53
139.77
140.09
165.25
147.48
128.06, 128.35, 129.05, 138.83
R1, 53.52, 174.43; R2, 53.21, 167.43
53.11, 173.53
R1, 53.03, 174.23; R2, 17.41; R3, 52.57, 167.24; R4, 30.65
R1, 14.39, 62.33, 173.75; R2, 17.36; R3, 14.27, 61.86, 167.49; R4, 30.62
14.19, 22.68, 31.80, 33.73, 47.25
58.79, 80.91
R2, 58.88, 72.25; R4, 59.05, 80.09
R1, 128.05, 128.46, 128.72, 134.11; R3, 52.99, 164.02; R4, 52.99, 173.79
R1, 53.51, 173.58; R3, 53.22, 167.84; R4, 53.54, 175.98
R1, 127.76, 128.14, 129.73, 132.12; R3, 16.73; R4, 29.43
R1, 52.38, 173.15; R2, R4, 127.25, 127.83, 127.98, 128.70, 129.07, 133.49, 133.91;
R3, 52.34, 165.77
2m
2n
2o
139.46
a
b
For numbering of the ring carbon atoms and substituents see Scheme 1 and Chart 2. All complexes exhibited Co-CO signals in the
region δ 198-205.
Chromatography was carried out using either silica gel (230-
400 mesh) or ungraded alumina, neutralized under ethyl
acetate, washed with ethanol and water, and oven-dried.
Melting points were obtained in open capillaries and are
Sch em e 2
1
uncorrected. H NMR and 13C NMR spectra were recorded at
frequencies of 400.134 and 100.614 MHz, respectively, both
being referenced to an internal solvent standard.
Rea ction s of Alk yn es w ith (Alk yn e)h exa ca r bon yld i-
coba lt Com p lexes (2). Gen er a l P r oced u r e. Equimolar
amounts (as specified in Table 1) of alkyne complex 2 and
anhydrous trimethylamine N-oxide in tetrahydrofuran/dichlo-
equiv, THF, 30 min in air) yields a complex mixture of
at least 10 metal-free organic compounds. This mixture
was analyzed by a combination of column and thin-layer
chromatography and LC-MS. Apart from the aforemen-
tioned compounds, products with stoichiometries (Ph-
C2H)3CO, (PhC2H)4, (PhC2H)4CO, and (PhC2H)4(CO)2
were detected, but not isolated pure. Since no free
phenylethyne is initially available, the formation of
triphenylbenzene is only possible if the organic portion
of the complex undergoes fission to give back free alkyne
and/or its cobalt complex under remarkably mild condi-
tions. Further reactions involving such cleavage are
included in the following paper.19
romethane (20 mL each on
a 3 mmol or smaller scale,
increased in proportion for larger scale reactions) were stirred
at room temperature for 45-90 min, experience (following
reactions by TLC) having shown that this time led generally
to the best yields of cobaltacyclopentadienes. Variations from
this procedure are noted below for individual products. All
were isolated by column chromatography. Known products
(including starting complexes and flyover complexes (3) de-
scribed in the following paper19) were identified by comparison
1
of IR and/or H NMR spectra with those of authentic samples.
Quantities of materials are given in Table 1, and 1H and 13C
NMR data for the cobaltacyclopentadienes are listed in Tables
4 and 5.
Typ ica l P r oced u r e. F or m a tion of P en ta ca r bon yl[µ-
(1,2,3,4-η:1,4-η)-1,2,3,4-t et r a k is(m et h oxyca r b on yl)-1,3-
bu ta d ien e-1,4-d iyl]d icoba lt (2e). Under nitrogen, hexacar-
bonyl(dimethyl acetylenedicarboxylate)dicobalt (593 mg, 1.38
mmol) was dissolved in a mixture of tetrahydrofuran (20 mL)
and dichloromethane (20 mL). Dimethyl acetylenedicarboxy-
late (197 mg, 1.38 mmol) was added to the stirred solution,
followed by solid trimethylamine N-oxide (81 mg, 1.38 mmol).
After being stirred at room temperature for 90 min, the
solution was filtered through Kieselguhr; the filtrate was
Exp er im en ta l Section
Gen er a l Rem a r k s. All reactions were carried out under a
dry nitrogen atmosphere, and solvents were purified and dried
under nitrogen before use. Anhydrous trimethylamine N-oxide
was obtained by recrystallization from hot DMF.26 Co2(CO)8
was obtained from Strem Chemicals and used as purchased.
(26) Soderquist, J . A.; Anderson, C. L. Tetrahedron Lett. 1986, 27,
3961.

204 Organometallics, Vol. 18, No. 2, 1999
Baxter et al.
evaporated on a rotary evaporator, and the solid was chro-
matographed on a silica gel column. Complex 2e was eluted
using ether/petroleum ether (4:1) and obtained as a brown
solid. IR (hexane, cm-1): νCO 2096, 2045, 1989; νCdO 1724. Anal.
Calcd for C17H12Co2O13 (Mr ) 541.8940): C, 37.6; H, 2.2.
Found: C, 37.7; H, 2.2. MS (m/z): 542.9079 (M+ + 1). It was
) 484.1042): C, 37.2; H, 2.1. Found: C, 37.2; H, 2.0. LRMS
(m/z): 456 (M+-CO).
P en ta ca r bon yl[µ-(1,2,3,4-η-1,4-η)-1,2-d im eth yl-4-p h en -
yl-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2n ). This compound was
eluted using hexane [following (Me2C2)Co2(CO)6 and (PhC2H)-
Co2(CO)6] and obtained as an orange-brown solid. IR (hexane,
cm-1): νCO 2077, 2056, 2025, 1972, 1940. Anal. Calcd for
followed by the flyover complex (3, R1 ) R2 ) R3 ) R4
)
COOMe) (117 mg) whose characterization is included in the
C
17H12Co2O5 (Mr ) 414.1457): C, 49.3; H, 2.9. Found: C, 49.35;
following paper.19
H, 2.4. MS (m/z): 386.1421 (M+ - CO).
P en ta ca r bon yl[µ-(1,2,3,4-η:1,4-η)-1,4-d ip h en yl-1,3-bu ta -
d ien e-1,4-d iyl]d icoba lt (2d ).7 This compound formed orange
crystals. IR (hexane, cm-1): νCO 2081, 2056, 2032, 1977. Anal.
Calcd for C21H12Co2O5: C, 54.6; H, 2.6. Found: C, 54.6; H, 2.7.
P en t a ca r b on yl[µ-(1,2,3,4-η:1,4-η)-1,3-b is(m et h oxyca r -
bon yl)-2,4-d ip h en yl-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2o).
This compound was obtained as a brown solid, eluted by
hexane/ether. IR (hexane, cm-1): νCO 2091, 2043, 2033, 1966;
P en t a ca r b on yl[µ-(1,2,3,4-η:1,4-η)-1,4-b is(m et h oxyca r -
bon yl)-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2f). This compound,
eluted from silica by petroleum ether/ether (2:1), was obtained
as a brown solid. IR (CH2Cl2, cm-1): νCO 2101, 2067, 2040;
νCdO 1707. Anal. Calcd for C13H8Co2O9 (Mr ) 425.8832): C,
37.6; H, 2.2. Found: C, 38.4; H, 2.7. MS (m/z): 425.8924 (M+).
P en t a ca r b on yl[µ-(1,2,3,4-η:1,4-η)-1,3-b is(m et h oxyca r -
bon yl)-2,4-d im eth yl-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2g).
This compound was obtained as a brown solid, eluted by
hexane/ether. IR (hexane, cm-1): νCO 2088, 2038, 1992; νCdO
1727. Anal. Calcd for C15H12Co2O9 (Mr ) 454.1213): C, 39.7;
H, 2.7. Found: C, 39.8; H, 2.7. LRMS (m/z): 426 (M+ - CO).
νCdO 1737, 1716. Anal. Calcd for C25H16Co2O9 (Mr
)
578.2631): C, 51.9; H, 2.8. Found: C, 52.0; H, 2.8. LRMS (m/
z): 578 (M+).
Nega tive Resu lts. Failure to observe more than trace
quantities of cobaltacyclopentadienes was recorded in the
following cases. The percentage of unreacted starting complex,
1, is given in brackets with the percentage of alkyne-exchange
product, 1, from “mixed” systems in parentheses: PhC2H with
the complexes of FcC2H [77] or HC2COOMe; Ph2C2 with the
complexes of Ph2C2 [98], C2(COOMe)2 [76], or PhC2H [87 (8)];
C2(COOMe)2 with the complexes of C2Me2 [56] or C2(SiMe3)2
[86]; C2Me2 with the complexes of C2Me2 [94], C2H2 [67 (16)],
ButC2H [86 (5)], or HC2CH2OMe [77 (5)]; HC2C5H11 with the
complex of PhC2H [93]; MeC2SiMe3 with the complexes of
MeC2SiMe3 [97] or PhC2H [81 (6)]; HC2COOMe with the
complex of PhC2H [78 (9)]; C2(CH2OH)2 with the complexes of
C2(CH2OH)2 [97], C2(COOMe)2 [83 (5)], or PhC2H; C2(SiMe3)2
with the complex of C2(COOMe)2 [92 (4)]; or any of the alkynes
C2(CH2Cl)2, HC2CH2Cl, HC2CH2OH [85], HC2CH(OH)C6H11
[96], HC2CH2NMe2 [71], HC2CH2OSO2Ph, and PhC2P(O)-
P en ta ca r bon yl[µ-(1,2,3,4-η:1,4-η)-1,3-bis(eth oxyca r bo-
n yl)-2,4-dim eth yl-1,3-bu tadien e-1,4-diyl]dicobalt (2h ). This
compound was eluted from silica by petroleum ether/ether (1:
1) and obtained as a brown solid. IR (CH2Cl2, cm-1): νCO 2086,
2033, 1982; νCdO 1712. Anal. Calcd for C17H16Co2O9 (Mr
)
481.9458). MS (m/z): 454.9729 (M+ + H - CO; calcd 454.9587).
P en t a ca r b on yl[µ-(1,2,3,4-η:1,4-η)-1,4-d i-n -p en t yl-1,3-
bu ta d ien e-1,4-d iyl]d icoba lt (2i). This compound was eluted
from silica by hexane and obtained as an orange solid. IR
(hexane, cm-1): νCO 2074, 2022, 2012, 1972. Anal. Calcd for
C19H24Co2O5 (Mr ) 450.2629). LRMS (m/z): 421 (M+ - H -
CO).
27
(OEt)2 with their own complexes. C2H2 yielded a very
unstable brown complex, possibly 2p , showing νCO (CH2Cl2,
cm-1) 2088, 2031, 1969 and 1H NMR (CDCl3) δ 3.75 (s), 3.50
(s) (ratio 1:1).
Rea ction of 3,3-Dim eth yl-1-bu tyn e w ith Its Hexa ca r -
b on yld icoba lt Com p lex. Carried out as for the above
preparations of complexes 2, the reaction of the alkyne (0.23
g, 2.84 mmol) with the complex (1.05 g, 2.84 mmol) in
dichloromethane (15 mL) and THF (15 mL) yielded unreacted
[(ButC2H)Co2(CO)6] (704 mg, 67%), eluted from silica by hexane
and followed by 2,5-di-tert-butylcyclopenta-2,4-dienone [71 mg,
15.4% (46% based on unrecovered complex)]: orange crystals;
P en ta ca r bon yl[µ-(1,2,3,4-η:1,4-η)-1,4-bis(m eth oxym eth -
yl)-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2j). This compound was
eluted with petroleum ether/ether (1:1) and obtained as an
orange solid. IR (CH2Cl2, cm-1): νCO 2092, 2023, 1968. Anal.
Calcd for C13H12Co2O7 (Mr ) 398.1005). LRMS (m/z): 397 (M+
- H). It was followed by p en ta ca r bon yl[µ-(1,2,3,4-η:1,4-η)-
1,3-bis(m eth oxym eth yl)-1,3-bu ta d ien e-1,4-d iyl]d icoba lt
(2k ). IR (CH2Cl2, cm-1): νCO 2085, 2026, 1968. Anal. Calcd
for C13H12Co2O7 (Mr ) 398.1005). LRMS (m/z): 371 (M+ + H
- CO).
P en t a ca r b on yl[µ-(1,2,3,4-η:1,4-η)-1,2-b is(m et h oxyca r -
bon yl)-4-p h en yl-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2l). This
compound was eluted from silica by petroleum ether/ether (1:
1) and obtained as brown crystals. IR (hexane, cm-1): νCO 2095,
2047, 1997; νCdO 1737, 1724. Anal. Calcd for C19H12Co2O9 (Mr
) 502.1653): C, 45.45; H, 2.4. Found: C, 45.8; H, 2.6. LRMS
(m/z): 502 (M+).
IR νCO 1715 cm-1
[lit.12 CO (CCl4) 1710 cm-1; δ (CCl4) 6.33 (s), 1.13 (s) (1:9)]; 13
NMR δ (CDCl3) 202, 140.4, 135.9, 32.4, 29.9.
;
1H NMR δ (CDCl3) 6.42 (s), 1.15 (s) (1:9)
ν
C
R ea ct ion s of t h e Dip h en ylcob a lt a cyclop en t a d ien e
Com p lex 2d . (a ) Th er m olysis u n d er Ca r bon Mon oxid e.
The complex (204 mg) in hexane (75 mL) was placed in a 200
mL steel autoclave under 110 atm of CO and heated to 80 °C
for 24 h. After cooling, the filtered solution was evaporated,
and the solid was chromatographed on silica. Hexane eluted
1,2,4-triphenylbenzene (101 mg, 75%), mp 119-120 °C, C24H18
(Mr ) 306). LRMS (m/z): 306 (M+), followed by a more polar
fraction (43 mg) containing a mixture of 2,5-diphenylfuran,
m/z 220 (M+), and 3,6-diphenyl-2H-pyran-2-one, m/z 248 (M+).
This complex was prepared in two ways (see Table 1). When
obtained from ethynylbenzene and the cobalt complex of
dimethyl acetylenedicarboxylate, the products, in order of
elution, were (i) hexacarbonyl(ethynylbenzene)dicobalt (5%),
(ii) complex 2d (3%), (iii) unstable complex 3 (R1 ) R4 ) Ph;
R2 ) R3 ) H) (trace), (iv) unreacted starting complex (17%),
and (v) complex 2l. When dimethyl acetylenedicarboxylate and
the complex of ethynylbenzene were the precursors, (i), (iv),
IR: νmax 1723, 1626, 1553, 1358 cm-1 1H NMR, δ (CDCl3):
.
7.91-7.71 (m, 4H), 7.59 (d, J ) 7.2 Hz, 1H), 7.51-7.38 (m,
6H), 6.81 (d, J ) 7.2 Hz, 1H).
(b) Th er m olysis in Aceton itr ile/Tolu en e. The complex
(414 mg) was refluxed in acetonitrile/toluene (1:1, 30 mL) for
24 h and worked up as above to yield 1,2,4-triphenylbenzene
(185 mg, 68%).
and (v) were followed by 3 (R1 ) Ph; R2 ) H; R3 ) R4
COOMe).
)
P e n t a c a r b o n y l[µ-(1,2,3,4-η:1,4-η)-1,2,4-t r is (c a r b o -
m eth oxy)-1,3-bu ta d ien e-1,4-d iyl]d icoba lt (2m ). This com-
pound was eluted using ether/petroleum ether (1:1) (following
the propiolate and dicarboxylate hexacarbonyldicobalt com-
plexes) and obtained as a brown solid. IR (CH2Cl2, cm-1): νCO
2104, 2053, 1999; νCdO 1723. Anal. Calcd for C15H10Co2O11 (Mr
(c) Oxid a tion . The complex (261 mg, 0.56 mmol) was
stirred in acetonitrile, and ceric ammonium nitrate (1.0 g, 1.8
(27) Baxter, R. J .; Knox, G. R.; McLaughlin, M.; Pauson, P. L.;
Spicer, M. D. J . Organomet. Chem., in press.

(Cobaltacyclopentadiene)cobalt Complexes
Organometallics, Vol. 18, No. 2, 1999 205
Ta ble 6. Cr ysta l Da ta a n d Exp er im en ta l Deta ils of
th e Cr ysta l Str u ctu r e Deter m in a tion for 2n
Cr ysta l Str u ctu r e Deter m in a tion of 2n . Red crystals of
C17H12Co2O5 (2n ) suitable for X-ray diffraction studies were
obtained by slow evaporation of an n-hexane/benzene solution
at 0 °C. A crystal having dimensions 0.45 × 0.15 × 0.05 mm
was mounted in a glass capillary. Data collection was per-
formed at 22 °C. The complex crystallizes in the triclinic space
group P1h, and accurate cell dimensions were obtained from
20 precisely centered reflections in the range 16.58 e 2θ e
22.68°. Data collection in the range 2.5 e θ e 24° was
performed using ω-2θ scans on a Rigaku AFC7S diffracto-
meter with graphite-monchromated Mo KR radiation. A total
of 5484 reflections were recorded, of which 5369 were unique
(Rint ) 0.031). An empirical absorption correction (ψ scans)
was applied to the data, and they were also corrected for
Lorentz and polarization effects. The structure was solved by
direct methods (SIR92)28 and expanded using Fourier tech-
niques (DIRDIF92).29 Full-matrix least-squares refinement
based on 3200 observed reflections (I > 1.00σ(I)) with all non-H
atoms anisotropic converged with R ) 0.062, Rw ) 0.044, and
GOF ) 1.25. A total of 434 parameters were refined, and the
residual electron density was in the range +0.57 to -0.47 e
Å-3. H atoms were included in calculated positions. Experi-
mental details are summarized in Table 6.
Crystal Data
empirical formula
mol wt
C17H12Co2O6
414.15
cryst color
cryst size, mm
space group
a, Å
b, Å
c, Å
red
0.45 × 0.15 × 0.05
P1h (No. 2)
7.146(2)
14.663(3)
16.502(5)
87.71(2)
85.05(3)
87.58(2)
1720.0(9)
4
R, deg
â, deg
γ, deg
V, Å3
Z
F
calcd, g cm-3
1.599
19.51
832.00
µ(Mo KR), cm-1
F(000)
Data Collection
diffractometer
radiation
wavelength, Å
T, K
Rigaku AFC7S
Mo KR
0.710 69
295
scan type
ω-2θ
scan rate, deg min-1
scan width, deg
θ range
16 (in ω)
0.94 + 0.35 tan θ
2.5-24
h, 0 to 8; k, -16 to 16;
l, 18 to -18
Ack n ow led gm en t. The authors thank the EPSRC
for a maintenance grant to R.J .B. and Dr. A. R. Kennedy
for X-ray data collection.
index ranges
Solution and Refinement
Su p p or tin g In for m a tion Ava ila ble: Text giving ad-
ditional experimental details for the structure determination
and tables of atomic coordinates, anisotropic displacement
parameters, interatomic distances, bond angles, and least-
squares planes (23 pages). Ordering information is given on
any current masthead page.
no. of reflcns measd
5484
no. of unique reflcns
no. of obsd relections (I > 1.00σ(I))
5369 (Rint ) 0.031)
3200
no. of variables
GOF
R
Rw
434
1.25
0.062
0.044
+0.57 to -0.47
residual electron density, e Å-3
OM980544Y
mmol) was added, causing a rapid color change. The major
product was again 1,2,4-triphenylbenzene.
(d ) Rea ction w ith Su lfu r . The complex (231 mg) was
heated in refluxing benzene (20 mL) for 12 h with a large
excess of sulfur (1 g). The only isolable product was 1,2,4-
triphenylbenzene.
(28) SIR-92: Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano,
M.; Giacovazzo, C.; Guagliardi, A.; Polidori, G. J . Appl. Crystallogr.
1994, 27, 435.
(29) DIRDIF92: Beurskens, P. T.; Admiraal, G.; Beurskens, G.;
Bosman, W. P.; Garcia-Granda, S.; Smits, J . M. M.; Smykalla, C.
DIRDIF Programme Systm; Technical Report of the Crystallography
Laboratory; University of Nijmegen: Nijmegen, The Netherlands.