Catalytic iron-mediated [4 + 1] cycloaddition of diallenes with carbon monoxide

Article abstract of DOI:10.1021/ja962908o

Conjugated diallenes undergo stereoselective [4 + 1] cycloadditions in a variety of solvents. Accelerated thermal exchange of CO in Fe(CO)₅ is catalyzed by the diallene substrate, consistent with the previous kinetic analysis. The Eyring activation parameters were determined to be ΔH(paragraph) = 16.2 (+0.7) kcal mol^-1 and ΔS(paragraph) = - 17.9 (± 2.3) cal mol^-1 K^-1. The negative entropy of activation is consistent with rate determining associative coordination of diallene to form a diallene-iron complex. The effect of coordinating solvents on this cycloaddition reaction suggests that the synthetic utility of these transformations can now be expanded to include diallenes with functional groups that may only weakly coordinate to the iron. Only one sterically demanding substituent on the diallene termini is necessary to give good π-facial selectivity, which also expands the synthetic utility of these transformations.

Full text of DOI:10.1021/ja962908o

J. Am. Chem. Soc. 1996, 118, 11783-11788
11783
Catalytic Iron-Mediated [4 + 1] Cycloaddition of Diallenes
with Carbon Monoxide
Matthew S. Sigman and Bruce E. Eaton*
Contribution from the Department of Chemistry, Washington State UniVersity,
Pullman, Washington 99164-4630
X
ReceiVed August 19, 1996
Abstract: Conjugated diallenes undergo stereoselective [4 + 1] cycloadditions in a variety of solvents. Accelerated
thermal exchange of CO in Fe(CO)5 is catalyzed by the diallene substrate, consistent with the previous kinetic analysis.
q
-1
q
The Eyring activation parameters were determined to be ∆H ) 16.2 ((0.7) kcal mol and ∆S ) -17.9 ((2.3)
-
1
-1
cal mol K . The negative entropy of activation is consistent with rate determining associative coordination of
diallene to form a diallene-iron complex. The effect of coordinating solvents on this cycloaddition reaction suggests
that the synthetic utility of these transformations can now be expanded to include diallenes with functional groups
that may only weakly coordinate to the iron. Only one sterically demanding substituent on the diallene termini is
necessary to give good π-facial selectivity, which also expands the synthetic utility of these transformations.
Introduction
monoxide to give 2,5-dialkylidenecyclopentenones.6 These
molecules have been important for the investigation of reactivity
and electronic structure in planar cross-conjugated π-systems.
The biological activity and material science applications of 2,5-
dialkylidenecyclopentenones have yet to be fully investigated.7
Until recently iron carbonyls catalytically mediating carbon-
carbon bond formation was unprecedented. Historically, carbon-
carbon bond chemistry accomplished with iron carbonyls had
Many small molecule drugs contain five-membered carbo-
cyclic rings, such as prostaglandins,1 methylenomycin antibiot-
ics, and clavirideneone antineoplastics.3 Transition-metal medi-
ated assembly of five-membered carbocycles could prove to be
an important methodology for the synthesis of these ring
systems. Syntheses of carbocyclic rings has been accomplished
,2
4
5
by [3 + 2] and the Pauson-Khand [2 + 2 + 1] cycloadditions.
Adding to the scope of transition-metal mediated synthesis of
five-membered rings was the discovery of catalytic iron-
mediated [4 + 1] cycloaddition of diallenes with carbon
8
been stoichiometric. In addition, catalytic iron carbonyl [4 +
1] cycloaddition reactions did not appear to require the extremes
of thermal and/or photochemical activation that usually ac-
company the use of Fe(CO)5. Herein we report a more detailed
study of these new cycloaddition reactions.
X
Abstract published in AdVance ACS Abstracts, November 15, 1996.
1) (a) Oats, J. A.; Fitzgerald, G. A.; branch, R. A.; Jackson, E. K.; Knapp,
(
H. R.; Roberts, L. J., II New Eng. J. Med. 1988, 319, 689. (b) Narumiya,
S.; Fukushima, M. J. Pharmacol. Exp. Ther. 1986, 239, 500. (c) Atsmon,
J.; Freeman, M. L.; Meredith, M. J.; Sweetman, B. J.; Roberts, L. J., II
AdVances in Prostaglandin, Thromboxane and Leukotriene Research,
Raven: New York, 1990; Vol. 21, p 887. (d) Hirata, M.; Ohno, K. AdVances
in Prostaglandin, Thromboxane and Leukotriene Research, Raven: New
York, 1990; Vol. 21, p 859. (e) Hada, A.; Yamamoto, Y.; Mori, Y.; Yamada,
Y.; Kikuchi, H. Biochem. Biophys. Res. Commun. 1985, 130, 515. (f) Kato,
T.; Fukushima, M.; Kurozumi, S.; Noyori, R. Cancer Res. 1986, 46, 3538.
(2) (a) Stork, G.; Isobe, M. J. Am. Chem. Soc. 1975, 97, 4745. (b) Stork,
G.; Isobe, M. J. Am. Chem. Soc. 1976, 97, 6260. (c) Sugiura, S.; Hazato,
A.; Tanaka, T.; Okamura, N.; Bannai, K; Manabe, K.; Kurozumi, S.; Suzuki,
M.; Noyori, R. Chem. Pharm. Bull. 1985, 33, 4120. (d) Nakazawa, M.;
Sakamoto, Y.; Takahashi,T.; Tomooka, K.; Ishikawa, K.; Nakai, T.
Tetrahedron Lett. 1993, 34, 5923.
(
3) (a) Smith, A. B., III; Branca, S. J.; Philla, N. N.; Guarciaro, M. A.
The discovery of the [4 + 1] cycloaddition of diallenes with
CO resulted from the treatment of 1 with stoichiometric amounts
of Fe2(CO)9 to yield cycloaddition product 5 in less than 9 min.
The catalytic [4 + 1] cycloaddition was achieved for 1-4 by
treatment with Fe(CO)5 (10 mol%, 17 mM CO) and gave highly
stereoselective product formation and good yields (72-81%)
of 5-8 (eq 1). Based on the previously published kinetic
analysis (Table 1), differences in reactivity of the diallene
J. Org. Chem. 1982, 47, 1855. (b) Corey, E. J.; Mehrota, M. M. J. Am.
Chem. Soc. 1984, 106, 3384. (c) Reference 2c. (d) Iwasaki, G.; Sano, M.;
Sodeoka, M.; Yoshida, K.; Shigasaki, M. I J. Org. Chem. 1988, 33, 4864.
(
1
e) Iguchi, K.; Kaneta, S.; Nagaoka, H.; Yamada, Y. Chem. Lett. 1989,
60, 157.
4) For recent transition-metal mediated [3 + 2] cycloadditions, see: (a)
Danheiser, R. L.; Dixon, B. R.; Gleason, R. W. J. Org. Chem. 1992, 57,
(
6
1
1
1
094. (b) Trost, B. M.; Grese, T. A.; Chan, D. M. T. J. Am. Chem. Soc.
991, 113, 7350. (c) Trost, B. M.; Matelich, M. C. J. Am. Chem. Soc. 1991,
13, 9007. (d) Boivin, J.; Tailhan, C.; Zard, S. Z. J. Am. Chem. Soc. 1991,
13, 5874. (e) Rathjen, H. J.; Margaretha, P.; Wolff, S.; Agosta, W. C. J.
Am. Chem. Soc. 1991, 113, 3904. (f) Trost, B. M. Angew. Chem., Int. Ed.
Engl. 1986, 25, 1. (g) Danheiser, R. L.; Carini, D. J.; Fink, D. M.; Basak,
A. Tetrahedron 1983, 39, 935. (h) Noyori, R. Acc. Chem. Res. 1979, 12,
(6) Eaton, B. E.; Rollman, B.; Kaduk, J. A. J. Am. Chem. Soc. 1992,
114, 6245.
(7) (a) Desiraju, G. R.; Bernstein, J.; Kishan, K. V. Radha; S.; Jagarlapudi
A. R. P. Tetrahedron Lett. 1989, 30, 3029. (b) Ten Hoeve, W.; Wynberg,
H. J. Org. Chem. 1980, 45, 2930. (c) Heller, H. G.; Piggott, R. D. J. Chem.
Soc., Perkin Trans. 1 1978, 989.
6
1.
(
5) For recent Pauson-Khand cycloadditions, see: (a) Jeong, N.; Hwang,
S. H.; Lee, Y.; Chung, Y. K. J. Am. Chem. Soc. 1994, 116, 3159. (b)
Pearson, A. J.; Shively, R. J. Jr.; Dubbert, R. A. Organometallics 1992,
(8) (a) Alper, H. In Organic Synthesis Via Metal Carbonyls; Wender,
P., Ed.; Wiley: New York, 1977; Vol. 2, pp 545-593. (b) Pearson, A. J.;
Perosa, A. Organometallics 1995, 14, 5178. (c) Pearson, A. J.; Dubbert, R.
A. Organometallics 1994, 13, 1656. (d) Fatiadi, A. J. J. Res. Natl. Stand.
Technol. 1990, 96, 1.
1
1, 4096. (c) Schore, N. E. In ComprehensiVe Organic Syntheis; Trost, B.
M., Ed.; Pergamon: Oxford, 1991; Vol. 5, p 1037. (d) Schore, N. E. Org.
React. (N.Y.) 1991, 41, 1. (e) Schore, N. E. Chem. ReV. 1988, 88, 1081.
S0002-7863(96)02908-3 CCC: $12.00 © 1996 American Chemical Society

11784 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Sigman and Eaton
Table 1. Relative Rate Dependence of Fe(CO)
Cycloaddition at 50 °C in THF
5
-Catalyzed [4 + 1]
obs, s-
1
k
rel
[1], mM
[Fe(CO)
5
], mM
[CO], mM
k
-
-
-
-
4
4
4
5
2
90
29
29
10
29
17
17
17
64
9.6 × 10
1.8 × 10
3.9 × 10
7.8 × 10
1
7
0
0.19
0.41
0.08
2
2
90
90
diastereomers, and concomitant alkylidene stereochemistry of
the cyclopentenone products, a mechanism was proposed as
depicted in Scheme 1.
The first order dependence of the observed rate on the
concentration of 1 and Fe(CO)5 and inverse dependence on CO
concentration is consistent with rate determining association of
Figure 1. Effect of solvent on the [4 + 1] cycloaddition of 1 with
CO.
1
and concomitant loss of carbon monoxide from Fe(CO)5 to
form A. The reactivity of 3 was observed to be significantly
greater than 4, apparently a consequence of the diallene
symmetry. Meso diallenes 2 and 3 have an s-cis conformation
distribution were of interest. The experiments described herein
were designed to address these unanswered questions.
(
π-faces are different) where the steric hinderance on coordina-
tion of the metal can be minimized. The diallene diastereomer
has equivalent π-faces and C2 symmetry for the s-cis
Results and Discussion
4
Solvent Effects. From the previous results it was not clear
if coordinating solvents assist the coordination of the diallene
to the iron or impede the reaction progress by competitively
binding the iron. Previously reported kinetic experiments were
only performed in THF-d8. To determine the effect of solvent
on the rate of the [4 + 1] cycloaddition of diallenes, the relative
rates were measured in THF-d8, C6D6, CD2Cl2, and CD3CN
conformation, necessitating coordination to a sterically hindered
π-face. Because the two diastereomers give different and
discrete products the cycloaddition must occur by a mechanism
whereby the relative stereochemistry of the diallene is transfered
to render the dialkylidenecyclopentenones stereospecifically.
4
This symmetry-dependent outcome is consistent with η -
coordination of the s-cis diallene and appears to rule out a
(Figure 1). The poorly coordinating solvent, CD2Cl2, and the
2
mechanism whereby facial selectivity is determined by η -
π-ligand solvent, C6D6, gave the fastest rates, while the σ-donor
solvent, THF-d8, gave slower rates. The strong σ-donor solvent,
CD3CN, gave significantly slower rates with only 2% conversion
after 60 h under standard reaction conditions. These results
are consistent with a mechanism where coordination of the
diallene to the iron is slowed by use of a competitive solvent
ligand, possibly involved in a nonproductive equilibrium. It
remains to be shown that iron solvent complexes are not
produced in the rate determining step in these coordinating
solvents.
coordination.
4
9
10
Cobalt η -bisketene and iron vinylallene complexes iso-
electronic to A have been isolated and characterized by X-ray
crystallography. Conversion of A to the metallacyclopentene
1
1
B may be proposed in analogy to the isolobal cobalt,
1
2
13
molybdenum, and zirconium diene complexes. Due to the
highly stereoselective cycloaddition of 4, the mechanism
proposed in Scheme 1 requires synchronous disrotatory move-
ment of both distal diallene carbons. Following a CO insertion
step, reductive elimination14 of C could form 8 and regenerate
the catalyst.
Selectivity of the [4 + 1] Cycloaddition. The highly
stereoselective cycloaddition for diallenes containing two tert-
butyls (or phenyls) made it important to test a diallene containing
only one sterically demanding substituent. As discussed in the
Many unanswered questions remain concerning how this iron
catalyzed reaction occurs under such mild conditions. For
4
example, a dissociative reaction to form the Fe(CO)3-η -
4
introduction, the iron putatively binds the diallene by η -
butadiene complex from Fe(CO)5, a complex isoelectronic to
A, required prolonged heating at 140 °C.15 Clearly the diallene
substrate plays an important role in the iron-mediated [4 + 1]
cycloaddition. Specifically, the preliminary data suggested an
unprecedented associative mechanism as the rate determining
step, which needed to be explored in greater detail. In addition,
the effect of solvent and nonsymmetric diallenes on the product
coordination, and, thereby, 4 gives only one product that results
from synchronous disrotatory movement of both terminal
diallene carbons. It was unclear whether the stereoselectivity
of alkylidene bond formation would be observed for a non-
symmetric diallene that did not contain two bulky groups on
the diallene termini. The possibility existed that in the absence
of two bulky substituents to dictate the π-face of iron coordina-
tion that alternative fluctional processes could yield a mixture
of bisalkylidene-cyclopentenone products.
(
9) Jewell, C. F., Jr.; Liebeskind, L. S.; Williamson, M. J. Am. Chem.
Soc. 1985, 107, 6715.
10) (a) Alcock, N. W.; Richards, D. J.; Thomas, S. E. Organometallics
(
1
1
991, 10, 231. (b) Kerr, C. E.; Eaton, B. E.; Kaduk, J. A. Organometallics
995, 14, 269.
(
11) Eaton, B. E.; King, J. A., Jr.; Vollhardt, K. P. C. J. Am. Chem. Soc.
1
986, 108, 1359.
(
(
12) Faller, J. W.; Rosen, A. M. J. Am. Chem. Soc. 1977, 99, 4858.
13) (a) Benn, R.; Schroth, G. J. Organomet. Chem. 1982, 228, 71. (b)
Yasuda, H.; Tatsumi, K.; Nakamura, A. Acc. Chem. Res. 1985, 18, 120.
c) Erker, G.; Engel, K.; Kruger, C.; Tsay, Y.-H.; Samuel, E.; Vogel, P. Z.
Naturforsch. B 1985, 40B, 150.
14) (a) Negishi, E.-i.; Cedarbaum, F. E.; Takahashi, T. Tetrahedron Lett.
(
(
1
986, 27, 2829. (b) Grubbs, R. H.; Miyashita, A. J. Organomet. Chem.
978, 161, 371.
1
(
15) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
Compound 12 was prepared to test the effect of a nonsym-
metric diallene in the [4 + 1] cycloaddition. Bromoallene 9
and Applications of Organotransition Metal Chemistry; University Sci-
ence: Mill Valley, 1987; p 848.

Fe-Mediated [4 + 1] Cycloaddition of Diallenes with CO
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11785
was prepared by methods previously reported.16 Treatment of
9
1
with granular zinc afforded 10 which was treated in situ with
1 in the presence of a catalytic amount of Pd(PPh3)4 (1 mol%)
17
to give the conjugated diallene 12 (eq 2). It should be noted
that in contrast to the symmetric diallenes studied previously
that 12 is formed as a racemic mixture and that the cycloaddition
experiment may be viewed as probing the ability of the iron to
coordinate preferentially to the enantiotopic re or si faces of
1
2.
Figure 2. Kinetics of the [4 + 1] cycloaddition of 1 with 13CO in
THF-d .
8
The diallene was treated with 10 mol% of Fe(CO)5 and
allowed to react at 50 °C (eq 3). The reaction was quenched
1
with silica gel/triethylamine. Analysis by H NMR showed only
2
,5-dialkylidenecyclopentenone products 13 and 14 together
with starting diallene 12. Ratios of 13:14 were calculated to
1
be 86:14 by H NMR. GCMS analysis of the same mixture
gave an identical product ratio. With only one sterically
demanding substituent on the terminal diallene carbons, Fe(CO)5
exhibits good π-facial selectivity. The diastereoselectivity is
similar to that observed in the photochemical Fe(CO)5 catalyzed
1
8
[
4 + 1] cycloadditions of allenyl imines, allenyl ketones, and
19
aldehydes. Allenyl ketones were recently shown to undergo
thermal cycloaddition with CO in 2-methyl-THF subsequent to
photochemical activation of Fe(CO)5 to give Fe(CO)4S (S )
Figure 3. Activation parameters of the Fe(CO)
cycloaddition of 1 with CO.
5
catalyzed [4 + 1]
20
2
-methyl-THF). However, the diallene cycloaddition reaction
proceeds by a purely thermal mechanism and the stereoselec-
tivity of alkylidene bond formation may be only superficially
related.
to the solvent internal standard, which is consistent with
1
3
incorporation of CO into the Fe(CO)5. No change was
1
3
observed for the C NMR CO resonance of Fe(CO)5 without
the diallene substrate under identical conditions.
Fe(CO)5 Carbon Monoxide Exchange Studies. The kinetic
data of Table 1 suggests that Fe(CO)5 associatively equilibrates
with A. Scheme 1 requires that the diallene catalyze CO
exchange via A converting to Fe(CO)5, which increases with
CO pressure. Because Fe(CO)5 is the resting state of the
catalyst, a rate determining equilibrium could be ruled out if
the diallene could not be shown to give any enhancement of
the CO exchange in the iron carbonyl. It is important to note
that the thermal exchange of CO from a pure sample of Fe(CO)5
IR experiments in THF at time zero showed bands assigned
-
1
to the CO stretching frequencies of Fe(CO)5 at 1993 cm (E)
-
1
and 2018 cm (A2). After cycloaddition completion a new IR
-
1
band was observed at 1967 cm (with unresolved shoulders)
which is consistent with the expected isotopic shift for iron
1
3
pentacarbonyl containing CO. In addition, mass spectral
analysis of the resulting cycloaddition product 5 showed two
21
13
is very slow (T1/2 > 4 yr).
isotopic peaks for the molecular ions with an increase in
C
It was of interest to determine if the diallene promoted
exchange of CO in Fe(CO)5. An experiment was performed
under standard conditions for [4 + 1] cycloaddition using
abundance of ∼30%. These results are consistent with a
diallene promoted exchange of CO in Fe(CO)5. The rate
1
3
enhancement by 1 observed for CO exchange in Fe(CO)5
cannot be accurately determined from these experiments because
rates of diffusion of CO in the absence of mixing could be rate
limiting.
1
3
diallene 1 (290 mM), Fe(CO)5 (29 mM, 10 mol%), and CO
1
3
(
99.3% C) at 64 mM in THF-d8. The Fe(CO)5 was carefully
purified by repeated bulb-to-bulb distillation. Carbon monoxide
exchange in Fe(CO)5 was monitored in separate experiments
Activation Parameters. All of the data thus far supports
rate determining association of the diallene to form A with
concomitant loss of 2 CO’s from Fe(CO)5. The production of
iron carbonyl complexes usually proceeds by either high
temperature or photochemical-promoted dissociation of one or
by C NMR and IR. At time zero, the 13C NMR resonance
for Fe(CO)5 was observed at 211 ppm. The NMR tube was
then placed in a constant temperature bath at 50 °C, and several
13
data points were obtained (Figure 2). After reaction completion
1
(
96% conversion measured by H NMR), the intensity of the
8a
two carbon monoxide ligands. In order to probe the transition
13
Fe(CO)5 peak in the C NMR increased by 1.6 times relative
state of this unusual reaction for Fe(CO)5, the activation
parameters were determined in C6D6 using a temperature range
from 30 °C to 80 °C (Figure 3). Kinetic data were collected in
triplicate to at least 80% conversion for five temperatures.
(
16) (a) Landor, S. L.; Patel, A. N.; Whiter, P. F.; Greaves, P. M.
Tetrahedron Lett. 1963, 483 and J. Chem. Soc. 1966, 1223. (b) Schiavelli,
M. D.; Germroth, T. C.; Stubbs, J. W. J. Org. Chem. 1976, 41, 681.
(17) Ruitenberg, K.; Kleijn, H.; Elsevier, C. J.; Meijer, J.; Vermee, P.
q
Using Eyring transition state theory, ∆H was calculated to be
Tetrahedron Lett. 1981, 22, 1451.
-
1
q
1
6.2 (( 0.7) kcal mol , and ∆S was calculated to be -17.9
q
(
(
18) Sigman, M. S.; Eaton, B. E. J. Org. Chem. 1994, 59, 7488.
19) Sigman, M. S.; Kerr, C. E.; Eaton, B. E. J. Am. Chem. Soc. 1993,
-
1
-1
(( 2.3) cal mol K . The large negative value for ∆S is
consistent with an ordered transition state as would be expected
for an associative coordination of diallene to Fe(CO)5 as depicted
in Scheme 1.
1
15, 7545.
20) Sigman, M. S.; Eaton, B. E.; Heise, J. D.; Kubiak, C. P. Organo-
metallics 1996, 15, 2829.
21) Keeley, D. F.; Johnson, R. E. J. Inorg. Nucl. Chem. 1959, 11, 33.
(
(

11786 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Sigman and Eaton
Scheme 1
Figure 4. Relative rates for the iron-mediated cycloaddition of 1 and
2
2.
Scheme 3
Scheme 2
Isotopically labeled diallenes 21 and 22 were synthesized as
shown in Scheme 2. Deprotonation of the acetylide proton of
2
4
1
5 with CH3CH2MgBr and quenching with D2O gave 16. The
16
allenyl bromide 17 was made by standard conditions with no
loss of deuterium incorporation. Allenyl zinc reagent 19 was
prepared in situ by treatment of 18 with ZnCl2. Deuterio-
diallenes were assembled by a palladium catalyzed coupling of
allenyl zinc reagent 19 with either bromoallene 17 or 20 with
From Benson22 table calculations it is clear that diallenes are
high energy molecules relative to other conjugated systems such
as dienes. It might be expected that the Hammond postulate
would be manifested in the coordination of the diallene to the
q
iron and that the low value of ∆H determined is consistent
17
no loss of deuterium incorporation.
with significant stabilization in 1 proceeding to A.23 However,
it may be expected that such stabilization in forming A would
be accompanied by a significant change in the hybridization of
the diallene π-system on coordination to the iron. It remained
unclear to what degree the diallene π-system had been affected
at the transition state in coordination to the iron.
Under standard conditions for cycloaddition (290 mM di-
allene, 29 mM Fe(CO)5), the reaction was monitored for >3
half-lives but showed no observable difference in rates between
diallene 1 and 22 (Figure 4). With the more sensitive diallene
2
1 there was also no rate difference measured. These data are
consistent with either of three possibilities; first, only minimal
Secondary Isotope Effects. The activation parameters and
2
changes in hybridization at the internal sp diallene carbons are
4
kinetic analysis indicate that formation of the η -complex A is
occurring in the transition state leading to A, second, diallene
coordination proceeds directly to B, where the internal diallene
carbons are not significantly affected, and third these results
do not rule out the possibility of rate determining coordination
of the iron to the terminal π-bonds of the diallene followed by
rapid isomerization to the internal π-bonds.
The results of this investigation have failed to rule out the
mechanism proposed previously. New insight into the mech-
anism of iron carbonyl catalyzed [4 + 1] cycloaddition of
diallenes with carbon monoxide has been revealed, however,
and Scheme 3 depicts the additional features of this chemistry
that we now propose. Consistent with the previously reported
kinetic results and the accelerated carbon monoxide exchange
in Fe(CO)5 facilitated by the diallene reported here, it now seems
reasonable to propose an equilibrium in the rate determining
rate determining. X-ray analysis of a Fe(CO)3-vinylallene
complex electronically related to intermediate A showed very
2
little change in hybridization of the internal allenic sp
carbons.1 It was appreciated at the outset of designing this
experiment that the secondary isotope effects would only be
observable for large changes in hybridization of the internal
diallene carbons at the transition state and that if the transition
state resembled A it was unlikely to give significantly different
rates for the deuteriodiallenes. Dideuteriodiallenes appropriately
0b
2
labeled would be expected to be most sensitive ({kD/kH} ) in
4
this regard when compared to 1 if rate determining η -
coordination was occurring and resulted in changes in hybrid-
ization of the π-system.
(
22) Benson, S. W.; Cruickshank, F. R.; Golden, D. M.; Haugen, G. R.;
O’Neal, H. E.; Rodgers, A. S.; Shaw, R.; Walsh, R. Chem. ReV. 1969, 69,
79.
23) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334.
2
(24) Landor, S. R.; Demetriou, B.; Evans, R. J.; Grzeskowiak, R.; Pavey,
P. J. Chem. Soc., Perkin II 1972, 1995.
(

Fe-Mediated [4 + 1] Cycloaddition of Diallenes with CO
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11787
by a literature method.26 THF-d
and benzene-d (Aldrich Chemical
Scheme 4
8
6
Co.) were distilled from benzophenone/potassium. Acetonitrile-d and
methylene chloride-d (Aldrich Chemical Co.) were distilled form CaH
Isotopic 13CO (99.3% C) was purchased from Isotec Inc. Fe(CO)
3
2
2
.
5
13
(
Strem Chemicals Inc.) was vacuum distilled and stored at -30 °C in
the dark.
Synthesis of 12. In a flame dried 50 mL flask, zinc (547 mg, 7.8
mmol) and 10 mL of freshly distilled THF (K/benzophenone) were
combined. To this flask, 1-bromo-3,4,4-trimethylpentadi-2,3-ene (9)
(1 mL, 6.0 mmol) was added and allowed to stir at ambient temperature
for 3 h until a gray color persisted. The gray solution was then added
via canula to a mixture of tetrakistriphenylphosphinepalladium (69 mg,
1
mol%) and 2-(2-methyl-3-butynyl)acetate (11) (795 mg, 6.3 mmol)
in 20 mL of THF. After 30 min, the reaction was quenched with MeOH
1 mL) and silica gel (1 g). The solvent was removed, and the resulting
(
residue was washed with 100 mL of pentane through a short pad of
silica. Column chromatography of the resulting oil followed by vacuum
distillation (35 °C, 1 mm) yielded 572 mg (54%) of 12 as a colorless
1
oil. H (300 MHz, C
1
6
D
6
) δ 1.01 (s, 9H), 1.56 (d, J ) 2.7 Hz, 6H),
.62 (d, J ) 2.6 Hz, 3H), 5.70 (dsep, J ) 2.7, 10.0 Hz, 1H), 5.78 (dq,
step, where the back reaction (Scheme 3) of A′ to Fe(CO)5 and
13 1
J ) 2.6, 10.0 Hz, 1H); C{ H} (75 MHz, C
D
6 6
) δ 15.10, 20.62, 29.23,
Cl ) 1960
20 calcd 176.1565, found 176.1542.
Cycloaddition of 12. To a low pressure 5 mm NMR tube were
combined 12 (48 mg, 0.284 mmol) and Fe(CO) (4 µL, 10 mol%) in
, and the tube charged with 10 psi CO. The reaction was allowed
1
is competitive with cycloaddition. The solvent effects on the
3
3.97, 89.69, 89.83, 91.16, 96.62, 203.01, 204.14; IR (CH
2
2
reaction appear to correlate with the σ-donor ability of the
solvent and may be explained by a nonproductive equilibrium
that diminishes the steady-state concentration of A′. No solvent
-
1
cm ; HRMS m/z for C13
H
5
1
complexes ML could be detected by H NMR analysis in the
6 6
C D
solvent that resulted in the slowest rate of cycloaddition (CD3CN
to heat for 60 h at 50 °C, and NMR analysis showed partial conversion
to product. The reaction was quenched with silica gel/triethylamine,
and the solvent was removed by a rotary evaporator. The crude product
9
9.5% D).
Cycloaddition with the nonsymmetric diallene 12 gave a
was washed with CH
evaporator, and the residue was loaded onto a prep-silica gel plate.
The plate was eluted with 15% CH Cl /hexane, and two small bands
2 2
Cl , the solvent was removed by a rotary
significant decrease in the stereoselectivity of alkylidene bond
formation as compared to its symmetric counterpart 2, which
may be attributed to either decreased π-facial selectivity in
2
2
were isolated together. NMR analysis of the bands showed a mixture
of products with a 86/14 isomeric ratio of E/Z. GCMS of the same
sample gave identical results for the isomeric ratio and confirmation
of the correct molecular weight.
4
forming A′ or isomerization between the two possible η -
diallene iron diastereomers A′ and A′′ (Scheme 4). The rapid
2
isomerization of η -allene-iron complexes has been reported
25
previously. Bulky groups on the diallene termini may impede
the formation of complexes related to D and thereby slow the
rate of equilibration between diastereomers.
1
1
3. 43% H (300 MHz, C
6 6
D ) δ 1.04 (s, 9H), 1.53 (s, 3H), 2.30 (s,
3H), 2.47 (s, 3H), 6.45 (d, J ) 7.3 Hz, 1H), 6.88 (d, J ) 7.3 Hz, 1H);
1
3
1
C{ H} (75 MHz, C
6
D
6
) δ 16.03, 20.11, 22.98, 30.35, 38.15, 127.77,
Cl ) 1693
20O calcd 204.1514, found 204.1512.
) δ 1.43 (s, 9H), 1.57 (s, 3H), 1.76 (s,
1
29.79, 132.92, 133.32, 141.93, 155.52, 195.72; IR (CH
2
2
Conclusions. It is now clear that conjugated diallenes
undergo stereoselective [4 + 1] cycloaddition in a variety of
solvents. Coordinating solvents slowed the rates of cyclo-
addition, probably due to competitive binding to the metal
center. No CO exchange in Fe(CO)5 is observed in the absence
of diallene under standard conditions for cycloaddition. Con-
sistent with rate determining associative coordination of diallene
to form complex A, a large negative entropy of activation was
determined. The synthetic utility of these transformations can
now be expanded to include diallenes with functional groups
that may only weakly coordinate to the iron. Only one sterically
demanding substituent on the diallene termini is necessary to
give good π-facial selectivity, which also expands the synthetic
utility of these transformations. The application of [4 + 1]
cycloaddition to the synthesis of 2,5-dialkylidenecyclopenten-
ones with biological activity and new material science applica-
tions is currently in progress.
-
1
cm ; HRMS m/z for C14
H
1
1
4. 8% H (300 MHz, C
D
6 6
3
H), 2.31 (s, 3H), 6.53 (d, J ) 7.2 Hz, 1H), 6.62 (d, J ) 7.2 Hz, 1H);
1
3
1
6 6
C{ H} (75 MHz, C D ) δ 20.08, 20.10, 23.01, 28.95, 37.24, 127.88,
1
2 2
31.19, 134.49, 135.56, 142.02, 155.80, 192.65; IR (CH Cl ) 1697
-1
cm ; MS m/z 204.
16. In a three necked 500 mL flask equipped with a reflux condenser
and an addition funnel, 80 mL of a 3 M (238 mmol) solution of ethyl
magnesium bromide in diethyl ether and 200 mL of freshly distilled
diethyl ether (NaK/benzophenone) were combined at room temperature.
The addition funnel was charged with 2-methyl-3-butyn-2-ol (15) (5.0
g, 59 mmol) and 20 mL of diethyl ether. This solution was added
slowly with vigorous stirring, the resulting mixture was refluxed for
9
0 min, and the solution was then cooled to ambient temperature. The
reaction was quenched by slow addition of a solution of D
O (99.96%,
59 mmol, 10 mL in 70 mL of THF). The resulting mixture was
2
5
vigorously stirred for 12 h and followed by addition of 50% HCl (10
mL). The salts were filtered and washed with diethyl ether (3×, 30
mL), and the resulting solution was washed with H
2
O (4×, 50 mL)
Experimental Section
and brine (1×, 50 mL) and dried over MgSO
4
. After filtration, the
diethyl ether was removed by a rotary evaporator, and the crude oil
General Procedures. All reactions and manipulations were con-
ducted under a dry argon atmosphere either using an inert atmosphere
glovebox or standard Schlenk techniques. All NMR data were recorded
was fractionally distilled (97 °C, uncorrected) to yield 2.73 g of a
1
colorless oil (54%) with >95% deuterium incorporation by NMR. H
[300 MHz, CDCl ] δ 2.40 (s, 0.05H), 2.08 (s, 1H), 1.51 (s, 6H).
3
1
on a Bruker AMX (300 MHz H). NMR splitting patterns are reported
1
7. In a 25 mL round bottom flask, a mixture of cuprous bromide
as follows: s ) singlet, d ) doublet, q ) quartet, h ) heptet, sep )
septet. IR were recorded on a Perkin Elmer 1600 FTIR. Mass spectral
data were obtained from the departmental facility at Washington State
University. Melting points were recorded on a Mel-Temp apparatus
and are uncorrected. Tetrakistriphenylphosphinepalladium was prepared
(
1.39 g, 9.67 mmol), ammonium bromide (1.10 g, 11.25 mmol), copper
powder (135 mg, 2.12 mmol) and concentrated hydrobromic acid (48%
w/w, 6.6 mL, 58 mmol) were sparged with argon for 10 min.
3
-Deuterio-2-methyl-3-butyn-2-ol (16) (2.25 g, 26.48 mmol) was added
to the mixture and allowed to stir for 30 min at room temperature. The
(25) Ben-Shoshan, R.; Petit, R. J. Am. Chem. Soc. 1967, 89, 2231.
(26) Coulson, D. R. Inorg. Syn. 1972, 13, 121.

11788 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Sigman and Eaton
1
mixture was put into a separatory funnel where the aqueous layer was
extracted with pentane (4×, 5 mL). The pentane extracts were
combined, filtered, washed with HBr (3×, 4 mL), washed with saturated
at 50 °C. The reaction was monitored by H and 13C NMR until 96%
conversion. The tube was then freeze-pumped-thawed four cycles
to remove the CO, and samples were prepared for FTIR and MS.
Solvent Effect Kinetics. To a silanized low pressure NMR tube
equipped with a J. Young valve, 1 (20 mg, 0.149 mM), 510 µL of
NaHCO
3
(2×, 5 mL) and dried with MgSO
4
. The pentane was removed
by a rotary evaporator and the resulting oil was vacuumed distilled
25 °C, 20 mmHg) to yield 2.59 g of a colorless oil (66%) with >93%
(
5
solvent, and Fe(CO) (1.9 µL, 10 mol%) were combined. The tube
was charged with 10 psi CO (16 mM) and placed into a 50 °C constant
temperature bath and monitored by H NMR using the residual solvent
resonance as an internal standard. The tube was repressurized with
CO after collecting each data point.
1
deuterium incorporation by NMR. H (300 MHz, CDCl
2 Hz, 0.07 H), 1.82 (s, 6H).
1. In a 100 mL round bottom flask 2.5 M n-butyllithium (1.5 mL,
.74 mmol) was slowly added to a solution of 1-bromo-1-deuterio-3-
3
) 5.80 (h, J
1
)
2
3
methylbuta-1,2-diene (553 mg, 3.74 mmol) in 35 mL of freshly distilled
diethyl ether (K/benzophenone) at -78 °C. The solution was allowed
to stir for 2 h at -78 °C and then slowly added via canula into a cooled
flask containing zinc chloride (510 mg, 3.74 mmol) and 5 mL of diethyl
ether. The resulting mixture was allowed to warm to ambient
temperature over 90 min where the solution turned gray. After stirring
for 30 min at ambient temperature, the mixture was added via canula
into a 250 mL flask containing tetrakistriphenylphosphinepalladium (22
mg, 18.7 µmol), 1-bromo-1-deuterio-3-methylbuta-1,2-diene (17) (570
mg, 3.85 mmol), and 100 mL of THF. The mixture was allowed to
stir for 30 min at ambient temperature, and the solvent was then
removed by a rotary evaporator. The residue was placed on a silica
gel pad (30 mL, 20 g), the pad was eluted with pentane (125 mL), and
the solvent was removed by a rotary evaporator. Column chromatog-
Activation Parameters. For each temperature, the reaction was
run in triplicate. To three silanized low pressure NMR tubes equipped
with J. Young valves was added 500 µL of a solution containing 290
mM of 1 (290 mM) and Fe(CO)
5 6 6
(29 mM) in C D . The tubes were
pressurized with 10 psi CO (16 mM) and placed into a constant
1
temperature bath. The reactions were monitored by H NMR until at
least 80% conversion using the residual benzene resonance as an internal
standard. The tube was repressurized with CO after collecting each
data point.
Secondary Isotope Effects. To silanized NMR tubes were added
5 8
Fe(CO) (29 mM) and diallene (1, 21, or 22, 290 mM) in THF-d . The
tubes were pressurized with 16 mM CO and placed into a constant
1
temperature bath at 50 °C. The reactions were monitored by H NMR
until at least 80% conversion using the residual THF resonance as an
internal standard. The tubes were repressurized with CO after collecting
each data point.
raphy on flash silica gel (hexane) yielded a white solid 234 mg (46%).
1
3
H [300 MHz, CDCl ] δ 5.42 (s, 0.14H), 1.69 (s, 12H); MS m/z 136.
2
2. Identical to method used for 21 except coupled with 20 40%
1
yield of a white solid. H (300 MHz, CDCl
s, 12H). MS m/z 135.
CO Exchange Study. To a silanized Wilmad 522-PP 5 mm high
pressure NMR tube, a solution of 290 mM 1 and 29 mM Fe(CO) in
THF-d was added. The tube was freeze-pumped-thawed three cycles
3
) δ 5.42 (s, 1.07H), 1.69
Acknowledgment. Support from the Donors of The Petro-
leum Research Fund and a Grant-in -Aid from Washington State
University is gratefully appreciated. NMR data were obtained
from the WSU NMR Center supported in part by the NIH
(
5
8
(RR0631401) and NSF (CHE-9115282).
1
3
13
and then charged with 40 psi CO (64 mM). An initial C NMR was
taken, and the tube was placed into a constant temperature water bath
JA962908O