Some allylic substituent effects in ring-closing metathesis reactions: allylic alcohol activation.

Article abstract of DOI:10.1021/ol990947+

[formula: see text] Dienes 2a-e were used to study allylic substituent effects in the ring-closing metathesis reaction (RCM). Both the steric and electronic character of the allylic substituents were found to influence alkene reactivities. Free allylic hydroxyl groups exert a large activating effect on the RCM reaction rates.

Full text of DOI:10.1021/ol990947+

ORGANIC
LETTERS
1
999
Vol. 1, No. 7
123-1125
Some Allylic Substituent Effects in
Ring-Closing Metathesis Reactions:
Allylic Alcohol Activation
1
Thomas R. Hoye* and Hongyu Zhao
Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
hoye@chem.umn.edu
Received August 13, 1999
ABSTRACT
Dienes 2a−e were used to study allylic substituent effects in the ring-closing metathesis reaction (RCM). Both the steric and electronic character
of the allylic substituents were found to influence alkene reactivities. Free allylic hydroxyl groups exert a large activating effect on the RCM
reaction rates.
The power of ring-closing metathesis reactions of dienes is
1
indisputable. Systematic studies that add definition to the
type of substrates that tolerate the cyclization conditions
provide valuable information that can be used in planning
the application of RCM reactions to more complex systems.
Ulman and Grubbs reported a very instructive and funda-
mental kinetic study in which they determined the relative
The first substrate we examined, linalool (2b), underwent
RCM unexpectedly quickly at room temperature using just
5 mol % of 1 to give 1-methylcyclopent-2-en-1-ol (3b) very
rates of initial reaction of a series of simple hydro-
carbons with the Grubbs ruthenium benzylidene carbene,
2
PhHCdRu(PCy
3
)
2
Cl
2
(1). These alkenes differed in the
4
efficiently. This was surprising for two reasons. It is known
degree of substitution on the double bond and at the allylic
and homoallylic centers.
that tert-butylethylene, containing a fully substituted allylic
2
center, is nearly inert to reaction with 1, and the ring-closing
We report here the ring-closing metathesis reactions of a
series of substrates 2, which differ in the nature and number
of allylic substituents. The skeleton, which bears one mono-
and one trisubstituted alkene, was chosen so that the initial
metathesis event would occur only at the less substituted
event involves reaction with a trisubstituted alkene, of which
5
there are few reported examples. We took advantage of the
opportunity that this diene framework provided to probe the
effect of various allylic substituents.
2
The substrates 2a-e were prepared, where necessary, by
alkene. The allylic carbon in these five substrates bears an
6
3
conventional routes. Various pairs were then subjected to
array of methyl, hydroxy, and/or methoxy substituents.
(
1) (a) Grubbs R. H.; Chang S Tetrahedron 1998, 54, 4413-4450. (b)
Schuster M.; Blechert S. Angew. Chem., Int. Ed. Engl. 1997, 36, 2037-
056.
2) Ulman, M.; Grubbs, R. H. Organometallics 1998, 17, 2484-2489.
(3) For a study of the effect of various substituents (including hydroxy-
methyl) on the ring-opening polymerization by 1 of a series of 3-substituted
cyclobutenes, see: Maughon, B. R.; Grubbs, R. H. Macromolecules 1997,
30, 3459-3469.
2
(
1
0.1021/ol990947+ CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/14/1999

competition experiments. Equimolar mixtures of, e.g., 2c and
d were treated with 4-10 mol % of 1 in CDCl at room
competition with 2b; indeed, the ether was inert even when
) with 1. Thus, the completely
substituted allylic carbon in 2e is sufficient to inhibit
alkylidene formation. The lack of reaction within 2e also
supports the assertion that all RCM events in 2a-2d are
initiated at the monosubstituted alkene.
2
3
heated (∼65 °C, CDCl
3
temperature. Reaction progress was monitored directly by
1
H NMR spectroscopy. In no case were any intermediate
alkylidene species detected, implying that the rate-limiting
event in the reaction of each substrate was the metal
exchange reaction between an external ruthenium alkylidene
To rule out the possibility that the reactivity difference
between 2b and 2e was primarily steric in nature, we
competed linalool (2b) with citronellene (2c) in which the
(either benzylidene or isopropylidene) and the terminal vinyl
group in 2. The results of the various pairwise competition
experiments are indicated in Figure 1 and are summarized
by the approximate relative reactivity values listed to the
left of each structure.
7
latter clearly has a less hindered vinyl group. Nonethe-
less, the tertiary alcohol 2b reacted faster than 2c, again
confirming the hydroxyl activating effect. There are several
possible reasons for the large activating effect of the hydroxy
group. For example, rapid and reversible ligand exchange
at the ruthenium center of alkoxy for chloride [to give a
species such as (RO)Cl(Cy
phosphine [to give species such as either Cl
ROH)RudCHR′ or the anionic complex [Cl (Cy
PH] ] could promote reaction by preasso-
3
P)
2
RudCHR′] or of alcohol for
(Cy P)-
P)(RO)-
2
3
(
2
3
-
+
RudCHR′] [Cy
3
ciation. Similarly, hydrogen bonding between the hydroxy
group and one of the chloride ligands could be the event
that favors subsequent reaction between the alkene (in R)
and carbene centers.
The outcome of the previously mentioned competition
between 2c and 2d suggests that the inductive effect of the
allylic ether reduces the reaction rate of the adjacent, less
electron-rich alkene by nearly an order of magnitude.
Finally, we studied the RCM reaction of the secondary
8
alcohol 2a. As we recently noted in a different context, an
unexpected reaction pathway was uncovered. Namely, in
addition to the ring-closed cyclopent-2-en-1-ol (3a), the
methyl ketone 4 was formed in a 1:1.5 ratio (Scheme 1). It
Scheme 1
Figure 1. Results of RCM competition reactions between equi-
molar amounts (∼0.02 M in each diene) of the bracketed pairs of
substrates in CDCl at room temperature in the presence of 4-10
3
mol % of 1. #x represents the product ratio (determined by
1
integration of appropriate H NMR resonances) at the percent
conversion of the more reactive substrate (as indicated in paren-
theses). In experiments involving the secondary alcohol 2a, 90-
100 mol % of 1 was used since carbene species are consumed by
competitive methyl ketone formation (see text). There is a consider-
able error bar in the measurement marked b, since only a small
amount of the cyclopent-2-en-1-ol methyl ether (3d) was present
even at high conversion of 2a.
is clear that ruthenium carbene species are being consumed
by this side reaction, since use of only 10 mol % of 1 results
1
in low conversion of 2a and all H NMR resonances due to
8
One early competition experiment convincingly demon-
strated the fact that the allylic free hydroxyl group in linalool
1 disappear. We suggest that the initial carbene 5 undergoes
tautomerization to the enolyl ruthenium hydride species 6,
which can further undergo reductive elimination, either
before or after tautomerization to the oxoalkyl ruthenium
(2b) has a significant activating effect on the reaction rate.
Namely, linalool methyl ether (2e) was unreactive in
1124
Org. Lett., Vol. 1, No. 7, 1999

hydride 7, to produce the methyl ketone 4. As a result of
this consumption of ruthenium carbenes, we elected to use
hancement is sufficient to overcome significant steric
deactivation. The corresponding methyl ethers are reduced
in reactivity relative to sterically similar methylated sub-
strates. Secondary allylic alcohols are a liability in RCM
reactions because of a net fragmentation reaction that
consumes ruthenium alkylidene species.
90-100 mol % of 1 in our competition experiments
involving 2a. Nonetheless, it is clear that the unhindered
secondary allylic alcohol 2a is the most reactive RCM
substrate among all of 2a-e. However, the designed use of
2
substrates containing the RCH(OH)CHdCH substructure
in productive RCM reactions is risky since only the most
rapid ring closures would likely compete with the unwanted
net fragmentation reaction like that involved in the 2a to 4
transformation. One instance in which an RCM substrate
Acknowledgment. We thank the National Institutes of
Health (CA76497) for funding this investigation.
OL990947+
containing a secondary allylic alcohol [i.e., CH
2
dCHCH(OH)-
(
6) The only new compound described in this paper, ether 2d, gave
CH(R)CH CHdCH ] was successfully ring-closed (to a
2
2
1
13
satisfactory H NMR, C NMR, IR, and HRMS data.
(7) A representative plot of change in product ratios with time (and %
conversion) is shown here for the 2b/2c competition experiment.
cyclopentene derivative) has been reported, but it is unknown
at which alkene the RCM was initiated.9
In conclusion, allylic hydroxyl groups greatly accelerate
the rate of carbene-exchange reaction between the adjacent
vinyl group and external ruthenium alkylidenes. This en-
(4) Three RCM reactions involving tertiary allylic alcohols were reported
recently: (a) Holt, D. J.; Barker, W. D.; Jenkins, P. R.; Davies, D. L.;
Garratt, S.; Fawcett, J.; Russell, D. R.; Ghosh, S. Angew. Chem., Int. Ed.
Engl. 1998, 37, 3298-3300. (b) Sellier, O.; Van de Weghe, P.; Eustache,
J. Tetrahedron Lett. 1999, 40, 5859-5860. (c) Wood, J. L.; Holubec, A.
A.; Stoltz, B. M.; Weiss, M. M.; Dixon, J. A.; Doan, B. D.; Shamji, M. F.;
Chen, J. M.; Heffron, T. P. J. Am. Chem. Soc. 1999, 121, 6326-6327.
(
5) E.g., citronellene (2c) has been previously ring-closed with the Dupont
tungsten metathesis catalyst. Nugent, W. A.; Feldman, J.; Calabrese, J. J.
Am. Chem. Soc. 1995, 117, 8992-8998.
(8) Hoye, T. R.; Zhao, H. Org. Lett. 1999, 1, 169-171.
(9) Crimmins, M. T.; King, B. W.; J. Org. Chem. 1996, 61, 4192-4193.
Org. Lett., Vol. 1, No. 7, 1999
1125