Selected substituent effects on the rate and efficiency of formation of an eight-membered ring by RCM

Article abstract of DOI:10.1021/jo702726b

(Chemical Equation Presented) Studies of a range of reactions forming cyclooctenones highlight a discrepancy between cyclization rate and cyclization efficiency. Cyclization rates change modestly as the oxygen function at the allylic position is varied, a

Full text of DOI:10.1021/jo702726b

Selected Substituent Effects on the Rate and Efficiency of Formation
of an Eight-Membered Ring by RCM
Lisa Mitchell,†,‡ John A. Parkinson,‡ Jonathan M. Percy,*,‡ and Kuldip Singh†
Department of Chemistry, UniVersity of Leicester, UniVersity Road, Leicester LE1 7RH, United Kingdom,
and Department of Pure and Applied Chemistry, WestCHEM, UniVersity of Strathclyde, Thomas Graham
Building, 295 Cathedral Street, Glasgow G1 1XL, United Kingdom
jonathan.percy@strath.ac.uk
ReceiVed December 21, 2007
Studies of a range of reactions forming cyclooctenones highlight a discrepancy between cyclization rate
and cyclization efficiency. Cyclization rates change modestly as the oxygen function at the allylic position
is varied, and increase upon gem-dimethylation. Cyclization efficiency has also been quantified for four
substrates, revealing a range of effective molarities (EMs) of 2 orders of magnitude that are substituent
dependent. The most efficient cyclization appears to result from suppression of the cross-metathesis
pathway through which oligomerization begins, rather than from a particularly rapid cyclization reaction.
In the presence of a Ti(IV) cocatalyst, diene monomers transform smoothly to eight-membered-ring
products without the intermediacy of dimers or other oligomers, indicating that the cyclizations are
kinetically and not thermodynamically controlled. The gem-dialkyl effect is also shown to be kinetic.
Introduction
RCM is limited.6 Some detailed computational studies of very
simple prototypical reactions7 and of relatively complex cy-
clization systems8 have been carried out, but there seem to be
relatively few general computational insights concerning cy-
clization.
The ring closing metathesis (RCM) reaction has changed the
way we think about the synthesis of cyclic molecules, leading
to the award of the Nobel Prize to the three chemists most
responsible for the major strategic advances in the area.1 The
direct synthetic connection of an R,ω-diene to a cyclic alkene
or unsaturated heterocycle in the presence of numerous func-
tional groups of a wide range of types and at high density can
be achieved routinely by using commercially available and easy-
to-handle Ru catalysts. Total syntheses of many complex natural
products2 have been planned and executed by using the RCM
as a strategic event.3 The tolerance of the reaction to variations
in ring size is remarkable, with even rings that are usually
difficult to make, like medium rings,4 being formed in high yield
(if appropriate substitution patterns are present). Though Fu¨rst-
ner5 inter alia has outlined a number of extremely useful general
ideas that explain the success of RCM reactions, our detailed
knowledge of the interplay of structure and efficiency in the
(2) For recent examples, see: (a) Klar, U.; Buchmann, B.; Schwede,
W.; Skuballa, W.; Hoffinann, J.; Lichtner, R. B. Angew. Chem., Int. Ed.
2006, 45, 7942. (b) Hong, Z. Y.; Liu, L.; Hsu, C. C.; Wong, C. H. Angew.
Chem., Int. Ed. 2006, 45, 7417. (c) Gradillas, A.; Perez-Castells, J. Angew.
Chem., Int. Ed. 2006, 45, 6086. (d) Furstner, A.; Nevado, C.; Tremblay,
M.; Chevrier, C.; Teply, F.; Aissa, C.; Waser, M. Angew. Chem., Int. Ed.
2006, 45, 5837. (e) Deiters, A.; Pettersson, M.; Martin, S. F. J. Org. Chem.
2006, 71, 6547. (f) Inoue, M.; Sato, T.; Hirama, M. Angew. Chem., Int.
Ed. 2006, 45, 4843. (g) Hoye, T. R.; Eklov, B. M.; Jeon, J.; Khoroosi, M.
Org. Lett. 2006, 8, 3383. (h) Hong, S. W.; Yang, J. H.; Weinreb, S. M. J.
Org. Chem. 2006, 71, 2078. (i) Bohrsch, V.; Neidhofer, J.; Blechert, S.
Angew. Chem., Int. Ed. 2006, 45, 1302.
(3) For an elegant and comprehensive recent review, see: Nicolaou, K.
C.; Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4490.
(4) (a) Maier, M. E. Angew. Chem., Int. Ed. 2000, 39, 2073. (b) Deiters,
A.; Martin, S. F. Chem. ReV. 2004, 104, 2199.
(5) Furstner, A. Angew. Chem., Int. Ed. 2000, 39, 3013.
(6) (a) For an excellent overview, see: Conrad, J. C.; Fogg, D. E. Curr.
Org. Chem. 2006, 10, 185. Enyne metathesis has received much more
detailed scrutiny; see: (b) Villar, H.; Frings, M.; Bolm, C. Chem. Soc. ReV.
2007, 36, 55. (c) Lloyd-Jones, G. C.; Margue, R. G.; de Vries, J. G. Angew.
Chem., Int. Ed. 2005, 44, 7442.
(7) (a) Cavallo, L. J. Am. Chem. Soc. 2002, 124, 8965. (b) Tsipis, A.
C.; Orpen, A. G.; Harvey, J. N. Dalton Trans. 2005, 2849. (c) Adlhart, C.;
Chen, P. J. Am. Chem. Soc. 2004, 126, 3496. (d) Fomine, S.; Vargas, S.
M.; Tlenkopatchev, M. A. Organometallics 2003, 22, 93.
† University of Leicester.
‡ University of Strathclyde.
(1) For the award lectures from the three Laureates, see: (a) Chauvin,
Y. Angew. Chem., Int. Edit. 2006, 45, 3740. (b) Schrock, R. R. Angew.
Chem.-Int. Ed. 2006, 45, 3748. (c) Grubbs, R. H. Angew. Chem., Int. Ed.
2006, 45, 3760. For a recent overview, see: Chem. Eng. News 2007, 85,
37. For a recent review, see: Hoveyda, A. H.; Zhugralin, A. R. Nature
2007, 450, 243.
10.1021/jo702726b CCC: $40.75 © 2008 American Chemical Society
Published on Web 02/14/2008
J. Org. Chem. 2008, 73, 2389-2395
2389

Mitchell et al.
Two classic reviews by Mandolini9 summarized a large set
of data derived from lactonizations, intramolecular etherifica-
tions, and C-C bond-forming reactions, and reported the
relationship between cyclization efficiency and ring size.
Medium rings represent some of the most difficult challenges
for cyclization strategies; the combination of developing ring
strain and the requirement to restrict rotations around 7-9
flexible bonds ensures unfavorable enthalpic and entropic10
contributions to ∆Gq; cyclization reactions that form medium
rings are therefore often relatively slow. High dilution (or
Ziegler) conditions are often employed to reduce the rate of
competing oligomerization pathways, because relatively inef-
ficient cyclization is anticipated.
SCHEME 1
The RCM reaction converts a diene precursor to two alkene
molecules, one of which is volatile ethylene in most synthetic
sequences, allowing the unfavorable ∆Sq (and ∆Hq) associated
with cyclization to be compensated for entropically. Neverthe-
less, RCM reactions that form medium rings, for example,
cyclooctannulation,11 which is severely enthalpically and en-
tropically disadvantaged, are reported to require high catalyst
loading, high dilution, long reaction times, and importantly some
degree of “gearing” of appropriately placed substituents12 to
deliver acceptable yields of cycloalkene products. These factors
combine to severely restrict scaleability (in principle). Almost
all our insights concerning RCM efficiency come from yield
measurements; there are very few kinetic studies of the RCM
reactions and those that are published involve the formation of
five-membered rings and cyclization is not rate-determining.13
There are no measured effective molarities in the literature; Fogg
has presented a number of expected values of RCM EMs based
on ring size but we are unaware of any experimental determina-
tions of EM for RCM reactions.14
Recently, we showed how we could use metalated difluoro-
alkene chemistry to advance trifluoroethanol rapidly to deliver
a number of sugar-like systems and a glycosyl phosphate
analogue; the synthesis of a cyclooctenone template 1 by RCM15
was a key step (Scheme 1).16
We wished to optimize the RCM reaction by varying the
reaction solvent and temperature and the loading of the
Ruthenium catalyst and by the correct choice of protection (R
in 1) for an allylic hydroxyl group, reporting the results of
qualitative studies in our full synthetic paper. We now wish to
report the results of a study in which substituent effects on RCM
are quantified for the first time, with the two fluorine atoms
acting as reporter groups for the various constituents within
complex reaction mixtures.
In the absence of quantitative information about a wide range
of systems, the optimization of RCM reactions remains a matter
of trial and error rather than one of rational design based on a
detailed understanding of the underlying principles.
A number of authors have reported that allylic substituents
can exert large effects on RCM reaction yield17 and regiochemi-
cal outcome.18 In the most cited paper in the area, Hoye and
(8) (a) Vyboishchikov, S. E.; Thiel, W. Chem.-Eur. J. 2005, 11, 3921.
(b) Castoldi, D.; Caggiano, L.; Panigada, L.; Sharon, O.; Costa, A. M.;
Gennari, C. Chem.- Eur. J. 2005, 12, 51.
(14) For a study that begins to examine a much wider range of systems
and discusses cyclization efficiency, see: Conrad, J. C.; Eelman, M. D.;
Duarte Silva, J. A.; Monfette, S.; Pamas, H. H.; Snelgrove, J. L.; Fogg, D.
E. J. Am. Chem. Soc. 2007, 129, 1024.
(15) (a) Miles, J. A. L.; Mitchell, L.; Percy, J. M.; Singh, K.; Uneyama,
E. J. Org. Chem., 2007, 72, 1575-1587. (b) For a related system, see:
Griffith, G. A.; Percy, J. M.; Pintat, S.; Smith, C. A.; Spencer, N.; Uneyama,
E. Org. Biomol. Chem. 2005, 3, 2701.
(9) (a) Illuminati, G.; Mandolini, L. Acc. Chem. Res. 1981, 14, 95. (b)
Galli, C.; Mandolini, L. Eur. J. Org. Chem. 2000, 3117.
(10) Buszek and co-workers have measured very small ∆Sq values for a
series of lactonisations forming eight-membered rings; see: Buszek, K. R.;
Jeong, Y.; Sato, N.; Still, P. C.; Muino, P. L.; Ghosh, I. Synth. Commun.
2001, 31, 1781.
(11) For a recent review, see: Michaut, A.; Rodriguez, J. Angew. Chem.,
Int. Ed. 2006, 45, 5740.
(16) For examples of RCM-based syntheses of selectively fluorinated
molecules, see: (a) Butt, A. H.; Percy, J. M.; Spencer, N. S. Chem. Commun.
2000, 1691. (b) Audouard, C.; Fawcett, J.; Griffiths, G. A.; Percy, J. M.;
Pintat, S.; Smith, C. A. Org. Biomol. Chem. 2004, 2, 528. (c) Audouard,
C.; Fawcett, J.; Griffith, G. A.; Kerouredan, E.; Miah, A.; Percy, J. M.;
Yang, H. L. Org. Lett. 2004, 6, 4269. (d) Fustero, S.; Catalan, S.; Piera, J.;
Sanz-Cervera, J. F.; Fernandez, B.; Acena, J. L. J. Org. Chem. 2006, 71,
4010. (e) Fustero, S.; Sanchez-Rosello, M.; Jimenez, D.; Sanz-Cervera, J.
F.; del Pozo, C.; Acena, J. L. J. Org. Chem. 2006, 71, 2706. (f) Fustero,
S.; Bartolome, A.; Sanz-Cervera, J. F.; Sanchez-Rosello, M.; Soler, J. G.;
de Arellano, C. R.; Fuentes, A. S. Org. Lett. 2003, 5, 2523. (g) De Matteis,
V.; van Delft, F. L.; Jakobi, H.; Lindell, S.; Tiebes, J.; Rutjes, F. J. Org.
Chem. 2006, 71, 7527. (h) De Matteis, V.; van Delft, F. L.; Tiebes, J.;
Rutjes, F. Eur. J. Org. Chem. 2006, 1166. (i) Yang, Y. Y.; Meng, W. D.;
Qing, F. L. Org. Lett. 2004, 6, 4257. (j) You, Z. W.; Wu, Y. Y.; Qing, F.
L. Tetrahedron Lett. 2004, 45, 9479.
(12) The literature describes a number of failed attempts to cyclize simple
(often geminally disubstituted) cyclooctene precursors; however, Taylor and
Crimmins found that precursors with appropriately placed vicinal substit-
uents could be cyclized successfully. For successful examples demonstrating
the importance of substituent patterns or gearing, see: (a) Crimmins, M.
T.; Choy, A. L. J. Org. Chem. 1997, 62, 7548. (b) Crimmins, M. T.; Tabet,
E. A. J. Am. Chem. Soc. 2000, 122, 5473. (c) Edwards, S. D.; Lewis, T.;
Taylor, R. J. K. Tetrahedron Lett. 1999, 40, 4267. For unsuccessful attempts
to cyclize less substituted systems, see: (d) Kirkland, T. A.; Grubbs, R. H.
J. Org. Chem. 1997, 62, 7310. (e) Hammer, K.; Undheim, K. Tetrahedron
1997, 53, 2309. For the ROMP of cyclooctene with tungsten alkylidene
catalysis, see: (f) Kress, J. J. Mol. Catal. A 1995, 102, 7. The copious
ROMP literature for cyclooctenes is well reviewed by Ivin; see: (g) Ivin,
K. J. Olefin Metathesis; Academic Press: New York, 1983.
(13) Most of the quantitative studies deal with simple 5- and 6-ring-
forming reactions: (a) Dias, E. L.; Nguyen, S. T.; Grubbs, R. H. J. Am.
Chem. Soc. 1997, 119, 3887. (b) Bassetti, M.; Centola, F.; Semeril, D.;
Bruneau, C.; Dixneuf, P. H. Organometallics 2003, 22, 4459. See also:
(c) Basu, K.; Cabral, J. A.; Paquette, L. A. Tetrahedron Lett. 2002, 43,
5453. (d) Guo, X.; Basu, K.; Cabral, J. A.; Paquette, L. A. Org. Lett. 2003,
5, 789. (e) Paquette, L. A.; Basu, K.; Eppich, J. C.; Hofferberth, J. E. HelV.
Chim. Acta 2002, 85, 3033.
(17) For reports of significant substitutent effects on RCM outcomes,
see: (a) Castoldi, D.; Caggiano, L.; Bayon, P.; Costa, A. M.; Cappella, P.;
Sharon, O.; Gennari, C. Tetrahedron 2005, 61, 2123. (b) Caggiano, L.;
Castoldi, D.; Beumer, R.; Bayon, P.; Tesler, J.; Gennari, C. Tetrahedron
Lett. 2003, 44, 7913. (c) Kaliappan, K. P.; Kumar, N. Tetrahedron 2005,
61, 7461. (d) Maishal, T. K.; Sinha-Mahapatra, D. K.; Paranjape, K.; Sarkar,
A. Tetrahedron Lett. 2002, 43, 2263. (e) Hyldtoft, L.; Madsen, R. J. Am.
Chem. Soc. 2000, 122, 8444.
2390 J. Org. Chem., Vol. 73, No. 6, 2008

Studies of a Range of Reactions Forming Cyclooctenones
SCHEME 2
CHART 1
FIGURE 1. Substrate consumption appears instantaneously in the
absence of sample treatment. RCM of 6c (0.01 M in DCM, 1 mol %
3, 30 mol % Ti(O-iPr)4, 25 ( 0.1 °C).
This range of substrates will allow us to begin to quantify
both the effects of the free hydroxyl group and commonly used
protected forms at the allylic position in path B and the gem-
dialkyl effect on the rate and efficiency of the cyclization
reaction.
Zhao19 studied the effect of allylic substitutents on the rate
ofcyclopentannulation (Scheme 2, path A) using Grubbs’ first
generation catalyst 4 and concluded that “free allylic hydroxyl
groups exerted a large actiVating effect upon the initial carbene
exchange reaction with an adjacent vinyl group...”. They also
observed that “secondary hydroxyl groups are a liability in RCM
reactions because of a net fragmentation reaction that consumes
ruthenium alkylidene species” (this arises via the isomerization
pathway shown in the scheme). Hoye’s system (5b) sterically
commits the active catalyst to react at the allylic alcohol terminus
(Path A).20 The published papers that cite Hoye and Zhao usually
refer to another order of events which is path B (substrate 5a
would react by this pathway).
Results and Discussion
Development of a Sampling Protocol for RCM Reactions.
We developed a protocol to follow RCM directly (2% precata-
lyst 3, 30% Ti(Oi-Pr)4 cocatalyst, 10 mM in substrate in dry
degassed CH2Cl2 at room temperature or above), in which
aliquots were withdrawn by syringe and passed through C-18
silanized solid-phase extraction (SPE) tubes that had been
preconditioned with wet acetonitrile. Each sample was eluted
with acetonitrile into a GC vial. This protocol ensured that the
alkylidene catalyst was destroyed very close to the time of
sampling and gave reproducible results for all substrates. Figure
1 shows the effects of aliquot treatment on the cyclization of
6c; in the untreated experiment, the consumption of precursor
appears to be instantaneous. Samples removed from reactions
which were transferred straight to GC vials for analysis without
treatment and the queued for GC analysis gave unreproducible
and misleading results as the RCM continued in the GC vial,
despite addition of wet CH2Cl2.22
The apparent rapid consumption of starting material can be
shown to be an experimental artifact by 19F NMR analysis, in
which starting material can be seen to be the major species
present in untreated aliquots taken close to the beginning of
the reaction. The profiles arising from the treated samples show
the smooth conversion of diene monomer into eight-membered-
ring product, without the formation of other products (detectable
by GC-MS or 19F NMR). The product formation curve mirrors
precursor consumption; response factors were determined for
all precursors and products. We also checked the composition
of product mixtures before and after passage through the SPE
media; there was no perturbation of the composition, indicating
that there was no selective retention of products or precursors
The large effects in Hoye’s system presumably arise because
of the orientation of the addition of the starting alkylidene across
the alkenyl group of the allylic fragment. To start the reaction,
the metal must add close to the substituent (the fragmentation
reaction can then proceed from this pathway) and significant
steric effects would be expected because the coordination sphere
around ruthenium is compressed by bulky ligands. Hoye
concludes that the large effect observed arises from a difference
in the rate of alkylidene transfer, rather than cyclization (a
conclusion overlooked by most of the authors citing the paper).
In the case of 1a-c and 6a-c (Chart 1), reaction would be
expected to initiate at the terminal Type I alkenyl group. Though
gem-dimethylation at the homoallylic position is likely to reduce
the rate of initiation in the more substituted species slightly,21
we would expect the rate retarding effect of an allylic oxygen
function to be bigger than the more remote steric effect.
(18) Allylic substituents appear to modulate the competition between
RCM reactions that form 5- and 6-membered rings; see: (a) Schmidt, B.;
Nave, S. Chem. Commun. 2006, 2489. (b) Quinn, K. J.; Isaacs, A. K.;
Arvary, R. A. Org. Lett. 2004, 6, 4143.
(19) Hoye, T. R.; Zhao, H. Org. Lett. 1999, 1, 1123.
(20) Chatterjee, A. K.; Choi, T. L.; Sanders, D. P.; Grubbs, R. H. J. Am.
Chem. Soc. 2003, 125, 11360.
(21) (a) Homoallylic methylation lowers the rate of alkylidene transfer
by ca. 50%; see: Ulman, M.; Grubbs, R. H. Organometallics 1998, 17,
2484. (b) For a recent investigation of related effects, see: Courchay, F.
C.; Baughman, T. W.; Wagener, K. B. J. Organomet. Chem. 2006, 691,
585.
(22) A landmark paper describes the scale-up of a synthetic campaign
in which RCM delivers a macrocycle. Analysis is secured by using a sulfur
nucleophile to sequester and deactivate the Ru-catalyst: Yee, N. K.; Farina,
V.; Houpis, I. N.; Haddad, N.; Frutos, R. P.; Gallou, F.; Wang, X. J.; Wei,
X. D.; Simpson, R. D.; Feng, X. W.; Fuchs, V.; Xu, Y. B.; Tan, J.; Zhang,
L.; Xu, J. H.; Smith-Keenan, L. L.; Vitous, J.; Ridges, M. D.; Spinelli, E.
M.; Johnson, M.; Donsbach, K.; Nicola, T.; Brenner, M.; Winter, E.; Kreye,
P.; Samstag, W. J. Org. Chem. 2006, 71, 7133.
J. Org. Chem, Vol. 73, No. 6, 2008 2391

Mitchell et al.
FIGURE 2. Competitive RCM between substrates 6a, 6b, and 6c (0.01
M in each substrate in DCM, 2 mol % 3, 30 mol % Ti(O-iPr)4 per
substrate, 25 ( 0.1 °C, treated aliquots).
FIGURE 3. Competitive RCM between gem-dimethylated 6c and less
substituted 1c (0.01 M in each substrate in DCM, 2 mol % 3, 30 mol
% Ti(O-iPr)4 per substrate, 25 ( 0.1 °C, treated aliquots).
(see the Supporting Information for details). Carrying out the
reactions at higher concentration leads to the formation of more
complex reaction mixtures, as discussed later in the paper
({1H}19F NMR spectra for simple and complex RCM product
mixtures and characterization studies are presented in the
Supporting Information).
Kinetic Profiling of the RCM Reactions. A competition
experiment was carried with 6a-c in the same reactor to
investigate the relationship between cyclization reactivity and
allylic functional group. The reactions were run in degassed
CH2Cl2 at 298 K with 3 at 2 mM (2 mol %) and Ti(O-iPr)4 (30
mol %) per substrate (each substrate present at 10 mM) to
maintain a constant ratio of catalyst to total substrate.
Data showed a relatively low degree of scatter suggesting
that the sampling and preparation routine was reliable and
reproducible. None of the reactions could be fitted to first-order
kinetic plots.23 The nature of the allylic substituent exerted a
decisive effect on the rate of consumption of precursor as shown
in Figure 2 (and on product formation which mirrors the
precursor consumption profile in each case); the order of
reactivity is 6a > 6b > 6c.24 The approximate half-lives of the
three substrates are 1200, 3500, and 5400 s. There is no obvious
explanation for this difference in reactivity but a number of
authors have commented on remote substituent effects on RCM
reactions.25,26 The steric sizes of -OH, -OBn, and -OBz
substituents as measured by their A-values are all rather similar
so significant steric effects seem unlikely.
We cannot explain these data by postulating the existence of
an equilibrium reaction between diene and a mixture of
cycloalkene and ethylene, because ethylene has very low
solubility in the reaction solvent and is free to depart the open
system.9f As cyclooctenol formation accurately mirrored diene
consumption, we are not observing an equilibrium reaction
between oligomers and cyclooctenol (several aliquots were
concentrated after analysis to look for oligomers by 19F NMR,
and none were found). Synthetic reactions run at higher initial
concentrations reach completion indicating that inhibition by
reaction product is not responsible for the establishment of an
equilibrium reaction under these conditions. Instead, it appears
that the loss of activity of 3 appears to be significant at room
temperature in CH2Cl2 even on the relatively short time scale
of 10 h, with major implications for the conduct of slow RCM
reactions with 3 (see the Supporting Information for the
procedure used to establish this and the experimental data). To
our knowledge, the rates of decomposition of 3, the active
benzylidene, or the methylidene formed upon turnover have not
been reported under synthetic conditions,28 though some com-
putational studies have been carried out.29
The gem-Dialkyl Effect on the RCM. Figure 3 shows the
competition between 1c and 6c in CH2Cl2 at 298 K under the
usual conditions and with the same sampling protocol as
described previously. The reaction of 6c is dominated by the
rapid cyclization with no visible induction period. The reaction
end-point lies within experimental error of 100% completion.
The failure of the slower reaction of 1c to complete is believed
to be due to the decomposition of catalyst on this time scale;
this reaction also appears to have a distinct induction period,
where relatively little substrate is turned over.
The reactions continue to different extents, with 6a reaching
100% conversion, 6b ca. 97%, and 6c 90%. Diene and
cycloalkene concentration can be measured reproducibly to (4%
so the slower reaction has clearly failed to reach completion.27
(23) Bassetti and co-workers (ref 9b) followed the formation of a
5-membered ring by RCM and noted that precursor consumption was not
first order in some circumstances. The kinetic regime can include catalyst
formation from precatalyst and a number of other events in addition to
cyclization. We have established that the range and order of reactivity is
correct by repeating the experiments with only a single substrate present.
The detailed kinetic analysis will be reported elsewhere.
(24) A referee suggested that the competition between the three substrates
may lead to perturbation of the profiles for the slower substrates. However,
the purpose of the experiment is to show the rank order of reactivity, and
the approximate range from highest to lowest, that the profiles from the
competition reaction do successfully. The order of reactivity is not changed
from single substrate experiments.
The exaggerated induction phase in the decay curve is
presumably due to competition between the two substrates. The
consumption of 1c speeds up significantly once most of 6c has
been consumed; this would be consistent with intermediate
alkylidene exiting more rapidly through cyclization in the gem-
dialkylated case. The approximate half-lives of the reactions
are 3900 and 21600 s, corresponding to an approximate rate
(25) (a) Aissa, C.; Riveiros, R.; Ragot, J.; Furstner, A. J. Am. Chem.
Soc. 2003, 125, 15512. (b) Meng, D. F.; Su, D. S.; Balog, A.; Bertinato,
P.; Sorensen, E. J.; Danishefsky, S. J.; Zheng, Y. H.; Chou, T. C.; He, L.
F.; Horwitz, S. B. J. Am. Chem. Soc. 1997, 119, 2733.
(26) For effects due to hydrogen bond formation, see: (a) Vassiliko-
giannakis, G.; Margaros, L.; Tofi, M. Org. Lett. 2004, 6, 205. Furstner, A.;
Thiel, O. R.; Blanda, G. Org. Lett. 2000, 2, 3731.
(27) Detection limits were estimated based on response factor measure-
ments and we estimate that (4% is a reasonable estimate of the accuray of
individual concentration measurements at 10 mM starting concentration.
(28) Ulman and Grubbs studied the lifetimes of a range of precatalyst
systems; the imidazolidinylidene-based precatalyst related to 3 was studied
but no data were reported for 3 (the dihydroimidazolidinylidene-based
precatalyst). See: Ulman, M.; Grubbs, R. H. J. Org. Chem. 1999, 64, 7202.
Grubbs and co-workers recently examined the decomposition of the relevant
phosphine complexes in the presence of ethene: Hong. S. H.; Wenzel, A.
G.; Salguero, T. T.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2007,
129, 7961.
(29) (a) van Rensburg, W. J.; Steynberg, P. J.; Kirk, M. M.; Meyer, W.
H.; Forman, G. S. J. Organomet. Chem. 2006, 691, 5312. (b)
2392 J. Org. Chem., Vol. 73, No. 6, 2008

Studies of a Range of Reactions Forming Cyclooctenones
SCHEME 3
SCHEME 4
difference of less than an order of magnitude, which is a typical
gem-dialkyl effect. The gem-dialkyl effect has been studied
extensively but the acceleration of medium ring formation has
not been quantified. The recent review by Jung and Piizzi30
includes a single example of an RCM subject to a gem-dialkyl
effect; gem-dimethylation on both sides of a midchain ketonic
carbonyl group (a major structural perturbation) changes the
outcome of a solvent-free metathesis reaction catalyzed by 10
completely from oligomerization to cycloheptannulation (Scheme
3).31
dilution was required for benzyl ether 1c (2.5 mM) and alcohol
1a (1 mM); significant quantities of cross metathesis products
were formed at 20 mM with these substrates, suggesting that
the allylic substituent modulates cyclization efficiency signifi-
cantly. We also noted that the synthesis of 6a could be carried
out at 10 mM without formation of other products, suggesting
that gem-dialkylation increases cyclization efficiency over cross
metathesis.
The effective molarity10,34 (EM, kintra/kinter) measures the
relative efficiency of the cyclization rate (kintra) compared to
the most chemically congruent dimerization (kinter). Scheme 4
shows how the total yields of cyclized and oligomerized
Forbes31 attributed the outcome to significant thermodynamic
destabilization of the acyclic ketone 8b relative to 9b caused
by the mutual proximity of two quaternary centers.32 As the
reactive Schrock catalyst is capable of catalyzing the ROMP
of neat 9b, the outcome would appear to be under thermody-
namic control. Our observed rate increase represents the only
observation of a kinetic gem-dialkyl effect on an RCM to our
knowledge; the origin of the effect will be explored elsewhere.
Measurement of Effective Molarities for Cyclooctannu-
lations. Scale-up requires reactions that can be run at concentra-
tions approaching 0.01 M or higher, or the volumes of solvent
required become prohibitive for normal laboratory equipment.22
We therefore sought to explore the effects of concentration upon
the metathesis outcomes quantitatively.33 In synthetic work, we
found that the RCM of benzoate 1b could be carried out
successfully up to 20 mM concentration (100% conversion, 46%
isolated purified yield, losses being incurred during removal of
ruthenium residues) with catalyst 3 (5 mol %) and a Ti(IV)
cocatalyst, with products resulting from cross metathesis forming
at higher concentrations. Discrete oligomer signals cannot be
identified by {1H}19F NMR under these conditions, as the
fluorine environments are too similar, resulting in overlapping
signals with a chemical shift similar to that of the starting diene,
indicating an acyclic species. The synthetic concentration for
RCM of 1b is a relatively high concentration compared to most
of those used in the literature for 8-membered-ring RCM. Higher
products were used to estimate kintra and kinter
.
This mechanism assumes that the RCM of 19 will always
initiate most rapidly on the Type I alkene leading to the
formation of new alkylidene 20, which closes affording met-
allocyclobutane 21. Varying the diene concentration and
exploiting the gem-difluoro group in the system by integrating
the {1H}19F NMR spectra should allow the EM to be determined
from a linear plot between the intramolecular:total intermo-
lecular product ratio and the reciprocal diene concentration,
assuming that this ratio accurately represents the partitioning
of 20 between oligomerization and cyclization pathways,
because
%
intra (yield of cyclic product 22) ) kintra[20]
and
%
inter (total yield of other products) ) kinter[diene 19][20]
so
%
intra/%inter ) kintra[20]/kinter[diene 19][20] )
EM(1/[diene 19])
The analysis assumes that ROMP (or ring opening then
(30) Jung, M. E.; Piizzi, G. Chem. ReV. 2005, 105, 1735.
ADMET) of the cycloalkene product 22 is slow under these
conditions. We have re-exposed 2b and 7b to 3 under synthetic
conditions (0.02 M) and failed to observe any change in the
19F NMR spectrum of the mixture after 24 h, vide supra.
However, under more concentrated conditions (1 M), we can
observe slow ring opening and oligomerization for 2b and 7b.
Alkenes within large rings are known to isomerize under
(31) (a) Forbes, M. D. E.; Patton, J. T.; Myers, T. L.; Maynard, H. D.;
Smith, D. W.; Schulz, G. R.; Wagener, K. B. J. Am. Chem. Soc. 1992,
114, 10978. (b) Murphy has suggested that a single methyl substituent
facilitates a macrocyclization by RCM; to our knowledge, these represent
the sole examples and neither has been quantified; see: Commeureuc, A.
G. J.; Murphy, J. A.; Dewis, M. L. Org. Lett. 2003, 5, 2785.
(32) Alder, R. W.; Allen, P. R.; Anderson, K. R.; Butts, C. P.; Khosravi,
E.; Martin, A.; Maunder, C. M.; Orpen, A. G.; St Pourcain, C. B. J. Chem.
Soc., Perkin Trans. 2 1998, 2083.
(33) An interesting qualitative study is described by: Yamamoto, K.;
Biswas, K.; Gaul, C.; Danishefsky, S. J. Tetrahedron Lett. 2003, 44, 3297.
(34) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183.
J. Org. Chem, Vol. 73, No. 6, 2008 2393

Mitchell et al.
FIGURE 5. Determination of effective molarities for the RCMs of
1b in the absence of the Ti(IV) cocatalyst (DCM, 5 mol % 3, reflux,
18 h).
the synthesis of 7a can be carried out at higher concentration
than that for 2a, we also determined an EM for 6b. The {1H}19F
NMR spectra at 218 K (the cyclooctenones are fluxional at 300
K) again showed good dispersion between an eight-membered
ring and side products, affording an EM of 1.09 M, which is
high for a reaction forming an eight-membered ring.
EMs of 0.017 and 0.008 M were obtained for the cyclizations
of 1c and 1a, respectively, using this method. These EM values
are more typical of reactions forming eight-membered rings.10a,34,37
EMs for the formation of eight-membered rings range from
0.001 to 0.1 M with the larger values obtaining for heteroan-
nelations of systems that have one or more rotors removed.38
The nature of the allylic functional group clearly exerts a
significant effect on the relatiVe efficiencies of the cyclization
and competing oligomerization.39
FIGURE 4. Determination of effective molarities for the RCMs of
(a) 1a and 1c and (b) 1b and 6b (DCM, 5 mol % 3, 30 mol % Ti(O-
iPr)4, reflux, 18 h).
We also determined the EM of the cyclization of 1b in the
absence of the Ti(IV) cocatalyst (Figure 5). The cyclization is
five times less efficient in the absence of the cocatalyst, but is
still more efficient than any of the cocatalyzed cyclizations of
the other substrates. Exploration of the role of the Ti(IV)
cocatalyst falls outside of the scope of this paper and will be
reported elsewhere.40
Most other discussions of substituent effects on RCM
outcomes derive exclusively from measurements of yields or
percent conversions that are assumed to correlate with reaction
thermodynamic control35 in the presence of 3 but our cycliza-
tions appear to be under kinetic control. These results will be
discussed more fully elsewhere.
Fortunately we observed excellent dispersion between 2b and
other products in the {1H}19F NMR spectra of product mixtures
from 1b, with the majority of the which, larger products (other
than oligomers) being formed having discrete signals, that could
be paired up using 2JF-F coupling constants and via correlation
spectra (19F-19F COSY).36 Spectra were determined initially
with a default relaxation delay, then re-recorded (duplicate
determinations) with a relaxation delay (D1) equal to 5T1 (T1
values were measured) to ensure reliable integration. Full sweep
width (0 to -300 ppm) spectra were also recorded to exclude
the possibility of peaks wrapped or folded into the window used
for integration.
In some cases, we were able to observe oligomers as distinct
ions in the electrospray mass spectra. We also synthesized an
acyclic heterodimer from 1a, and (16-membered) cyclic dimers
of 1a and 1b; these results are described fully in the Supporting
Information.
(37) Fogg and co-workers (ref 14) quote EMs for the formation of rings
of a range of sizes which are for etherifications and C-C bond forming
reactions of various types, rather than RCMs.9f While it is true to say that
EMs for a given ring size fall within a particular band, some of those bands
are quite wide. There are a number of significant changes in geometry on
the pathway from acyclic diene to cycloalkene that may be represented
well or poorly by the enthalpic and entropic properties of the product
cycloalkene. We would argue that the EMs for non-RCM reactions are at
absolute best a guideline, rather than a reliable predictor of cyclization
efficiency.
(38) For a remarkably effective RCM forming an 8-membered ring,
see: Chavan, S. P.; Thakkar, M.; Jogdand, G. F.; Kalkote, U. R. J. Org.
Chem. 2006, 71, 8986.
(39) The relative effect of ether and ester allylic functionality on cross-
metathesis rate has not been quantified but there are a number of publications
that claim that chelation between ester carbonyl oxygen and Ru disables
cross-metathesis events. For example, see: (a) McNaughton, B. R.;
Bucholtz, K. M.; Camaano-Moure, A.; Miller, B. L. Org. Lett. 2005, 7,
733. (b) Michaelis, S.; Blechert, S. Org. Lett. 2005, 7, 5513. (c) Sheddan,
N. A.; Arion, V. B.; Mulzer, J. Tetrahedron Lett. 2006, 47, 6689.
(40) Lewis acids are known to facilitate RCM; see: (a) Marsella, M. J.;
Maynard, H. D.; Grubbs, R. H. Angew. Chem., Int. Ed. Engl. 1997, 36,
1101. (b) Furstner, A.; Langemann, K. J. Am. Chem. Soc. 1997, 119, 9130.
The two papers describe quite different scenarios. In the former, a crown
ether is synthesized efficiently via a Group I metal-templated RCM. The
The correlation between intramolecular:intermolecular prod-
uct ratio and the reciprocal concentration is excellent (Figure
4) affording an EM of 0.25 M, which is high for the formation
of an eight-membered ring from a highly flexible system. Since
(35) Lee, C. W.; Grubbs, R. H. Org. Lett. 2000, 2, 2145.
(36) The possibility of alkene (positional) isomerization before cyclization
was also examined explicitly; we found no evidence for the formation of
seven-membered-ring products under conditions that might promote isomer-
ization (see the Supporting Information for details).
2394 J. Org. Chem., Vol. 73, No. 6, 2008

Studies of a Range of Reactions Forming Cyclooctenones
mL) was removed from the reaction and transferred to a GC vial
before the addition of a solution of Grubbs’ second generation
precatalyst 3 (2 mol %) in CH2Cl2 (0.2 mL); aliquots were then
withdrawn from the reaction at varying intervals throughout. Prior
to starting the experiment, Supelco 1 mL C-18 solid-phase
extraction tubes were preconditioned with 1 mL each of 20% water/
acetonitrile. Each aliquot subsequently withdrawn from the reaction
was passed through one of these tubes. Samples were eluted with
1 mL each of acetonitrile into GC vials for analysis.
Procedure for Determining Effective Molarity. Titanium
isopropoxide (0.3 equiv) was added to solutions of 1a in dry,
degassed CH2Cl2 in dry reactor tubes and the solutions were
refluxed under nitrogen for 30 min; 3 (5 mol %) was added and
the solution refluxed overnight. These samples were not SPE treated
on the basis that catalyst activity is likely to be minimal after >6
× 104 s at reflux in CH2Cl2 so there is no issue arising from further
reaction during sample concentration.
All spectra for the experiment upon repetition were run with D1
) 5 s (5T1, longest T1 measured at 1 s) to ensure no slow relaxing
signals were unresolved within the spectral time scale. The spectral
window was set to 40 ppm (between -100 and -140 ppm, centered
at -120 ppm), and 1024 scans were recorded. Folding/wrap-
ping was interrogated recording the full spectral window (0 to
-300 ppm) in 100 ppm increments (D1 ) 5 s, 256 scans per
experiment). No folding or wrapping was observed into the product
spectral window.
Each solution was cooled and the solvent carefully distilled off
at atmospheric pressure to prevent loss of volatile 2a. The crude
products were analyzed by {1H}19F NMR and ES-MS. Larger
products were not characterized from these mixtures but a series
of larger rings were identified in the ES- mass spectrum; 16-
membered-ring species were synthesized by another route (vide
infra).
rates.
However, since allylic protecting groups slow down the
rate of cyclization modestly, the assumption that these substrates
are less efficient or do not work may be misleading if analyses
were made on the same time scale as the reaction of the free
allylic alcohol. It is possible that many of the reactions for which
very low yields have been reported with protected allylic
alcohols were not left for an adequate amount of time before
carrying out a conversion or yield analysis, or that inappropriate
reaction concentrations were employed.
Conclusions
At the appropriate concentrations, all the diene substrates
studied converted smoothly and directly to their corresponding
cyclooctenones, apparently without the intermediacy of oligo-
mers, though cyclic and acyclic oligomers were detected at
higher concentrations (and characterized fully in some cases).
Reaction rate is affected modestly by gem-dialkylation and the
nature of allylic oxygen functions, in stark contrast to many
qualitative observations from the synthetic literature where
reaction yields can vary from >50% for free alcohols to zero
for derivatives.
The first EMs have been measured for RCM reactions; while
the values are typical for cycloctannelation, there is remarkable
sensitivity to the nature of the allylic substituent. The high EM
for benzoate 1b appears to arise from relatively low CM rates
for this substrate; the substrate that cyclizes most rapidly is
not necessarily the one that cyclizes most efficiently.
These results indicate strongly that substituent effects on CM
must be considered in addition to the likely rate of cyclization
when planning or evaluating an RCM reaction or outcomes.
Quantification of the interplay of structure and reactivity will
aid rational activity and efficiency-based design of RCM
substrates, especially for challenging synthetic targets. The
overall equilibrium constant for the process is important, but
the reaction outcome is determined by a wider range of subtle
factors, which we hope to understand more fully through
electronic structure calculations and detailed study of reaction
kinetics.
Duplicate EM Determination for 1a with Modified Product
Isolation. Experiments were run between 5 and 40 mM, this time
leaving weaker solutions (5-10 mM) to evaporate at room
temperature at atmospheric pressure. 19F NMR spectra were
obtained directly from the stronger solutions (15-40 mM) to
confirm that 2a was not lost during slow evaporation. All spectra
for these experiments were run with D1 ) 5 s (vide infra) to ensure
no slow relaxing signals were unresolved within the spectral time
scale.
Acknowledgment. We thank the EPSRC and the University
of Leicester (DTG studentship for L.M.), and the University of
Strathclyde for affording visitor status to L.M. We also wish to
acknowledge the use of the EPSRC’s (Engineering and Physical
Sciences Research Council) National Mass Spectrometry Service
Centre, University of Wales Swansea.
Experimental Section
Full preparative and characterization details for all compounds
are reported in the Supporting Information.
Procedure for Determining Reaction Concentration/Time
Profiles. Titanium(IV) isopropoxide (30 mol %) was added to a
0.01 M solution of dienes 1a-c and 6a-c in dry, degassed
dichloromethane and the solution was stirred in a jacketed flask
connected to a circulating chiller to maintain constant temperature
(25.0 ( 0.1 °C) under an atmosphere of nitrogen. An aliquot (0.2
Supporting Information Available: Synthetic procedures and
characterization data for kinetic substrates, kinetic data and details
of EM determination, oligomer characterization, and synthesis of
16-membered products. This material is available free of charge
via the Internet at http://pubs.acs.org.
latter paper describes the facilitation by a Ti(IV) cocatalyst of an RCM
catalyzed by 4. However, the lower Lewis acidity of 3 is expected to
diminish the effect of chelate formation on RCM; see: Furstner, A.; Thiel,
O. R.; Lehmann, C. W. Organometallics 2002, 21, 331.
JO702726B
J. Org. Chem, Vol. 73, No. 6, 2008 2395